Authors: Yixuan Wang, Paul Stynes, Pramod Pathak, Cristina Muntean
Abstract: Ensuring the safety and well-being of elderly and vulnerable populations in assisted living environments is a critical concern. Computer vision presents an innovative and powerful approach to predicting health risks through video monitoring, employing human action recognition (HAR) technology. However, real-time prediction of human actions with high performance and efficiency is a challenge. This research proposes a real-time human action recognition model that combines a deep learning model and a live video prediction and alert system, in order to predict falls, staggering and chest pain for residents in assisted living. Six thousand RGB video samples from the NTU RGB+D 60 dataset were selected to create a dataset with four classes: Falling, Staggering, Chest Pain, and Normal, with the Normal class comprising 40 daily activities. Transfer learning technique was applied to train four state-of-the-art HAR models on a GPU server, namely, UniFormerV2, TimeSformer, I3D, and SlowFast. Results of the four models are presented in this paper based on class-wise and macro performance metrics, inference efficiency, model complexity and computational costs. TimeSformer is proposed for developing the real-time human action recognition model, leveraging its leading macro F1 score (95.33%), recall (95.49%), and precision (95.19%) along with significantly higher inference throughput compared to the others. This research provides insights to enhance safety and health of the elderly and people with chronic illnesses in assisted living environments, fostering sustainable care, smarter communities and industry innovation.
Authors: Haoliang Shang, Hanyu Wu, Guangyao Zhai, Boyang Sun, Fangjinhua Wang, Federico Tombari, Marc Pollefeys
Abstract: Scene graphs capture complex relationships among objects, serving as strong priors for content generation and manipulation. Yet, reasonably manipulating scene graphs -- whether by adding nodes or modifying edges -- remains a challenging and untouched task. Tasks such as adding a node to the graph or reasoning about a node's relationships with all others are computationally intractable, as even a single edge modification can trigger conflicts due to the intricate interdependencies within the graph. To address these challenges, we introduce SG-Tailor, an autoregressive model that predicts the conflict-free relationship between any two nodes. SG-Tailor not only infers inter-object relationships, including generating commonsense edges for newly added nodes but also resolves conflicts arising from edge modifications to produce coherent, manipulated graphs for downstream tasks. For node addition, the model queries the target node and other nodes from the graph to predict the appropriate relationships. For edge modification, SG-Tailor employs a Cut-And-Stitch strategy to solve the conflicts and globally adjust the graph. Extensive experiments demonstrate that SG-Tailor outperforms competing methods by a large margin and can be seamlessly integrated as a plug-in module for scene generation and robotic manipulation tasks.
Authors: Tonmoy Ghosh, Edward Sazonov
Abstract: Food image recognition is a challenging task in computer vision due to the high variability and complexity of food images. In this study, we investigate the potential of Noisy Vision Transformers (NoisyViT) for improving food classification performance. By introducing noise into the learning process, NoisyViT reduces task complexity and adjusts the entropy of the system, leading to enhanced model accuracy. We fine-tune NoisyViT on three benchmark datasets: Food2K (2,000 categories, ~1M images), Food-101 (101 categories, ~100K images), and CNFOOD-241 (241 categories, ~190K images). The performance of NoisyViT is evaluated against state-of-the-art food recognition models. Our results demonstrate that NoisyViT achieves Top-1 accuracies of 95%, 99.5%, and 96.6% on Food2K, Food-101, and CNFOOD-241, respectively, significantly outperforming existing approaches. This study underscores the potential of NoisyViT for dietary assessment, nutritional monitoring, and healthcare applications, paving the way for future advancements in vision-based food computing. Code for reproducing NoisyViT for food recognition is available at NoisyViT_Food.
Authors: Kangwei Liu, Junwu Liu, Yun Cao, Jinlin Guo, Xiaowei Yi
Abstract: Recent advances in talking face generation have significantly improved facial animation synthesis. However, existing approaches face fundamental limitations: 3DMM-based methods maintain temporal consistency but lack fine-grained regional control, while Stable Diffusion-based methods enable spatial manipulation but suffer from temporal inconsistencies. The integration of these approaches is hindered by incompatible control mechanisms and semantic entanglement of facial representations. This paper presents DisentTalk, introducing a data-driven semantic disentanglement framework that decomposes 3DMM expression parameters into meaningful subspaces for fine-grained facial control. Building upon this disentangled representation, we develop a hierarchical latent diffusion architecture that operates in 3DMM parameter space, integrating region-aware attention mechanisms to ensure both spatial precision and temporal coherence. To address the scarcity of high-quality Chinese training data, we introduce CHDTF, a Chinese high-definition talking face dataset. Extensive experiments show superior performance over existing methods across multiple metrics, including lip synchronization, expression quality, and temporal consistency. Project Page: https://kangweiiliu.github.io/DisentTalk.
Authors: Arun Reddy, Alexander Martin, Eugene Yang, Andrew Yates, Kate Sanders, Kenton Murray, Reno Kriz, Celso M. de Melo, Benjamin Van Durme, Rama Chellappa
Abstract: In this work, we tackle the problem of text-to-video retrieval (T2VR). Inspired by the success of late interaction techniques in text-document, text-image, and text-video retrieval, our approach, Video-ColBERT, introduces a simple and efficient mechanism for fine-grained similarity assessment between queries and videos. Video-ColBERT is built upon 3 main components: a fine-grained spatial and temporal token-wise interaction, query and visual expansions, and a dual sigmoid loss during training. We find that this interaction and training paradigm leads to strong individual, yet compatible, representations for encoding video content. These representations lead to increases in performance on common text-to-video retrieval benchmarks compared to other bi-encoder methods.
Authors: Yifei Feng, Mingxin Yang, Shuhui Yang, Sheng Zhang, Jiaao Yu, Zibo Zhao, Yuhong Liu, Jie Jiang, Chunchao Guo
Abstract: Painting textures for existing geometries is a critical yet labor-intensive process in 3D asset generation. Recent advancements in text-to-image (T2I) models have led to significant progress in texture generation. Most existing research approaches this task by first generating images in 2D spaces using image diffusion models, followed by a texture baking process to achieve UV texture. However, these methods often struggle to produce high-quality textures due to inconsistencies among the generated multi-view images, resulting in seams and ghosting artifacts. In contrast, 3D-based texture synthesis methods aim to address these inconsistencies, but they often neglect 2D diffusion model priors, making them challenging to apply to real-world objects To overcome these limitations, we propose RomanTex, a multiview-based texture generation framework that integrates a multi-attention network with an underlying 3D representation, facilitated by our novel 3D-aware Rotary Positional Embedding. Additionally, we incorporate a decoupling characteristic in the multi-attention block to enhance the model's robustness in image-to-texture task, enabling semantically-correct back-view synthesis. Furthermore, we introduce a geometry-related Classifier-Free Guidance (CFG) mechanism to further improve the alignment with both geometries and images. Quantitative and qualitative evaluations, along with comprehensive user studies, demonstrate that our method achieves state-of-the-art results in texture quality and consistency.
Authors: Lingyan Ran, Lidong Wang, Guangcong Wang, Peng Wang, Yanning Zhang
Abstract: The task of translating visible-to-infrared images (V2IR) is inherently challenging due to three main obstacles: 1) achieving semantic-aware translation, 2) managing the diverse wavelength spectrum in infrared imagery, and 3) the scarcity of comprehensive infrared datasets. Current leading methods tend to treat V2IR as a conventional image-to-image synthesis challenge, often overlooking these specific issues. To address this, we introduce DiffV2IR, a novel framework for image translation comprising two key elements: a Progressive Learning Module (PLM) and a Vision-Language Understanding Module (VLUM). PLM features an adaptive diffusion model architecture that leverages multi-stage knowledge learning to infrared transition from full-range to target wavelength. To improve V2IR translation, VLUM incorporates unified Vision-Language Understanding. We also collected a large infrared dataset, IR-500K, which includes 500,000 infrared images compiled by various scenes and objects under various environmental conditions. Through the combination of PLM, VLUM, and the extensive IR-500K dataset, DiffV2IR markedly improves the performance of V2IR. Experiments validate DiffV2IR's excellence in producing high-quality translations, establishing its efficacy and broad applicability. The code, dataset, and DiffV2IR model will be available at https://github.com/LidongWang-26/DiffV2IR.
Authors: Alexander Lobashev, Maria Larchenko, Dmitry Guskov
Abstract: We propose SW-Guidance, a training-free approach for image generation conditioned on the color distribution of a reference image. While it is possible to generate an image with fixed colors by first creating an image from a text prompt and then applying a color style transfer method, this approach often results in semantically meaningless colors in the generated image. Our method solves this problem by modifying the sampling process of a diffusion model to incorporate the differentiable Sliced 1-Wasserstein distance between the color distribution of the generated image and the reference palette. Our method outperforms state-of-the-art techniques for color-conditional generation in terms of color similarity to the reference, producing images that not only match the reference colors but also maintain semantic coherence with the original text prompt. Our source code is available at https://github.com/alobashev/sw-guidance/.
Authors: Maria Larchenko, Alexander Lobashev, Dmitry Guskov, Vladimir Vladimirovich Palyulin
Abstract: In this work, we introduce Modulated Flows (ModFlows), a novel approach for color transfer between images based on rectified flows. The primary goal of the color transfer is to adjust the colors of a target image to match the color distribution of a reference image. Our technique is based on optimal transport and executes color transfer as an invertible transformation within the RGB color space. The ModFlows utilizes the bijective property of flows, enabling us to introduce a common intermediate color distribution and build a dataset of rectified flows. We train an encoder on this dataset to predict the weights of a rectified model for new images. After training on a set of optimal transport plans, our approach can generate plans for new pairs of distributions without additional fine-tuning. We additionally show that the trained encoder provides an image embedding, associated only with its color style. The presented method is capable of processing 4K images and achieves the state-of-the-art performance in terms of content and style similarity. Our source code is available at https://github.com/maria-larchenko/modflows
Authors: Zhongyu Yang, Jun Chen, Dannong Xu, Junjie Fei, Xiaoqian Shen, Liangbing Zhao, Chun-Mei Feng, Mohamed Elhoseiny
Abstract: Knowledge discovery and collection are intelligence-intensive tasks that traditionally require significant human effort to ensure high-quality outputs. Recent research has explored multi-agent frameworks for automating Wikipedia-style article generation by retrieving and synthesizing information from the internet. However, these methods primarily focus on text-only generation, overlooking the importance of multimodal content in enhancing informativeness and engagement. In this work, we introduce WikiAutoGen, a novel system for automated multimodal Wikipedia-style article generation. Unlike prior approaches, WikiAutoGen retrieves and integrates relevant images alongside text, enriching both the depth and visual appeal of generated content. To further improve factual accuracy and comprehensiveness, we propose a multi-perspective self-reflection mechanism, which critically assesses retrieved content from diverse viewpoints to enhance reliability, breadth, and coherence, etc. Additionally, we introduce WikiSeek, a benchmark comprising Wikipedia articles with topics paired with both textual and image-based representations, designed to evaluate multimodal knowledge generation on more challenging topics. Experimental results show that WikiAutoGen outperforms previous methods by 8%-29% on our WikiSeek benchmark, producing more accurate, coherent, and visually enriched Wikipedia-style articles. We show some of our generated examples in https://wikiautogen.github.io/ .
Authors: Axel Descamps, S\'el\`ene Forget, Ali\'enor Lahlou, Claire Lavergne, Camille Berthelot, Guillaume Stirnemann, Rodolphe Vuilleumier, Nicolas Ch\'eron
Abstract: Grouping elements into families to analyse them separately is a standard analysis procedure in many areas of sciences. We propose herein a new algorithm based on the simple idea that members from a family look like each other, and don't resemble elements foreign to the family. After reordering the data according to the distance between elements, the analysis is automatically performed with easily-understandable parameters. Noise is explicitly taken into account to deal with the variety of problems of a data-driven world. We applied the algorithm to sort biomolecules conformations, gene sequences, cells, images, and experimental conditions.
Authors: Sina Ditzel, Achref Jaziri, Iuliia Pliushch, Visvanathan Ramesh
Abstract: The robustness of image recognition algorithms remains a critical challenge, as current models often depend on large quantities of labeled data. In this paper, we propose a hybrid approach that combines the adaptability of neural networks with the interpretability, transparency, and robustness of domain-specific quasi-invariant operators. Our method decomposes the recognition into multiple task-specific operators that focus on different characteristics, supported by a novel confidence measurement tailored to these operators. This measurement enables the network to prioritize reliable features and accounts for noise. We argue that our design enhances transparency and robustness, leading to improved performance, particularly in low-data regimes. Experimental results in traffic sign detection highlight the effectiveness of the proposed method, especially in semi-supervised and unsupervised scenarios, underscoring its potential for data-constrained applications.
Authors: Md. Barkat Ullah Tusher, Shartaz Khan Akash, Amirul Islam Showmik
Abstract: This paper showcases an experimental study on anomaly detection using computer vision. The study focuses on class distinction and performance evaluation, combining OpenCV with deep learning techniques while employing a TensorFlow-based convolutional neural network for real-time face recognition and classification. The system effectively distinguishes among three classes: authorized personnel (admin), intruders, and non-human entities. A MobileNetV2-based deep learning model is utilized to optimize real-time performance, ensuring high computational efficiency without compromising accuracy. Extensive dataset preprocessing, including image augmentation and normalization, enhances the models generalization capabilities. Our analysis demonstrates classification accuracies of 90.20% for admin, 98.60% for intruders, and 75.80% for non-human detection, while maintaining an average processing rate of 30 frames per second. The study leverages transfer learning, batch normalization, and Adam optimization to achieve stable and robust learning, and a comparative analysis of class differentiation strategies highlights the impact of feature extraction techniques and training methodologies. The results indicate that advanced feature selection and data augmentation significantly enhance detection performance, particularly in distinguishing human from non-human scenes. As an experimental study, this research provides critical insights into optimizing deep learning-based surveillance systems for high-security environments and improving the accuracy and efficiency of real-time anomaly detection.
Authors: Tommie Kerssies, Niccol\`o Cavagnero, Alexander Hermans, Narges Norouzi, Giuseppe Averta, Bastian Leibe, Gijs Dubbelman, Daan de Geus
Abstract: Vision Transformers (ViTs) have shown remarkable performance and scalability across various computer vision tasks. To apply single-scale ViTs to image segmentation, existing methods adopt a convolutional adapter to generate multi-scale features, a pixel decoder to fuse these features, and a Transformer decoder that uses the fused features to make predictions. In this paper, we show that the inductive biases introduced by these task-specific components can instead be learned by the ViT itself, given sufficiently large models and extensive pre-training. Based on these findings, we introduce the Encoder-only Mask Transformer (EoMT), which repurposes the plain ViT architecture to conduct image segmentation. With large-scale models and pre-training, EoMT obtains a segmentation accuracy similar to state-of-the-art models that use task-specific components. At the same time, EoMT is significantly faster than these methods due to its architectural simplicity, e.g., up to 4x faster with ViT-L. Across a range of model sizes, EoMT demonstrates an optimal balance between segmentation accuracy and prediction speed, suggesting that compute resources are better spent on scaling the ViT itself rather than adding architectural complexity. Code: https://www.tue-mps.org/eomt/.
Authors: Marco Garosi, Alessandro Conti, Gaowen Liu, Elisa Ricci, Massimiliano Mancini
Abstract: Attribute detection is crucial for many computer vision tasks, as it enables systems to describe properties such as color, texture, and material. Current approaches often rely on labor-intensive annotation processes which are inherently limited: objects can be described at an arbitrary level of detail (e.g., color vs. color shades), leading to ambiguities when the annotators are not instructed carefully. Furthermore, they operate within a predefined set of attributes, reducing scalability and adaptability to unforeseen downstream applications. We present Compositional Caching (ComCa), a training-free method for open-vocabulary attribute detection that overcomes these constraints. ComCa requires only the list of target attributes and objects as input, using them to populate an auxiliary cache of images by leveraging web-scale databases and Large Language Models to determine attribute-object compatibility. To account for the compositional nature of attributes, cache images receive soft attribute labels. Those are aggregated at inference time based on the similarity between the input and cache images, refining the predictions of underlying Vision-Language Models (VLMs). Importantly, our approach is model-agnostic, compatible with various VLMs. Experiments on public datasets demonstrate that ComCa significantly outperforms zero-shot and cache-based baselines, competing with recent training-based methods, proving that a carefully designed training-free approach can successfully address open-vocabulary attribute detection.
Authors: Mingzhen Huang, Fu-Jen Chu, Bugra Tekin, Kevin J Liang, Haoyu Ma, Weiyao Wang, Xingyu Chen, Pierre Gleize, Hongfei Xue, Siwei Lyu, Kris Kitani, Matt Feiszli, Hao Tang
Abstract: We introduce HOIGPT, a token-based generative method that unifies 3D hand-object interactions (HOI) perception and generation, offering the first comprehensive solution for captioning and generating high-quality 3D HOI sequences from a diverse range of conditional signals (\eg text, objects, partial sequences). At its core, HOIGPT utilizes a large language model to predict the bidrectional transformation between HOI sequences and natural language descriptions. Given text inputs, HOIGPT generates a sequence of hand and object meshes; given (partial) HOI sequences, HOIGPT generates text descriptions and completes the sequences. To facilitate HOI understanding with a large language model, this paper introduces two key innovations: (1) a novel physically grounded HOI tokenizer, the hand-object decomposed VQ-VAE, for discretizing HOI sequences, and (2) a motion-aware language model trained to process and generate both text and HOI tokens. Extensive experiments demonstrate that HOIGPT sets new state-of-the-art performance on both text generation (+2.01% R Precision) and HOI generation (-2.56 FID) across multiple tasks and benchmarks.
Authors: Yufan Ren, Zicong Jiang, Tong Zhang, S{\o}ren Forchhammer, Sabine S\"usstrunk
Abstract: Text-guided image editing using Text-to-Image (T2I) models often fails to yield satisfactory results, frequently introducing unintended modifications, such as the loss of local detail and color changes. In this paper, we analyze these failure cases and attribute them to the indiscriminate optimization across all frequency bands, even though only specific frequencies may require adjustment. To address this, we introduce a simple yet effective approach that enables the selective optimization of specific frequency bands within localized spatial regions for precise edits. Our method leverages wavelets to decompose images into different spatial resolutions across multiple frequency bands, enabling precise modifications at various levels of detail. To extend the applicability of our approach, we provide a comparative analysis of different frequency-domain techniques. Additionally, we extend our method to 3D texture editing by performing frequency decomposition on the triplane representation, enabling frequency-aware adjustments for 3D textures. Quantitative evaluations and user studies demonstrate the effectiveness of our method in producing high-quality and precise edits.
Authors: Chenyangguang Zhang, Alexandros Delitzas, Fangjinhua Wang, Ruida Zhang, Xiangyang Ji, Marc Pollefeys, Francis Engelmann
Abstract: We introduce the task of predicting functional 3D scene graphs for real-world indoor environments from posed RGB-D images. Unlike traditional 3D scene graphs that focus on spatial relationships of objects, functional 3D scene graphs capture objects, interactive elements, and their functional relationships. Due to the lack of training data, we leverage foundation models, including visual language models (VLMs) and large language models (LLMs), to encode functional knowledge. We evaluate our approach on an extended SceneFun3D dataset and a newly collected dataset, FunGraph3D, both annotated with functional 3D scene graphs. Our method significantly outperforms adapted baselines, including Open3DSG and ConceptGraph, demonstrating its effectiveness in modeling complex scene functionalities. We also demonstrate downstream applications such as 3D question answering and robotic manipulation using functional 3D scene graphs. See our project page at https://openfungraph.github.io
Authors: Sara Al-Emadi, Yin Yang, Ferda Ofli
Abstract: Object detectors have achieved remarkable performance in many applications; however, these deep learning models are typically designed under the i.i.d. assumption, meaning they are trained and evaluated on data sampled from the same (source) distribution. In real-world deployment, however, target distributions often differ from source data, leading to substantial performance degradation. Domain Generalisation (DG) seeks to bridge this gap by enabling models to generalise to Out-Of-Distribution (OOD) data without access to target distributions during training, enhancing robustness to unseen conditions. In this work, we examine the generalisability and robustness of state-of-the-art object detectors under real-world distribution shifts, focusing particularly on spatial domain shifts. Despite the need, a standardised benchmark dataset specifically designed for assessing object detection under realistic DG scenarios is currently lacking. To address this, we introduce Real-World Distribution Shifts (RWDS), a suite of three novel DG benchmarking datasets that focus on humanitarian and climate change applications. These datasets enable the investigation of domain shifts across (i) climate zones and (ii) various disasters and geographic regions. To our knowledge, these are the first DG benchmarking datasets tailored for object detection in real-world, high-impact contexts. We aim for these datasets to serve as valuable resources for evaluating the robustness and generalisation of future object detection models. Our datasets and code are available at https://github.com/RWGAI/RWDS.
Authors: Rong Wang, Fabian Prada, Ziyan Wang, Zhongshi Jiang, Chengxiang Yin, Junxuan Li, Shunsuke Saito, Igor Santesteban, Javier Romero, Rohan Joshi, Hongdong Li, Jason Saragih, Yaser Sheikh
Abstract: We present a novel method for reconstructing personalized 3D human avatars with realistic animation from only a few images. Due to the large variations in body shapes, poses, and cloth types, existing methods mostly require hours of per-subject optimization during inference, which limits their practical applications. In contrast, we learn a universal prior from over a thousand clothed humans to achieve instant feedforward generation and zero-shot generalization. Specifically, instead of rigging the avatar with shared skinning weights, we jointly infer personalized avatar shape, skinning weights, and pose-dependent deformations, which effectively improves overall geometric fidelity and reduces deformation artifacts. Moreover, to normalize pose variations and resolve coupled ambiguity between canonical shapes and skinning weights, we design a 3D canonicalization process to produce pixel-aligned initial conditions, which helps to reconstruct fine-grained geometric details. We then propose a multi-frame feature aggregation to robustly reduce artifacts introduced in canonicalization and fuse a plausible avatar preserving person-specific identities. Finally, we train the model in an end-to-end framework on a large-scale capture dataset, which contains diverse human subjects paired with high-quality 3D scans. Extensive experiments show that our method generates more authentic reconstruction and animation than state-of-the-arts, and can be directly generalized to inputs from casually taken phone photos. Project page and code is available at https://github.com/rongakowang/FRESA.
Authors: Bilal Alsallakh, Timothy Wroge, Vivek Miglani, Narine Kokhlikyan
Abstract: We explore the symmetry of the mean k x k weight kernel in each layer of various convolutional neural networks. Unlike individual neurons, the mean kernels in internal layers tend to be symmetric about their centers instead of favoring specific directions. We investigate why this symmetry emerges in various datasets and models, and how it is impacted by certain architectural choices. We show how symmetry correlates with desirable properties such as shift and flip consistency, and might constitute an inherent inductive bias in convolutional neural networks.
Authors: Najeebullah, Maaz Salman, Zar Nawab Khan Swati
Abstract: Digital image spoofing has emerged as a significant security threat in biometric authentication systems, particularly those relying on facial recognition. This study evaluates the performance of three vision based models, MobileNetV2, ResNET50, and Vision Transformer, ViT, for spoof detection in image classification, utilizing a dataset of 150,986 images divided into training , 140,002, testing, 10,984, and validation ,39,574, sets. Spoof detection is critical for enhancing the security of image recognition systems, and this research compares the models effectiveness through accuracy, precision, recall, and F1 score metrics. Results reveal that MobileNetV2 outperforms other architectures on the test dataset, achieving an accuracy of 91.59%, precision of 91.72%, recall of 91.59%, and F1 score of 91.58%, compared to ViT 86.54%, 88.28%, 86.54%, and 86.39%, respectively. On the validation dataset, MobileNetV2, and ViT excel, with MobileNetV2 slightly ahead at 97.17% accuracy versus ViT 96.36%. MobileNetV2 demonstrates faster convergence during training and superior generalization to unseen data, despite both models showing signs of overfitting. These findings highlight MobileNetV2 balanced performance and robustness, making it the preferred choice for spoof detection applications where reliability on new data is essential. The study underscores the importance of model selection in security sensitive contexts and suggests MobileNetV2 as a practical solution for real world deployment.
Authors: Hao Guo, Jianfei Zhu, Wei Fan, Chunzhi Yi, Feng Jiang
Abstract: Referring expression comprehension (REC) aims at achieving object localization based on natural language descriptions. However, existing REC approaches are constrained by object category descriptions and single-attribute intention descriptions, hindering their application in real-world scenarios. In natural human-robot interactions, users often express their desires through individual states and intentions, accompanied by guiding gestures, rather than detailed object descriptions. To address this challenge, we propose Multi-ref EC, a novel task framework that integrates state descriptions, derived intentions, and embodied gestures to locate target objects. We introduce the State-Intention-Gesture Attributes Reference (SIGAR) dataset, which combines state and intention expressions with embodied references. Through extensive experiments with various baseline models on SIGAR, we demonstrate that properly ordered multi-attribute references contribute to improved localization performance, revealing that single-attribute reference is insufficient for natural human-robot interaction scenarios. Our findings underscore the importance of multi-attribute reference expressions in advancing visual-language understanding.
Authors: Hui Chen, Liangyu Liu, Xianchao Xiu, Wanquan Liu
Abstract: Hyperspectral unmixing (HU) is a critical yet challenging task in remote sensing. However, existing nonnegative matrix factorization (NMF) methods with graph learning mostly focus on first-order or second-order nearest neighbor relationships and usually require manual parameter tuning, which fails to characterize intrinsic data structures. To address the above issues, we propose a novel adaptive multi-order graph regularized NMF method (MOGNMF) with three key features. First, multi-order graph regularization is introduced into the NMF framework to exploit global and local information comprehensively. Second, these parameters associated with the multi-order graph are learned adaptively through a data-driven approach. Third, dual sparsity is embedded to obtain better robustness, i.e., $\ell_{1/2}$-norm on the abundance matrix and $\ell_{2,1}$-norm on the noise matrix. To solve the proposed model, we develop an alternating minimization algorithm whose subproblems have explicit solutions, thus ensuring effectiveness. Experiments on simulated and real hyperspectral data indicate that the proposed method delivers better unmixing results.
Authors: Ruiyi Wang, Yushuo Zheng, Zicheng Zhang, Chunyi Li, Shuaicheng Liu, Guangtao Zhai, Xiaohong Liu
Abstract: Existing real-world image dehazing methods primarily attempt to fine-tune pre-trained models or adapt their inference procedures, thus heavily relying on the pre-trained models and associated training data. Moreover, restoring heavily distorted information under dense haze requires generative diffusion models, whose potential in dehazing remains underutilized partly due to their lengthy sampling processes. To address these limitations, we introduce a novel hazing-dehazing pipeline consisting of a Realistic Hazy Image Generation framework (HazeGen) and a Diffusion-based Dehazing framework (DiffDehaze). Specifically, HazeGen harnesses robust generative diffusion priors of real-world hazy images embedded in a pre-trained text-to-image diffusion model. By employing specialized hybrid training and blended sampling strategies, HazeGen produces realistic and diverse hazy images as high-quality training data for DiffDehaze. To alleviate the inefficiency and fidelity concerns associated with diffusion-based methods, DiffDehaze adopts an Accelerated Fidelity-Preserving Sampling process (AccSamp). The core of AccSamp is the Tiled Statistical Alignment Operation (AlignOp), which can provide a clean and faithful dehazing estimate within a small fraction of sampling steps to reduce complexity and enable effective fidelity guidance. Extensive experiments demonstrate the superior dehazing performance and visual quality of our approach over existing methods. The code is available at https://github.com/ruiyi-w/Learning-Hazing-to-Dehazing.
URLs: https://github.com/ruiyi-w/Learning-Hazing-to-Dehazing.
Authors: Fucai Ke, Vijay Kumar B G, Xingjian Leng, Zhixi Cai, Zaid Khan, Weiqing Wang, Pari Delir Haghighi, Hamid Rezatofighi, Manmohan Chandraker
Abstract: Visual reasoning (VR), which is crucial in many fields for enabling human-like visual understanding, remains highly challenging. Recently, compositional visual reasoning approaches, which leverage the reasoning abilities of large language models (LLMs) with integrated tools to solve problems, have shown promise as more effective strategies than end-to-end VR methods. However, these approaches face limitations, as frozen LLMs lack tool awareness in VR, leading to performance bottlenecks. While leveraging LLMs for reasoning is widely used in other domains, they are not directly applicable to VR due to limited training data, imperfect tools that introduce errors and reduce data collection efficiency in VR, and challenging in fine-tuning on noisy workflows. To address these challenges, we propose DWIM: i) Discrepancy-aware training Workflow generation, which assesses tool usage and extracts more viable workflows for training; and ii) Instruct-Masking fine-tuning, which guides the model to only clone effective actions, enabling the generation of more practical solutions. Our experiments demonstrate that DWIM achieves state-of-the-art performance across various VR tasks, exhibiting strong generalization on multiple widely-used datasets.
Authors: Ben Rahman
Abstract: Semantic segmentation has made significant strides in pixel-level image understanding, yet it remains limited in capturing contextual and semantic relationships between objects. Current models, such as CNN and Transformer-based architectures, excel at identifying pixel-level features but fail to distinguish semantically similar objects (e.g., "doctor" vs. "nurse" in a hospital scene) or understand complex contextual scenarios (e.g., differentiating a running child from a regular pedestrian in autonomous driving). To address these limitations, we proposed a novel Context-Aware Semantic Segmentation framework that integrates Large Language Models (LLMs) with state-of-the-art vision backbones. Our hybrid model leverages the Swin Transformer for robust visual feature extraction and GPT-4 for enriching semantic understanding through text embeddings. A Cross-Attention Mechanism is introduced to align vision and language features, enabling the model to reason about context more effectively. Additionally, Graph Neural Networks (GNNs) are employed to model object relationships within the scene, capturing dependencies that are overlooked by traditional models. Experimental results on benchmark datasets (e.g., COCO, Cityscapes) demonstrate that our approach outperforms the existing methods in both pixel-level accuracy (mIoU) and contextual understanding (mAP). This work bridges the gap between vision and language, paving the path for more intelligent and context-aware vision systems in applications including autonomous driving, medical imaging, and robotics.
Authors: Junle Liu, Yun Zhang, Zixi Guo
Abstract: Feature Coding for Machines (FCM) aims to compress intermediate features effectively for remote intelligent analytics, which is crucial for future intelligent visual applications. In this paper, we propose a Multiscale Feature Importance-based Bit Allocation (MFIBA) for end-to-end FCM. First, we find that the importance of features for machine vision tasks varies with the scales, object size, and image instances. Based on this finding, we propose a Multiscale Feature Importance Prediction (MFIP) module to predict the importance weight for each scale of features. Secondly, we propose a task loss-rate model to establish the relationship between the task accuracy losses of using compressed features and the bitrate of encoding these features. Finally, we develop a MFIBA for end-to-end FCM, which is able to assign coding bits of multiscale features more reasonably based on their importance. Experimental results demonstrate that when combined with a retained Efficient Learned Image Compression (ELIC), the proposed MFIBA achieves an average of 38.202% bitrate savings in object detection compared to the anchor ELIC. Moreover, the proposed MFIBA achieves an average of 17.212% and 36.492% feature bitrate savings for instance segmentation and keypoint detection, respectively. When the proposed MFIBA is applied to the LIC-TCM, it achieves an average of 18.103%, 19.866% and 19.597% bit rate savings on three machine vision tasks, respectively, which validates the proposed MFIBA has good generalizability and adaptability to different machine vision tasks and FCM base codecs.
Authors: Yang Ren, Hai Jiang, Menglong Yang, Wei Li, Shuaicheng Liu
Abstract: RAW-to-sRGB mapping, or the simulation of the traditional camera image signal processor (ISP), aims to generate DSLR-quality sRGB images from raw data captured by smartphone sensors. Despite achieving comparable results to sophisticated handcrafted camera ISP solutions, existing learning-based methods still struggle with detail disparity and color distortion. In this paper, we present ISPDiffuser, a diffusion-based decoupled framework that separates the RAW-to-sRGB mapping into detail reconstruction in grayscale space and color consistency mapping from grayscale to sRGB. Specifically, we propose a texture-aware diffusion model that leverages the generative ability of diffusion models to focus on local detail recovery, in which a texture enrichment loss is further proposed to prompt the diffusion model to generate more intricate texture details. Subsequently, we introduce a histogram-guided color consistency module that utilizes color histogram as guidance to learn precise color information for grayscale to sRGB color consistency mapping, with a color consistency loss designed to constrain the learned color information. Extensive experimental results show that the proposed ISPDiffuser outperforms state-of-the-art competitors both quantitatively and visually. The code is available at https://github.com/RenYangSCU/ISPDiffuser.
Authors: Guanglu Dong, Xiangyu Liao, Mingyang Li, Guihuan Guo, Chao Ren
Abstract: Generative Adversarial Networks (GANs) have been widely applied to image super-resolution (SR) to enhance the perceptual quality. However, most existing GAN-based SR methods typically perform coarse-grained discrimination directly on images and ignore the semantic information of images, making it challenging for the super resolution networks (SRN) to learn fine-grained and semantic-related texture details. To alleviate this issue, we propose a semantic feature discrimination method, SFD, for perceptual SR. Specifically, we first design a feature discriminator (Feat-D), to discriminate the pixel-wise middle semantic features from CLIP, aligning the feature distributions of SR images with that of high-quality images. Additionally, we propose a text-guided discrimination method (TG-D) by introducing learnable prompt pairs (LPP) in an adversarial manner to perform discrimination on the more abstract output feature of CLIP, further enhancing the discriminative ability of our method. With both Feat-D and TG-D, our SFD can effectively distinguish between the semantic feature distributions of low-quality and high-quality images, encouraging SRN to generate more realistic and semantic-relevant textures. Furthermore, based on the trained Feat-D and LPP, we propose a novel opinion-unaware no-reference image quality assessment (OU NR-IQA) method, SFD-IQA, greatly improving OU NR-IQA performance without any additional targeted training. Extensive experiments on classical SISR, real-world SISR, and OU NR-IQA tasks demonstrate the effectiveness of our proposed methods.
Authors: Haoqiang Lin, Haokun Wen, Xuemeng Song, Meng Liu, Yupeng Hu, Liqiang Nie
Abstract: Composed Image Retrieval (CIR) allows users to search target images with a multimodal query, comprising a reference image and a modification text that describes the user's modification demand over the reference image. Nevertheless, due to the expensive labor cost of training data annotation, recent researchers have shifted to the challenging task of zero-shot CIR (ZS-CIR), which targets fulfilling CIR without annotated triplets. The pioneer ZS-CIR studies focus on converting the CIR task into a standard text-to-image retrieval task by pre-training a textual inversion network that can map a given image into a single pseudo-word token. Despite their significant progress, their coarse-grained textual inversion may be insufficient to capture the full content of the image accurately. To overcome this issue, in this work, we propose a novel Fine-grained Textual Inversion Network for ZS-CIR, named FTI4CIR. In particular, FTI4CIR comprises two main components: fine-grained pseudo-word token mapping and tri-wise caption-based semantic regularization. The former maps the image into a subject-oriented pseudo-word token and several attribute-oriented pseudo-word tokens to comprehensively express the image in the textual form, while the latter works on jointly aligning the fine-grained pseudo-word tokens to the real-word token embedding space based on a BLIP-generated image caption template. Extensive experiments conducted on three benchmark datasets demonstrate the superiority of our proposed method.
Authors: Hanshuo Qiu, Jie Jiang, Ruoli Yang, Lixin Zhan, Jizhao Liu
Abstract: RGB-T road scene semantic segmentation enhances visual scene understanding in complex environments characterized by inadequate illumination or occlusion by fusing information from RGB and thermal images. Nevertheless, existing RGB-T semantic segmentation models typically depend on simple addition or concatenation strategies or ignore the differences between information at different levels. To address these issues, we proposed a novel RGB-T road scene semantic segmentation network called Brain-Inspired Multi-Iteration Interaction Network (BIMII-Net). First, to meet the requirements of accurate texture and local information extraction in road scenarios like autonomous driving, we proposed a deep continuous-coupled neural network (DCCNN) architecture based on a brain-inspired model. Second, to enhance the interaction and expression capabilities among multi-modal information, we designed a cross explicit attention-enhanced fusion module (CEAEF-Module) in the feature fusion stage of BIMII-Net to effectively integrate features at different levels. Finally, we constructed a complementary interactive multi-layer decoder structure, incorporating the shallow-level feature iteration module (SFI-Module), the deep-level feature iteration module (DFI-Module), and the multi-feature enhancement module (MFE-Module) to collaboratively extract texture details and global skeleton information, with multi-module joint supervision further optimizing the segmentation results. Experimental results demonstrate that BIMII-Net achieves state-of-the-art (SOTA) performance in the brain-inspired computing domain and outperforms most existing RGB-T semantic segmentation methods. It also exhibits strong generalization capabilities on multiple RGB-T datasets, proving the effectiveness of brain-inspired computer models in multi-modal image segmentation tasks.
Authors: Zhuoran Zhao, Linlin Yang, Pengzhan Sun, Pan Hui, Angela Yao
Abstract: Recent synthetic 3D human datasets for the face, body, and hands have pushed the limits on photorealism. Face recognition and body pose estimation have achieved state-of-the-art performance using synthetic training data alone, but for the hand, there is still a large synthetic-to-real gap. This paper presents the first systematic study of the synthetic-to-real gap of 3D hand pose estimation. We analyze the gap and identify key components such as the forearm, image frequency statistics, hand pose, and object occlusions. To facilitate our analysis, we propose a data synthesis pipeline to synthesize high-quality data. We demonstrate that synthetic hand data can achieve the same level of accuracy as real data when integrating our identified components, paving the path to use synthetic data alone for hand pose estimation. Code and data are available at: https://github.com/delaprada/HandSynthesis.git.
Authors: Chaohan Wang, Yutong Xie, Qi Chen, Yuyin Zhou, Qi Wu
Abstract: Mamba, with its selective State Space Models (SSMs), offers a more computationally efficient solution than Transformers for long-range dependency modeling. However, there is still a debate about its effectiveness in high-resolution 3D medical image segmentation. In this study, we present a comprehensive investigation into Mamba's capabilities in 3D medical image segmentation by tackling three pivotal questions: Can Mamba replace Transformers? Can it elevate multi-scale representation learning? Is complex scanning necessary to unlock its full potential? We evaluate Mamba's performance across three large public benchmarks-AMOS, TotalSegmentator, and BraTS. Our findings reveal that UlikeMamba, a U-shape Mamba-based network, consistently surpasses UlikeTrans, a U-shape Transformer-based network, particularly when enhanced with custom-designed 3D depthwise convolutions, boosting accuracy and computational efficiency. Further, our proposed multi-scale Mamba block demonstrates superior performance in capturing both fine-grained details and global context, especially in complex segmentation tasks, surpassing Transformer-based counterparts. We also critically assess complex scanning strategies, finding that simpler methods often suffice, while our Tri-scan approach delivers notable advantages in the most challenging scenarios. By integrating these advancements, we introduce a new network for 3D medical image segmentation, positioning Mamba as a transformative force that outperforms leading models such as nnUNet, CoTr, and U-Mamba, offering competitive accuracy with superior computational efficiency. This study provides key insights into Mamba's unique advantages, paving the way for more efficient and accurate approaches to 3D medical imaging.
Authors: Weizhi Chen, Jingbo Chen, Yupeng Deng, Jiansheng Chen, Yuman Feng, Zhihao Xi, Diyou Liu, Kai Li, Yu Meng
Abstract: This study addresses the technical bottlenecks in handling long text and the "hallucination" issue caused by insufficient short text information in remote sensing vision-language foundation models (VLFM). We propose a novel vision-language foundation model, LRSCLIP, and a multimodal dataset, LRS2M. The main contributions are as follows: (1) By integrating multi-source remote sensing data and adopting a large language model labeling strategy, we construct the LRS2M dataset, which contains 2 million image-text pairs, providing both short and long texts for the first time, thus solving the problem of semantic granularity limitations in existing datasets; (2) The design of the LRSCLIP architecture based on Long-CLIP's KPS module, which extends CLIP's text processing capacity and achieves fine-grained cross-modal feature alignment through a dual-text loss weighting mechanism. Experimental results show that LRSCLIP improves retrieval accuracy by 10\%-20\% over the Long-CLIP baseline in the zero-shot long-text cross-modal retrieval task. For the zero-shot short-text cross-modal retrieval task, LRSCLIP achieves improvements over the current best model, GeoRSCLIP, with increases of 0.17\%, 0.67\%, and 0.92\% in Text to Image R@1, Image to Text R@1, and mR on RSITMD, respectively, and 0.04\%, 2.93\%, and 1.28\% on RSICD. In the zero-shot image classification task (average accuracy=75.75\%) and semantic localization task (Rmi=0.7653), LRSCLIP achieves state-of-the-art performance. These results validate the dual advantages of fine-grained semantic understanding and global feature matching in LRSCLIP. This work provides a new benchmark model and data support for remote sensing multimodal learning. The related code has been open source and is available at https://github.com/MitsuiChen14/LRSCLIP.
Authors: Jiaqi Liao, Zhengyuan Yang, Linjie Li, Dianqi Li, Kevin Lin, Yu Cheng, Lijuan Wang
Abstract: In this work, we study the problem of Text-to-Image In-Context Learning (T2I-ICL). While Unified Multimodal LLMs (MLLMs) have advanced rapidly in recent years, they struggle with contextual reasoning in T2I-ICL scenarios. To address this limitation, we propose a novel framework that incorporates a thought process called ImageGen-CoT prior to image generation. To avoid generating unstructured ineffective reasoning steps, we develop an automatic pipeline to curate a high-quality ImageGen-CoT dataset. We then fine-tune MLLMs using this dataset to enhance their contextual reasoning capabilities. To further enhance performance, we explore test-time scale-up strategies and propose a novel hybrid scaling approach. This approach first generates multiple ImageGen-CoT chains and then produces multiple images for each chain via sampling. Extensive experiments demonstrate the effectiveness of our proposed method. Notably, fine-tuning with the ImageGen-CoT dataset leads to a substantial 80\% performance gain for SEED-X on T2I-ICL tasks. See our project page at https://ImageGen-CoT.github.io/. Code and model weights will be open-sourced.
Authors: Yuchao Gu, Weijia Mao, Mike Zheng Shou
Abstract: Long-context autoregressive modeling has significantly advanced language generation, but video generation still struggles to fully utilize extended temporal contexts. To investigate long-context video modeling, we introduce Frame AutoRegressive (FAR), a strong baseline for video autoregressive modeling. Just as language models learn causal dependencies between tokens (i.e., Token AR), FAR models temporal causal dependencies between continuous frames, achieving better convergence than Token AR and video diffusion transformers. Building on FAR, we observe that long-context vision modeling faces challenges due to visual redundancy. Existing RoPE lacks effective temporal decay for remote context and fails to extrapolate well to long video sequences. Additionally, training on long videos is computationally expensive, as vision tokens grow much faster than language tokens. To tackle these issues, we propose balancing locality and long-range dependency. We introduce FlexRoPE, an test-time technique that adds flexible temporal decay to RoPE, enabling extrapolation to 16x longer vision contexts. Furthermore, we propose long short-term context modeling, where a high-resolution short-term context window ensures fine-grained temporal consistency, while an unlimited long-term context window encodes long-range information using fewer tokens. With this approach, we can train on long video sequences with a manageable token context length. We demonstrate that FAR achieves state-of-the-art performance in both short- and long-video generation, providing a simple yet effective baseline for video autoregressive modeling.
Authors: Chau Pham, Juan C. Caicedo, Bryan A. Plummer
Abstract: Prior work using Masked Autoencoders (MAEs) typically relies on random patch masking based on the assumption that images have significant redundancies across different channels, allowing for the reconstruction of masked content using cross-channel correlations. However, this assumption does not hold in Multi-Channel Imaging (MCI), where channels may provide complementary information with minimal feature overlap. Thus, these MAEs primarily learn local structures within individual channels from patch reconstruction, failing to fully leverage cross-channel interactions and limiting their MCI effectiveness. In this paper, we present ChA-MAEViT, an MAE-based method that enhances feature learning across MCI channels via four key strategies: (1) dynamic channel-patch masking, which compels the model to reconstruct missing channels in addition to masked patches, thereby enhancing cross-channel dependencies and improving robustness to varying channel configurations; (2) memory tokens, which serve as long-term memory aids to promote information sharing across channels, addressing the challenges of reconstructing structurally diverse channels; (3) hybrid token fusion module, which merges fine-grained patch tokens with a global class token to capture richer representations; and (4) Channel-Aware Decoder, a lightweight decoder utilizes channel tokens to effectively reconstruct image patches. Experiments on satellite and microscopy datasets, CHAMMI, JUMP-CP, and So2Sat, show that ChA-MAEViT significantly outperforms state-of-the-art MCI-ViTs by 3.0-21.5%, highlighting the importance of cross-channel interactions in MCI.
Authors: Zhiying Yan, Yiyuan Liang, Shilv Cai, Tao Zhang, Sheng Zhong, Luxin Yan, Xu Zou
Abstract: Semantic 4D Gaussians can be used for reconstructing and understanding dynamic scenes, with temporal variations than static scenes. Directly applying static methods to understand dynamic scenes will fail to capture the temporal features. Few works focus on dynamic scene understanding based on Gaussian Splatting, since once the same update strategy is employed for both dynamic and static parts, regardless of the distinction and interaction between Gaussians, significant artifacts and noise appear. We propose Dual-Hierarchical Optimization (DHO), which consists of Hierarchical Gaussian Flow and Hierarchical Gaussian Guidance in a divide-and-conquer manner. The former implements effective division of static and dynamic rendering and features. The latter helps to mitigate the issue of dynamic foreground rendering distortion in textured complex scenes. Extensive experiments show that our method consistently outperforms the baselines on both synthetic and real-world datasets, and supports various downstream tasks. Project Page: https://sweety-yan.github.io/DHO.
Authors: Yuguang Li, Ivaylo Boyadzhiev, Zixuan Liu, Linda Shapiro, Alex Colburn
Abstract: Reconstructing precise camera poses and floor plan layouts from wide-baseline RGB panoramas is a difficult and unsolved problem. We introduce BADGR, a novel diffusion model that jointly performs reconstruction and bundle adjustment (BA) to refine poses and layouts from a coarse state, using 1D floor boundary predictions from dozens of images of varying input densities. Unlike a guided diffusion model, BADGR is conditioned on dense per-entity outputs from a single-step Levenberg Marquardt (LM) optimizer and is trained to predict camera and wall positions while minimizing reprojection errors for view-consistency. The objective of layout generation from denoising diffusion process complements BA optimization by providing additional learned layout-structural constraints on top of the co-visible features across images. These constraints help BADGR to make plausible guesses on spatial relations which help constrain pose graph, such as wall adjacency, collinearity, and learn to mitigate errors from dense boundary observations with global contexts. BADGR trains exclusively on 2D floor plans, simplifying data acquisition, enabling robust augmentation, and supporting variety of input densities. Our experiments and analysis validate our method, which significantly outperforms the state-of-the-art pose and floor plan layout reconstruction with different input densities.
Authors: Philip Doldo, Derek Everett, Amol Khanna, Andre T Nguyen, Edward Raff
Abstract: Projected Gradient Descent (PGD) under the $L_\infty$ ball has become one of the defacto methods used in adversarial robustness evaluation for computer vision (CV) due to its reliability and efficacy, making a strong and easy-to-implement iterative baseline. However, PGD is computationally demanding to apply, especially when using thousands of iterations is the current best-practice recommendation to generate an adversarial example for a single image. In this work, we introduce a simple novel method for early termination of PGD based on cycle detection by exploiting the geometry of how PGD is implemented in practice and show that it can produce large speedup factors while providing the \emph{exact} same estimate of model robustness as standard PGD. This method substantially speeds up PGD without sacrificing any attack strength, enabling evaluations of robustness that were previously computationally intractable.
Authors: Jingyu Liu, Zijie Xin, Yuhan Fu, Ruixiang Zhao, Bangxiang Lan, Xirong Li
Abstract: Sketch animation, which brings static sketches to life by generating dynamic video sequences, has found widespread applications in GIF design, cartoon production, and daily entertainment. While current sketch animation methods perform well in single-object sketch animation, they struggle in multi-object scenarios. By analyzing their failures, we summarize two challenges of transitioning from single-object to multi-object sketch animation: object-aware motion modeling and complex motion optimization. For multi-object sketch animation, we propose MoSketch based on iterative optimization through Score Distillation Sampling (SDS), without any other data for training. We propose four modules: LLM-based scene decomposition, LLM-based motion planning, motion refinement network and compositional SDS, to tackle the two challenges in a divide-and-conquer strategy. Extensive qualitative and quantitative experiments demonstrate the superiority of our method over existing sketch animation approaches. MoSketch takes a pioneering step towards multi-object sketch animation, opening new avenues for future research and applications. The code will be released.
Authors: Dohwan Ko, Sihyeon Kim, Yumin Suh, Vijay Kumar B. G, Minseo Yoon, Manmohan Chandraker, Hyunwoo J. Kim
Abstract: Spatio-temporal reasoning is essential in understanding real-world environments in various fields, eg, autonomous driving and sports analytics. Recent advances have improved the spatial reasoning ability of Vision-Language Models (VLMs) by introducing large-scale data, but these models still struggle to analyze kinematic elements like traveled distance and speed of moving objects. To bridge this gap, we construct a spatio-temporal reasoning dataset and benchmark involving kinematic instruction tuning, referred to as STKit and STKit-Bench. They consist of real-world videos with 3D annotations, detailing object motion dynamics: traveled distance, speed, movement direction, inter-object distance comparisons, and relative movement direction. To further scale such data construction to videos without 3D labels, we propose an automatic pipeline to generate pseudo-labels using 4D reconstruction in real-world scale. With our kinematic instruction tuning data for spatio-temporal reasoning, we present ST-VLM, a VLM enhanced for spatio-temporal reasoning, which exhibits outstanding performance on STKit-Bench. Furthermore, we show that ST-VLM generalizes robustly across diverse domains and tasks, outperforming baselines on other spatio-temporal benchmarks (eg, ActivityNet, TVQA+). Finally, by integrating learned spatio-temporal reasoning with existing abilities, ST-VLM enables complex multi-step reasoning. Project page: https://ikodoh.github.io/ST-VLM.
Authors: Reza Pourreza, Rishit Dagli, Apratim Bhattacharyya, Sunny Panchal, Guillaume Berger, Roland Memisevic
Abstract: AI models have made significant strides in recent years in their ability to describe and answer questions about real-world images. They have also made progress in the ability to converse with users in real-time using audio input. This raises the question: have we reached the point where AI models, connected to a camera and microphone, can converse with users in real-time about scenes and events that are unfolding live in front of the camera? This has been a long-standing goal in AI and is a prerequisite for real-world AI assistants and humanoid robots to interact with humans in everyday situations. In this work, we introduce a new dataset and benchmark, the Qualcomm Interactive Video Dataset (IVD), which allows us to assess the extent to which existing models can support these abilities, and to what degree these capabilities can be instilled through fine-tuning. The dataset is based on a simple question-answering setup, where users ask questions that the system has to answer, in real-time, based on the camera and audio input. We show that existing models fall far behind human performance on this task, and we identify the main sources for the performance gap. However, we also show that for many of the required perceptual skills, fine-tuning on this form of data can significantly reduce this gap.
Authors: Farzad Beizaee, Gregory A. Lodygensky, Christian Desrosiers, Jose Dolz
Abstract: Recent advances in diffusion models have spurred research into their application for Reconstruction-based unsupervised anomaly detection. However, these methods may struggle with maintaining structural integrity and recovering the anomaly-free content of abnormal regions, especially in multi-class scenarios. Furthermore, diffusion models are inherently designed to generate images from pure noise and struggle to selectively alter anomalous regions of an image while preserving normal ones. This leads to potential degradation of normal regions during reconstruction, hampering the effectiveness of anomaly detection. This paper introduces a reformulation of the standard diffusion model geared toward selective region alteration, allowing the accurate identification of anomalies. By modeling anomalies as noise in the latent space, our proposed \textbf{Deviation correction diffusion} (\Ours) model preserves the normal regions and encourages transformations exclusively on anomalous areas. This selective approach enhances the reconstruction quality, facilitating effective unsupervised detection and localization of anomaly regions. Comprehensive evaluations demonstrate the superiority of our method in accurately identifying and localizing anomalies in complex images, with pixel-level AUPRC improvements of 11-14\% over state-of-the-art models on well known anomaly detection datasets. The code is available at https://github.com/farzad-bz/DeCo-Diff
Authors: Zhiwei Huang, Hailin Yu, Yichun Shentu, Jin Yuan, Guofeng Zhang
Abstract: This paper presents a novel camera relocalization method, STDLoc, which leverages Feature Gaussian as scene representation. STDLoc is a full relocalization pipeline that can achieve accurate relocalization without relying on any pose prior. Unlike previous coarse-to-fine localization methods that require image retrieval first and then feature matching, we propose a novel sparse-to-dense localization paradigm. Based on this scene representation, we introduce a novel matching-oriented Gaussian sampling strategy and a scene-specific detector to achieve efficient and robust initial pose estimation. Furthermore, based on the initial localization results, we align the query feature map to the Gaussian feature field by dense feature matching to enable accurate localization. The experiments on indoor and outdoor datasets show that STDLoc outperforms current state-of-the-art localization methods in terms of localization accuracy and recall.
Authors: Yunhe Gao, Di Liu, Zhuowei Li, Yunsheng Li, Dongdong Chen, Mu Zhou, Dimitris N. Metaxas
Abstract: Medical image segmentation remains challenging due to the vast diversity of anatomical structures, imaging modalities, and segmentation tasks. While deep learning has made significant advances, current approaches struggle to generalize as they require task-specific training or fine-tuning on unseen classes. We present Iris, a novel In-context Reference Image guided Segmentation framework that enables flexible adaptation to novel tasks through the use of reference examples without fine-tuning. At its core, Iris features a lightweight context task encoding module that distills task-specific information from reference context image-label pairs. This rich context embedding information is used to guide the segmentation of target objects. By decoupling task encoding from inference, Iris supports diverse strategies from one-shot inference and context example ensemble to object-level context example retrieval and in-context tuning. Through comprehensive evaluation across twelve datasets, we demonstrate that Iris performs strongly compared to task-specific models on in-distribution tasks. On seven held-out datasets, Iris shows superior generalization to out-of-distribution data and unseen classes. Further, Iris's task encoding module can automatically discover anatomical relationships across datasets and modalities, offering insights into medical objects without explicit anatomical supervision.
Authors: Piera Riccio, Francesco Galati, Kajetan Schweighofer, Noa Garcia, Nuria Oliver
Abstract: We introduce ImageSet2Text, a novel approach that leverages vision-language foundation models to automatically create natural language descriptions of image sets. Inspired by concept bottleneck models (CBMs) and based on visual-question answering (VQA) chains, ImageSet2Text iteratively extracts key concepts from image subsets, encodes them into a structured graph, and refines insights using an external knowledge graph and CLIP-based validation. This iterative process enhances interpretability and enables accurate and detailed set-level summarization. Through extensive experiments, we evaluate ImageSet2Text's descriptions on accuracy, completeness, readability and overall quality, benchmarking it against existing vision-language models and introducing new datasets for large-scale group image captioning.
Authors: Zizhi Chen, Minghao Han, Xukun Zhang, Shuwei Ma, Tao Liu, Xing Wei, Lihua Zhang
Abstract: Multimodal learning combining pathology images and genomic sequences enhances cancer survival analysis but faces clinical implementation barriers due to limited access to genomic sequencing in under-resourced regions. To enable survival prediction using only whole-slide images (WSI), we propose the Visual-Genomic Answering-Guided Transformer (VGAT), a framework integrating Visual Question Answering (VQA) techniques for genomic modality reconstruction. By adapting VQA's text feature extraction approach, we derive stable genomic representations that circumvent dimensionality challenges in raw genomic data. Simultaneously, a cluster-based visual prompt module selectively enhances discriminative WSI patches, addressing noise from unfiltered image regions. Evaluated across five TCGA datasets, VGAT outperforms existing WSI-only methods, demonstrating the viability of genomic-informed inference without sequencing. This approach bridges multimodal research and clinical feasibility in resource-constrained settings. The code link is https://github.com/CZZZZZZZZZZZZZZZZZ/VGAT.
Authors: Yufei Cai, Hu Han, Yuxiang Wei, Shiguang Shan, Xilin Chen
Abstract: The progress on generative models has led to significant advances on text-to-video (T2V) generation, yet the motion controllability of generated videos remains limited. Existing motion transfer methods explored the motion representations of reference videos to guide generation. Nevertheless, these methods typically rely on sample-specific optimization strategy, resulting in high computational burdens. In this paper, we propose \textbf{EfficientMT}, a novel and efficient end-to-end framework for video motion transfer. By leveraging a small set of synthetic paired motion transfer samples, EfficientMT effectively adapts a pretrained T2V model into a general motion transfer framework that can accurately capture and reproduce diverse motion patterns. Specifically, we repurpose the backbone of the T2V model to extract temporal information from reference videos, and further propose a scaler module to distill motion-related information. Subsequently, we introduce a temporal integration mechanism that seamlessly incorporates reference motion features into the video generation process. After training on our self-collected synthetic paired samples, EfficientMT enables general video motion transfer without requiring test-time optimization. Extensive experiments demonstrate that our EfficientMT outperforms existing methods in efficiency while maintaining flexible motion controllability. Our code will be available https://github.com/PrototypeNx/EfficientMT.
Authors: Hyeongjin Nam, Donghwan Kim, Jeongtaek Oh, Kyoung Mu Lee
Abstract: Most existing methods of 3D clothed human reconstruction from a single image treat the clothed human as a single object without distinguishing between cloth and human body. In this regard, we present DeClotH, which separately reconstructs 3D cloth and human body from a single image. This task remains largely unexplored due to the extreme occlusion between cloth and the human body, making it challenging to infer accurate geometries and textures. Moreover, while recent 3D human reconstruction methods have achieved impressive results using text-to-image diffusion models, directly applying such an approach to this problem often leads to incorrect guidance, particularly in reconstructing 3D cloth. To address these challenges, we propose two core designs in our framework. First, to alleviate the occlusion issue, we leverage 3D template models of cloth and human body as regularizations, which provide strong geometric priors to prevent erroneous reconstruction by the occlusion. Second, we introduce a cloth diffusion model specifically designed to provide contextual information about cloth appearance, thereby enhancing the reconstruction of 3D cloth. Qualitative and quantitative experiments demonstrate that our proposed approach is highly effective in reconstructing both 3D cloth and the human body. More qualitative results are provided at https://hygenie1228.github.io/DeClotH/.
Authors: Akshay Kulkarni, Ge Yan, Chung-En Sun, Tuomas Oikarinen, Tsui-Wei Weng
Abstract: Concept bottleneck models (CBM) aim to produce inherently interpretable models that rely on human-understandable concepts for their predictions. However, existing approaches to design interpretable generative models based on CBMs are not yet efficient and scalable, as they require expensive generative model training from scratch as well as real images with labor-intensive concept supervision. To address these challenges, we present two novel and low-cost methods to build interpretable generative models through post-hoc techniques and we name our approaches: concept-bottleneck autoencoder (CB-AE) and concept controller (CC). Our proposed approaches enable efficient and scalable training without the need of real data and require only minimal to no concept supervision. Additionally, our methods generalize across modern generative model families including generative adversarial networks and diffusion models. We demonstrate the superior interpretability and steerability of our methods on numerous standard datasets like CelebA, CelebA-HQ, and CUB with large improvements (average ~25%) over the prior work, while being 4-15x faster to train. Finally, a large-scale user study is performed to validate the interpretability and steerability of our methods.
Authors: Yukang Lin, Hokit Fung, Jianjin Xu, Zeping Ren, Adela S. M. Lau, Guosheng Yin, Xiu Li
Abstract: Recent portrait animation methods have made significant strides in generating realistic lip synchronization. However, they often lack explicit control over head movements and facial expressions, and cannot produce videos from multiple viewpoints, resulting in less controllable and expressive animations. Moreover, text-guided portrait animation remains underexplored, despite its user-friendly nature. We present a novel two-stage text-guided framework, MVPortrait (Multi-view Vivid Portrait), to generate expressive multi-view portrait animations that faithfully capture the described motion and emotion. MVPortrait is the first to introduce FLAME as an intermediate representation, effectively embedding facial movements, expressions, and view transformations within its parameter space. In the first stage, we separately train the FLAME motion and emotion diffusion models based on text input. In the second stage, we train a multi-view video generation model conditioned on a reference portrait image and multi-view FLAME rendering sequences from the first stage. Experimental results exhibit that MVPortrait outperforms existing methods in terms of motion and emotion control, as well as view consistency. Furthermore, by leveraging FLAME as a bridge, MVPortrait becomes the first controllable portrait animation framework that is compatible with text, speech, and video as driving signals.
Authors: Jaihoon Kim, Taehoon Yoon, Jisung Hwang, Minhyuk Sung
Abstract: We propose an inference-time scaling approach for pretrained flow models. Recently, inference-time scaling has gained significant attention in LLMs and diffusion models, improving sample quality or better aligning outputs with user preferences by leveraging additional computation. For diffusion models, particle sampling has allowed more efficient scaling due to the stochasticity at intermediate denoising steps. On the contrary, while flow models have gained popularity as an alternative to diffusion models--offering faster generation and high-quality outputs in state-of-the-art image and video generative models--efficient inference-time scaling methods used for diffusion models cannot be directly applied due to their deterministic generative process. To enable efficient inference-time scaling for flow models, we propose three key ideas: 1) SDE-based generation, enabling particle sampling in flow models, 2) Interpolant conversion, broadening the search space and enhancing sample diversity, and 3) Rollover Budget Forcing (RBF), an adaptive allocation of computational resources across timesteps to maximize budget utilization. Our experiments show that SDE-based generation, particularly variance-preserving (VP) interpolant-based generation, improves the performance of particle sampling methods for inference-time scaling in flow models. Additionally, we demonstrate that RBF with VP-SDE achieves the best performance, outperforming all previous inference-time scaling approaches.
Authors: Peishan Huang, Dong Li
Abstract: In recent years, the rapid development of machine learning has brought reforms and challenges to traditional communication systems. Semantic communication has appeared as an effective strategy to effectively extract relevant semantic signals semantic segmentation labels and image features for image transmission. However, the insufficient number of extracted semantic features of images will potentially result in a low reconstruction accuracy, which hinders the practical applications and still remains challenging for solving. In order to fill this gap, this letter proposes a multi-text transmission semantic communication (Multi-SC) system, which uses the visual language model (VLM) to assist in the transmission of image semantic signals. Unlike previous image transmission semantic communication systems, the proposed system divides the image into multiple blocks and extracts multiple text information from the image using a modified large language and visual assistant (LLaVA), and combines semantic segmentation tags with semantic text for image recovery. Simulation results show that the proposed text semantics diversity scheme can significantly improve the reconstruction accuracy compared with related works.
Authors: Zhiying Song, Lei Yang, Fuxi Wen, Jun Li
Abstract: Cooperative perception presents significant potential for enhancing the sensing capabilities of individual vehicles, however, inter-agent latency remains a critical challenge. Latencies cause misalignments in both spatial and semantic features, complicating the fusion of real-time observations from the ego vehicle with delayed data from others. To address these issues, we propose TraF-Align, a novel framework that learns the flow path of features by predicting the feature-level trajectory of objects from past observations up to the ego vehicle's current time. By generating temporally ordered sampling points along these paths, TraF-Align directs attention from the current-time query to relevant historical features along each trajectory, supporting the reconstruction of current-time features and promoting semantic interaction across multiple frames. This approach corrects spatial misalignment and ensures semantic consistency across agents, effectively compensating for motion and achieving coherent feature fusion. Experiments on two real-world datasets, V2V4Real and DAIR-V2X-Seq, show that TraF-Align sets a new benchmark for asynchronous cooperative perception.
Authors: Jiaqi Liao, Yuwei Niu, Fanqing Meng, Hao Li, Changyao Tian, Yinuo Du, Yuwen Xiong, Dianqi Li, Xizhou Zhu, Li Yuan, Jifeng Dai, Yu Cheng
Abstract: Recent years have witnessed remarkable advances in Large Vision-Language Models (LVLMs), which have achieved human-level performance across various complex vision-language tasks. Following LLaVA's paradigm, mainstream LVLMs typically employ a shallow MLP for visual-language alignment through a two-stage training process: pretraining for cross-modal alignment followed by instruction tuning. While this approach has proven effective, the underlying mechanisms of how MLPs bridge the modality gap remain poorly understood. Although some research has explored how LLMs process transformed visual tokens, few studies have investigated the fundamental alignment mechanism. Furthermore, the MLP adapter requires retraining whenever switching LLM backbones. To address these limitations, we first investigate the working principles of MLP adapters and discover that they learn to project visual embeddings into subspaces spanned by corresponding text embeddings progressively. Based on this insight, we propose LangBridge, a novel adapter that explicitly maps visual tokens to linear combinations of LLM vocabulary embeddings. This innovative design enables pretraining-free adapter transfer across different LLMs while maintaining performance. Our experimental results demonstrate that a LangBridge adapter pre-trained on Qwen2-0.5B can be directly applied to larger models such as LLaMA3-8B or Qwen2.5-14B while maintaining competitive performance. Overall, LangBridge enables interpretable vision-language alignment by grounding visual representations in LLM vocab embedding, while its plug-and-play design ensures efficient reuse across multiple LLMs with nearly no performance degradation. See our project page at https://LangBridge.github.io/
Authors: Mingxiao Tu, Hoijoon Jung, Alireza Moghadam, Jineel Raythatha, Lachlan Allan, Jeremy Hsu, Andre Kyme, Jinman Kim
Abstract: In perioperative care, precise in-bed 3D patient pose and shape estimation (PSE) can be vital in optimizing patient positioning in preoperative planning, enabling accurate overlay of medical images for augmented reality-based surgical navigation, and mitigating risks of prolonged immobility during recovery. Conventional PSE methods relying on modalities such as RGB-D, infrared, or pressure maps often struggle with occlusions caused by bedding and complex patient positioning, leading to inaccurate estimation that can affect clinical outcomes. To address these challenges, we present the first multi-modal in-bed patient 3D PSE network that fuses detailed geometric features extracted from routinely acquired computed tomography (CT) scans with depth maps (mPSE-CT). mPSE-CT incorporates a shape estimation module that utilizes probabilistic correspondence alignment, a pose estimation module with a refined neural network, and a final parameters mixing module. This multi-modal network robustly reconstructs occluded body regions and enhances the accuracy of the estimated 3D human mesh model. We validated mPSE-CT using proprietary whole-body rigid phantom and volunteer datasets in clinical scenarios. mPSE-CT outperformed the best-performing prior method by 23% and 49.16% in pose and shape estimation respectively, demonstrating its potential for improving clinical outcomes in challenging perioperative environments.
Authors: Ziyuan Liu, Jiawei Zhang, Wenyu Wang, Yuantao Gu
Abstract: Most existing change detection (CD) methods focus on optical images captured at different times, and deep learning (DL) has achieved remarkable success in this domain. However, in extreme scenarios such as disaster response, synthetic aperture radar (SAR), with its active imaging capability, is more suitable for providing post-event data. This introduces new challenges for CD methods, as existing weight-sharing Siamese networks struggle to effectively learn the cross-modal data distribution between optical and SAR images. To address this challenge, we propose a unified MultiModal CD framework, M$^2$CD. We integrate Mixture of Experts (MoE) modules into the backbone to explicitly handle diverse modalities, thereby enhancing the model's ability to learn multimodal data distributions. Additionally, we innovatively propose an Optical-to-SAR guided path (O2SP) and implement self-distillation during training to reduce the feature space discrepancy between different modalities, further alleviating the model's learning burden. We design multiple variants of M$^2$CD based on both CNN and Transformer backbones. Extensive experiments validate the effectiveness of the proposed framework, with the MiT-b1 version of M$^2$CD outperforming all state-of-the-art (SOTA) methods in optical-SAR CD tasks.
Authors: Bingjian Yao, Weiping Lin, Yan He, Zheng Wang, Liangsheng Wang
Abstract: The fine-grained annotations in whole slide images (WSIs) show the boundaries of various pathological regions. However, generating such detailed annotation is often costly, whereas the coarse annotations are relatively simpler to produce. Existing methods for refining coarse annotations often rely on extensive training samples or clean datasets, and fail to capture both intra-slide and inter-slide latent sematic patterns, limiting their precision. In this paper, we propose a prototype-guided approach. Specifically, we introduce a local-to-global approach to construct non-redundant representative prototypes by jointly modeling intra-slide local semantics and inter-slide contextual relationships. Then a prototype-guided pseudo-labeling module is proposed for refining coarse annotations. Finally, we employ dynamic data sampling and re-finetuning strategy to train a patch classifier. Extensive experiments on three publicly available WSI datasets, covering lymph, liver, and colorectal cancers, demonstrate that our method significantly outperforms existing state-of-the-art (SOTA) methods. The code will be available.
Authors: Xuli Shen, Hua Cai, Dingding Yu, Weilin Shen, Qing Xu, Xiangyang Xue
Abstract: Generating emotion-specific talking head videos from audio input is an important and complex challenge for human-machine interaction. However, emotion is highly abstract concept with ambiguous boundaries, and it necessitates disentangled expression parameters to generate emotionally expressive talking head videos. In this work, we present EmoHead to synthesize talking head videos via semantic expression parameters. To predict expression parameter for arbitrary audio input, we apply an audio-expression module that can be specified by an emotion tag. This module aims to enhance correlation from audio input across various emotions. Furthermore, we leverage pre-trained hyperplane to refine facial movements by probing along the vertical direction. Finally, the refined expression parameters regularize neural radiance fields and facilitate the emotion-consistent generation of talking head videos. Experimental results demonstrate that semantic expression parameters lead to better reconstruction quality and controllability.
Authors: Jiaxin Zhang, Junjun Jiang, Youyu Chen, Kui Jiang, Xianming Liu
Abstract: Accurate object segmentation is crucial for high-quality scene understanding in the 3D vision domain. However, 3D segmentation based on 3D Gaussian Splatting (3DGS) struggles with accurately delineating object boundaries, as Gaussian primitives often span across object edges due to their inherent volume and the lack of semantic guidance during training. In order to tackle these challenges, we introduce Clear Object Boundaries for 3DGS Segmentation (COB-GS), which aims to improve segmentation accuracy by clearly delineating blurry boundaries of interwoven Gaussian primitives within the scene. Unlike existing approaches that remove ambiguous Gaussians and sacrifice visual quality, COB-GS, as a 3DGS refinement method, jointly optimizes semantic and visual information, allowing the two different levels to cooperate with each other effectively. Specifically, for the semantic guidance, we introduce a boundary-adaptive Gaussian splitting technique that leverages semantic gradient statistics to identify and split ambiguous Gaussians, aligning them closely with object boundaries. For the visual optimization, we rectify the degraded suboptimal texture of the 3DGS scene, particularly along the refined boundary structures. Experimental results show that COB-GS substantially improves segmentation accuracy and robustness against inaccurate masks from pre-trained model, yielding clear boundaries while preserving high visual quality. Code is available at https://github.com/ZestfulJX/COB-GS.
Authors: Changyong He, Jin Zeng, Jiawei Zhang, Jiajie Guo
Abstract: Time-of-Flight (ToF) sensors efficiently capture scene depth, but the nonlinear depth construction procedure often results in extremely large noise variance or even invalid areas. Recent methods based on deep neural networks (DNNs) achieve enhanced ToF denoising accuracy but tend to struggle when presented with severe noise corruption due to limited prior knowledge of ToF data distribution. In this paper, we propose DepthCAD, a novel ToF denoising approach that ensures global structural smoothness by leveraging the rich prior knowledge in Stable Diffusion and maintains local metric accuracy by steering the diffusion process with confidence guidance. To adopt the pretrained image diffusion model to ToF depth denoising, we apply the diffusion on raw ToF correlation measurements with dynamic range normalization before converting to depth maps. Experimental results validate the state-of-the-art performance of the proposed scheme, and the evaluation on real data further verifies its robustness against real-world ToF noise.
Authors: Yiqing Li, Xuan Wang, Jiawei Wu, Yikun Ma, Zhi Jin
Abstract: Synthesizing novel views of large-scale scenes from unconstrained in-the-wild images is an important but challenging task in computer vision. Existing methods, which optimize per-image appearance and transient occlusion through implicit neural networks from dense training views (approximately 1000 images), struggle to perform effectively under sparse input conditions, resulting in noticeable artifacts. To this end, we propose SparseGS-W, a novel framework based on 3D Gaussian Splatting that enables the reconstruction of complex outdoor scenes and handles occlusions and appearance changes with as few as five training images. We leverage geometric priors and constrained diffusion priors to compensate for the lack of multi-view information from extremely sparse input. Specifically, we propose a plug-and-play Constrained Novel-View Enhancement module to iteratively improve the quality of rendered novel views during the Gaussian optimization process. Furthermore, we propose an Occlusion Handling module, which flexibly removes occlusions utilizing the inherent high-quality inpainting capability of constrained diffusion priors. Both modules are capable of extracting appearance features from any user-provided reference image, enabling flexible modeling of illumination-consistent scenes. Extensive experiments on the PhotoTourism and Tanks and Temples datasets demonstrate that SparseGS-W achieves state-of-the-art performance not only in full-reference metrics, but also in commonly used non-reference metrics such as FID, ClipIQA, and MUSIQ.
Authors: Juntao Jian, Xiuping Liu, Zixuan Chen, Manyi Li, Jian Liu, Ruizhen Hu
Abstract: Recent advances in dexterous grasping synthesis have demonstrated significant progress in producing reasonable and plausible grasps for many task purposes. But it remains challenging to generalize to unseen object categories and diverse task instructions. In this paper, we propose G-DexGrasp, a retrieval-augmented generation approach that can produce high-quality dexterous hand configurations for unseen object categories and language-based task instructions. The key is to retrieve generalizable grasping priors, including the fine-grained contact part and the affordance-related distribution of relevant grasping instances, for the following synthesis pipeline. Specifically, the fine-grained contact part and affordance act as generalizable guidance to infer reasonable grasping configurations for unseen objects with a generative model, while the relevant grasping distribution plays as regularization to guarantee the plausibility of synthesized grasps during the subsequent refinement optimization. Our comparison experiments validate the effectiveness of our key designs for generalization and demonstrate the remarkable performance against the existing approaches. Project page: https://g-dexgrasp.github.io/
Authors: Shujuan Li, Yu-Shen Liu, Zhizhong Han
Abstract: Reconstructing open surfaces from multi-view images is vital in digitalizing complex objects in daily life. A widely used strategy is to learn unsigned distance functions (UDFs) by checking if their appearance conforms to the image observations through neural rendering. However, it is still hard to learn continuous and implicit UDF representations through 3D Gaussians splatting (3DGS) due to the discrete and explicit scene representation, i.e., 3D Gaussians. To resolve this issue, we propose a novel approach to bridge the gap between 3D Gaussians and UDFs. Our key idea is to overfit thin and flat 2D Gaussian planes on surfaces, and then, leverage the self-supervision and gradient-based inference to supervise unsigned distances in both near and far area to surfaces. To this end, we introduce novel constraints and strategies to constrain the learning of 2D Gaussians to pursue more stable optimization and more reliable self-supervision, addressing the challenges brought by complicated gradient field on or near the zero level set of UDFs. We report numerical and visual comparisons with the state-of-the-art on widely used benchmarks and real data to show our advantages in terms of accuracy, efficiency, completeness, and sharpness of reconstructed open surfaces with boundaries. Project page: https://lisj575.github.io/GaussianUDF/
Authors: Haiyu Zhang, Xinyuan Chen, Yaohui Wang, Xihui Liu, Yunhong Wang, Yu Qiao
Abstract: Diffusion models have achieved remarkable progress in the field of video generation. However, their iterative denoising nature requires a large number of inference steps to generate a video, which is slow and computationally expensive. In this paper, we begin with a detailed analysis of the challenges present in existing diffusion distillation methods and propose a novel efficient method, namely AccVideo, to reduce the inference steps for accelerating video diffusion models with synthetic dataset. We leverage the pretrained video diffusion model to generate multiple valid denoising trajectories as our synthetic dataset, which eliminates the use of useless data points during distillation. Based on the synthetic dataset, we design a trajectory-based few-step guidance that utilizes key data points from the denoising trajectories to learn the noise-to-video mapping, enabling video generation in fewer steps. Furthermore, since the synthetic dataset captures the data distribution at each diffusion timestep, we introduce an adversarial training strategy to align the output distribution of the student model with that of our synthetic dataset, thereby enhancing the video quality. Extensive experiments demonstrate that our model achieves 8.5x improvements in generation speed compared to the teacher model while maintaining comparable performance. Compared to previous accelerating methods, our approach is capable of generating videos with higher quality and resolution, i.e., 5-seconds, 720x1280, 24fps.
Authors: Nadja Gruber, Johannes Schwab, Markus Haltmeier, Ander Biguri, Clemens Dlaska, Gyeongha Hwang
Abstract: We propose Noisier2Inverse, a correction-free self-supervised deep learning approach for general inverse prob- lems. The proposed method learns a reconstruction function without the need for ground truth samples and is ap- plicable in cases where measurement noise is statistically correlated. This includes computed tomography, where detector imperfections or photon scattering create correlated noise patterns, as well as microscopy and seismic imaging, where physical interactions during measurement introduce dependencies in the noise structure. Similar to Noisier2Noise, a key step in our approach is the generation of noisier data from which the reconstruction net- work learns. However, unlike Noisier2Noise, the proposed loss function operates in measurement space and is trained to recover an extrapolated image instead of the original noisy one. This eliminates the need for an extrap- olation step during inference, which would otherwise suffer from ill-posedness. We numerically demonstrate that our method clearly outperforms previous self-supervised approaches that account for correlated noise.
Authors: Yaomin Shen, Xiaojian Lin, Wei Fan
Abstract: In the domain of multimodal intent recognition (MIR), the objective is to recognize human intent by integrating a variety of modalities, such as language text, body gestures, and tones. However, existing approaches face difficulties adequately capturing the intrinsic connections between the modalities and overlooking the corresponding semantic representations of intent. To address these limitations, we present the Anchor-based Mul- timodal Embedding with Semantic Synchronization (A-MESS) framework. We first design an Anchor-based Multimodal Embed- ding (A-ME) module that employs an anchor-based embedding fusion mechanism to integrate multimodal inputs. Furthermore, we develop a Semantic Synchronization (SS) strategy with the Triplet Contrastive Learning pipeline, which optimizes the pro- cess by synchronizing multimodal representation with label de- scriptions produced by the large language model. Comprehensive experiments indicate that our A-MESS achieves state-of-the-art and provides substantial insight into multimodal representation and downstream tasks.
Authors: Saverio Cavasin, Pietro Biasetton, Mattia Tamiazzo, Mauro Conti, Simone Milani
Abstract: During criminal investigations, images of persons of interest directly influence the success of identification procedures. However, law enforcement agencies often face challenges related to the scarcity of high-quality images or their obsolescence, which can affect the accuracy and success of people searching processes. This paper introduces a novel forensic mugshot augmentation framework aimed at addressing these limitations. Our approach enhances the identification probability of individuals by generating additional, high-quality images through customizable data augmentation techniques, while maintaining the biometric integrity and consistency of the original data. Several experimental results show that our method significantly improves identification accuracy and robustness across various forensic scenarios, demonstrating its effectiveness as a trustworthy tool law enforcement applications. Index Terms: Digital Forensics, Person re-identification, Feature extraction, Data augmentation, Visual-Language models.
Authors: Shijie Ma, Yuying Ge, Teng Wang, Yuxin Guo, Yixiao Ge, Ying Shan
Abstract: The synergy between generative and discriminative models receives growing attention. While discriminative Contrastive Language-Image Pre-Training (CLIP) excels in high-level semantics, it struggles with perceiving fine-grained visual details. Generally, to enhance representations, generative models take CLIP's visual features as conditions for reconstruction. However, the underlying principle remains underexplored. In this work, we empirically found that visually perfect generations are not always optimal for representation enhancement. The essence lies in effectively extracting fine-grained knowledge from generative models while mitigating irrelevant information. To explore critical factors, we delve into three aspects: (1) Conditioning mechanisms: We found that even a small number of local tokens can drastically reduce the difficulty of reconstruction, leading to collapsed training. We thus conclude that utilizing only global visual tokens as conditions is the most effective strategy. (2) Denoising configurations: We observed that end-to-end training introduces extraneous information. To address this, we propose a two-stage training strategy to prioritize learning useful visual knowledge. Additionally, we demonstrate that lightweight denoisers can yield remarkable improvements. (3) Generation paradigms: We explore both continuous and discrete denoisers with desirable outcomes, validating the versatility of our method. Through our in-depth explorations, we have finally arrived at an effective method, namely GenHancer, which consistently outperforms prior arts on the MMVP-VLM benchmark, e.g., 6.0% on OpenAICLIP. The enhanced CLIP can be further plugged into multimodal large language models for better vision-centric performance. All the models and codes are made publicly available.
Authors: Zhengwentai Sun, Heyuan Li, Xihe Yang, Keru Zheng, Shuliang Ning, Yihao Zhi, Hongjie Liao, Chenghong Li, Shuguang Cui, Xiaoguang Han
Abstract: Achieving fine-grained controllability in human image synthesis is a long-standing challenge in computer vision. Existing methods primarily focus on either facial synthesis or near-frontal body generation, with limited ability to simultaneously control key factors such as viewpoint, pose, clothing, and identity in a disentangled manner. In this paper, we introduce a new disentangled and controllable human synthesis task, which explicitly separates and manipulates these four factors within a unified framework. We first develop an end-to-end generative model trained on MVHumanNet for factor disentanglement. However, the domain gap between MVHumanNet and in-the-wild data produce unsatisfacotry results, motivating the exploration of virtual try-on (VTON) dataset as a potential solution. Through experiments, we observe that simply incorporating the VTON dataset as additional data to train the end-to-end model degrades performance, primarily due to the inconsistency in data forms between the two datasets, which disrupts the disentanglement process. To better leverage both datasets, we propose a stage-by-stage framework that decomposes human image generation into three sequential steps: clothed A-pose generation, back-view synthesis, and pose and view control. This structured pipeline enables better dataset utilization at different stages, significantly improving controllability and generalization, especially for in-the-wild scenarios. Extensive experiments demonstrate that our stage-by-stage approach outperforms end-to-end models in both visual fidelity and disentanglement quality, offering a scalable solution for real-world tasks. Additional demos are available on the project page: https://taited.github.io/discohuman-project/.
Authors: Vinayak Mali, Saurabh Jaiswal
Abstract: Falls among elderly residents in assisted living homes pose significant health risks, often leading to injuries and a decreased quality of life. Current fall detection solutions typically rely on sensor-based systems that require dedicated hardware, or on video-based models that demand high computational resources and GPUs for real-time processing. In contrast, this paper presents a robust fall detection system that does not require any additional sensors or high-powered hardware. The system uses pose estimation techniques, combined with threshold-based analysis and a voting mechanism, to effectively distinguish between fall and non-fall activities. For pose detection, we leverage MediaPipe, a lightweight and efficient framework that enables real-time processing on standard CPUs with minimal computational overhead. By analyzing motion, body position, and key pose points, the system processes pose features with a 20-frame buffer, minimizing false positives and maintaining high accuracy even in real-world settings. This unobtrusive, resource-efficient approach provides a practical solution for enhancing resident safety in old age homes, without the need for expensive sensors or high-end computational resources.
Authors: Juncen Guo, Xiaoguang Zhu, Liangyu Teng, Hao Yang, Jing Liu, Yang Liu, Liang Song
Abstract: Class-incremental Learning (CIL) enables the model to incrementally absorb knowledge from new classes and build a generic classifier across all previously encountered classes. When the model optimizes with new classes, the knowledge of previous classes is inevitably erased, leading to catastrophic forgetting. Addressing this challenge requires making a trade-off between retaining old knowledge and accommodating new information. However, this balancing process often requires sacrificing some information, which can lead to a partial loss in the model's ability to discriminate between classes. To tackle this issue, we design the adaptive weighted parameter fusion with Contrastive Language-Image Pre-training (CLIP), which not only takes into account the variability of the data distribution of different tasks, but also retains all the effective information of the parameter matrix to the greatest extent. In addition, we introduce a balance factor that can balance the data distribution alignment and distinguishability of adjacent tasks. Experimental results on several traditional benchmarks validate the superiority of the proposed method.
Authors: Kartik Jangra, Aman Kumar Singh, Yashwani Mann, Geetanjali Rathee
Abstract: Recent advancements in vision-language models have achieved remarkable results in making language models understand vision inputs. However, a unified approach to align these models across diverse tasks such as image captioning and visual question answering remains a challenge. Existing methods either require very big language models or very big datasets which is not efficient in utilizing existing models. This paper addresses this gap and devises a training strategy of auto-regressive vision-language models, to unify vision-language tasks like image-captioning and visual question answering. We propose four training stages for aligning the vision model with the language model, in other words, the language model is given an ability to process visual inputs. We also devise different attention masks for training transformer-based language models that improve the quality of visual features. Further, we introduce some findings, 1) the attention mask should not be applied on visual inputs, 2) the Language model converges faster on AI- generated data, 3) More work should be done in the alignment stage during the pre-training of the model, 4) the model can easily adapt to any downstream tasks like visual question answering on healthcare datasets like PathVQA. After training the model for one epoch for all the stages, it outperforms large models like VILA-13 billion models on common benchmarks like CIDEr scores on COCO and Flickr30k datasets and achieves very close scores to GIT-2 on the same dataset despite being a much smaller model trained on a much smaller dataset. All of the training is done using best practices available like multi- GPU parallel training, lower-precision training with 16-bit float numbers, faster attention (SDPA), and gradient accumulation, and completed the training within 12 hours.
Authors: Junwei Zheng, Ruiping Liu, Yufan Chen, Zhenfang Chen, Kailun Yang, Jiaming Zhang, Rainer Stiefelhagen
Abstract: Absolute Pose Regression (APR) predicts 6D camera poses but lacks the adaptability to unknown environments without retraining, while Relative Pose Regression (RPR) generalizes better yet requires a large image retrieval database. Visual Odometry (VO) generalizes well in unseen environments but suffers from accumulated error in open trajectories. To address this dilemma, we introduce a new task, Scene-agnostic Pose Regression (SPR), which can achieve accurate pose regression in a flexible way while eliminating the need for retraining or databases. To benchmark SPR, we created a large-scale dataset, 360SPR, with over 200K photorealistic panoramas, 3.6M pinhole images and camera poses in 270 scenes at three different sensor heights. Furthermore, a SPR-Mamba model is initially proposed to address SPR in a dual-branch manner. Extensive experiments and studies demonstrate the effectiveness of our SPR paradigm, dataset, and model. In the unknown scenes of both 360SPR and 360Loc datasets, our method consistently outperforms APR, RPR and VO. The dataset and code are available at https://junweizheng93.github.io/publications/SPR/SPR.html.
URLs: https://junweizheng93.github.io/publications/SPR/SPR.html.
Authors: Elena Buglakova, Anwai Archit, Edoardo D'Imprima, Julia Mahamid, Constantin Pape, Anna Kreshuk
Abstract: Segmentation of very large images is a common problem in microscopy, medical imaging or remote sensing. The problem is usually addressed by sliding window inference, which can theoretically lead to seamlessly stitched predictions. However, in practice many of the popular pipelines still suffer from tiling artifacts. We investigate the root cause of these issues and show that they stem from the normalization layers within the neural networks. We propose indicators to detect normalization issues and further explore the trade-offs between artifact-free and high-quality predictions, using three diverse microscopy datasets as examples. Finally, we propose to use BatchRenorm as the most suitable normalization strategy, which effectively removes tiling artifacts and enhances transfer performance, thereby improving the reusability of trained networks for new datasets.
Authors: Jan Koh\'ut, Michal Hradi\v{s}
Abstract: A common use case for OCR applications involves users uploading documents and progressively correcting automatic recognition to obtain the final transcript. This correction phase presents an opportunity for progressive adaptation of the OCR model, making it crucial to adapt early, while ensuring stability and reliability. We demonstrate that state-of-the-art transformer-based models can effectively support this adaptation, gradually reducing the annotator's workload. Our results show that fine-tuning can reliably start with just 16 lines, yielding a 10% relative improvement in CER, and scale up to 40% with 256 lines. We further investigate the impact of model components, clarifying the roles of the encoder and decoder in the fine-tuning process. To guide adaptation, we propose reliable stopping criteria, considering both direct approaches and global trend analysis. Additionally, we show that OCR models can be leveraged to cut annotation costs by half through confidence-based selection of informative lines, achieving the same performance with fewer annotations.
Authors: Haim Sawdayee, Chuan Guo, Guy Tevet, Bing Zhou, Jian Wang, Amit H. Bermano
Abstract: Text-to-motion generative models span a wide range of 3D human actions but struggle with nuanced stylistic attributes such as a "Chicken" style. Due to the scarcity of style-specific data, existing approaches pull the generative prior towards a reference style, which often results in out-of-distribution low quality generations. In this work, we introduce LoRA-MDM, a lightweight framework for motion stylization that generalizes to complex actions while maintaining editability. Our key insight is that adapting the generative prior to include the style, while preserving its overall distribution, is more effective than modifying each individual motion during generation. Building on this idea, LoRA-MDM learns to adapt the prior to include the reference style using only a few samples. The style can then be used in the context of different textual prompts for generation. The low-rank adaptation shifts the motion manifold in a semantically meaningful way, enabling realistic style infusion even for actions not present in the reference samples. Moreover, preserving the distribution structure enables advanced operations such as style blending and motion editing. We compare LoRA-MDM to state-of-the-art stylized motion generation methods and demonstrate a favorable balance between text fidelity and style consistency.
Authors: Olgica Zaric, Carmen Leser, Vladimir Juras, Alex Farr, Malina Gologan, Stanislas Rapacchi, Laura Villazan Garcia, Christian Singer, Siegfried Trattnig, Christian Licht, Ramona Woitek
Abstract: Introduction: In sodium (23Na) MRI, partial volume effects (PVE) are one of the most common causes of errors in the quantification of tissue sodium concentration (TSC) in vivo. Advanced image reconstruction algorithms, such as compressed sensing (CS), have been shown to potentially reduce PVE. Therefore, we investigated the feasibility of CS-based methods for image quality and TSC quantification accuracy improvement in patients with breast cancer (BC). Subjects and Methods: Three healthy participants and 12 female participants with BC were examined on a 7T MRI scanner in this study. We reconstructed 23Na-MRI images using the weighted total variation (wTV) and directional total variation (dTV), anatomically guided total variation (AG-TV), and adaptive combine (ADC) reconstruction and performed image quality assessment. We evaluated agreement in tumor volumes delineated on sodium data using the Dice score and performed TSC quantification for different image reconstruction approaches. Results: All methods provided sodium images of the breast with good quality. The mean Dice scores for wTV, dTV, and AG-TV were 65%, 72%, and 75%, respectively. In the breast tumors, average TSC values were 83.0, 72.0, 80.0, and 84.0 mmol/L, respectively. There was a significant difference between dTV and wTV (p<0.001), as well as between dTV and AG-TV (p<0.001) and dTV and ADC algorithm (p<0.001). Conclusion: The results of this study showed that there are differences in tumor appearance and TSC estimations that might be depending on the type of image reconstruction and parameters used, most likely due to differences in their robustness in reducing PVE.
Authors: Mia Siemon, Ivan Nikolov, Thomas B. Moeslund, Kamal Nasrollahi
Abstract: In Pose-based Video Anomaly Detection prior art is rooted on the assumption that abnormal events can be mostly regarded as a result of uncommon human behavior. Opposed to utilizing skeleton representations of humans, however, we investigate the potential of learning recurrent motion patterns of normal human behavior using 2D contours. Keeping all advantages of pose-based methods, such as increased object anonymization, the shift from human skeletons to contours is hypothesized to leave the opportunity to cover more object categories open for future research. We propose formulating the problem as a regression and a classification task, and additionally explore two distinct data representation techniques for contours. To further reduce the computational complexity of Pose-based Video Anomaly Detection solutions, all methods in this study are based on shallow Neural Networks from the field of Deep Learning, and evaluated on the three most prominent benchmark datasets within Video Anomaly Detection and their human-related counterparts, totaling six datasets. Our results indicate that this novel perspective on Pose-based Video Anomaly Detection marks a promising direction for future research.
Authors: Xinxing Cheng, Tianyang Zhang, Wenqi Lu, Qingjie Meng, Alejandro F. Frangi, Jinming Duan
Abstract: Deep learning-based image registration methods have shown state-of-the-art performance and rapid inference speeds. Despite these advances, many existing approaches fall short in capturing spatially varying information in non-local regions of feature maps due to the reliance on spatially-shared convolution kernels. This limitation leads to suboptimal estimation of deformation fields. In this paper, we propose a 3D Spatial-Awareness Convolution Block (SACB) to enhance the spatial information within feature representations. Our SACB estimates the spatial clusters within feature maps by leveraging feature similarity and subsequently parameterizes the adaptive convolution kernels across diverse regions. This adaptive mechanism generates the convolution kernels (weights and biases) tailored to spatial variations, thereby enabling the network to effectively capture spatially varying information. Building on SACB, we introduce a pyramid flow estimator (named SACB-Net) that integrates SACBs to facilitate multi-scale flow composition, particularly addressing large deformations. Experimental results on the brain IXI and LPBA datasets as well as Abdomen CT datasets demonstrate the effectiveness of SACB and the superiority of SACB-Net over the state-of-the-art learning-based registration methods. The code is available at https://github.com/x-xc/SACB_Net .
Authors: Hongcheng Gao, Jiashu Qu, Jingyi Tang, Baolong Bi, Yue Liu, Hongyu Chen, Li Liang, Li Su, Qingming Huang
Abstract: The hallucination of large multimodal models (LMMs), providing responses that appear correct but are actually incorrect, limits their reliability and applicability. This paper aims to study the hallucination problem of LMMs in video modality, which is dynamic and more challenging compared to static modalities like images and text. From this motivation, we first present a comprehensive benchmark termed HAVEN for evaluating hallucinations of LMMs in video understanding tasks. It is built upon three dimensions, i.e., hallucination causes, hallucination aspects, and question formats, resulting in 6K questions. Then, we quantitatively study 7 influential factors on hallucinations, e.g., duration time of videos, model sizes, and model reasoning, via experiments of 16 LMMs on the presented benchmark. In addition, inspired by recent thinking models like OpenAI o1, we propose a video-thinking model to mitigate the hallucinations of LMMs via supervised reasoning fine-tuning (SRFT) and direct preference optimization (TDPO)-- where SRFT enhances reasoning capabilities while TDPO reduces hallucinations in the thinking process. Extensive experiments and analyses demonstrate the effectiveness. Remarkably, it improves the baseline by 7.65% in accuracy on hallucination evaluation and reduces the bias score by 4.5%. The code and data are public at https://github.com/Hongcheng-Gao/HAVEN.
Authors: Xiangting Meng, Jiaqi Yang, Mingshu Chen, Chenxin Yan, Yujiao Shi, Wenchao Ding, Laurent Kneip
Abstract: In the realm of object pose estimation, scenarios involving both dynamic objects and moving cameras are prevalent. However, the scarcity of corresponding real-world datasets significantly hinders the development and evaluation of robust pose estimation models. This is largely attributed to the inherent challenges in accurately annotating object poses in dynamic scenes captured by moving cameras. To bridge this gap, this paper presents a novel dataset DynOPETs and a dedicated data acquisition and annotation pipeline tailored for object pose estimation and tracking in such unconstrained environments. Our efficient annotation method innovatively integrates pose estimation and pose tracking techniques to generate pseudo-labels, which are subsequently refined through pose graph optimization. The resulting dataset offers accurate pose annotations for dynamic objects observed from moving cameras. To validate the effectiveness and value of our dataset, we perform comprehensive evaluations using 18 state-of-the-art methods, demonstrating its potential to accelerate research in this challenging domain. The dataset will be made publicly available to facilitate further exploration and advancement in the field.
Authors: Ozan Unal, Steven Marty, Dengxin Dai
Abstract: Burst image super-resolution (BISR) aims to enhance the resolution of a keyframe by leveraging information from multiple low-resolution images captured in quick succession. In the deep learning era, BISR methods have evolved from fully convolutional networks to transformer-based architectures, which, despite their effectiveness, suffer from the quadratic complexity of self-attention. We see Mamba as the next natural step in the evolution of this field, offering a comparable global receptive field and selective information routing with only linear time complexity. In this work, we introduce BurstMamba, a Mamba-based architecture for BISR. Our approach decouples the task into two specialized branches: a spatial module for keyframe super-resolution and a temporal module for subpixel prior extraction, striking a balance between computational efficiency and burst information integration. To further enhance burst processing with Mamba, we propose two novel strategies: (i) optical flow-based serialization, which aligns burst sequences only during state updates to preserve subpixel details, and (ii) a wavelet-based reparameterization of the state-space update rules, prioritizing high-frequency features for improved burst-to-keyframe information passing. Our framework achieves SOTA performance on public benchmarks of SyntheticSR, RealBSR-RGB, and RealBSR-RAW.
Authors: Niccolo Avogaro, Thomas Frick, Mattia Rigotti, Andrea Bartezzaghi, Filip Janicki, Cristiano Malossi, Konrad Schindler, Roy Assaf
Abstract: Large Vision-Language Models (VLMs) are increasingly being regarded as foundation models that can be instructed to solve diverse tasks by prompting, without task-specific training. We examine the seemingly obvious question: how to effectively prompt VLMs for semantic segmentation. To that end, we systematically evaluate the segmentation performance of several recent models guided by either text or visual prompts on the out-of-distribution MESS dataset collection. We introduce a scalable prompting scheme, few-shot prompted semantic segmentation, inspired by open-vocabulary segmentation and few-shot learning. It turns out that VLMs lag far behind specialist models trained for a specific segmentation task, by about 30% on average on the Intersection-over-Union metric. Moreover, we find that text prompts and visual prompts are complementary: each one of the two modes fails on many examples that the other one can solve. Our analysis suggests that being able to anticipate the most effective prompt modality can lead to a 11% improvement in performance. Motivated by our findings, we propose PromptMatcher, a remarkably simple training-free baseline that combines both text and visual prompts, achieving state-of-the-art results outperforming the best text-prompted VLM by 2.5%, and the top visual-prompted VLM by 3.5% on few-shot prompted semantic segmentation.
Authors: Yabin Wang, Zhiwu Huang, Xiaopeng Hong
Abstract: This paper identifies OpenSDI, a challenge for spotting diffusion-generated images in open-world settings. In response to this challenge, we define a new benchmark, the OpenSDI dataset (OpenSDID), which stands out from existing datasets due to its diverse use of large vision-language models that simulate open-world diffusion-based manipulations. Another outstanding feature of OpenSDID is its inclusion of both detection and localization tasks for images manipulated globally and locally by diffusion models. To address the OpenSDI challenge, we propose a Synergizing Pretrained Models (SPM) scheme to build up a mixture of foundation models. This approach exploits a collaboration mechanism with multiple pretrained foundation models to enhance generalization in the OpenSDI context, moving beyond traditional training by synergizing multiple pretrained models through prompting and attending strategies. Building on this scheme, we introduce MaskCLIP, an SPM-based model that aligns Contrastive Language-Image Pre-Training (CLIP) with Masked Autoencoder (MAE). Extensive evaluations on OpenSDID show that MaskCLIP significantly outperforms current state-of-the-art methods for the OpenSDI challenge, achieving remarkable relative improvements of 14.23% in IoU (14.11% in F1) and 2.05% in accuracy (2.38% in F1) compared to the second-best model in localization and detection tasks, respectively. Our dataset and code are available at https://github.com/iamwangyabin/OpenSDI.
Authors: Mehdi Moshtaghi, Siavash H. Khajavi, Joni Pajarinen
Abstract: We introduce RGB-Th-Bench, the first benchmark designed to evaluate the ability of Vision-Language Models (VLMs) to comprehend RGB-Thermal image pairs. While VLMs have demonstrated remarkable progress in visual reasoning and multimodal understanding, their evaluation has been predominantly limited to RGB-based benchmarks, leaving a critical gap in assessing their capabilities in infrared vision tasks. Existing visible-infrared datasets are either task-specific or lack high-quality annotations necessary for rigorous model evaluation. To address these limitations, RGB-Th-Bench provides a comprehensive evaluation framework covering 14 distinct skill dimensions, with a total of 1,600+ expert-annotated Yes/No questions. The benchmark employs two accuracy metrics: a standard question-level accuracy and a stricter skill-level accuracy, which evaluates model robustness across multiple questions within each skill dimension. This design ensures a thorough assessment of model performance, including resilience to adversarial and hallucinated responses. We conduct extensive evaluations on 19 state-of-the-art VLMs, revealing significant performance gaps in RGB-Thermal understanding. Our results show that even the strongest models struggle with thermal image comprehension, with performance heavily constrained by their RGB-based capabilities. Additionally, the lack of large-scale application-specific and expert-annotated thermal-caption-pair datasets in pre-training is an important reason of the observed performance gap. RGB-Th-Bench highlights the urgent need for further advancements in multimodal learning to bridge the gap between visible and thermal image understanding. The dataset is available through this link, and the evaluation code will also be made publicly available.
Authors: Jan Koh\'ut, Martin Do\v{c}ekal, Michal Hradi\v{s}, Marek Va\v{s}ko
Abstract: Manual digitization of bibliographic metadata is time consuming and labor intensive, especially for historical and real-world archives with highly variable formatting across documents. Despite advances in machine learning, the absence of dedicated datasets for metadata extraction hinders automation. To address this gap, we introduce BiblioPage, a dataset of scanned title pages annotated with structured bibliographic metadata. The dataset consists of approximately 2,000 monograph title pages collected from 14 Czech libraries, spanning a wide range of publication periods, typographic styles, and layout structures. Each title page is annotated with 16 bibliographic attributes, including title, contributors, and publication metadata, along with precise positional information in the form of bounding boxes. To extract structured information from this dataset, we valuated object detection models such as YOLO and DETR combined with transformer-based OCR, achieving a maximum mAP of 52 and an F1 score of 59. Additionally, we assess the performance of various visual large language models, including LlamA 3.2-Vision and GPT-4o, with the best model reaching an F1 score of 67. BiblioPage serves as a real-world benchmark for bibliographic metadata extraction, contributing to document understanding, document question answering, and document information extraction. Dataset and evaluation scripts are availible at: https://github.com/DCGM/biblio-dataset
Authors: Rupak Bose, Chinedu Innocent Nwoye, Aditya Bhat, Nicolas Padoy
Abstract: The acquisition of annotated datasets with paired images and segmentation masks is a critical challenge in domains such as medical imaging, remote sensing, and computer vision. Manual annotation demands significant resources, faces ethical constraints, and depends heavily on domain expertise. Existing generative models often target single-modality outputs, either images or segmentation masks, failing to address the need for high-quality, simultaneous image-mask generation. Additionally, these models frequently lack adaptable conditioning mechanisms, restricting control over the generated outputs and limiting their applicability for dataset augmentation and rare scenario simulation. We propose CoSimGen, a diffusion-based framework for controllable simultaneous image and mask generation. Conditioning is intuitively achieved through (1) text prompts grounded in class semantics, (2) spatial embedding of context prompts to provide spatial coherence, and (3) spectral embedding of timestep information to model noise levels during diffusion. To enhance controllability and training efficiency, the framework incorporates contrastive triplet loss between text and class embeddings, alongside diffusion and adversarial losses. Initial low-resolution outputs 128 x 128 are super-resolved to 512 x 512, producing high-fidelity images and masks with strict adherence to conditions. We evaluate CoSimGen on metrics such as FID, KID, LPIPS, Class FID, Positive predicted value for image fidelity and semantic alignment of generated samples over 4 diverse datasets. CoSimGen achieves state-of-the-art performance across all datasets, achieving the lowest KID of 0.11 and LPIPS of 0.53 across datasets.
Authors: Saurav Sharma, Didier Mutter, Nicolas Padoy
Abstract: While vision-language models like CLIP have advanced zero-shot surgical phase recognition, they struggle with fine-grained surgical activities, especially action triplets. This limitation arises because current CLIP formulations rely on global image features, which overlook the fine-grained semantics and contextual details crucial for complex tasks like zero-shot triplet recognition. Furthermore, these models do not explore the hierarchical structure inherent in triplets, reducing their ability to generalize to novel triplets. To address these challenges, we propose fine-CLIP, which learns object-centric features and lever- ages the hierarchy in triplet formulation. Our approach integrates three components: hierarchical prompt modeling to capture shared semantics, LoRA-based vision backbone adaptation for enhanced feature extraction, and a graph-based condensation strategy that groups similar patch features into meaningful object clusters. Since triplet classification is a challenging task, we introduce an alternative yet meaningful base-to-novel generalization benchmark with two settings on the CholecT50 dataset: Unseen-Target, assessing adaptability to triplets with novel anatomical structures, and Unseen-Instrument-Verb, where models need to generalize to novel instrument-verb interactions. fine-CLIP shows significant improvements in F1 and mAP, enhancing zero-shot recognition of novel surgical triplets.
Authors: Andrii Yermakov, Jan Cech, Jiri Matas
Abstract: This paper tackles the challenge of detecting partially manipulated facial deepfakes, which involve subtle alterations to specific facial features while retaining the overall context, posing a greater detection difficulty than fully synthetic faces. We leverage the Contrastive Language-Image Pre-training (CLIP) model, specifically its ViT-L/14 visual encoder, to develop a generalizable detection method that performs robustly across diverse datasets and unknown forgery techniques with minimal modifications to the original model. The proposed approach utilizes parameter-efficient fine-tuning (PEFT) techniques, such as LN-tuning, to adjust a small subset of the model's parameters, preserving CLIP's pre-trained knowledge and reducing overfitting. A tailored preprocessing pipeline optimizes the method for facial images, while regularization strategies, including L2 normalization and metric learning on a hyperspherical manifold, enhance generalization. Trained on the FaceForensics++ dataset and evaluated in a cross-dataset fashion on Celeb-DF-v2, DFDC, FFIW, and others, the proposed method achieves competitive detection accuracy comparable to or outperforming much more complex state-of-the-art techniques. This work highlights the efficacy of CLIP's visual encoder in facial deepfake detection and establishes a simple, powerful baseline for future research, advancing the field of generalizable deepfake detection. The code is available at: https://github.com/yermandy/deepfake-detection
Authors: Boyi Li, Ye Yuan, Wenjun Tan
Abstract: The MedSAM model, built upon the SAM framework, enhances medical image segmentation through generalizable training but still exhibits notable limitations. First, constraints in the perturbation window settings during training can cause MedSAM to incorrectly segment small tissues or organs together with adjacent structures, leading to segmentation errors. Second, when dealing with medical image targets characterized by irregular shapes and complex structures, segmentation often relies on narrowing the bounding box to refine segmentation intent. However, MedSAM's performance under reduced bounding box prompts remains suboptimal. To address these challenges, this study proposes a bounding box adaptive perturbation algorithm to optimize the training process. The proposed approach aims to reduce segmentation errors for small targets and enhance the model's accuracy when processing reduced bounding box prompts, ultimately improving the robustness and reliability of the MedSAM model for complex medical imaging tasks.
Authors: Qian Wang, Zhihao Zhan, Jialei He, Zhituo Tu, Xiang Zhu, Jie Yuan
Abstract: Highly accurate geometric precision and dense image features characterize True Digital Orthophoto Maps (TDOMs), which are in great demand for applications such as urban planning, infrastructure management, and environmental monitoring. Traditional TDOM generation methods need sophisticated processes, such as Digital Surface Models (DSM) and occlusion detection, which are computationally expensive and prone to errors. This work presents an alternative technique rooted in 2D Gaussian Splatting (2DGS), free of explicit DSM and occlusion detection. With depth map generation, spatial information for every pixel within the TDOM is retrieved and can reconstruct the scene with high precision. Divide-and-conquer strategy achieves excellent GS training and rendering with high-resolution TDOMs at a lower resource cost, which preserves higher quality of rendering on complex terrain and thin structure without a decrease in efficiency. Experimental results demonstrate the efficiency of large-scale scene reconstruction and high-precision terrain modeling. This approach provides accurate spatial data, which assists users in better planning and decision-making based on maps.
Authors: Jungin Park, Jiyoung Lee, Kwanghoon Sohn
Abstract: View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov.
Authors: Ilias Stogiannidis, Steven McDonagh, Sotirios A. Tsaftaris
Abstract: Vision-Language Models (VLMs) have recently emerged as powerful tools, excelling in tasks that integrate visual and textual comprehension, such as image captioning, visual question answering, and image-text retrieval. However, existing benchmarks for VLMs include spatial components, which often fail to isolate spatial reasoning from related tasks such as object detection or semantic comprehension. In this paper, we address these deficiencies with a multi-faceted approach towards understanding spatial reasoning. Informed by the diverse and multi-dimensional nature of human spatial reasoning abilities, we present a detailed analysis that first delineates the core elements of spatial reasoning: spatial relations, orientation and navigation, mental rotation, and spatial visualization, and then assesses the performance of these models in both synthetic and real-world images, bridging controlled and naturalistic contexts. We analyze 13 state-of-the-art Vision-Language Models, uncovering pivotal insights into their spatial reasoning performance. Our results reveal profound shortcomings in current VLMs, with average accuracy across the 13 models approximating random chance, highlighting spatial reasoning as a persistent obstacle. This work not only exposes the pressing need to advance spatial reasoning within VLMs but also establishes a solid platform for future exploration. Code available on GitHub (https://github.com/stogiannidis/srbench) and dataset available on HuggingFace (https://huggingface.co/datasets/stogiannidis/srbench).
URLs: https://github.com/stogiannidis/srbench), https://huggingface.co/datasets/stogiannidis/srbench).
Authors: Chengjie Ge, Xueyang Fu, Peng He, Kunyu Wang, Chengzhi Cao, Zheng-Jun Zha
Abstract: Leveraging its robust linear global modeling capability, Mamba has notably excelled in computer vision. Despite its success, existing Mamba-based vision models have overlooked the nuances of event-driven tasks, especially in video reconstruction. Event-based video reconstruction (EBVR) demands spatial translation invariance and close attention to local event relationships in the spatio-temporal domain. Unfortunately, conventional Mamba algorithms apply static window partitions and standard reshape scanning methods, leading to significant losses in local connectivity. To overcome these limitations, we introduce EventMamba--a specialized model designed for EBVR tasks. EventMamba innovates by incorporating random window offset (RWO) in the spatial domain, moving away from the restrictive fixed partitioning. Additionally, it features a new consistent traversal serialization approach in the spatio-temporal domain, which maintains the proximity of adjacent events both spatially and temporally. These enhancements enable EventMamba to retain Mamba's robust modeling capabilities while significantly preserving the spatio-temporal locality of event data. Comprehensive testing on multiple datasets shows that EventMamba markedly enhances video reconstruction, drastically improving computation speed while delivering superior visual quality compared to Transformer-based methods.
Authors: Yuli Zhou, Guolei Sun, Yawei Li, Yuqian Fu, Luca Benini, Ender Konukoglu
Abstract: Video camouflaged object segmentation (VCOS), aiming at segmenting camouflaged objects that seamlessly blend into their environment, is a fundamental vision task with various real-world applications. With the release of SAM2, video segmentation has witnessed significant progress. However, SAM2's capability of segmenting camouflaged videos is suboptimal, especially when given simple prompts such as point and box. To address the problem, we propose Camouflaged SAM2 (CamSAM2), which enhances SAM2's ability to handle camouflaged scenes without modifying SAM2's parameters. Specifically, we introduce a decamouflaged token to provide the flexibility of feature adjustment for VCOS. To make full use of fine-grained and high-resolution features from the current frame and previous frames, we propose implicit object-aware fusion (IOF) and explicit object-aware fusion (EOF) modules, respectively. Object prototype generation (OPG) is introduced to abstract and memorize object prototypes with informative details using high-quality features from previous frames. Extensive experiments are conducted to validate the effectiveness of our approach. While CamSAM2 only adds negligible learnable parameters to SAM2, it substantially outperforms SAM2 on three VCOS datasets, especially achieving 12.2 mDice gains with click prompt on MoCA-Mask and 19.6 mDice gains with mask prompt on SUN-SEG-Hard, with Hiera-T as the backbone. The code will be available at \href{https://github.com/zhoustan/CamSAM2}{github.com/zhoustan/CamSAM2}.
Authors: Junhyuk So, Jiwoong Shin, Chaeyeon Jang, Eunhyeok Park
Abstract: Recently, diffusion models have achieved significant advances in vision, text, and robotics. However, they still face slow generation speeds due to sequential denoising processes. To address this, a parallel sampling method based on Picard iteration was introduced, effectively reducing sequential steps while ensuring exact convergence to the original output. Nonetheless, Picard iteration does not guarantee faster convergence, which can still result in slow generation in practice. In this work, we propose a new parallelization scheme, the Picard Consistency Model (PCM), which significantly reduces the number of generation steps in Picard iteration. Inspired by the consistency model, PCM is directly trained to predict the fixed-point solution, or the final output, at any stage of the convergence trajectory. Additionally, we introduce a new concept called model switching, which addresses PCM's limitations and ensures exact convergence. Extensive experiments demonstrate that PCM achieves up to a 2.71x speedup over sequential sampling and a 1.77x speedup over Picard iteration across various tasks, including image generation and robotic control.
Authors: Pihai Sun (Faculty of Computing, Harbin Institute of Technology), Junjun Jiang (Faculty of Computing, Harbin Institute of Technology), Yuanqi Yao (Faculty of Computing, Harbin Institute of Technology), Youyu Chen (Faculty of Computing, Harbin Institute of Technology), Wenbo Zhao (Faculty of Computing, Harbin Institute of Technology), Kui Jiang (Faculty of Computing, Harbin Institute of Technology), Xianming Liu (Faculty of Computing, Harbin Institute of Technology)
Abstract: Image-event joint depth estimation methods leverage complementary modalities for robust perception, yet face challenges in generalizability stemming from two factors: 1) limited annotated image-event-depth datasets causing insufficient cross-modal supervision, and 2) inherent frequency mismatches between static images and dynamic event streams with distinct spatiotemporal patterns, leading to ineffective feature fusion. To address this dual challenge, we propose Frequency-decoupled Unified Self-supervised Encoder (FUSE) with two synergistic components: The Parameter-efficient Self-supervised Transfer (PST) establishes cross-modal knowledge transfer through latent space alignment with image foundation models, effectively mitigating data scarcity by enabling joint encoding without depth ground truth.Complementing this, we propose the Frequency-Decoupled Fusion module (FreDFuse) to explicitly decouple high-frequency edge features from low-frequency structural components, resolving modality-specific frequency mismatches through physics-aware fusion. This combined approach enables FUSE to construct a universal image-event encoder that only requires lightweight decoder adaptation for target datasets. Extensive experiments demonstrate state-of-the-art performance with 14% and 24.9% improvements in Abs.Rel on MVSEC and DENSE datasets. The framework exhibits remarkable zero-shot adaptability to challenging scenarios including extreme lighting and motion blur, significantly advancing real-world deployment capabilities. The source code for our method is publicly available at: https://github.com/sunpihai-up/FUSE
Authors: Chengan Che, Chao Wang, Tom Vercauteren, Sophia Tsoka, Luis C. Garcia-Peraza-Herrera
Abstract: Advancements in computer-assisted surgical procedures heavily rely on accurate visual data interpretation from camera systems used during surgeries. Traditional open-access datasets focusing on surgical procedures are often limited by their small size, typically consisting of fewer than 100 videos with less than 100K images. To address these constraints, a new dataset called Surg-3M has been compiled using a novel aggregation pipeline that collects high-resolution videos from online sources. Featuring an extensive collection of over 4K surgical videos and more than 3 million high-quality images from multiple procedure types, Surg-3M offers a comprehensive resource surpassing existing alternatives in size and scope, including two novel tasks. To demonstrate the effectiveness of this dataset, we present SurgFM, a self-supervised foundation model pretrained on Surg-3M that achieves impressive results in downstream tasks such as surgical phase recognition, action recognition, and tool presence detection. Combining key components from ConvNeXt, DINO, and an innovative augmented distillation method, SurgFM exhibits exceptional performance compared to specialist architectures across various benchmarks. Our experimental results show that SurgFM outperforms state-of-the-art models in multiple downstream tasks, including significant gains in surgical phase recognition (+8.9pp, +4.7pp, and +3.9pp of Jaccard in AutoLaparo, M2CAI16, and Cholec80), action recognition (+3.1pp of mAP in CholecT50) and tool presence detection (+4.6pp of mAP in Cholec80). Moreover, even when using only half of the data, SurgFM outperforms state-of-the-art models in AutoLaparo and achieves state-of-the-art performance in Cholec80. Both Surg-3M and SurgFM have significant potential to accelerate progress towards developing autonomous robotic surgery systems.
Authors: Haoyu Fu, Diankun Zhang, Zongchuang Zhao, Jianfeng Cui, Dingkang Liang, Chong Zhang, Dingyuan Zhang, Hongwei Xie, Bing Wang, Xiang Bai
Abstract: End-to-end (E2E) autonomous driving methods still struggle to make correct decisions in interactive closed-loop evaluation due to limited causal reasoning capability. Current methods attempt to leverage the powerful understanding and reasoning abilities of Vision-Language Models (VLMs) to resolve this dilemma. However, the problem is still open that few VLMs for E2E methods perform well in the closed-loop evaluation due to the gap between the semantic reasoning space and the purely numerical trajectory output in the action space. To tackle this issue, we propose ORION, a holistic E2E autonomous driving framework by vision-language instructed action generation. ORION uniquely combines a QT-Former to aggregate long-term history context, a Large Language Model (LLM) for driving scenario reasoning, and a generative planner for precision trajectory prediction. ORION further aligns the reasoning space and the action space to implement a unified E2E optimization for both visual question-answering (VQA) and planning tasks. Our method achieves an impressive closed-loop performance of 77.74 Driving Score (DS) and 54.62% Success Rate (SR) on the challenge Bench2Drive datasets, which outperforms state-of-the-art (SOTA) methods by a large margin of 14.28 DS and 19.61% SR.
Authors: Christina Kassab, Sacha Morin, Martin B\"uchner, Mat\'ias Mattamala, Kumaraditya Gupta, Abhinav Valada, Liam Paull, Maurice Fallon
Abstract: 3D scene understanding has been transformed by open-vocabulary language models that enable interaction via natural language. However, the evaluation of these representations is limited to closed-set semantics that do not capture the richness of language. This work presents OpenLex3D, a dedicated benchmark to evaluate 3D open-vocabulary scene representations. OpenLex3D provides entirely new label annotations for 23 scenes from Replica, ScanNet++, and HM3D, which capture real-world linguistic variability by introducing synonymical object categories and additional nuanced descriptions. By introducing an open-set 3D semantic segmentation task and an object retrieval task, we provide insights on feature precision, segmentation, and downstream capabilities. We evaluate various existing 3D open-vocabulary methods on OpenLex3D, showcasing failure cases, and avenues for improvement. The benchmark is publicly available at: https://openlex3d.github.io/.
Authors: Suzhe Xu, Jialin Peng, Chengyuan Zhang
Abstract: Segmentation is a fundamental task in computer vision, with prompt-driven methods gaining prominence due to their flexibility. The recent Segment Anything Model (SAM) has demonstrated powerful point-prompt segmentation capabilities, while text-based segmentation models offer rich semantic understanding. However, existing approaches rarely explore how to effectively combine these complementary modalities for optimal segmentation performance. This paper presents BiPrompt-SAM, a novel dual-modal prompt segmentation framework that fuses the advantages of point and text prompts through an explicit selection mechanism. Specifically, we leverage SAM's inherent ability to generate multiple mask candidates, combined with a semantic guidance mask from text prompts, and explicitly select the most suitable candidate based on similarity metrics. This approach can be viewed as a simplified Mixture of Experts (MoE) system, where the point and text modules act as distinct "experts," and the similarity scoring serves as a rudimentary "gating network." We conducted extensive evaluations on both the Endovis17 medical dataset and RefCOCO series natural image datasets. On Endovis17, BiPrompt-SAM achieved 89.55\% mDice and 81.46\% mIoU, comparable to state-of-the-art specialized medical segmentation models. On the RefCOCO series datasets, our method attained 87.1\%, 86.5\%, and 85.8\% IoU, significantly outperforming existing approaches. Experiments demonstrate that our explicit dual-selection method effectively combines the spatial precision of point prompts with the semantic richness of text prompts, particularly excelling in scenarios involving semantically complex objects, multiple similar objects, and partial occlusions. BiPrompt-SAM not only provides a simple yet effective implementation but also offers a new perspective on multi-modal prompt fusion.
Authors: Konyul Park, Yecheol Kim, Daehun Kim, Jun Won Choi
Abstract: Modern autonomous driving perception systems utilize complementary multi-modal sensors, such as LiDAR and cameras. Although sensor fusion architectures enhance performance in challenging environments, they still suffer significant performance drops under severe sensor failures, such as LiDAR beam reduction, LiDAR drop, limited field of view, camera drop, and occlusion. This limitation stems from inter-modality dependencies in current sensor fusion frameworks. In this study, we introduce an efficient and robust LiDAR-camera 3D object detector, referred to as MoME, which can achieve robust performance through a mixture of experts approach. Our MoME fully decouples modality dependencies using three parallel expert decoders, which use camera features, LiDAR features, or a combination of both to decode object queries, respectively. We propose Multi-Expert Decoding (MED) framework, where each query is decoded selectively using one of three expert decoders. MoME utilizes an Adaptive Query Router (AQR) to select the most appropriate expert decoder for each query based on the quality of camera and LiDAR features. This ensures that each query is processed by the best-suited expert, resulting in robust performance across diverse sensor failure scenarios. We evaluated the performance of MoME on the nuScenes-R benchmark. Our MoME achieved state-of-the-art performance in extreme weather and sensor failure conditions, significantly outperforming the existing models across various sensor failure scenarios.
Authors: Vladan Stojni\'c, Yannis Kalantidis, Ji\v{r}\'i Matas, Giorgos Tolias
Abstract: We propose a training-free method for open-vocabulary semantic segmentation using Vision-and-Language Models (VLMs). Our approach enhances the initial per-patch predictions of VLMs through label propagation, which jointly optimizes predictions by incorporating patch-to-patch relationships. Since VLMs are primarily optimized for cross-modal alignment and not for intra-modal similarity, we use a Vision Model (VM) that is observed to better capture these relationships. We address resolution limitations inherent to patch-based encoders by applying label propagation at the pixel level as a refinement step, significantly improving segmentation accuracy near class boundaries. Our method, called LPOSS+, performs inference over the entire image, avoiding window-based processing and thereby capturing contextual interactions across the full image. LPOSS+ achieves state-of-the-art performance among training-free methods, across a diverse set of datasets. Code: https://github.com/vladan-stojnic/LPOSS
Authors: Kartik Thakral, Tamar Glaser, Tal Hassner, Mayank Vatsa, Richa Singh
Abstract: Existing unlearning algorithms in text-to-image generative models often fail to preserve the knowledge of semantically related concepts when removing specific target concepts: a challenge known as adjacency. To address this, we propose FADE (Fine grained Attenuation for Diffusion Erasure), introducing adjacency aware unlearning in diffusion models. FADE comprises two components: (1) the Concept Neighborhood, which identifies an adjacency set of related concepts, and (2) Mesh Modules, employing a structured combination of Expungement, Adjacency, and Guidance loss components. These enable precise erasure of target concepts while preserving fidelity across related and unrelated concepts. Evaluated on datasets like Stanford Dogs, Oxford Flowers, CUB, I2P, Imagenette, and ImageNet1k, FADE effectively removes target concepts with minimal impact on correlated concepts, achieving atleast a 12% improvement in retention performance over state-of-the-art methods.
Authors: Jingdan Kang, Haoxin Yang, Yan Cai, Huaidong Zhang, Xuemiao Xu, Yong Du, Shengfeng He
Abstract: Image generation technology has brought significant advancements across various fields but has also raised concerns about data misuse and potential rights infringements, particularly with respect to creating visual artworks. Current methods aimed at safeguarding artworks often employ adversarial attacks. However, these methods face challenges such as poor transferability, high computational costs, and the introduction of noticeable noise, which compromises the aesthetic quality of the original artwork. To address these limitations, we propose a Structurally Imperceptible and Transferable Adversarial (SITA) attacks. SITA leverages a CLIP-based destylization loss, which decouples and disrupts the robust style representation of the image. This disruption hinders style extraction during stylized image generation, thereby impairing the overall stylization process. Importantly, SITA eliminates the need for a surrogate diffusion model, leading to significantly reduced computational overhead. The method's robust style feature disruption ensures high transferability across diverse models. Moreover, SITA introduces perturbations by embedding noise within the imperceptible structural details of the image. This approach effectively protects against style extraction without compromising the visual quality of the artwork. Extensive experiments demonstrate that SITA offers superior protection for artworks against unauthorized use in stylized generation. It significantly outperforms existing methods in terms of transferability, computational efficiency, and noise imperceptibility. Code is available at https://github.com/A-raniy-day/SITA.
Authors: Vitaly Gnatyuk, Valeriia Koriukina Ilya Levoshevich, Pavel Nurminskiy, Guenter Wallner
Abstract: With video games steadily increasing in complexity, automated generation of game content has found widespread interest. However, the task of 3D gaming map art creation remains underexplored to date due to its unique complexity and domain-specific challenges. While recent works have addressed related topics such as retro-style level generation and procedural terrain creation, these works primarily focus on simpler data distributions. To the best of our knowledge, we are the first to demonstrate the application of modern AI techniques for high-resolution texture manipulation in complex, highly detailed AAA 3D game environments. We introduce a novel Smart Brush for map editing, designed to assist artists in seamlessly modifying selected areas of a game map with minimal effort. By leveraging generative adversarial networks and diffusion models we propose two variants of the brush that enable efficient and context-aware generation. Our hybrid workflow aims to enhance both artistic flexibility and production efficiency, enabling the refinement of environments without manually reworking every detail, thus helping to bridge the gap between automation and creative control in game development. A comparative evaluation of our two methods with adapted versions of several state-of-the art models shows that our GAN-based brush produces the sharpest and most detailed outputs while preserving image context while the evaluated state-of-the-art models tend towards blurrier results and exhibit difficulties in maintaining contextual consistency.
Authors: Zhuoming Liu, Yiquan Li, Khoi Duc Nguyen, Yiwu Zhong, Yin Li
Abstract: Pre-trained video large language models (Video LLMs) exhibit remarkable reasoning capabilities, yet adapting these models to new tasks involving additional modalities or data types (e.g., audio or 3D information) remains challenging. In this paper, we present PAVE, a flexible framework for adapting pre-trained Video LLMs to downstream tasks with side-channel signals, such as audio, 3D cues, or multi-view videos. PAVE introduces lightweight adapters, referred to as "patches," which add a small number of parameters and operations to a base model without modifying its architecture or pre-trained weights. In doing so, PAVE can effectively adapt the pre-trained base model to support diverse downstream tasks, including audio-visual question answering, 3D reasoning, multi-view video recognition, and high frame rate video understanding. Across these tasks, PAVE significantly enhances the performance of the base model, surpassing state-of-the-art task-specific models while incurring a minor cost of ~0.1% additional FLOPs and parameters. Further, PAVE supports multi-task learning and generalizes well across different Video LLMs. Our code is available at https://github.com/dragonlzm/PAVE.
Authors: Ruixi You, Hecheng Jia, Feng Xu
Abstract: Synthetic Aperture Radar (SAR) imagery provides all-weather, all-day, and high-resolution imaging capabilities but its unique imaging mechanism makes interpretation heavily reliant on expert knowledge, limiting interpretability, especially in complex target tasks. Translating SAR images into optical images is a promising solution to enhance interpretation and support downstream tasks. Most existing research focuses on scene-level translation, with limited work on object-level translation due to the scarcity of paired data and the challenge of accurately preserving contour and texture details. To address these issues, this study proposes a keypoint-guided diffusion model (KeypointDiff) for SAR-to-optical image translation of unpaired aircraft targets. This framework introduces supervision on target class and azimuth angle via keypoints, along with a training strategy for unpaired data. Based on the classifier-free guidance diffusion architecture, a class-angle guidance module (CAGM) is designed to integrate class and angle information into the diffusion generation process. Furthermore, adversarial loss and consistency loss are employed to improve image fidelity and detail quality, tailored for aircraft targets. During sampling, aided by a pre-trained keypoint detector, the model eliminates the requirement for manually labeled class and azimuth information, enabling automated SAR-to-optical translation. Experimental results demonstrate that the proposed method outperforms existing approaches across multiple metrics, providing an efficient and effective solution for object-level SAR-to-optical translation and downstream tasks. Moreover, the method exhibits strong zero-shot generalization to untrained aircraft types with the assistance of the keypoint detector.
Authors: Zhiyang Liu, Dong Yang, Minghao Zhang, Hanyu Sun, Hong Wu, Huiying Wang, Wen Shen, Chao Chai, Shuang Xia
Abstract: Despite that deep learning (DL) methods have presented tremendous potential in many medical image analysis tasks, the practical applications of medical DL models are limited due to the lack of enough data samples with manual annotations. By noting that the clinical radiology examinations are associated with radiology reports that describe the images, we propose to develop a foundation model for multi-model head MRI by using contrastive learning on the images and the corresponding radiology findings. In particular, a contrastive learning framework is proposed, where a mixed syntax and semantic similarity matching metric is integrated to reduce the thirst of extreme large dataset in conventional contrastive learning framework. Our proposed similarity enhanced contrastive language image pretraining (SeLIP) is able to effectively extract more useful features. Experiments revealed that our proposed SeLIP performs well in many downstream tasks including image-text retrieval task, classification task, and image segmentation, which highlights the importance of considering the similarities among texts describing different images in developing medical image foundation models.
Authors: Manjushree Aithal, Rosaura G. VidalMata, Manikandtan Kartha, Gong Chen, Eashan Adhikarla, Lucas N. Kirsten, Zhicheng Fu, Nikhil A. Madhusudhana, Joe Nasti
Abstract: Low-light image enhancement is crucial for a myriad of applications, from night vision and surveillance, to autonomous driving. However, due to the inherent limitations that come in hand with capturing images in low-illumination environments, the task of enhancing such scenes still presents a formidable challenge. To advance research in this field, we introduce our Low Exposure Night Vision (LENVIZ) Dataset, a comprehensive multi-exposure benchmark dataset for low-light image enhancement comprising of over 230K frames showcasing 24K real-world indoor and outdoor, with-and without human, scenes. Captured using 3 different camera sensors, LENVIZ offers a wide range of lighting conditions, noise levels, and scene complexities, making it the largest publicly available up-to 4K resolution benchmark in the field. LENVIZ includes high quality human-generated ground truth, for which each multi-exposure low-light scene has been meticulously curated and edited by expert photographers to ensure optimal image quality. Furthermore, we also conduct a comprehensive analysis of current state-of-the-art low-light image enhancement techniques on our dataset and highlight potential areas of improvement.
Authors: Jun Zhou, Jiahao Li, Zunnan Xu, Hanhui Li, Yiji Cheng, Fa-Ting Hong, Qin Lin, Qinglin Lu, Xiaodan Liang
Abstract: Currently, instruction-based image editing methods have made significant progress by leveraging the powerful cross-modal understanding capabilities of vision language models (VLMs). However, they still face challenges in three key areas: 1) complex scenarios; 2) semantic consistency; and 3) fine-grained editing. To address these issues, we propose FireEdit, an innovative Fine-grained Instruction-based image editing framework that exploits a REgion-aware VLM. FireEdit is designed to accurately comprehend user instructions and ensure effective control over the editing process. Specifically, we enhance the fine-grained visual perception capabilities of the VLM by introducing additional region tokens. Relying solely on the output of the LLM to guide the diffusion model may lead to suboptimal editing results. Therefore, we propose a Time-Aware Target Injection module and a Hybrid Visual Cross Attention module. The former dynamically adjusts the guidance strength at various denoising stages by integrating timestep embeddings with the text embeddings. The latter enhances visual details for image editing, thereby preserving semantic consistency between the edited result and the source image. By combining the VLM enhanced with fine-grained region tokens and the time-dependent diffusion model, FireEdit demonstrates significant advantages in comprehending editing instructions and maintaining high semantic consistency. Extensive experiments indicate that our approach surpasses the state-of-the-art instruction-based image editing methods. Our project is available at https://zjgans.github.io/fireedit.github.io.
Authors: Aaron Serianni, Tyler Zhu, Vikram V. Ramaswamy, Olga Russakovsky
Abstract: Computer vision models have been shown to exhibit and amplify biases across a wide array of datasets and tasks. Existing methods for quantifying bias in classification models primarily focus on dataset distribution and model performance on subgroups, overlooking the internal workings of a model. We introduce the Attention-IoU (Attention Intersection over Union) metric and related scores, which use attention maps to reveal biases within a model's internal representations and identify image features potentially causing the biases. First, we validate Attention-IoU on the synthetic Waterbirds dataset, showing that the metric accurately measures model bias. We then analyze the CelebA dataset, finding that Attention-IoU uncovers correlations beyond accuracy disparities. Through an investigation of individual attributes through the protected attribute of Male, we examine the distinct ways biases are represented in CelebA. Lastly, by subsampling the training set to change attribute correlations, we demonstrate that Attention-IoU reveals potential confounding variables not present in dataset labels.
Authors: Carlos Plou, Cesar Borja, Ruben Martinez-Cantin, Ana C. Murillo
Abstract: Information retrieval in hour-long videos presents a significant challenge, even for state-of-the-art Vision-Language Models (VLMs), particularly when the desired information is localized within a small subset of frames. Long video data presents challenges for VLMs due to context window limitations and the difficulty of pinpointing frames containing the answer. Our novel video agent, FALCONEye, combines a VLM and a Large Language Model (LLM) to search relevant information along the video, and locate the frames with the answer. FALCONEye novelty relies on 1) the proposed meta-architecture, which is better suited to tackle hour-long videos compared to short video approaches in the state-of-the-art; 2) a new efficient exploration algorithm to locate the information using short clips, captions and answer confidence; and 3) our state-of-the-art VLMs calibration analysis for the answer confidence. Our agent is built over a small-size VLM and a medium-size LLM being accessible to run on standard computational resources. We also release FALCON-Bench, a benchmark to evaluate long (average > 1 hour) Video Answer Search challenges, highlighting the need for open-ended question evaluation. Our experiments show FALCONEye's superior performance than the state-of-the-art in FALCON-Bench, and similar or better performance in related benchmarks.
Authors: Xinpeng Li, Shijian Deng, Bolin Lai, Weiguo Pian, James M. Rehg, Yapeng Tian
Abstract: Multimodal social interaction understanding (MMSI) is critical in human-robot interaction systems. In real-world scenarios, AI agents are required to provide real-time feedback. However, existing models often depend on both past and future contexts, which hinders them from applying to real-world problems. To bridge this gap, we propose an online MMSI setting, where the model must resolve MMSI tasks using only historical information, such as recorded dialogues and video streams. To address the challenges of missing the useful future context, we develop a novel framework, named Online-MMSI-VLM, that leverages two complementary strategies: multi-party conversation forecasting and social-aware visual prompting with multi-modal large language models. First, to enrich linguistic context, the multi-party conversation forecasting simulates potential future utterances in a coarse-to-fine manner, anticipating upcoming speaker turns and then generating fine-grained conversational details. Second, to effectively incorporate visual social cues like gaze and gesture, social-aware visual prompting highlights the social dynamics in video with bounding boxes and body keypoints for each person and frame. Extensive experiments on three tasks and two datasets demonstrate that our method achieves state-of-the-art performance and significantly outperforms baseline models, indicating its effectiveness on Online-MMSI. The code and pre-trained models will be publicly released at: https://github.com/Sampson-Lee/OnlineMMSI.
Authors: Tianhao Qi, Jianlong Yuan, Wanquan Feng, Shancheng Fang, Jiawei Liu, SiYu Zhou, Qian He, Hongtao Xie, Yongdong Zhang
Abstract: Sora has unveiled the immense potential of the Diffusion Transformer (DiT) architecture in single-scene video generation. However, the more challenging task of multi-scene video generation, which offers broader applications, remains relatively underexplored. To bridge this gap, we propose Mask$^2$DiT, a novel approach that establishes fine-grained, one-to-one alignment between video segments and their corresponding text annotations. Specifically, we introduce a symmetric binary mask at each attention layer within the DiT architecture, ensuring that each text annotation applies exclusively to its respective video segment while preserving temporal coherence across visual tokens. This attention mechanism enables precise segment-level textual-to-visual alignment, allowing the DiT architecture to effectively handle video generation tasks with a fixed number of scenes. To further equip the DiT architecture with the ability to generate additional scenes based on existing ones, we incorporate a segment-level conditional mask, which conditions each newly generated segment on the preceding video segments, thereby enabling auto-regressive scene extension. Both qualitative and quantitative experiments confirm that Mask$^2$DiT excels in maintaining visual consistency across segments while ensuring semantic alignment between each segment and its corresponding text description. Our project page is https://tianhao-qi.github.io/Mask2DiTProject.
Authors: Lifu Wang, Daqing Liu, Xinchen Liu, Xiaodong He
Abstract: Text encoders in diffusion models have rapidly evolved, transitioning from CLIP to T5-XXL. Although this evolution has significantly enhanced the models' ability to understand complex prompts and generate text, it also leads to a substantial increase in the number of parameters. Despite T5 series encoders being trained on the C4 natural language corpus, which includes a significant amount of non-visual data, diffusion models with T5 encoder do not respond to those non-visual prompts, indicating redundancy in representational power. Therefore, it raises an important question: "Do we really need such a large text encoder?" In pursuit of an answer, we employ vision-based knowledge distillation to train a series of T5 encoder models. To fully inherit its capabilities, we constructed our dataset based on three criteria: image quality, semantic understanding, and text-rendering. Our results demonstrate the scaling down pattern that the distilled T5-base model can generate images of comparable quality to those produced by T5-XXL, while being 50 times smaller in size. This reduction in model size significantly lowers the GPU requirements for running state-of-the-art models such as FLUX and SD3, making high-quality text-to-image generation more accessible.
Authors: Hao Yu, Zhuokai Zhao, Shen Yan, Lukasz Korycki, Jianyu Wang, Baosheng He, Jiayi Liu, Lizhu Zhang, Xiangjun Fan, Hanchao Yu
Abstract: The rapid advancement of large vision-language models (LVLMs) has driven significant progress in multimodal tasks, enabling models to interpret, reason, and generate outputs across both visual and textual domains. While excelling in generative tasks, existing LVLMs often face limitations in tasks requiring high-fidelity representation learning, such as generating image or text embeddings for retrieval. Recent work has proposed finetuning LVLMs for representational learning, but the fine-tuned model often loses its generative capabilities due to the representational learning training paradigm. To address this trade-off, we introduce CAFe, a contrastive-autoregressive fine-tuning framework that enhances LVLMs for both representation and generative tasks. By integrating a contrastive objective with autoregressive language modeling, our approach unifies these traditionally separate tasks, achieving state-of-the-art results in both multimodal retrieval and multimodal generative benchmarks, including object hallucination (OH) mitigation. CAFe establishes a novel framework that synergizes embedding and generative functionalities in a single model, setting a foundation for future multimodal models that excel in both retrieval precision and coherent output generation.
Authors: Liang Pan, Zeshi Yang, Zhiyang Dou, Wenjia Wang, Buzhen Huang, Bo Dai, Taku Komura, Jingbo Wang
Abstract: Synthesizing diverse and physically plausible Human-Scene Interactions (HSI) is pivotal for both computer animation and embodied AI. Despite encouraging progress, current methods mainly focus on developing separate controllers, each specialized for a specific interaction task. This significantly hinders the ability to tackle a wide variety of challenging HSI tasks that require the integration of multiple skills, e.g., sitting down while carrying an object. To address this issue, we present TokenHSI, a single, unified transformer-based policy capable of multi-skill unification and flexible adaptation. The key insight is to model the humanoid proprioception as a separate shared token and combine it with distinct task tokens via a masking mechanism. Such a unified policy enables effective knowledge sharing across skills, thereby facilitating the multi-task training. Moreover, our policy architecture supports variable length inputs, enabling flexible adaptation of learned skills to new scenarios. By training additional task tokenizers, we can not only modify the geometries of interaction targets but also coordinate multiple skills to address complex tasks. The experiments demonstrate that our approach can significantly improve versatility, adaptability, and extensibility in various HSI tasks. Website: https://liangpan99.github.io/TokenHSI/
Authors: Fernando Julio Cendra, Kai Han
Abstract: The inherent ambiguity in defining visual concepts poses significant challenges for modern generative models, such as the diffusion-based Text-to-Image (T2I) models, in accurately learning concepts from a single image. Existing methods lack a systematic way to reliably extract the interpretable underlying intrinsic concepts. To address this challenge, we present ICE, short for Intrinsic Concept Extraction, a novel framework that exclusively utilizes a T2I model to automatically and systematically extract intrinsic concepts from a single image. ICE consists of two pivotal stages. In the first stage, ICE devises an automatic concept localization module to pinpoint relevant text-based concepts and their corresponding masks within the image. This critical stage streamlines concept initialization and provides precise guidance for subsequent analysis. The second stage delves deeper into each identified mask, decomposing the object-level concepts into intrinsic concepts and general concepts. This decomposition allows for a more granular and interpretable breakdown of visual elements. Our framework demonstrates superior performance on intrinsic concept extraction from a single image in an unsupervised manner. Project page: https://visual-ai.github.io/ice
Authors: Baifeng Shi, Boyi Li, Han Cai, Yao Lu, Sifei Liu, Marco Pavone, Jan Kautz, Song Han, Trevor Darrell, Pavlo Molchanov, Hongxu Yin
Abstract: High-resolution perception of visual details is crucial for daily tasks. Current vision pre-training, however, is still limited to low resolutions (e.g., 378 x 378 pixels) due to the quadratic cost of processing larger images. We introduce PS3 that scales CLIP-style vision pre-training to 4K resolution with a near-constant cost. Instead of contrastive learning on global image representation, PS3 is pre-trained by selectively processing local regions and contrasting them with local detailed captions, enabling high-resolution representation learning with greatly reduced computational overhead. The pre-trained PS3 is able to both encode the global image at low resolution and selectively process local high-resolution regions based on their saliency or relevance to a text prompt. When applying PS3 to multi-modal LLM (MLLM), the resulting model, named VILA-HD, significantly improves high-resolution visual perception compared to baselines without high-resolution vision pre-training such as AnyRes and S^2 while using up to 4.3x fewer tokens. PS3 also unlocks appealing scaling properties of VILA-HD, including scaling up resolution for free and scaling up test-time compute for better performance. Compared to state of the arts, VILA-HD outperforms previous MLLMs such as NVILA and Qwen2-VL across multiple benchmarks and achieves better efficiency than latest token pruning approaches. Finally, we find current benchmarks do not require 4K-resolution perception, which motivates us to propose 4KPro, a new benchmark of image QA at 4K resolution, on which VILA-HD outperforms all previous MLLMs, including a 14.5% improvement over GPT-4o, and a 3.2% improvement and 2.96x speedup over Qwen2-VL.
Authors: Zihang Lai, Andrea Vedaldi
Abstract: Temporal consistency is critical in video prediction to ensure that outputs are coherent and free of artifacts. Traditional methods, such as temporal attention and 3D convolution, may struggle with significant object motion and may not capture long-range temporal dependencies in dynamic scenes. To address this gap, we propose the Tracktention Layer, a novel architectural component that explicitly integrates motion information using point tracks, i.e., sequences of corresponding points across frames. By incorporating these motion cues, the Tracktention Layer enhances temporal alignment and effectively handles complex object motions, maintaining consistent feature representations over time. Our approach is computationally efficient and can be seamlessly integrated into existing models, such as Vision Transformers, with minimal modification. It can be used to upgrade image-only models to state-of-the-art video ones, sometimes outperforming models natively designed for video prediction. We demonstrate this on video depth prediction and video colorization, where models augmented with the Tracktention Layer exhibit significantly improved temporal consistency compared to baselines.
Authors: Hongyu Liu, Xuan Wang, Ziyu Wan, Yue Ma, Jingye Chen, Yanbo Fan, Yujun Shen, Yibing Song, Qifeng Chen
Abstract: This work focuses on open-domain 4D avatarization, with the purpose of creating a 4D avatar from a portrait image in an arbitrary style. We select parametric triplanes as the intermediate 4D representation and propose a practical training paradigm that takes advantage of both generative adversarial networks (GANs) and diffusion models. Our design stems from the observation that 4D GANs excel at bridging images and triplanes without supervision yet usually face challenges in handling diverse data distributions. A robust 2D diffusion prior emerges as the solution, assisting the GAN in transferring its expertise across various domains. The synergy between these experts permits the construction of a multi-domain image-triplane dataset, which drives the development of a general 4D avatar creator. Extensive experiments suggest that our model, AvatarArtist, is capable of producing high-quality 4D avatars with strong robustness to various source image domains. The code, the data, and the models will be made publicly available to facilitate future studies..
Authors: Xuan Ju, Weicai Ye, Quande Liu, Qiulin Wang, Xintao Wang, Pengfei Wan, Di Zhang, Kun Gai, Qiang Xu
Abstract: Current video generative foundation models primarily focus on text-to-video tasks, providing limited control for fine-grained video content creation. Although adapter-based approaches (e.g., ControlNet) enable additional controls with minimal fine-tuning, they encounter challenges when integrating multiple conditions, including: branch conflicts between independently trained adapters, parameter redundancy leading to increased computational cost, and suboptimal performance compared to full fine-tuning. To address these challenges, we introduce FullDiT, a unified foundation model for video generation that seamlessly integrates multiple conditions via unified full-attention mechanisms. By fusing multi-task conditions into a unified sequence representation and leveraging the long-context learning ability of full self-attention to capture condition dynamics, FullDiT reduces parameter overhead, avoids conditions conflict, and shows scalability and emergent ability. We further introduce FullBench for multi-task video generation evaluation. Experiments demonstrate that FullDiT achieves state-of-the-art results, highlighting the efficacy of full-attention in complex multi-task video generation.
Authors: Chuong Huynh, Jinyu Yang, Ashish Tawari, Mubarak Shah, Son Tran, Raffay Hamid, Trishul Chilimbi, Abhinav Shrivastava
Abstract: Composed Image Retrieval (CIR) is a complex task that aims to retrieve images based on a multimodal query. Typical training data consists of triplets containing a reference image, a textual description of desired modifications, and the target image, which are expensive and time-consuming to acquire. The scarcity of CIR datasets has led to zero-shot approaches utilizing synthetic triplets or leveraging vision-language models (VLMs) with ubiquitous web-crawled image-caption pairs. However, these methods have significant limitations: synthetic triplets suffer from limited scale, lack of diversity, and unnatural modification text, while image-caption pairs hinder joint embedding learning of the multimodal query due to the absence of triplet data. Moreover, existing approaches struggle with complex and nuanced modification texts that demand sophisticated fusion and understanding of vision and language modalities. We present CoLLM, a one-stop framework that effectively addresses these limitations. Our approach generates triplets on-the-fly from image-caption pairs, enabling supervised training without manual annotation. We leverage Large Language Models (LLMs) to generate joint embeddings of reference images and modification texts, facilitating deeper multimodal fusion. Additionally, we introduce Multi-Text CIR (MTCIR), a large-scale dataset comprising 3.4M samples, and refine existing CIR benchmarks (CIRR and Fashion-IQ) to enhance evaluation reliability. Experimental results demonstrate that CoLLM achieves state-of-the-art performance across multiple CIR benchmarks and settings. MTCIR yields competitive results, with up to 15% performance improvement. Our refined benchmarks provide more reliable evaluation metrics for CIR models, contributing to the advancement of this important field.
Authors: Xiang Xu, Lingdong Kong, Hui Shuai, Wenwei Zhang, Liang Pan, Kai Chen, Ziwei Liu, Qingshan Liu
Abstract: LiDAR representation learning has emerged as a promising approach to reducing reliance on costly and labor-intensive human annotations. While existing methods primarily focus on spatial alignment between LiDAR and camera sensors, they often overlook the temporal dynamics critical for capturing motion and scene continuity in driving scenarios. To address this limitation, we propose SuperFlow++, a novel framework that integrates spatiotemporal cues in both pretraining and downstream tasks using consecutive LiDAR-camera pairs. SuperFlow++ introduces four key components: (1) a view consistency alignment module to unify semantic information across camera views, (2) a dense-to-sparse consistency regularization mechanism to enhance feature robustness across varying point cloud densities, (3) a flow-based contrastive learning approach that models temporal relationships for improved scene understanding, and (4) a temporal voting strategy that propagates semantic information across LiDAR scans to improve prediction consistency. Extensive evaluations on 11 heterogeneous LiDAR datasets demonstrate that SuperFlow++ outperforms state-of-the-art methods across diverse tasks and driving conditions. Furthermore, by scaling both 2D and 3D backbones during pretraining, we uncover emergent properties that provide deeper insights into developing scalable 3D foundation models. With strong generalizability and computational efficiency, SuperFlow++ establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving. The code is publicly available at https://github.com/Xiangxu-0103/SuperFlow
Authors: Mingju Gao, Yike Pan, Huan-ang Gao, Zongzheng Zhang, Wenyi Li, Hao Dong, Hao Tang, Li Yi, Hao Zhao
Abstract: As interest grows in world models that predict future states from current observations and actions, accurately modeling part-level dynamics has become increasingly relevant for various applications. Existing approaches, such as Puppet-Master, rely on fine-tuning large-scale pre-trained video diffusion models, which are impractical for real-world use due to the limitations of 2D video representation and slow processing times. To overcome these challenges, we present PartRM, a novel 4D reconstruction framework that simultaneously models appearance, geometry, and part-level motion from multi-view images of a static object. PartRM builds upon large 3D Gaussian reconstruction models, leveraging their extensive knowledge of appearance and geometry in static objects. To address data scarcity in 4D, we introduce the PartDrag-4D dataset, providing multi-view observations of part-level dynamics across over 20,000 states. We enhance the model's understanding of interaction conditions with a multi-scale drag embedding module that captures dynamics at varying granularities. To prevent catastrophic forgetting during fine-tuning, we implement a two-stage training process that focuses sequentially on motion and appearance learning. Experimental results show that PartRM establishes a new state-of-the-art in part-level motion learning and can be applied in manipulation tasks in robotics. Our code, data, and models are publicly available to facilitate future research.
Authors: Sangwon Beak, Hyeonwoo Kim, Hanbyul Joo
Abstract: We present a method for learning 3D spatial relationships between object pairs, referred to as object-object spatial relationships (OOR), by leveraging synthetically generated 3D samples from pre-trained 2D diffusion models. We hypothesize that images synthesized by 2D diffusion models inherently capture plausible and realistic OOR cues, enabling efficient ways to collect a 3D dataset to learn OOR for various unbounded object categories. Our approach begins by synthesizing diverse images that capture plausible OOR cues, which we then uplift into 3D samples. Leveraging our diverse collection of plausible 3D samples for the object pairs, we train a score-based OOR diffusion model to learn the distribution of their relative spatial relationships. Additionally, we extend our pairwise OOR to multi-object OOR by enforcing consistency across pairwise relations and preventing object collisions. Extensive experiments demonstrate the robustness of our method across various object-object spatial relationships, along with its applicability to real-world 3D scene arrangement tasks using the OOR diffusion model.
Authors: Lingdong Kong, Dongyue Lu, Xiang Xu, Lai Xing Ng, Wei Tsang Ooi, Benoit R. Cottereau
Abstract: Cross-platform adaptation in event-based dense perception is crucial for deploying event cameras across diverse settings, such as vehicles, drones, and quadrupeds, each with unique motion dynamics, viewpoints, and class distributions. In this work, we introduce EventFly, a framework for robust cross-platform adaptation in event camera perception. Our approach comprises three key components: i) Event Activation Prior (EAP), which identifies high-activation regions in the target domain to minimize prediction entropy, fostering confident, domain-adaptive predictions; ii) EventBlend, a data-mixing strategy that integrates source and target event voxel grids based on EAP-driven similarity and density maps, enhancing feature alignment; and iii) EventMatch, a dual-discriminator technique that aligns features from source, target, and blended domains for better domain-invariant learning. To holistically assess cross-platform adaptation abilities, we introduce EXPo, a large-scale benchmark with diverse samples across vehicle, drone, and quadruped platforms. Extensive experiments validate our effectiveness, demonstrating substantial gains over popular adaptation methods. We hope this work can pave the way for more adaptive, high-performing event perception across diverse and complex environments.
Authors: Vincent C. M\"uller
Abstract: This paper investigates the prospects of AI without representation in general, and the proposals of Rodney Brooks in particular. What turns out to be characteristic of Brooks' proposal is the rejection of central control in intelligent agents; his systems has as much or as little representation as traditional AI. The traditional view that representation is necessary for intelligence presupposes that intelligence requires central control. However, much of recent cognitive science suggests that we should dispose of the image of intelligent agents as central representation processors. If this paradigm shift is achieved, Brooks' proposal for non-centralized cognition without representation appears promising for full-blown intelligent agents - though not for conscious agents and thus not for human-like AI.
Authors: Muhammad Ahmad, Sardar Usman, Ildar Batyrshin, Muhammad Muzammil, K. Sajid, M. Hasnain, Muhammad Jalal, Grigori Sidorov
Abstract: Background: Lung disease is a significant health issue, particularly in children and elderly individuals. It often results from lung infections and is one of the leading causes of mortality in children. Globally, lung-related diseases claim many lives each year, making early and accurate diagnoses crucial. Radiographs are valuable tools for the diagnosis of such conditions. The most prevalent lung diseases, including pneumonia, asthma, allergies, chronic obstructive pulmonary disease (COPD), bronchitis, emphysema, and lung cancer, represent significant public health challenges. Early prediction of these conditions is critical, as it allows for the identification of risk factors and implementation of preventive measures to reduce the likelihood of disease onset Methods: In this study, we utilized a dataset comprising 3,475 chest X-ray images sourced from from Mendeley Data provided by Talukder, M. A. (2023) [14], categorized into three classes: normal, lung opacity, and pneumonia. We applied five pre-trained deep learning models, including CNN, ResNet50, DenseNet, CheXNet, and U-Net, as well as two transfer learning algorithms such as Vision Transformer (ViT) and Shifted Window (Swin) to classify these images. This approach aims to address diagnostic issues in lung abnormalities by reducing reliance on human intervention through automated classification systems. Our analysis was conducted in both binary and multiclass settings. Results: In the binary classification, we focused on distinguishing between normal and viral pneumonia cases, whereas in the multi-class classification, all three classes (normal, lung opacity, and viral pneumonia) were included. Our proposed methodology (ViT) achieved remarkable performance, with accuracy rates of 99% for binary classification and 95.25% for multiclass classification.
Authors: Xuan Liu, Xiaobin Chang
Abstract: In continual learning (CL), catastrophic forgetting often arises due to feature drift. This challenge is particularly prominent in the exemplar-free continual learning (EFCL) setting, where samples from previous tasks cannot be retained, making it difficult to preserve prior knowledge. To address this issue, some EFCL methods aim to identify feature spaces that minimize the impact on previous tasks while accommodating new ones. However, they rely on static features or outdated statistics stored from old tasks, which prevents them from capturing the dynamic evolution of the feature space in CL, leading to performance degradation over time. In this paper, we introduce the Drift-Resistant Space (DRS), which effectively handles feature drifts without requiring explicit feature modeling or the storage of previous tasks. A novel parameter-efficient fine-tuning approach called Low-Rank Adaptation Subtraction (LoRA-) is proposed to develop the DRS. This method subtracts the LoRA weights of old tasks from the initial pre-trained weight before processing new task data to establish the DRS for model training. Therefore, LoRA- enhances stability, improves efficiency, and simplifies implementation. Furthermore, stabilizing feature drifts allows for better plasticity by learning with a triplet loss. Our method consistently achieves state-of-the-art results, especially for long task sequences, across multiple datasets.
Authors: Haiqi Liu, C. L. Philip Chen, Tong Zhang
Abstract: Cross-subject EEG emotion recognition is challenged by significant inter-subject variability and intricately entangled intra-subject variability. Existing works have primarily addressed these challenges through domain adaptation or generalization strategies. However, they typically require extensive target subject data or demonstrate limited generalization performance to unseen subjects. Recent few-shot learning paradigms attempt to address these limitations but often encounter catastrophic overfitting during subject-specific adaptation with limited samples. This article introduces the few-shot adapter with a cross-view fusion method called FACE for cross-subject EEG emotion recognition, which leverages dynamic multi-view fusion and effective subject-specific adaptation. Specifically, FACE incorporates a cross-view fusion module that dynamically integrates global brain connectivity with localized patterns via subject-specific fusion weights to provide complementary emotional information. Moreover, the few-shot adapter module is proposed to enable rapid adaptation for unseen subjects while reducing overfitting by enhancing adapter structures with meta-learning. Experimental results on three public EEG emotion recognition benchmarks demonstrate FACE's superior generalization performance over state-of-the-art methods. FACE provides a practical solution for cross-subject scenarios with limited labeled data.
Authors: Abdul Qayyum, Moona Mazher, Devran Ugurlu, Jose Alonso Solis Lemus, Cristobal Rodero, Steven A Niederer
Abstract: Whole-heart segmentation from CT and MRI scans is crucial for cardiovascular disease analysis, yet existing methods struggle with modality-specific biases and the need for extensive labeled datasets. To address these challenges, we propose a foundation model for whole-heart segmentation using a self-supervised learning (SSL) framework based on a student-teacher architecture. Our model is pretrained on a large, unlabeled dataset of CT and MRI scans, leveraging the xLSTM backbone to capture long-range spatial dependencies and complex anatomical structures in 3D medical images. By incorporating multi-modal pretraining, our approach ensures strong generalization across both CT and MRI modalities, mitigating modality-specific variations and improving segmentation accuracy in diverse clinical settings. The use of large-scale unlabeled data significantly reduces the dependency on manual annotations, enabling robust performance even with limited labeled data. We further introduce an xLSTM-UNet-based architecture for downstream whole-heart segmentation tasks, demonstrating its effectiveness on few-label CT and MRI datasets. Our results validate the robustness and adaptability of the proposed model, highlighting its potential for advancing automated whole-heart segmentation in medical imaging.
Authors: Kangwei Liu, Mengru Wang, Yujie Luo, Lin Yuan, Mengshu Sun, Ningyu Zhang, Lei Liang, Zhiqiang Zhang, Jun Zhou, Huajun Chen
Abstract: Fine-tuning enables large language models (LLMs) to adapt to specific domains, but often undermines their previously established safety alignment. To mitigate the degradation of model safety during fine-tuning, we introduce LookAhead Tuning, which comprises two simple, low-resource, and effective data-driven methods that modify training data by previewing partial answer prefixes. Both methods aim to preserve the model's inherent safety mechanisms by minimizing perturbations to initial token distributions. Comprehensive experiments demonstrate that LookAhead Tuning effectively maintains model safety without sacrificing robust performance on downstream tasks. Our findings position LookAhead Tuning as a reliable and efficient solution for the safe and effective adaptation of LLMs. Code is released at https://github.com/zjunlp/LookAheadTuning.
Authors: Swati Sharma, Fabian A. Braeu, Thanadet Chuangsuwanich, Tin A. Tun, Quan V Hoang, Rachel Chong, Shamira Perera, Ching-Lin Ho, Rahat Husain, Martin L. Buist, Tin Aung, Micha\"el J. A. Girard
Abstract: Purpose: To characterize the 3D structural phenotypes of the optic nerve head (ONH) in patients with glaucoma, high myopia, and concurrent high myopia and glaucoma, and to evaluate their variations across these conditions. Participants: A total of 685 optical coherence tomography (OCT) scans from 754 subjects of Singapore-Chinese ethnicity, including 256 healthy (H), 94 highly myopic (HM), 227 glaucomatous (G), and 108 highly myopic with glaucoma (HMG) cases. Methods: We segmented the retinal and connective tissues from OCT volumes and their boundary edges were converted into 3D point clouds. To classify the 3D point clouds into four ONH conditions, i.e., H, HM, G, and HMG, a specialized ensemble network was developed, consisting of an encoder to transform high-dimensional input data into a compressed latent vector, a decoder to reconstruct point clouds from the latent vector, and a classifier to categorize the point clouds into the four ONH conditions. Results: The classification network achieved high accuracy, distinguishing H, HM, G, and HMG classes with a micro-average AUC of 0.92 $\pm$ 0.03 on an independent test set. The decoder effectively reconstructed point clouds, achieving a Chamfer loss of 0.013 $\pm$ 0.002. Dimensionality reduction clustered ONHs into four distinct groups, revealing structural variations such as changes in retinal and connective tissue thickness, tilting and stretching of the disc and scleral canal opening, and alterations in optic cup morphology, including shallow or deep excavation, across the four conditions. Conclusions: This study demonstrated that ONHs exhibit distinct structural signatures across H, HM, G, and HMG conditions. The findings further indicate that ONH morphology provides sufficient information for classification into distinct clusters, with principal components capturing unique structural patterns within each group.
Authors: Yiling Wang, Elia Lombardo, Adrian Thummerer, Tom Bl\"ocker, Yu Fan, Yue Zhao, Christianna Iris Papadopoulou, Coen Hurkmans, Rob H. N. Tijssen, Pia A. W. G\"orts, Shyama U. Tetar, Davide Cusumano, Martijn P. W. Intven, Pim Borman, Marco Riboldi, Denis Dud\'a\v{s}, Hilary Byrne, Lorenzo Placidi, Marco Fusella, Michael Jameson, Miguel Palacios, Paul Cobussen, Tobias Finazzi, Cornelis J. A. Haasbeek, Paul Keall, Christopher Kurz, Guillaume Landry, Matteo Maspero
Abstract: Purpose: Magnetic resonance imaging (MRI) to visualize anatomical motion is becoming increasingly important when treating cancer patients with radiotherapy. Hybrid MRI-linear accelerator (MRI-linac) systems allow real-time motion management during irradiation. This paper presents a multi-institutional real-time MRI time series dataset from different MRI-linac vendors. The dataset is designed to support developing and evaluating real-time tumor localization (tracking) algorithms for MRI-guided radiotherapy within the TrackRAD2025 challenge (https://trackrad2025.grand-challenge.org/). Acquisition and validation methods: The dataset consists of sagittal 2D cine MRIs in 585 patients from six centers (3 Dutch, 1 German, 1 Australian, and 1 Chinese). Tumors in the thorax, abdomen, and pelvis acquired on two commercially available MRI-linacs (0.35 T and 1.5 T) were included. For 108 cases, irradiation targets or tracking surrogates were manually segmented on each temporal frame. The dataset was randomly split into a public training set of 527 cases (477 unlabeled and 50 labeled) and a private testing set of 58 cases (all labeled). Data Format and Usage Notes: The data is publicly available under the TrackRAD2025 collection: https://doi.org/10.57967/hf/4539. Both the images and segmentations for each patient are available in metadata format. Potential Applications: This novel clinical dataset will enable the development and evaluation of real-time tumor localization algorithms for MRI-guided radiotherapy. By enabling more accurate motion management and adaptive treatment strategies, this dataset has the potential to advance the field of radiotherapy significantly.
URLs: https://trackrad2025.grand-challenge.org/)., https://doi.org/10.57967/hf/4539.
Authors: Sidhanth Holalkere, David S. Bindel, Silvia Sell\'an, Alexander Terenin
Abstract: Poisson Surface Reconstruction is a widely-used algorithm for reconstructing a surface from an oriented point cloud. To facilitate applications where only partial surface information is available, or scanning is performed sequentially, a recent line of work proposes to incorporate uncertainty into the reconstructed surface via Gaussian process models. The resulting algorithms first perform Gaussian process interpolation, then solve a set of volumetric partial differential equations globally in space, resulting in a computationally expensive two-stage procedure. In this work, we apply recently-developed techniques from geometric Gaussian processes to combine interpolation and surface reconstruction into a single stage, requiring only one linear solve per sample. The resulting reconstructed surface samples can be queried locally in space, without the use of problem-dependent volumetric meshes or grids. These capabilities enable one to (a) perform probabilistic collision detection locally around the region of interest, (b) perform ray casting without evaluating points not on the ray's trajectory, and (c) perform next-view planning on a per-slice basis. They also improve reconstruction quality, by not requiring one to approximate kernel matrix inverses with diagonal matrices as part of intermediate computations. Results show that our approach provides a cleaner, more-principled, and more-flexible stochastic surface reconstruction pipeline.
Authors: Yorick Estievenart, Sukanya Patra, Souhaib Ben Taieb
Abstract: Efficient and reliable operation of Concentrated Solar Power (CSP) plants is essential for meeting the growing demand for sustainable energy. However, high-temperature solar receivers face severe operational risks, such as freezing, deformation, and corrosion, resulting in costly downtime and maintenance. To monitor CSP plants, cameras mounted on solar receivers record infrared images at irregular intervals ranging from one to five minutes throughout the day. Anomalous images can be detected by thresholding an anomaly score, where the threshold is chosen to optimize metrics such as the F1-score on a validation set. This work proposes a framework for generating more reliable decision thresholds with finite-sample coverage guarantees on any chosen risk function. Our framework also incorporates an abstention mechanism, allowing high-risk predictions to be deferred to domain experts. Second, we propose a density forecasting method to estimate the likelihood of an observed image given a sequence of previously observed images, using this likelihood as its anomaly score. Third, we analyze the deployment results of our framework across multiple training scenarios over several months for two CSP plants. This analysis provides valuable insights to our industry partner for optimizing maintenance operations. Finally, given the confidential nature of our dataset, we provide an extended simulated dataset, leveraging recent advancements in generative modeling to create diverse thermal images that simulate multiple CSP plants. Our code is publicly available.
Authors: Christian John Hurry, Jinjie Zhang, Olubukola Ishola, Emma Slade, Cuong Q. Nguyen
Abstract: Developing computer vision for high-content screening is challenging due to various sources of distribution-shift caused by changes in experimental conditions, perturbagens, and fluorescent markers. The impact of different sources of distribution-shift are confounded in typical evaluations of models based on transfer learning, which limits interpretations of how changes to model design and training affect generalisation. We propose an evaluation scheme that isolates sources of distribution-shift using the JUMP-CP dataset, allowing researchers to evaluate generalisation with respect to specific sources of distribution-shift. We then present a channel-agnostic masked autoencoder $\mathbf{Campfire}$ which, via a shared decoder for all channels, scales effectively to datasets containing many different fluorescent markers, and show that it generalises to out-of-distribution experimental batches, perturbagens, and fluorescent markers, and also demonstrates successful transfer learning from one cell type to another.
Authors: Shoffan Saifullah, Rafa{\l} Dre\.zewski
Abstract: Medical image segmentation, particularly for brain tumor analysis, demands precise and computationally efficient models due to the complexity of multimodal MRI datasets and diverse tumor morphologies. This study introduces PSO-UNet, which integrates Particle Swarm Optimization (PSO) with the U-Net architecture for dynamic hyperparameter optimization. Unlike traditional manual tuning or alternative optimization approaches, PSO effectively navigates complex hyperparameter search spaces, explicitly optimizing the number of filters, kernel size, and learning rate. PSO-UNet substantially enhances segmentation performance, achieving Dice Similarity Coefficients (DSC) of 0.9578 and 0.9523 and Intersection over Union (IoU) scores of 0.9194 and 0.9097 on the BraTS 2021 and Figshare datasets, respectively. Moreover, the method reduces computational complexity significantly, utilizing only 7.8 million parameters and executing in approximately 906 seconds, markedly faster than comparable U-Net-based frameworks. These outcomes underscore PSO-UNet's robust generalization capabilities across diverse MRI modalities and tumor classifications, emphasizing its clinical potential and clear advantages over conventional hyperparameter tuning methods. Future research will explore hybrid optimization strategies and validate the framework against other bio-inspired algorithms to enhance its robustness and scalability.
Authors: Xinpeng Liu, Zeyi Huang, Fumio Okura, Yasuyuki Matsushita
Abstract: Novel view synthesis has demonstrated impressive progress recently, with 3D Gaussian splatting (3DGS) offering efficient training time and photorealistic real-time rendering. However, reliance on Cartesian coordinates limits 3DGS's performance on distant objects, which is important for reconstructing unbounded outdoor environments. We found that, despite its ultimate simplicity, using homogeneous coordinates, a concept on the projective geometry, for the 3DGS pipeline remarkably improves the rendering accuracies of distant objects. We therefore propose Homogeneous Gaussian Splatting (HoGS) incorporating homogeneous coordinates into the 3DGS framework, providing a unified representation for enhancing near and distant objects. HoGS effectively manages both expansive spatial positions and scales particularly in outdoor unbounded environments by adopting projective geometry principles. Experiments show that HoGS significantly enhances accuracy in reconstructing distant objects while maintaining high-quality rendering of nearby objects, along with fast training speed and real-time rendering capability. Our implementations are available on our project page https://kh129.github.io/hogs/.
Authors: Chonghang Zhao, Mingyuan Ge, Xiaogang Yang, Yong S. Chu, Hanfei Yan
Abstract: A long-standing challenge in tomography is the 'missing wedge' problem, which arises when the acquisition of projection images within a certain angular range is restricted due to geometrical constraints. This incomplete dataset results in significant artifacts and poor resolution in the reconstructed image. To tackle this challenge, we propose an approach dubbed Perception Fused Iterative Tomography Reconstruction Engine, which integrates a convolutional neural network (CNN) with perceptional knowledge as a smart regularizer into an iterative solving engine. We employ the Alternating Direction Method of Multipliers to optimize the solution in both physics and image domains, thereby achieving a physically coherent and visually enhanced result. We demonstrate the effectiveness of the proposed approach using various experimental datasets obtained with different x-ray microscopy techniques. All show significantly improved reconstruction even with a missing wedge of over 100 degrees - a scenario where conventional methods fail. Notably, it also improves the reconstruction in case of sparse projections, despite the network not being specifically trained for that. This demonstrates the robustness and generality of our method of addressing commonly occurring challenges in 3D x-ray imaging applications for real-world problems.
Authors: Zeqiang Wei, Kai Jin, Zeyi Hou, Kuan Song, Xiuzhuang Zhou
Abstract: Transformers bring significantly improved performance to the light field image super-resolution task due to their long-range dependency modeling capability. However, the inherently high computational complexity of their core self-attention mechanism has increasingly hindered their advancement in this task. To address this issue, we first introduce the LF-VSSM block, a novel module inspired by progressive feature extraction, to efficiently capture critical long-range spatial-angular dependencies in light field images. LF-VSSM successively extracts spatial features within sub-aperture images, spatial-angular features between sub-aperture images, and spatial-angular features between light field image pixels. On this basis, we propose a lightweight network, $L^2$FMamba (Lightweight Light Field Mamba), which integrates the LF-VSSM block to leverage light field features for super-resolution tasks while overcoming the computational challenges of Transformer-based approaches. Extensive experiments on multiple light field datasets demonstrate that our method reduces the number of parameters and complexity while achieving superior super-resolution performance with faster inference speed.
Authors: Xuechen Liang, Meiling Tao, Yinghui Xia, Jianhui Wang, Kun Li, Yijin Wang, Jingsong Yang, Tianyu Shi, Yuantao Wang, Miao Zhang, Xueqian Wang
Abstract: Large language models (LLMs) have made significant advances in the field of natural language processing, but they still face challenges such as continuous decision-making, lack of long-term memory, and limited context windows in dynamic environments. To address these issues, this paper proposes an innovative framework Memory-Enhanced Agents with Reflective Self-improvement. The MARS framework comprises three agents: the User, the Assistant, and the Checker. By integrating iterative feedback, reflective mechanisms, and a memory optimization mechanism based on the Ebbinghaus forgetting curve, it significantly enhances the agents capabilities in handling multi-tasking and long-span information.
Authors: Xiaoqing Zhang, Hanfeng Shi, Xiangyu Li, Haili Ye, Tao Xu, Na Li, Yan Hu, Fan Lv, Jiangfan Chen, Jiang Liu
Abstract: Parkinson's disease (PD) is a prevalent neurodegenerative disorder globally. The eye's retina is an extension of the brain and has great potential in PD screening. Recent studies have suggested that texture features extracted from retinal layers can be adopted as biomarkers for PD diagnosis under optical coherence tomography (OCT) images. Frequency domain learning techniques can enhance the feature representations of deep neural networks (DNNs) by decomposing frequency components involving rich texture features. Additionally, previous works have not exploited texture features for automated PD screening in OCT. Motivated by the above analysis, we propose a novel Adaptive Wavelet Filter (AWF) that serves as the Practical Texture Feature Amplifier to fully leverage the merits of texture features to boost the PD screening performance of DNNs with the aid of frequency domain learning. Specifically, AWF first enhances texture feature representation diversities via channel mixer, then emphasizes informative texture feature representations with the well-designed adaptive wavelet filtering token mixer. By combining the AWFs with the DNN stem, AWFNet is constructed for automated PD screening. Additionally, we introduce a novel Balanced Confidence (BC) Loss by mining the potential of sample-wise predicted probabilities of all classes and class frequency prior, to further boost the PD screening performance and trustworthiness of AWFNet. The extensive experiments manifest the superiority of our AWFNet and BC over state-of-the-art methods in terms of PD screening performance and trustworthiness.
Authors: Yongting Hu, Yuxin Lin, Chengliang Liu, Xiaoling Luo, Xiaoyan Dou, Qihao Xu, Yong Xu
Abstract: Multi-view diabetic retinopathy (DR) detection has recently emerged as a promising method to address the issue of incomplete lesions faced by single-view DR. However, it is still challenging due to the variable sizes and scattered locations of lesions. Furthermore, existing multi-view DR methods typically merge multiple views without considering the correlations and redundancies of lesion information across them. Therefore, we propose a novel method to overcome the challenges of difficult lesion information learning and inadequate multi-view fusion. Specifically, we introduce a two-branch network to obtain both local lesion features and their global dependencies. The high-frequency component of the wavelet transform is used to exploit lesion edge information, which is then enhanced by global semantic to facilitate difficult lesion learning. Additionally, we present a cross-view fusion module to improve multi-view fusion and reduce redundancy. Experimental results on large public datasets demonstrate the effectiveness of our method. The code is open sourced on https://github.com/HuYongting/WGLIN.
Authors: Jee Won Lee, Hansol Lim, SooYeun Yang, Jongseong Brad Choi
Abstract: This paper presents a novel masked attention-based 3D Gaussian Splatting (3DGS) approach to enhance robotic perception and object detection in industrial and smart factory environments. U2-Net is employed for background removal to isolate target objects from raw images, thereby minimizing clutter and ensuring that the model processes only relevant data. Additionally, a Sobel filter-based attention mechanism is integrated into the 3DGS framework to enhance fine details - capturing critical features such as screws, wires, and intricate textures essential for high-precision tasks. We validate our approach using quantitative metrics, including L1 loss, SSIM, PSNR, comparing the performance of the background-removed and attention-incorporated 3DGS model against the ground truth images and the original 3DGS training baseline. The results demonstrate significant improves in visual fidelity and detail preservation, highlighting the effectiveness of our method in enhancing robotic vision for object recognition and manipulation in complex industrial settings.
Authors: Muyi Bao, Shuchang Lyu, Zhaoyang Xu, Qi Zhao, Changyu Zeng, Wenpei Bai, Guangliang Cheng
Abstract: Skin lesion segmentation is a critical challenge in computer vision, and it is essential to separate pathological features from healthy skin for diagnostics accurately. Traditional Convolutional Neural Networks (CNNs) are limited by narrow receptive fields, and Transformers face significant computational burdens. This paper presents a novel skin lesion segmentation framework, the Atrous Shifted Parallel Vision Mamba UNet (ASP-VMUNet), which integrates the efficient and scalable Mamba architecture to overcome limitations in traditional CNNs and computationally demanding Transformers. The framework introduces an atrous scan technique that minimizes background interference and expands the receptive field, enhancing Mamba's scanning capabilities. Additionally, the inclusion of a Parallel Vision Mamba (PVM) layer and a shift round operation optimizes feature segmentation and fosters rich inter-segment information exchange. A supplementary CNN branch with a Selective-Kernel (SK) Block further refines the segmentation by blending local and global contextual information. Tested on four benchmark datasets (ISIC16/17/18 and PH2), ASP-VMUNet demonstrates superior performance in skin lesion segmentation, validated by comprehensive ablation studies. This approach not only advances medical image segmentation but also highlights the benefits of hybrid architectures in medical imaging technology. Our code is available at https://github.com/BaoBao0926/ASP-VMUNet/tree/main.
Authors: Masaya Hasegawa, Koji Yasuda
Abstract: Diffusion models, which have been advancing rapidly in recent years, may generate samples that closely resemble the training data. This phenomenon, known as memorization, may lead to copyright issues. In this study, we propose a method to quantify the ease of reproducing training data in unconditional diffusion models. The average of a sample population following the Langevin equation in the reverse diffusion process moves according to a first-order ordinary differential equation (ODE). This ODE establishes a 1-to-1 correspondence between images and their noisy counterparts in the latent space. Since the ODE is reversible and the initial noisy images are sampled randomly, the volume of an image's projected area represents the probability of generating those images. We examined the ODE, which projects images to latent space, and succeeded in quantifying the ease of reproducing training data by measuring the volume growth rate in this process. Given the relatively low computational complexity of this method, it allows us to enhance the quality of training data by detecting and modifying the easily memorized training samples.
Authors: Stefano Della Fiore, Alessandro Gnutti, Marco Dalai, Pierangelo Migliorati, Riccardo Leonardi
Abstract: Traditional image compression methods aim to faithfully reconstruct images for human perception. In contrast, Coding for Machines focuses on compressing images to preserve information relevant to a specific machine task. In this paper, we present an image compression system designed to retain text-specific features for subsequent Optical Character Recognition (OCR). Our encoding process requires half the time needed by the OCR module, making it especially suitable for devices with limited computational capacity. In scenarios where on-device OCR is computationally prohibitive, images are compressed and later processed to recover the text content. Experimental results demonstrate that our method achieves significant improvements in text extraction accuracy at low bitrates, even improving over the accuracy of OCR performed on uncompressed images, thus acting as a local pre-processing step.
Authors: Xiaohui Sun, Jiangwei Mo, Hanlin Wu, Jie Ma
Abstract: Recent advancements in diffusion models (DMs) have greatly advanced remote sensing image super-resolution (RSISR). However, their iterative sampling processes often result in slow inference speeds, limiting their application in real-time tasks. To address this challenge, we propose the latent consistency model for super-resolution (LCMSR), a novel single-step diffusion approach designed to enhance both efficiency and visual quality in RSISR tasks. Our proposal is structured into two distinct stages. In the first stage, we pretrain a residual autoencoder to encode the differential information between high-resolution (HR) and low-resolution (LR) images, transitioning the diffusion process into a latent space to reduce computational costs. The second stage focuses on consistency diffusion learning, which aims to learn the distribution of residual encodings in the latent space, conditioned on LR images. The consistency constraint enforces that predictions at any two timesteps along the reverse diffusion trajectory remain consistent, enabling direct mapping from noise to data. As a result, the proposed LCMSR reduces the iterative steps of traditional diffusion models from 50-1000 or more to just a single step, significantly improving efficiency. Experimental results demonstrate that LCMSR effectively balances efficiency and performance, achieving inference times comparable to non-diffusion models while maintaining high-quality output.
Authors: Sheng Wang
Abstract: As robotic technologies advancing towards more complex multimodal interactions and manipulation tasks, the integration of advanced Vision-Language Models (VLMs) has become a key driver in the field. Despite progress with current methods, challenges persist in fusing depth and RGB information within 3D environments and executing tasks guided by linguistic instructions. In response to these challenges, we have enhanced the existing RoboFlamingo framework by introducing RoboFlamingo-Plus, which incorporates depth data into VLMs to significantly improve robotic manipulation performance. Our research achieves a nuanced fusion of RGB and depth information by integrating a pre-trained Vision Transformer (ViT) with a resampling technique, closely aligning this combined data with linguistic cues for superior multimodal understanding. The novelty of RoboFlamingo-Plus lies in its adaptation of inputs for depth data processing, leveraging a pre-trained resampler for depth feature extraction, and employing cross-attention mechanisms for optimal feature integration. These improvements allow RoboFlamingo-Plus to not only deeply understand 3D environments but also easily perform complex, language-guided tasks in challenging settings. Experimental results show that RoboFlamingo-Plus boosts robotic manipulation by 10-20% over current methods, marking a significant advancement. Codes and model weights are public at RoboFlamingo-Plus.
Authors: Xin Cai
Abstract: In this article, we primarily examine a variety of RL-based and RL-free methods designed to address Reinforcement Learning from Human Feedback (RLHF) and Large Reasoning Models (LRMs). We begin with a concise overview of the typical steps involved in RLHF and LRMs. Next, we reinterpret several RL-based and RL-free algorithms through the perspective of neural structured bandit prediction, providing a clear conceptual framework that uncovers a deeper connection between these seemingly distinct approaches. Following this, we briefly review some core principles of reinforcement learning, drawing attention to an often-overlooked aspect in existing RLHF studies. This leads to a detailed derivation of the standard RLHF objective within a full RL context, demonstrating its equivalence to neural structured bandit prediction. Finally, by reinvestigating the principles behind Proximal Policy Optimization (PPO), we pinpoint areas needing adjustment, which culminates in the introduction of the Generalized Reinforce Optimization (GRO) framework, seamlessly integrating RL-based and RL-free methods in RLHF. We look forward to the community's efforts to empirically validate GRO and invite constructive feedback.
Authors: Dhananjaya Jayasundara, Sudarshan Rajagopalan, Yasiru Ranasinghe, Trac D. Tran, Vishal M. Patel
Abstract: Implicit Neural Representations (INRs) are increasingly recognized as a versatile data modality for representing discretized signals, offering benefits such as infinite query resolution and reduced storage requirements. Existing signal compression approaches for INRs typically employ one of two strategies: 1. direct quantization with entropy coding of the trained INR; 2. deriving a latent code on top of the INR through a learnable transformation. Thus, their performance is heavily dependent on the quantization and entropy coding schemes employed. In this paper, we introduce SINR, an innovative compression algorithm that leverages the patterns in the vector spaces formed by weights of INRs. We compress these vector spaces using a high-dimensional sparse code within a dictionary. Further analysis reveals that the atoms of the dictionary used to generate the sparse code do not need to be learned or transmitted to successfully recover the INR weights. We demonstrate that the proposed approach can be integrated with any existing INR-based signal compression technique. Our results indicate that SINR achieves substantial reductions in storage requirements for INRs across various configurations, outperforming conventional INR-based compression baselines. Furthermore, SINR maintains high-quality decoding across diverse data modalities, including images, occupancy fields, and Neural Radiance Fields.
Authors: Shaolei Zhang, Jinyan Liu, Tianyi Qian, Xuesong Li
Abstract: Convolutional neural networks (CNNs) and transformers are widely employed in constructing UNet architectures for medical image segmentation tasks. However, CNNs struggle to model long-range dependencies, while transformers suffer from quadratic computational complexity. Recently, Mamba, a type of State Space Models, has gained attention for its exceptional ability to model long-range interactions while maintaining linear computational complexity. Despite the emergence of several Mamba-based methods, they still present the following limitations: first, their network designs generally lack perceptual capabilities for the original input data; second, they primarily focus on capturing global information, while often neglecting local details. To address these challenges, we propose a prompt-guided CNN-Mamba dual-path UNet, termed PGM-UNet, for medical image segmentation. Specifically, we introduce a prompt-guided residual Mamba module that adaptively extracts dynamic visual prompts from the original input data, effectively guiding Mamba in capturing global information. Additionally, we design a local-global information fusion network, comprising a local information extraction module, a prompt-guided residual Mamba module, and a multi-focus attention fusion module, which effectively integrates local and global information. Furthermore, inspired by Kolmogorov-Arnold Networks (KANs), we develop a multi-scale information extraction module to capture richer contextual information without altering the resolution. We conduct extensive experiments on the ISIC-2017, ISIC-2018, DIAS, and DRIVE. The results demonstrate that the proposed method significantly outperforms state-of-the-art approaches in multiple medical image segmentation tasks.
Authors: Ge Gao, Siyue Teng, Tianhao Peng, Fan Zhang, David Bull
Abstract: While video compression based on implicit neural representations (INRs) has recently demonstrated great potential, existing INR-based video codecs still cannot achieve state-of-the-art (SOTA) performance compared to their conventional or autoencoder-based counterparts given the same coding configuration. In this context, we propose a Generative Implicit Video Compression framework, GIViC, aiming at advancing the performance limits of this type of coding methods. GIViC is inspired by the characteristics that INRs share with large language and diffusion models in exploiting long-term dependencies. Through the newly designed implicit diffusion process, GIViC performs diffusive sampling across coarse-to-fine spatiotemporal decompositions, gradually progressing from coarser-grained full-sequence diffusion to finer-grained per-token diffusion. A novel Hierarchical Gated Linear Attention-based transformer (HGLA), is also integrated into the framework, which dual-factorizes global dependency modeling along scale and sequential axes. The proposed GIViC model has been benchmarked against SOTA conventional and neural codecs using a Random Access (RA) configuration (YUV 4:2:0, GOPSize=32), and yields BD-rate savings of 15.94%, 22.46% and 8.52% over VVC VTM, DCVC-FM and NVRC, respectively. As far as we are aware, GIViC is the first INR-based video codec that outperforms VTM based on the RA coding configuration. The source code will be made available.
Authors: Deepti Madurai Muthu, Priyanka S, Lalitha Rani N, P. G. Kubendran Amos
Abstract: Reliable quantification of Ki-67, a key proliferation marker in breast cancer, is essential for molecular subtyping and informed treatment planning. Conventional approaches, including visual estimation and manual counting, suffer from interobserver variability and limited reproducibility. This study introduces an AI-assisted method using the YOLOv8 object detection framework for automated Ki-67 scoring. High-resolution digital images (40x magnification) of immunohistochemically stained tumor sections were captured from Ki-67 hotspot regions and manually annotated by a domain expert to distinguish Ki-67-positive and negative tumor cells. The dataset was augmented and divided into training (80%), validation (10%), and testing (10%) subsets. Among the YOLOv8 variants tested, the Medium model achieved the highest performance, with a mean Average Precision at 50% Intersection over Union (mAP50) exceeding 85% for Ki-67-positive cells. The proposed approach offers an efficient, scalable, and objective alternative to conventional scoring methods, supporting greater consistency in Ki-67 evaluation. Future directions include developing user-friendly clinical interfaces and expanding to multi-institutional datasets to enhance generalizability and facilitate broader adoption in diagnostic practice.
Authors: Federico Lincetto, Gianluca Agresti, Mattia Rossi, Pietro Zanuttigh
Abstract: Neural Radiance Fields (NeRF) have shown impressive performances in the rendering of 3D scenes from arbitrary viewpoints. While RGB images are widely preferred for training volume rendering models, the interest in other radiance modalities is also growing. However, the capability of the underlying implicit neural models to learn and transfer information across heterogeneous imaging modalities has seldom been explored, mostly due to the limited training data availability. For this purpose, we present MultimodalStudio (MMS): it encompasses MMS-DATA and MMS-FW. MMS-DATA is a multimodal multi-view dataset containing 32 scenes acquired with 5 different imaging modalities: RGB, monochrome, near-infrared, polarization and multispectral. MMS-FW is a novel modular multimodal NeRF framework designed to handle multimodal raw data and able to support an arbitrary number of multi-channel devices. Through extensive experiments, we demonstrate that MMS-FW trained on MMS-DATA can transfer information between different imaging modalities and produce higher quality renderings than using single modalities alone. We publicly release the dataset and the framework, to promote the research on multimodal volume rendering and beyond.
Authors: Yusen Xie, Zhengmin Huang, Shaojie Shen, Jun Ma
Abstract: In this paper, we introduce Semi-SD, a novel metric depth estimation framework tailored for surrounding cameras equipment in autonomous driving. In this work, the input data consists of adjacent surrounding frames and camera parameters. We propose a unified spatial-temporal-semantic fusion module to construct the visual fused features. Cross-attention components for surrounding cameras and adjacent frames are utilized to focus on metric scale information refinement and temporal feature matching. Building on this, we propose a pose estimation framework using surrounding cameras, their corresponding estimated depths, and extrinsic parameters, which effectively address the scale ambiguity in multi-camera setups. Moreover, semantic world model and monocular depth estimation world model are integrated to supervised the depth estimation, which improve the quality of depth estimation. We evaluate our algorithm on DDAD and nuScenes datasets, and the results demonstrate that our method achieves state-of-the-art performance in terms of surrounding camera based depth estimation quality. The source code will be available on https://github.com/xieyuser/Semi-SD.
Authors: Francisco Mena, Diego Arenas, Miro Miranda, Andreas Dengel
Abstract: In recent years, the development of robust multi-source models has emerged in the Earth Observation (EO) field. These are models that leverage data from diverse sources to improve predictive accuracy when there is missing data. Despite these advancements, the factors influencing the varying effectiveness of such models remain poorly understood. In this study, we evaluate the predictive performance of six state-of-the-art multi-source models in predicting scenarios where either a single data source is missing or only a single source is available. Our analysis reveals that the efficacy of these models is intricately tied to the nature of the task, the complementarity among data sources, and the model design. Surprisingly, we observe instances where the removal of certain data sources leads to improved predictive performance, challenging the assumption that incorporating all available data is always beneficial. These findings prompt critical reflections on model complexity and the necessity of all collected data sources, potentially shaping the way for more streamlined approaches in EO applications.
Authors: Zixue Zeng, Matthew Cartier, Xiaoyan Zhao, Pengyu Chen, Xin Meng, Zhiyu Sheng, Maryam Satarpour, John M Cormack, Allison C. Bean, Ryan P. Nussbaum, Maya Maurer, Emily Landis-Walkenhorst, Kang Kim, Ajay D. Wasan, Jiantao Pu
Abstract: Available studies on chronic lower back pain (cLBP) typically focus on one or a few specific tissues rather than conducting a comprehensive layer-by-layer analysis. Since three-dimensional (3-D) images often contain hundreds of slices, manual annotation of these anatomical structures is both time-consuming and error-prone. We aim to develop and validate a novel approach called InterSliceBoost to enable the training of a segmentation model on a partially annotated dataset without compromising segmentation performance. The architecture of InterSliceBoost includes two components: an inter-slice generator and a segmentation model. The generator utilizes residual block-based encoders to extract features from adjacent image-mask pairs (IMPs). Differential features are calculated and input into a decoder to generate inter-slice IMPs. The segmentation model is trained on partially annotated datasets (e.g., skipping 1, 2, 3, or 7 images) and the generated inter-slice IMPs. To validate the performance of InterSliceBoost, we utilized a dataset of 76 B-mode ultrasound scans acquired on 29 subjects enrolled in an ongoing cLBP study. InterSliceBoost, trained on only 33% of the image slices, achieved a mean Dice coefficient of 80.84% across all six layers on the independent test set, with Dice coefficients of 73.48%, 61.11%, 81.87%, 95.74%, 83.52% and 88.74% for segmenting dermis, superficial fat, superficial fascial membrane, deep fat, deep fascial membrane, and muscle. This performance is significantly higher than the conventional model trained on fully annotated images (p<0.05). InterSliceBoost can effectively segment the six tissue layers depicted on 3-D B-model ultrasound images in settings with partial annotations.
Authors: Zixue Zeng, Xiaoyan Zhao, Matthew Cartier, Xin Meng, Jiantao Pu
Abstract: 3D ultrasound delivers high-resolution, real-time images of soft tissues, which is essential for pain research. However, manually distinguishing various tissues for quantitative analysis is labor-intensive. To streamline this process, we developed and validated GRN+, a novel multi-model framework that automates layer segmentation with minimal annotated data. GRN+ combines a ResNet-based generator and a U-Net segmentation model. Through a method called Segmentation-guided Enhancement (SGE), the generator produces new images and matching masks under the guidance of the segmentation model, with its weights adjusted according to the segmentation loss gradient. To prevent gradient explosion and secure stable training, a two-stage backpropagation strategy was implemented: the first stage propagates the segmentation loss through both the generator and segmentation model, while the second stage concentrates on optimizing the segmentation model alone, thereby refining mask prediction using the generated images. Tested on 69 fully annotated 3D ultrasound scans from 29 subjects with six manually labeled tissue layers, GRN+ outperformed all other semi-supervised methods in terms of the Dice coefficient using only 5% labeled data, despite not using unlabeled data for unsupervised training. Additionally, when applied to fully annotated datasets, GRN+ with SGE achieved a 2.16% higher Dice coefficient while incurring lower computational costs compared to other models. Overall, GRN+ provides accurate tissue segmentation while reducing both computational expenses and the dependency on extensive annotations, making it an effective tool for 3D ultrasound analysis in cLBP patients.
Authors: Chuanzhi Xu, Haoxian Zhou, Haodong Chen, Vera Chung, Qiang Qu
Abstract: Event cameras have gained increasing attention for 3D reconstruction due to their high temporal resolution, low latency, and high dynamic range. They capture per-pixel brightness changes asynchronously, allowing accurate reconstruction under fast motion and challenging lighting conditions. In this survey, we provide a comprehensive review of event-driven 3D reconstruction methods, including stereo, monocular, and multimodal systems. We further categorize recent developments based on geometric, learning-based, and hybrid approaches. Emerging trends, such as neural radiance fields and 3D Gaussian splatting with event data, are also covered. The related works are structured chronologically to illustrate the innovations and progression within the field. To support future research, we also highlight key research gaps and future research directions in dataset, experiment, evaluation, event representation, etc.
Authors: Zhi Hou, Tianyi Zhang, Yuwen Xiong, Haonan Duan, Hengjun Pu, Ronglei Tong, Chengyang Zhao, Xizhou Zhu, Yu Qiao, Jifeng Dai, Yuntao Chen
Abstract: While recent vision-language-action models trained on diverse robot datasets exhibit promising generalization capabilities with limited in-domain data, their reliance on compact action heads to predict discretized or continuous actions constrains adaptability to heterogeneous action spaces. We present Dita, a scalable framework that leverages Transformer architectures to directly denoise continuous action sequences through a unified multimodal diffusion process. Departing from prior methods that condition denoising on fused embeddings via shallow networks, Dita employs in-context conditioning -- enabling fine-grained alignment between denoised actions and raw visual tokens from historical observations. This design explicitly models action deltas and environmental nuances. By scaling the diffusion action denoiser alongside the Transformer's scalability, Dita effectively integrates cross-embodiment datasets across diverse camera perspectives, observation scenes, tasks, and action spaces. Such synergy enhances robustness against various variances and facilitates the successful execution of long-horizon tasks. Evaluations across extensive benchmarks demonstrate state-of-the-art or comparative performance in simulation. Notably, Dita achieves robust real-world adaptation to environmental variances and complex long-horizon tasks through 10-shot finetuning, using only third-person camera inputs. The architecture establishes a versatile, lightweight and open-source baseline for generalist robot policy learning. Project Page: https://robodita.github.io.
Authors: Pratibha Kumari, Afshin Bozorgpour, Daniel Reisenb\"uchler, Edgar Jost, Martina Crysandt, Christian Matek, Dorit Merhof
Abstract: White blood cell (WBC) classification plays a vital role in hematology for diagnosing various medical conditions. However, it faces significant challenges due to domain shifts caused by variations in sample sources (e.g., blood or bone marrow) and differing imaging conditions across hospitals. Traditional deep learning models often suffer from catastrophic forgetting in such dynamic environments, while foundation models, though generally robust, experience performance degradation when the distribution of inference data differs from that of the training data. To address these challenges, we propose a generative replay-based Continual Learning (CL) strategy designed to prevent forgetting in foundation models for WBC classification. Our method employs lightweight generators to mimic past data with a synthetic latent representation to enable privacy-preserving replay. To showcase the effectiveness, we carry out extensive experiments with a total of four datasets with different task ordering and four backbone models including ResNet50, RetCCL, CTransPath, and UNI. Experimental results demonstrate that conventional fine-tuning methods degrade performance on previously learned tasks and struggle with domain shifts. In contrast, our continual learning strategy effectively mitigates catastrophic forgetting, preserving model performance across varying domains. This work presents a practical solution for maintaining reliable WBC classification in real-world clinical settings, where data distributions frequently evolve.
Authors: Yan Zhuang, Minheng Chen, Chao Cao, Tong Chen, Jing Zhang, Xiaowei Yu, Yanjun Lyu, Lu Zhang, Tianming Liu, Dajiang Zhu
Abstract: Understanding the structural and functional organization of the human brain requires a detailed examination of cortical folding patterns, among which the three-hinge gyrus (3HG) has been identified as a key structural landmark. GyralNet, a network representation of cortical folding, models 3HGs as nodes and gyral crests as edges, highlighting their role as critical hubs in cortico-cortical connectivity. However, existing methods for analyzing 3HGs face significant challenges, including the sub-voxel scale of 3HGs at typical neuroimaging resolutions, the computational complexity of establishing cross-subject correspondences, and the oversimplification of treating 3HGs as independent nodes without considering their community-level relationships. To address these limitations, we propose a fully differentiable subnetwork partitioning framework that employs a spectral modularity maximization optimization strategy to modularize the organization of 3HGs within GyralNet. By incorporating topological structural similarity and DTI-derived connectivity patterns as attribute features, our approach provides a biologically meaningful representation of cortical organization. Extensive experiments on the Human Connectome Project (HCP) dataset demonstrate that our method effectively partitions GyralNet at the individual level while preserving the community-level consistency of 3HGs across subjects, offering a robust foundation for understanding brain connectivity.
Authors: Jiazhi Guan, Kaisiyuan Wang, Zhiliang Xu, Quanwei Yang, Yasheng Sun, Shengyi He, Borong Liang, Yukang Cao, Yingying Li, Haocheng Feng, Errui Ding, Jingdong Wang, Youjian Zhao, Hang Zhou, Ziwei Liu
Abstract: Despite the recent progress of audio-driven video generation, existing methods mostly focus on driving facial movements, leading to non-coherent head and body dynamics. Moving forward, it is desirable yet challenging to generate holistic human videos with both accurate lip-sync and delicate co-speech gestures w.r.t. given audio. In this work, we propose AudCast, a generalized audio-driven human video generation framework adopting a cascade Diffusion-Transformers (DiTs) paradigm, which synthesizes holistic human videos based on a reference image and a given audio. 1) Firstly, an audio-conditioned Holistic Human DiT architecture is proposed to directly drive the movements of any human body with vivid gesture dynamics. 2) Then to enhance hand and face details that are well-knownly difficult to handle, a Regional Refinement DiT leverages regional 3D fitting as the bridge to reform the signals, producing the final results. Extensive experiments demonstrate that our framework generates high-fidelity audio-driven holistic human videos with temporal coherence and fine facial and hand details. Resources can be found at https://guanjz20.github.io/projects/AudCast.
Authors: Junzhi Ning, Dominic Marshall, Yijian Gao, Xiaodan Xing Yang Nan, Yingying Fang, Sheng Zhang, Matthieu Komorowski, Guang Yang
Abstract: Chest X-ray radiographs (CXRs) play a pivotal role in diagnosing and monitoring cardiopulmonary diseases. However, lung opac- ities in CXRs frequently obscure anatomical structures, impeding clear identification of lung borders and complicating the localization of pathology. This challenge significantly hampers segmentation accuracy and precise lesion identification, which are crucial for diagnosis. To tackle these issues, our study proposes an unpaired CXR translation framework that converts CXRs with lung opacities into counterparts without lung opacities while preserving semantic features. Central to our approach is the use of adaptive activation masks to selectively modify opacity regions in lung CXRs. Cross-domain alignment ensures translated CXRs without opacity issues align with feature maps and prediction labels from a pre-trained CXR lesion classifier, facilitating the interpretability of the translation process. We validate our method using RSNA, MIMIC-CXR-JPG and JSRT datasets, demonstrating superior translation quality through lower Frechet Inception Distance (FID) and Kernel Inception Distance (KID) scores compared to existing meth- ods (FID: 67.18 vs. 210.4, KID: 0.01604 vs. 0.225). Evaluation on RSNA opacity, MIMIC acute respiratory distress syndrome (ARDS) patient CXRs and JSRT CXRs show our method enhances segmentation accuracy of lung borders and improves lesion classification, further underscoring its potential in clinical settings (RSNA: mIoU: 76.58% vs. 62.58%, Sensitivity: 85.58% vs. 77.03%; MIMIC ARDS: mIoU: 86.20% vs. 72.07%, Sensitivity: 92.68% vs. 86.85%; JSRT: mIoU: 91.08% vs. 85.6%, Sensitivity: 97.62% vs. 95.04%). Our approach advances CXR imaging analysis, especially in investigating segmentation impacts through image translation techniques.
Authors: Sungyeon Kim, Xinliang Zhu, Xiaofan Lin, Muhammet Bastan, Douglas Gray, Suha Kwak
Abstract: Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency.
Authors: Lukas Mack, Felix Gr\"uninger, Benjamin A. Richardson, Regine Lendway, Katherine J. Kuchenbecker, Joerg Stueckler
Abstract: Accurate 3D pose estimation of grasped objects is an important prerequisite for robots to perform assembly or in-hand manipulation tasks, but object occlusion by the robot's own hand greatly increases the difficulty of this perceptual task. Here, we propose that combining visual information and proprioception with binary, low-resolution tactile contact measurements from across the interior surface of an articulated robotic hand can mitigate this issue. The visuo-tactile object-pose-estimation problem is formulated probabilistically in a factor graph. The pose of the object is optimized to align with the three kinds of measurements using a robust cost function to reduce the influence of visual or tactile outlier readings. The advantages of the proposed approach are first demonstrated in simulation: a custom 15-DoF robot hand with one binary tactile sensor per link grasps 17 YCB objects while observed by an RGB-D camera. This low-resolution in-hand tactile sensing significantly improves object-pose estimates under high occlusion and also high visual noise. We also show these benefits through grasping tests with a preliminary real version of our tactile hand, obtaining reasonable visuo-tactile estimates of object pose at approximately 13.3 Hz on average.
Authors: Weide Liu, Zhonghua Wu, Henghui Ding, Fayao Liu, Jie Lin, Guosheng Lin, Wei Zhou
Abstract: In this work, we address the challenging task of few-shot segmentation. Previous few-shot segmentation methods mainly employ the information of support images as guidance for query image segmentation. Although some works propose to build cross-reference between support and query images, their extraction of query information still depends on the support images. We here propose to extract the information from the query itself independently to benefit the few-shot segmentation task. To this end, we first propose a prior extractor to learn the query information from the unlabeled images with our proposed global-local contrastive learning. Then, we extract a set of predetermined priors via this prior extractor. With the obtained priors, we generate the prior region maps for query images, which locate the objects, as guidance to perform cross interaction with support features. In such a way, the extraction of query information is detached from the support branch, overcoming the limitation by support, and could obtain more informative query clues to achieve better interaction. Without bells and whistles, the proposed approach achieves new state-of-the-art performance for the few-shot segmentation task on PASCAL-5$^{i}$ and COCO datasets.
Authors: Peng Gao, Shijie Geng, Renrui Zhang, Teli Ma, Rongyao Fang, Yongfeng Zhang, Hongsheng Li, Yu Qiao
Abstract: Large-scale contrastive vision-language pre-training has shown significant progress in visual representation learning. Unlike traditional visual systems trained by a fixed set of discrete labels, a new paradigm was introduced in \cite{radford2021learning} to directly learn to align images with raw texts in an open-vocabulary setting. On downstream tasks, a carefully chosen text prompt is employed to make zero-shot predictions.~To avoid non-trivial prompt engineering, context optimization \cite{zhou2021coop} has been proposed to learn continuous vectors as task-specific prompts with few-shot training examples.~In this paper, we show that there is an alternative path to achieve better vision-language models other than prompt tuning.~While prompt tuning is for the textual inputs, we propose CLIP-Adapter to conduct fine-tuning with feature adapters on either visual or language branch. Specifically, CLIP-Adapter adopts an additional bottleneck layer to learn new features and performs residual-style feature blending with the original pre-trained features.~As a consequence, CLIP-Adapter is able to outperform context optimization while maintains a simple design. Experiments and extensive ablation studies on various visual classification tasks demonstrate the effectiveness of our approach. Code is released at t https://github.com/gaopengcuhk/CLIP-Adapter.
Authors: Oliver Boyne, James Charles, Roberto Cipolla
Abstract: In this paper we present a high fidelity and articulated 3D human foot model. The model is parameterised by a disentangled latent code in terms of shape, texture and articulated pose. While high fidelity models are typically created with strong supervision such as 3D keypoint correspondences or pre-registration, we focus on the difficult case of little to no annotation. To this end, we make the following contributions: (i) we develop a Foot Implicit Neural Deformation field model, named FIND, capable of tailoring explicit meshes at any resolution i.e. for low or high powered devices; (ii) an approach for training our model in various modes of weak supervision with progressively better disentanglement as more labels, such as pose categories, are provided; (iii) a novel unsupervised part-based loss for fitting our model to 2D images which is better than traditional photometric or silhouette losses; (iv) finally, we release a new dataset of high resolution 3D human foot scans, Foot3D. On this dataset, we show our model outperforms a strong PCA implementation trained on the same data in terms of shape quality and part correspondences, and that our novel unsupervised part-based loss improves inference on images.
Authors: Bin Wang, Wenbin Pei, Bing Xue, Mengjie Zhang
Abstract: Deep convolutional neural networks have proven their effectiveness, and have been acknowledged as the most dominant method for image classification. However, a severe drawback of deep convolutional neural networks is poor explainability. Unfortunately, in many real-world applications, users need to understand the rationale behind the predictions of deep convolutional neural networks when determining whether they should trust the predictions or not. To resolve this issue, a novel genetic algorithm-based method is proposed for the first time to automatically evolve local explanations that can assist users to assess the rationality of the predictions. Furthermore, the proposed method is model-agnostic, i.e., it can be utilised to explain any deep convolutional neural network models. In the experiments, ResNet is used as an example model to be explained, and the ImageNet dataset is selected as the benchmark dataset. DenseNet and MobileNet are further explained to demonstrate the model-agnostic characteristic of the proposed method. The evolved local explanations on four images, randomly selected from ImageNet, are presented, which show that the evolved local explanations are straightforward to be recognised by humans. Moreover, the evolved explanations can explain the predictions of deep convolutional neural networks on all four images very well by successfully capturing meaningful interpretable features of the sample images. Further analysis based on the 30 runs of the experiments exhibits that the evolved local explanations can also improve the probabilities/confidences of the deep convolutional neural network models in making the predictions. The proposed method can obtain local explanations within one minute, which is more than ten times faster than LIME (the state-of-the-art method).
Authors: Chuang Zhu, Kebin Liu, Wenqi Tang, Ke Mei, Jiaqi Zou, Tiejun Huang
Abstract: The divergence between labeled training data and unlabeled testing data is a significant challenge for recent deep learning models. Unsupervised domain adaptation (UDA) attempts to solve such problem. Recent works show that self-training is a powerful approach to UDA. However, existing methods have difficulty in balancing the scalability and performance. In this paper, we propose a hard-aware instance adaptive self-training framework for UDA on the task of semantic segmentation. To effectively improve the quality and diversity of pseudo-labels, we develop a novel pseudo-label generation strategy with an instance adaptive selector. We further enrich the hard class pseudo-labels with inter-image information through a skillfully designed hard-aware pseudo-label augmentation. Besides, we propose the region-adaptive regularization to smooth the pseudo-label region and sharpen the non-pseudo-label region. For the non-pseudo-label region, consistency constraint is also constructed to introduce stronger supervision signals during model optimization. Our method is so concise and efficient that it is easy to be generalized to other UDA methods. Experiments on GTA5 to Cityscapes, SYNTHIA to Cityscapes, and Cityscapes to Oxford RobotCar demonstrate the superior performance of our approach compared with the state-of-the-art methods. Our codes are available at https://github.com/bupt-ai-cz/HIAST.
Authors: Mingjie Li, Ben Beck, Tharindu Rathnayake, Lingheng Meng, Zijue Chen, Akansel Cosgun, Xiaojun Chang, Dana Kuli\'c
Abstract: Cycling is a healthy and sustainable mode of transport. However, interactions with motor vehicles remain a key barrier to increased cycling participation. The ability to detect potentially dangerous interactions from on-bike sensing could provide important information to riders and policymakers. A key influence on rider comfort and safety is close passes, i.e., when a vehicle narrowly passes a cyclist. In this paper, we introduce a novel benchmark, called Cyc-CP, towards close pass (CP) event detection from video streams. The task is formulated into two problem categories: scene-level and instance-level. Scene-level detection ascertains the presence of a CP event within the provided video clip. Instance-level detection identifies the specific vehicle within the scene that precipitates a CP event. To address these challenges, we introduce four benchmark models, each underpinned by advanced deep-learning methodologies. For training and evaluating those models, we have developed a synthetic dataset alongside the acquisition of a real-world dataset. The benchmark evaluations reveal that the models achieve an accuracy of 88.13\% for scene-level detection and 84.60\% for instance-level detection on the real-world dataset. We envision this benchmark as a test-bed to accelerate CP detection and facilitate interaction between the fields of road safety, intelligent transportation systems and artificial intelligence. Both the benchmark datasets and detection models will be available at https://github.com/SustainableMobility/cyc-cp to facilitate experimental reproducibility and encourage more in-depth research in the field.
Authors: Ze-Xin Yin, Peng-Yi Jiao, Jiaxiong Qiu, Ming-Ming Cheng, Bo Ren
Abstract: Existing Neural Radiance Fields (NeRF) methods suffer from the existence of reflective objects, often resulting in blurry or distorted rendering. Instead of calculating a single radiance field, we propose a multi-space neural radiance field (MS-NeRF) that represents the scene using a group of feature fields in parallel sub-spaces, which leads to a better understanding of the neural network toward the existence of reflective and refractive objects. Our multi-space scheme works as an enhancement to existing NeRF methods, with only small computational overheads needed for training and inferring the extra-space outputs. We design different multi-space modules for representative MLP-based and grid-based NeRF methods, which improve Mip-NeRF 360 by 4.15 dB in PSNR with 0.5% extra parameters and further improve TensoRF by 2.71 dB with 0.046% extra parameters on reflective regions without degrading the rendering quality on other regions. We further construct a novel dataset consisting of 33 synthetic scenes and 7 real captured scenes with complex reflection and refraction, where we design complex camera paths to fully benchmark the robustness of NeRF-based methods. Extensive experiments show that our approach significantly outperforms the existing single-space NeRF methods for rendering high-quality scenes concerned with complex light paths through mirror-like objects. The source code, dataset, and results are available via our project page: https://zx-yin.github.io/msnerf/.
Authors: Byeongho Heo, Taekyung Kim, Sangdoo Yun, Dongyoon Han
Abstract: Pre-training with random masked inputs has emerged as a novel trend in self-supervised training. However, supervised learning still faces a challenge in adopting masking augmentations, primarily due to unstable training. In this paper, we propose a novel way to involve masking augmentations dubbed Masked Sub-branch (MaskSub). MaskSub consists of the main-branch and sub-branch, the latter being a part of the former. The main-branch undergoes conventional training recipes, while the sub-branch merits intensive masking augmentations, during training. MaskSub tackles the challenge by mitigating adverse effects through a relaxed loss function similar to a self-distillation loss. Our analysis shows that MaskSub improves performance, with the training loss converging faster than in standard training, which suggests our method stabilizes the training process. We further validate MaskSub across diverse training scenarios and models, including DeiT-III training, MAE finetuning, CLIP finetuning, BERT training, and hierarchical architectures (ResNet and Swin Transformer). Our results show that MaskSub consistently achieves impressive performance gains across all the cases. MaskSub provides a practical and effective solution for introducing additional regularization under various training recipes. Code available at https://github.com/naver-ai/augsub
Authors: Xianjie Liu, Keren Fu, Yao Jiang, Qijun Zhao
Abstract: The Segment Anything Model (SAM) represents a significant breakthrough into foundation models for computer vision, providing a large-scale image segmentation model. However, despite SAM's zero-shot performance, its segmentation masks lack fine-grained details, particularly in accurately delineating object boundaries. Therefore, it is both interesting and valuable to explore whether SAM can be improved towards highly accurate object segmentation, which is known as the dichotomous image segmentation (DIS) task. To address this issue, we propose DIS-SAM, which advances SAM towards DIS with extremely accurate details. DIS-SAM is a framework specifically tailored for highly accurate segmentation, maintaining SAM's promptable design. DIS-SAM employs a two-stage approach, integrating SAM with a modified advanced network that was previously designed to handle the prompt-free DIS task. To better train DIS-SAM, we employ a ground truth enrichment strategy by modifying original mask annotations. Despite its simplicity, DIS-SAM significantly advances the SAM, HQ-SAM, and Pi-SAM ~by 8.5%, ~6.9%, and ~3.7% maximum F-measure. Our code at https://github.com/Tennine2077/DIS-SAM
Authors: Yi Xin, Jianjiang Yang, Siqi Luo, Haodi Zhou, Junlong Du, Xiaohong Liu, Yue Fan, Qing Li, Yuntao Du
Abstract: Large-scale pre-trained vision models (PVMs) have shown great potential for adaptability across various downstream vision tasks. However, with state-of-the-art PVMs growing to billions or even trillions of parameters, the standard full fine-tuning paradigm is becoming unsustainable due to high computational and storage demands. In response, researchers are exploring parameter-efficient fine-tuning (PEFT), which seeks to exceed the performance of full fine-tuning with minimal parameter modifications. This survey provides a comprehensive overview and future directions for visual PEFT, offering a systematic review of the latest advancements. First, we provide a formal definition of PEFT and discuss model pre-training methods. We then categorize existing methods into three categories: addition-based, partial-based, and unified-based. Finally, we introduce the commonly used datasets and applications and suggest potential future research challenges. A comprehensive collection of resources is available at https://github.com/synbol/Awesome-Parameter-Efficient-Transfer-Learning.
URLs: https://github.com/synbol/Awesome-Parameter-Efficient-Transfer-Learning.
Authors: Kun Li, George Vosselman, Michael Ying Yang
Abstract: Visual Question Answering (VQA) is a challenging task of predicting the answer to a question about the content of an image. It requires deep understanding of both the textual question and visual image. Prior works directly evaluate the answering models by simply calculating the accuracy of the predicted answers. However, the inner reasoning behind the prediction is disregarded in such a "black box" system, and we do not even know if one can trust the predictions. In some cases, the models still get the correct answers even when they focus on irrelevant visual regions or textual tokens, which makes the models unreliable and illogical. To generate both visual and textual rationales next to the predicted answer to the given image/question pair, we propose Multimodal Rationales for VQA, MRVQA. Considering the extra annotations brought by the new outputs, MRVQA is trained and evaluated by samples converted from some existing VQA datasets and their visual labels. The extensive experiments demonstrate that the visual and textual rationales support the prediction of the answers, and further improve the accuracy. Furthermore, MRVQA achieves competitive performance on generic VQA datatsets in the zero-shot evaluation setting. The dataset and source code will be released under https://github.com/lik1996/CRVQA2024.
Authors: Yiming Ma, Victor Sanchez, Tanaya Guha
Abstract: We propose CLIP-EBC, the first fully CLIP-based model for accurate crowd density estimation. While the CLIP model has demonstrated remarkable success in addressing recognition tasks such as zero-shot image classification, its potential for counting has been largely unexplored due to the inherent challenges in transforming a regression problem, such as counting, into a recognition task. In this work, we investigate and enhance CLIP's ability to count, focusing specifically on the task of estimating crowd sizes from images. Existing classification-based crowd-counting frameworks have significant limitations, including the quantization of count values into bordering real-valued bins and the sole focus on classification errors. These practices result in label ambiguity near the shared borders and inaccurate prediction of count values. Hence, directly applying CLIP within these frameworks may yield suboptimal performance. To address these challenges, we first propose the Enhanced Blockwise Classification (EBC) framework. Unlike previous methods, EBC utilizes integer-valued bins, effectively reducing ambiguity near bin boundaries. Additionally, it incorporates a regression loss based on density maps to improve the prediction of count values. Within our backbone-agnostic EBC framework, we then introduce CLIP-EBC to fully leverage CLIP's recognition capabilities for this task. Extensive experiments demonstrate the effectiveness of EBC and the competitive performance of CLIP-EBC. Specifically, our EBC framework can improve existing classification-based methods by up to 44.5% on the UCF-QNRF dataset, and CLIP-EBC achieves state-of-the-art performance on the NWPU-Crowd test set, with an MAE of 58.2 and an RMSE of 268.5, representing improvements of 8.6% and 13.3% over the previous best method, STEERER. The code and weights are available at https://github.com/Yiming-M/CLIP-EBC.
Authors: Marvin Anas Hahn, Kathl\'en Kohn, Orlando Marigliano, Tomas Pajdla
Abstract: Rolling shutter (RS) cameras dominate consumer and smartphone markets. Several methods for computing the absolute pose of RS cameras have appeared in the last 20 years, but the relative pose problem has not been fully solved yet. We provide a unified theory for the important class of order-one rolling shutter (RS$_1$) cameras. These cameras generalize the perspective projection to RS cameras, projecting a generic space point to exactly one image point via a rational map. We introduce a new back-projection RS camera model, characterize RS$_1$ cameras, construct explicit parameterizations of such cameras, and determine the image of a space line. We classify all minimal problems for solving the relative camera pose problem with linear RS$_1$ cameras and discover new practical cases. Finally, we show how the theory can be used to explain RS models previously used for absolute pose computation.
Authors: Stella Bounareli, Christos Tzelepis, Vasileios Argyriou, Ioannis Patras, Georgios Tzimiropoulos
Abstract: Video-driven neural face reenactment aims to synthesize realistic facial images that successfully preserve the identity and appearance of a source face, while transferring the target head pose and facial expressions. Existing GAN-based methods suffer from either distortions and visual artifacts or poor reconstruction quality, i.e., the background and several important appearance details, such as hair style/color, glasses and accessories, are not faithfully reconstructed. Recent advances in Diffusion Probabilistic Models (DPMs) enable the generation of high-quality realistic images. To this end, in this paper we present DiffusionAct, a novel method that leverages the photo-realistic image generation of diffusion models to perform neural face reenactment. Specifically, we propose to control the semantic space of a Diffusion Autoencoder (DiffAE), in order to edit the facial pose of the input images, defined as the head pose orientation and the facial expressions. Our method allows one-shot, self, and cross-subject reenactment, without requiring subject-specific fine-tuning. We compare against state-of-the-art GAN-, StyleGAN2-, and diffusion-based methods, showing better or on-par reenactment performance.
Authors: Wencheng Zhu, Xin Zhou, Pengfei Zhu, Yu Wang, Qinghua Hu
Abstract: In this paper, we propose a simple yet effective contrastive knowledge distillation framework that achieves sample-wise logit alignment while preserving semantic consistency. Conventional knowledge distillation approaches exhibit over-reliance on feature similarity per sample, which risks overfitting, and contrastive approaches focus on inter-class discrimination at the expense of intra-sample semantic relationships. Our approach transfers "dark knowledge" through teacher-student contrastive alignment at the sample level. Specifically, our method first enforces intra-sample alignment by directly minimizing teacher-student logit discrepancies within individual samples. Then, we utilize inter-sample contrasts to preserve semantic dissimilarities across samples. By redefining positive pairs as aligned teacher-student logits from identical samples and negative pairs as cross-sample logit combinations, we reformulate these dual constraints into an InfoNCE loss framework, reducing computational complexity lower than sample squares while eliminating dependencies on temperature parameters and large batch sizes. We conduct comprehensive experiments across three benchmark datasets, including the CIFAR-100, ImageNet-1K, and MS COCO datasets, and experimental results clearly confirm the effectiveness of the proposed method on image classification, object detection, and instance segmentation tasks.
Authors: Tunhou Zhang, Shiyu Li, Hsin-Pai Cheng, Feng Yan, Hai Li, Yiran Chen
Abstract: Exploring dense connectivity of convolutional operators establishes critical "synapses" to communicate feature vectors from different levels and enriches the set of transformations on Computer Vision applications. Yet, even with heavy-machinery approaches such as Neural Architecture Search (NAS), discovering effective connectivity patterns requires tremendous efforts due to either constrained connectivity design space or a sub-optimal exploration process induced by an unconstrained search space. In this paper, we propose CSCO, a novel paradigm that fabricates effective connectivity of convolutional operators with minimal utilization of existing design motifs and further utilizes the discovered wiring to construct high-performing ConvNets. CSCO guides the exploration via a neural predictor as a surrogate of the ground-truth performance. We introduce Graph Isomorphism as data augmentation to improve sample efficiency and propose a Metropolis-Hastings Evolutionary Search (MH-ES) to evade locally optimal architectures and advance search quality. Results on ImageNet show ~0.6% performance improvement over hand-crafted and NAS-crafted dense connectivity. Our code is publicly available.
Authors: Chongwei Liu, Haojie Li, Zhihui Wang, Rui Xu
Abstract: Recent advances in Multi-Object Tracking (MOT) have demonstrated significant success in short-term association within the separated tracking-by-detection online paradigm. However, long-term tracking remains challenging. While graph-based approaches address this by modeling trajectories as global graphs, these methods are unsuitable for real-time applications due to their non-online nature. In this paper, we review the concept of trajectory graphs and propose a novel perspective by representing them as directed acyclic graphs. This representation can be described using frame-ordered object sequences and binary adjacency matrices. We observe that this structure naturally aligns with Transformer attention mechanisms, enabling us to model the association problem using a classic Transformer architecture. Based on this insight, we introduce a concise Pure Transformer (PuTR) to validate the effectiveness of Transformer in unifying short- and long-term tracking for separated online MOT. Extensive experiments on four diverse datasets (SportsMOT, DanceTrack, MOT17, and MOT20) demonstrate that PuTR effectively establishes a solid baseline compared to existing foundational online methods while exhibiting superior domain adaptation capabilities. Furthermore, the separated nature enables efficient training and inference, making it suitable for practical applications. Implementation code and trained models are available at https://github.com/chongweiliu/PuTR .
Authors: Laura Fieback, Jakob Spiegelberg, Hanno Gottschalk
Abstract: Large Vision Language Models (LVLMs) have shown remarkable capabilities in multimodal tasks like visual question answering or image captioning. However, inconsistencies between the visual information and the generated text, a phenomenon referred to as hallucinations, remain an unsolved problem with regard to the trustworthiness of LVLMs. To address this problem, recent works proposed to incorporate computationally costly Large (Vision) Language Models in order to detect hallucinations on a sentence- or subsentence-level. In this work, we introduce MetaToken, a lightweight binary classifier to detect hallucinations on the token-level at negligible cost. Based on a statistical analysis, we reveal key factors of hallucinations in LVLMs. MetaToken can be applied to any open-source LVLM without any knowledge about ground truth data providing a calibrated detection of hallucinations. We evaluate our method on four state-of-the-art LVLMs demonstrating the effectiveness of our approach.
Authors: Chun-Hung Wu, Shih-Hong Chen, Chih-Yao Hu, Hsin-Yu Wu, Kai-Hsin Chen, Yu-You Chen, Chih-Hai Su, Chih-Kuo Lee, Yu-Lun Liu
Abstract: This paper presents Deformable Neural Vessel Representations (DeNVeR), an unsupervised approach for vessel segmentation in X-ray angiography videos without annotated ground truth. DeNVeR utilizes optical flow and layer separation techniques, enhancing segmentation accuracy and adaptability through test-time training. Key contributions include a novel layer separation bootstrapping technique, a parallel vessel motion loss, and the integration of Eulerian motion fields for modeling complex vessel dynamics. A significant component of this research is the introduction of the XACV dataset, the first X-ray angiography coronary video dataset with high-quality, manually labeled segmentation ground truth. Extensive evaluations on both XACV and CADICA datasets demonstrate that DeNVeR outperforms current state-of-the-art methods in vessel segmentation accuracy and generalization capability while maintaining temporal coherency.
Authors: Zhanhao Liang, Yuhui Yuan, Shuyang Gu, Bohan Chen, Tiankai Hang, Mingxi Cheng, Ji Li, Liang Zheng
Abstract: Generating visually appealing images is fundamental to modern text-to-image generation models. A potential solution to better aesthetics is direct preference optimization (DPO), which has been applied to diffusion models to improve general image quality including prompt alignment and aesthetics. Popular DPO methods propagate preference labels from clean image pairs to all the intermediate steps along the two generation trajectories. However, preference labels provided in existing datasets are blended with layout and aesthetic opinions, which would disagree with aesthetic preference. Even if aesthetic labels were provided (at substantial cost), it would be hard for the two-trajectory methods to capture nuanced visual differences at different steps. To improve aesthetics economically, this paper uses existing generic preference data and introduces step-by-step preference optimization (SPO) that discards the propagation strategy and allows fine-grained image details to be assessed. Specifically, at each denoising step, we 1) sample a pool of candidates by denoising from a shared noise latent, 2) use a step-aware preference model to find a suitable win-lose pair to supervise the diffusion model, and 3) randomly select one from the pool to initialize the next denoising step. This strategy ensures that diffusion models focus on the subtle, fine-grained visual differences instead of layout aspect. We find that aesthetics can be significantly enhanced by accumulating these improved minor differences. When fine-tuning Stable Diffusion v1.5 and SDXL, SPO yields significant improvements in aesthetics compared with existing DPO methods while not sacrificing image-text alignment compared with vanilla models. Moreover, SPO converges much faster than DPO methods due to the use of more correct preference labels provided by the step-aware preference model.
Authors: Leigang Qu, Haochuan Li, Tan Wang, Wenjie Wang, Yongqi Li, Liqiang Nie, Tat-Seng Chua
Abstract: How humans can effectively and efficiently acquire images has always been a perennial question. A classic solution is text-to-image retrieval from an existing database; however, the limited database typically lacks creativity. By contrast, recent breakthroughs in text-to-image generation have made it possible to produce attractive and counterfactual visual content, but it faces challenges in synthesizing knowledge-intensive images. In this work, we rethink the relationship between text-to-image generation and retrieval, proposing a unified framework for both tasks with one single Large Multimodal Model (LMM). Specifically, we first explore the intrinsic discriminative abilities of LMMs and introduce an efficient generative retrieval method for text-to-image retrieval in a training-free manner. Subsequently, we unify generation and retrieval autoregressively and propose an autonomous decision mechanism to choose the best-matched one between generated and retrieved images as the response to the text prompt. To standardize the evaluation of unified text-to-image generation and retrieval, we construct TIGeR-Bench, a benchmark spanning both creative and knowledge-intensive domains. Extensive experiments on TIGeR-Bench and two retrieval benchmarks, i.e., Flickr30K and MS-COCO, demonstrate the superiority of our proposed framework.
Authors: Arman Zarei, Keivan Rezaei, Samyadeep Basu, Mehrdad Saberi, Mazda Moayeri, Priyatham Kattakinda, Soheil Feizi
Abstract: Text-to-image diffusion-based generative models have the stunning ability to generate photo-realistic images and achieve state-of-the-art low FID scores on challenging image generation benchmarks. However, one of the primary failure modes of these text-to-image generative models is in composing attributes, objects, and their associated relationships accurately into an image. In our paper, we investigate compositional attribute binding failures, where the model fails to correctly associate descriptive attributes (such as color, shape, or texture) with the corresponding objects in the generated images, and highlight that imperfect text conditioning with CLIP text-encoder is one of the primary reasons behind the inability of these models to generate high-fidelity compositional scenes. In particular, we show that (i) there exists an optimal text-embedding space that can generate highly coherent compositional scenes showing that the output space of the CLIP text-encoder is sub-optimal, and (ii) the final token embeddings in CLIP are erroneous as they often include attention contributions from unrelated tokens in compositional prompts. Our main finding shows that significant compositional improvements can be achieved (without harming the model's FID score) by fine-tuning only a simple and parameter-efficient linear projection on CLIP's representation space in Stable-Diffusion variants using a small set of compositional image-text pairs.
Authors: Hong-Xing Yu, Haoyi Duan, Charles Herrmann, William T. Freeman, Jiajun Wu
Abstract: We present WonderWorld, a novel framework for interactive 3D scene generation that enables users to interactively specify scene contents and layout and see the created scenes in low latency. The major challenge lies in achieving fast generation of 3D scenes. Existing scene generation approaches fall short of speed as they often require (1) progressively generating many views and depth maps, and (2) time-consuming optimization of the scene geometry representations. We introduce the Fast Layered Gaussian Surfels (FLAGS) as our scene representation and an algorithm to generate it from a single view. Our approach does not need multiple views, and it leverages a geometry-based initialization that significantly reduces optimization time. Another challenge is generating coherent geometry that allows all scenes to be connected. We introduce the guided depth diffusion that allows partial conditioning of depth estimation. WonderWorld generates connected and diverse 3D scenes in less than 10 seconds on a single A6000 GPU, enabling real-time user interaction and exploration. We demonstrate the potential of WonderWorld for user-driven content creation and exploration in virtual environments. We release full code and software for reproducibility. Project website: https://kovenyu.com/WonderWorld/.
Authors: Alex Hanson, Allen Tu, Vasu Singla, Mayuka Jayawardhana, Matthias Zwicker, Tom Goldstein
Abstract: Recent advances in novel view synthesis have enabled real-time rendering speeds with high reconstruction accuracy. 3D Gaussian Splatting (3D-GS), a foundational point-based parametric 3D scene representation, models scenes as large sets of 3D Gaussians. However, complex scenes can consist of millions of Gaussians, resulting in high storage and memory requirements that limit the viability of 3D-GS on devices with limited resources. Current techniques for compressing these pretrained models by pruning Gaussians rely on combining heuristics to determine which Gaussians to remove. At high compression ratios, these pruned scenes suffer from heavy degradation of visual fidelity and loss of foreground details. In this paper, we propose a principled sensitivity pruning score that preserves visual fidelity and foreground details at significantly higher compression ratios than existing approaches. It is computed as a second-order approximation of the reconstruction error on the training views with respect to the spatial parameters of each Gaussian. Additionally, we propose a multi-round prune-refine pipeline that can be applied to any pretrained 3D-GS model without changing its training pipeline. After pruning 90% of Gaussians, a substantially higher percentage than previous methods, our PUP 3D-GS pipeline increases average rendering speed by 3.56$\times$ while retaining more salient foreground information and achieving higher image quality metrics than existing techniques on scenes from the Mip-NeRF 360, Tanks & Temples, and Deep Blending datasets.
Authors: Rohit Bharadwaj, Hanan Gani, Muzammal Naseer, Fahad Shahbaz Khan, Salman Khan
Abstract: The recent developments in Large Multi-modal Video Models (Video-LMMs) have significantly enhanced our ability to interpret and analyze video data. Despite their impressive capabilities, current Video-LMMs have not been evaluated for anomaly detection tasks, which is critical to their deployment in practical scenarios e.g., towards identifying deepfakes, manipulated video content, traffic accidents and crimes. In this paper, we introduce VANE-Bench, a benchmark designed to assess the proficiency of Video-LMMs in detecting and localizing anomalies and inconsistencies in videos. Our dataset comprises an array of videos synthetically generated using existing state-of-the-art text-to-video generation models, encompassing a variety of subtle anomalies and inconsistencies grouped into five categories: unnatural transformations, unnatural appearance, pass-through, disappearance and sudden appearance. Additionally, our benchmark features real-world samples from existing anomaly detection datasets, focusing on crime-related irregularities, atypical pedestrian behavior, and unusual events. The task is structured as a visual question-answering challenge to gauge the models' ability to accurately detect and localize the anomalies within the videos. We evaluate nine existing Video-LMMs, both open and closed sources, on this benchmarking task and find that most of the models encounter difficulties in effectively identifying the subtle anomalies. In conclusion, our research offers significant insights into the current capabilities of Video-LMMs in the realm of anomaly detection, highlighting the importance of our work in evaluating and improving these models for real-world applications. Our code and data is available at https://hananshafi.github.io/vane-benchmark/
Authors: Yongting Zhang, Lu Chen, Guodong Zheng, Yifeng Gao, Rui Zheng, Jinlan Fu, Zhenfei Yin, Senjie Jin, Yu Qiao, Xuanjing Huang, Feng Zhao, Tao Gui, Jing Shao
Abstract: The emergence of Vision Language Models (VLMs) has brought unprecedented advances in understanding multimodal information. The combination of textual and visual semantics in VLMs is highly complex and diverse, making the safety alignment of these models challenging. Furthermore, due to the limited study on the safety alignment of VLMs, there is a lack of large-scale, high-quality datasets. To address these limitations, we propose a Safety Preference Alignment dataset for Vision Language Models named SPA-VL. In terms of breadth, SPA-VL covers 6 harmfulness domains, 13 categories, and 53 subcategories, and contains 100,788 samples of the quadruple (question, image, chosen response, rejected response). In terms of depth, the responses are collected from 12 open-source (e.g., QwenVL) and closed-source (e.g., Gemini) VLMs to ensure diversity. The construction of preference data is fully automated, and the experimental results indicate that models trained with alignment techniques on the SPA-VL dataset exhibit substantial improvements in harmlessness and helpfulness while maintaining core capabilities. SPA-VL, as a large-scale, high-quality, and diverse dataset, represents a significant milestone in ensuring that VLMs achieve both harmlessness and helpfulness.
Authors: Tianyu Wei, Shanmin Pang, Qi Guo, Yizhuo Ma, Xiaofeng Cao, Ming-Ming Cheng, Qing Guo
Abstract: Text-to-image diffusion models can generate realistic images based on textual inputs, enabling users to convey their opinions visually through language. Meanwhile, within language, emotion plays a crucial role in expressing personal opinions in our daily lives and the inclusion of maliciously negative content can lead users astray, exacerbating negative emotions. Recognizing the success of diffusion models and the significance of emotion, we investigate a previously overlooked risk associated with text-to-image diffusion models, that is, utilizing emotion in the input texts to introduce negative content and provoke unfavorable emotions in users. Specifically, we identify a new backdoor attack, i.e., emotion-aware backdoor attack (EmoAttack), which introduces malicious negative content triggered by emotional texts during image generation. We formulate such an attack as a diffusion personalization problem to avoid extensive model retraining and propose the EmoBooth. Unlike existing personalization methods, our approach fine-tunes a pre-trained diffusion model by establishing a mapping between a cluster of emotional words and a given reference image containing malicious negative content. To validate the effectiveness of our method, we built a dataset and conducted extensive analysis and discussion about its effectiveness. Given consumers' widespread use of diffusion models, uncovering this threat is critical for society.
Authors: Chang-Han Yeh, Chin-Yang Lin, Zhixiang Wang, Chi-Wei Hsiao, Ting-Hsuan Chen, Hau-Shiang Shiu, Yu-Lun Liu
Abstract: We present DiffIR2VR-Zero, a zero-shot framework that enables any pre-trained image restoration diffusion model to perform high-quality video restoration without additional training. While image diffusion models have shown remarkable restoration capabilities, their direct application to video leads to temporal inconsistencies, and existing video restoration methods require extensive retraining for different degradation types. Our approach addresses these challenges through two key innovations: a hierarchical latent warping strategy that maintains consistency across both keyframes and local frames, and a hybrid token merging mechanism that adaptively combines optical flow and feature matching. Through extensive experiments, we demonstrate that our method not only maintains the high-quality restoration of base diffusion models but also achieves superior temporal consistency across diverse datasets and degradation conditions, including challenging scenarios like 8$\times$ super-resolution and severe noise. Importantly, our framework works with any image restoration diffusion model, providing a versatile solution for video enhancement without task-specific training or modifications.
Authors: Ali Hatamizadeh, Jan Kautz
Abstract: We propose a novel hybrid Mamba-Transformer backbone, MambaVision, specifically tailored for vision applications. Our core contribution includes redesigning the Mamba formulation to enhance its capability for efficient modeling of visual features. Through a comprehensive ablation study, we demonstrate the feasibility of integrating Vision Transformers (ViT) with Mamba. Our results show that equipping the Mamba architecture with self-attention blocks in the final layers greatly improves its capacity to capture long-range spatial dependencies. Based on these findings, we introduce a family of MambaVision models with a hierarchical architecture to meet various design criteria. For classification on the ImageNet-1K dataset, MambaVision variants achieve state-of-the-art (SOTA) performance in terms of both Top-1 accuracy and throughput. In downstream tasks such as object detection, instance segmentation, and semantic segmentation on MS COCO and ADE20K datasets, MambaVision outperforms comparably sized backbones while demonstrating favorable performance. Code: https://github.com/NVlabs/MambaVision
Authors: Sarah Rockstroh, Patrick Frenzel, Daniel Matthes, Kay Schubert, David Wollburg, Mirco Fuchs
Abstract: Assessing an athlete's performance in canoe sprint is often established by measuring a variety of kinematic parameters during training sessions. Many of these parameters are related to single or multiple paddle stroke cycles. Determining on- and offset of these cycles in force sensor signals is usually not straightforward and requires human interaction. This paper explores convolutional neural networks (CNNs) and recurrent neural networks (RNNs) in terms of their ability to automatically predict these events. In addition, our work proposes an extension to the recently published SoftED metric for event detection in order to properly assess the model performance on time windows. In our results, an RNN based on bidirectional gated recurrent units (BGRUs) turned out to be the most suitable model for paddle stroke detection.
Authors: Yingxue Xu, Yihui Wang, Fengtao Zhou, Jiabo Ma, Cheng Jin, Shu Yang, Jinbang Li, Zhengyu Zhang, Chenglong Zhao, Huajun Zhou, Zhenhui Li, Huangjing Lin, Xin Wang, Jiguang Wang, Anjia Han, Ronald Cheong Kin Chan, Li Liang, Xiuming Zhang, Hao Chen
Abstract: Remarkable strides in computational pathology have been made in the task-agnostic foundation model that advances the performance of a wide array of downstream clinical tasks. Despite the promising performance, there are still several challenges. First, prior works have resorted to either vision-only or image-caption data, disregarding pathology reports with more clinically authentic information from pathologists and gene expression profiles which respectively offer distinct knowledge for versatile clinical applications. Second, the current progress in pathology FMs predominantly concentrates on the patch level, where the restricted context of patch-level pretraining fails to capture whole-slide patterns. Even recent slide-level FMs still struggle to provide whole-slide context for patch representation. In this study, for the first time, we develop a pathology foundation model incorporating three levels of modalities: pathology slides, pathology reports, and gene expression data, which resulted in 26,169 slide-level modality pairs from 10,275 patients across 32 cancer types, amounting to over 116 million pathological patch images. To leverage these data for CPath, we propose a novel whole-slide pretraining paradigm that injects the multimodal whole-slide context into the patch representation, called Multimodal Self-TAught PRetraining (mSTAR). The proposed paradigm revolutionizes the pretraining workflow for CPath, enabling the pathology FM to acquire the whole-slide context. To the best of our knowledge, this is the first attempt to incorporate three modalities at the whole-slide context for enhancing pathology FMs. To systematically evaluate the capabilities of mSTAR, we built the largest spectrum of oncological benchmark, spanning 7 categories of oncological applications in 15 types of 97 practical oncological tasks.
Authors: Hinako Mitsuoka, Kazuhiro Hotta
Abstract: Semantic segmentation of microscopy cell images by deep learning is a significant technique. We considered that the Transformers, which have recently outperformed CNNs in image recognition, could also be improved and developed for cell image segmentation. Transformers tend to focus more on contextual information than on detailed information. This tendency leads to a lack of detailed information for segmentation. Therefore, to supplement or reinforce the missing detailed information, we hypothesized that feedback processing in the human visual cortex should be effective. Our proposed Feedback Former is a novel architecture for semantic segmentation, in which Transformers is used as an encoder and has a feedback processing mechanism. Feature maps with detailed information are fed back to the lower layers from near the output of the model to compensate for the lack of detailed information which is the weakness of Transformers and improve the segmentation accuracy. By experiments on three cell image datasets, we confirmed that our method surpasses methods without feedback, demonstrating its superior accuracy in cell image segmentation. Our method achieved higher segmentation accuracy while consuming less computational cost than conventional feedback approaches. Moreover, our method offered superior precision without simply increasing the model size of Transformer encoder, demonstrating higher accuracy with lower computational cost.
Authors: Andrzej Perzanowski, Tony Lindeberg
Abstract: This paper presents an in-depth analysis of the scale generalisation properties of the scale-covariant and scale-invariant Gaussian derivative networks, complemented with both conceptual and algorithmic extensions. For this purpose, Gaussian derivative networks (GaussDerNets) are evaluated on new rescaled versions of the Fashion-MNIST and the CIFAR-10 datasets, with spatial scaling variations over a factor of 4 in the testing data, that are not present in the training data. Additionally, evaluations on the previously existing STIR datasets show that the GaussDerNets achieve better scale generalisation than previously reported for these datasets for other types of deep networks. We first experimentally demonstrate that the GaussDerNets have quite good scale generalisation properties on the new datasets, and that average pooling of feature responses over scales may sometimes also lead to better results than the previously used approach of max pooling over scales. Then, we demonstrate that using a spatial max pooling mechanism after the final layer enables localisation of non-centred objects in image domain, with maintained scale generalisation properties. We also show that regularisation during training, by applying dropout across the scale channels, referred to as scale-channel dropout, improves both the performance and the scale generalisation. In additional ablation studies, we demonstrate that discretisations of GaussDerNets, based on the discrete analogue of the Gaussian kernel in combination with central difference operators, perform best or among the best, compared to a set of other discrete approximations of the Gaussian derivative kernels. Finally, by visualising the activation maps and the learned receptive fields, we demonstrate that the GaussDerNets have very good explainability properties.
Authors: Yazhou Zhu, Minxian Li, Qiaolin Ye, Shidong Wang, Tong Xin, Haofeng Zhang
Abstract: Few-shot medical image segmentation (FSMIS) aims to perform the limited annotated data learning in the medical image analysis scope. Despite the progress has been achieved, current FSMIS models are all trained and deployed on the same data domain, as is not consistent with the clinical reality that medical imaging data is always across different data domains (e.g. imaging modalities, institutions and equipment sequences). How to enhance the FSMIS models to generalize well across the different specific medical imaging domains? In this paper, we focus on the matching mechanism of the few-shot semantic segmentation models and introduce an Earth Mover's Distance (EMD) calculation based domain robust matching mechanism for the cross-domain scenario. Specifically, we formulate the EMD transportation process between the foreground support-query features, the texture structure aware weights generation method, which proposes to perform the sobel based image gradient calculation over the nodes, is introduced in the EMD matching flow to restrain the domain relevant nodes. Besides, the point set level distance measurement metric is introduced to calculated the cost for the transportation from support set nodes to query set nodes. To evaluate the performance of our model, we conduct experiments on three scenarios (i.e., cross-modal, cross-sequence and cross-institution), which includes eight medical datasets and involves three body regions, and the results demonstrate that our model achieves the SoTA performance against the compared models.
Authors: Daniel Cores, Michael Dorkenwald, Manuel Mucientes, Cees G. M. Snoek, Yuki M. Asano
Abstract: Large language models have demonstrated impressive performance when integrated with vision models even enabling video understanding. However, evaluating video models presents its own unique challenges, for which several benchmarks have been proposed. In this paper, we show that the currently most used video-language benchmarks can be solved without requiring much temporal reasoning. We identified three main issues in existing datasets: (i) static information from single frames is often sufficient to solve the tasks (ii) the text of the questions and candidate answers is overly informative, allowing models to answer correctly without relying on any visual input (iii) world knowledge alone can answer many of the questions, making the benchmarks a test of knowledge replication rather than video reasoning. In addition, we found that open-ended question-answering benchmarks for video understanding suffer from similar issues while the automatic evaluation process with LLMs is unreliable, making it an unsuitable alternative. As a solution, we propose TVBench, a novel open-source video multiple-choice question-answering benchmark, and demonstrate through extensive evaluations that it requires a high level of temporal understanding. Surprisingly, we find that most recent state-of-the-art video-language models perform similarly to random performance on TVBench, with only a few models such as Qwen2-VL, and Tarsier clearly surpassing this baseline.
Authors: Jian Huang, Chengrui Dong, Xuanhua Chen, Peidong Liu
Abstract: Implicit neural representation and explicit 3D Gaussian Splatting (3D-GS) for novel view synthesis have achieved remarkable progress with frame-based camera (e.g. RGB and RGB-D cameras) recently. Compared to frame-based camera, a novel type of bio-inspired visual sensor, i.e. event camera, has demonstrated advantages in high temporal resolution, high dynamic range, low power consumption and low latency. Due to its unique asynchronous and irregular data capturing process, limited work has been proposed to apply neural representation or 3D Gaussian splatting for an event camera. In this work, we present IncEventGS, an incremental 3D Gaussian Splatting reconstruction algorithm with a single event camera. To recover the 3D scene representation incrementally, we exploit the tracking and mapping paradigm of conventional SLAM pipelines for IncEventGS. Given the incoming event stream, the tracker firstly estimates an initial camera motion based on prior reconstructed 3D-GS scene representation. The mapper then jointly refines both the 3D scene representation and camera motion based on the previously estimated motion trajectory from the tracker. The experimental results demonstrate that IncEventGS delivers superior performance compared to prior NeRF-based methods and other related baselines, even we do not have the ground-truth camera poses. Furthermore, our method can also deliver better performance compared to state-of-the-art event visual odometry methods in terms of camera motion estimation. Code is publicly available at: https://github.com/wu-cvgl/IncEventGS.
Authors: Yujian Zhao, Chengru Wu, Yinong Xu, Xuanzheng Du, Ruiyu Li, Guanglin Niu
Abstract: Cloth-changing person re-identification (CC-ReID), also known as Long-Term Person Re-Identification (LT-ReID) is a critical and challenging research topic in computer vision that has recently garnered significant attention. However, due to the high cost of constructing CC-ReID data, the existing data-driven models are hard to train efficiently on limited data, causing overfitting issue. To address this challenge, we propose a low-cost and efficient pipeline for generating controllable and high-quality synthetic data simulating the surveillance of real scenarios specific to the CC-ReID task. Particularly, we construct a new self-annotated CC-ReID dataset named Cloth-Changing Unreal Person (CCUP), containing 6,000 IDs, 1,179,976 images, 100 cameras, and 26.5 outfits per individual. Based on this large-scale dataset, we introduce an effective and scalable pretrain-finetune framework for enhancing the generalization capabilities of the traditional CC-ReID models. The extensive experiments demonstrate that two typical models namely TransReID and FIRe^2, when integrated into our framework, outperform other state-of-the-art models after pretraining on CCUP and finetuning on the benchmarks such as PRCC, VC-Clothes and NKUP. The CCUP is available at: https://github.com/yjzhao1019/CCUP.
Authors: Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum, Daniel Barath, Andreas Geiger, Marc Pollefeys
Abstract: Gaussian splatting and single-view depth estimation are typically studied in isolation. In this paper, we present DepthSplat to connect Gaussian splatting and depth estimation and study their interactions. More specifically, we first contribute a robust multi-view depth model by leveraging pre-trained monocular depth features, leading to high-quality feed-forward 3D Gaussian splatting reconstructions. We also show that Gaussian splatting can serve as an unsupervised pre-training objective for learning powerful depth models from large-scale multi-view posed datasets. We validate the synergy between Gaussian splatting and depth estimation through extensive ablation and cross-task transfer experiments. Our DepthSplat achieves state-of-the-art performance on ScanNet, RealEstate10K and DL3DV datasets in terms of both depth estimation and novel view synthesis, demonstrating the mutual benefits of connecting both tasks. In addition, DepthSplat enables feed-forward reconstruction from 12 input views (512x960 resolutions) in 0.6 seconds.
Authors: Li Chaorong, Ling Xudong, Yang Qiang, Qin Fengqing, Huang Yuanyuan
Abstract: Deep learning models have achieved remarkable progress in precipitation prediction. However, they still face significant challenges in accurately capturing spatial details of radar images, particularly in regions of high precipitation intensity. This limitation results in reduced spatial localization accuracy when predicting radar echo images across varying precipitation intensities. To address this challenge, we propose an innovative precipitation prediction approach termed the Multi-Task Latent Diffusion Model (MTLDM). The core idea of MTLDM lies in the recognition that precipitation radar images represent a combination of multiple components, each corresponding to different precipitation intensities. Thus, we adopt a divide-and-conquer strategy, decomposing radar images into several sub-images based on their precipitation intensities and individually modeling these components. During the prediction stage, MTLDM integrates these sub-image representations by utilizing a trained latent-space rainfall diffusion model, followed by decoding through a multi-task decoder to produce the final precipitation prediction. Experimental evaluations conducted on the MRMS dataset demonstrate that the proposed MTLDM method surpasses state-of-the-art techniques, achieving a Critical Success Index (CSI) improvement of 13-26%.
Authors: Josiah Aklilu, Xiaohan Wang, Serena Yeung-Levy
Abstract: Precise action localization in untrimmed video is vital for fields such as professional sports and minimally invasive surgery, where the delineation of particular motions in recordings can dramatically enhance analysis. But in many cases, large scale datasets with video-label pairs for localization are unavailable, limiting the opportunity to fine-tune video-understanding models. Recent developments in large vision-language models (LVLM) address this need with impressive zero-shot capabilities in a variety of video understanding tasks. However, the adaptation of LVLMs, with their powerful visual question answering capabilities, to zero-shot localization in long-form video is still relatively unexplored. To this end, we introduce a true Zero-shot Action Localization method (ZEAL). Specifically, we leverage the built-in action knowledge of a large language model (LLM) to inflate actions into detailed descriptions of the archetypal start and end of the action. These descriptions serve as queries to LVLM for generating frame-level confidence scores which can be aggregated to produce localization outputs. The simplicity and flexibility of our method lends it amenable to more capable LVLMs as they are developed, and we demonstrate remarkable results in zero-shot action localization on a challenging benchmark, without any training. Our code is publicly available at $\href{https://github.com/josaklil-ai/zeal}{github.com/josaklil-ai/zeal}$.
Authors: Chin-Yang Lin, Chung-Ho Wu, Chang-Han Yeh, Shih-Han Yen, Cheng Sun, Yu-Lun Liu
Abstract: Neural Radiance Fields (NeRF) face significant challenges in extreme few-shot scenarios, primarily due to overfitting and long training times. Existing methods, such as FreeNeRF and SparseNeRF, use frequency regularization or pre-trained priors but struggle with complex scheduling and bias. We introduce FrugalNeRF, a novel few-shot NeRF framework that leverages weight-sharing voxels across multiple scales to efficiently represent scene details. Our key contribution is a cross-scale geometric adaptation scheme that selects pseudo ground truth depth based on reprojection errors across scales. This guides training without relying on externally learned priors, enabling full utilization of the training data. It can also integrate pre-trained priors, enhancing quality without slowing convergence. Experiments on LLFF, DTU, and RealEstate-10K show that FrugalNeRF outperforms other few-shot NeRF methods while significantly reducing training time, making it a practical solution for efficient and accurate 3D scene reconstruction.
Authors: Cheng-De Fan, Chen-Wei Chang, Yi-Ruei Liu, Jie-Ying Lee, Jiun-Long Huang, Yu-Chee Tseng, Yu-Lun Liu
Abstract: We present SpectroMotion, a novel approach that combines 3D Gaussian Splatting (3DGS) with physically-based rendering (PBR) and deformation fields to reconstruct dynamic specular scenes. Previous methods extending 3DGS to model dynamic scenes have struggled to represent specular surfaces accurately. Our method addresses this limitation by introducing a residual correction technique for accurate surface normal computation during deformation, complemented by a deformable environment map that adapts to time-varying lighting conditions. We implement a coarse-to-fine training strategy significantly enhancing scene geometry and specular color prediction. It is the only existing 3DGS method capable of synthesizing photorealistic real-world dynamic specular scenes, outperforming state-of-the-art methods in rendering complex, dynamic, and specular scenes.
Authors: Lehan Wang, Haonan Wang, Honglong Yang, Jiaji Mao, Zehong Yang, Jun Shen, Xiaomeng Li
Abstract: Several medical Multimodal Large Languange Models (MLLMs) have been developed to address tasks involving visual images with textual instructions across various medical modalities, achieving impressive results. Most current medical generalist models are region-agnostic, treating the entire image as a holistic representation. However, they struggle to identify which specific regions they are focusing on when generating a sentence. To mimic the behavior of doctors, who typically begin by reviewing the entire image before concentrating on specific regions for a thorough evaluation, we aim to enhance the capability of medical MLLMs in understanding anatomical regions within entire medical scans. To achieve it, we first formulate Region-Centric tasks and construct a large-scale dataset, MedRegInstruct, to incorporate regional information into training. Combining our collected dataset with other medical multimodal corpora for training, we propose a Region-Aware medical MLLM, MedRegA, which is the first bilingual generalist medical AI system to simultaneously handle image-level and region-level medical vision-language tasks across a broad range of modalities. Our MedRegA not only enables three region-centric tasks, but also achieves the best performance for visual question answering, report generation and medical image classification over 8 modalities, showcasing significant versatility. Experiments demonstrate that our model can not only accomplish powerful performance across various medical vision-language tasks in bilingual settings, but also recognize and detect structures in multimodal medical scans, boosting the interpretability and user interactivity of medical MLLMs. Our project page is https://medrega.github.io.
Authors: Shehan Munasinghe, Hanan Gani, Wenqi Zhu, Jiale Cao, Eric Xing, Fahad Shahbaz Khan, Salman Khan
Abstract: Fine-grained alignment between videos and text is challenging due to complex spatial and temporal dynamics in videos. Existing video-based Large Multimodal Models (LMMs) handle basic conversations but struggle with precise pixel-level grounding in videos. To address this, we introduce VideoGLaMM, a LMM designed for fine-grained pixel-level grounding in videos based on user-provided textual inputs. Our design seamlessly connects three key components: a Large Language Model, a dual vision encoder that emphasizes both spatial and temporal details, and a spatio-temporal decoder for accurate mask generation. This connection is facilitated via tunable V-L and L-V adapters that enable close Vision-Language (VL) alignment. The architecture is trained to synchronize both spatial and temporal elements of video content with textual instructions. To enable fine-grained grounding, we curate a multimodal dataset featuring detailed visually-grounded conversations using a semiautomatic annotation pipeline, resulting in a diverse set of 38k video-QA triplets along with 83k objects and 671k masks. We evaluate VideoGLaMM on three challenging tasks: Grounded Conversation Generation, Visual Grounding, and Referring Video Segmentation. Experimental results show that our model consistently outperforms existing approaches across all three tasks.
Authors: Yudong Han, Qingpei Guo, Liyuan Pan, Liu Liu, Yu Guan, Ming Yang
Abstract: The challenge in LLM-based video understanding lies in preserving visual and semantic information in long videos while maintaining a memory-affordable token count. However, redundancy and correspondence in videos have hindered the performance potential of existing methods. Through statistical learning on current datasets, we observe that redundancy occurs in both repeated and answer-irrelevant frames, and the corresponding frames vary with different questions. This suggests the possibility of adopting dynamic encoding to balance detailed video information preservation with token budget reduction. To this end, we propose a dynamic cooperative network, DynFocus, for memory-efficient video encoding in this paper. Specifically, i) a Dynamic Event Prototype Estimation (DPE) module to dynamically select meaningful frames for question answering; (ii) a Compact Cooperative Encoding (CCE) module that encodes meaningful frames with detailed visual appearance and the remaining frames with sketchy perception separately. We evaluate our method on five publicly available benchmarks, and experimental results consistently demonstrate that our method achieves competitive performance.
Authors: Rohith Peddi, Saurabh, Ayush Abhay Shrivastava, Parag Singla, Vibhav Gogate
Abstract: Spatio-Temporal Scene Graphs (STSGs) provide a concise and expressive representation of dynamic scenes by modeling objects and their evolving relationships over time. However, real-world visual relationships often exhibit a long-tailed distribution, causing existing methods for tasks like Video Scene Graph Generation (VidSGG) and Scene Graph Anticipation (SGA) to produce biased scene graphs. To this end, we propose ImparTail, a novel training framework that leverages loss masking and curriculum learning to mitigate bias in the generation and anticipation of spatio-temporal scene graphs. Unlike prior methods that add extra architectural components to learn unbiased estimators, we propose an impartial training objective that reduces the dominance of head classes during learning and focuses on underrepresented tail relationships. Our curriculum-driven mask generation strategy further empowers the model to adaptively adjust its bias mitigation strategy over time, enabling more balanced and robust estimations. To thoroughly assess performance under various distribution shifts, we also introduce two new tasks Robust Spatio-Temporal Scene Graph Generation and Robust Scene Graph Anticipation offering a challenging benchmark for evaluating the resilience of STSG models. Extensive experiments on the Action Genome dataset demonstrate the superior unbiased performance and robustness of our method compared to existing baselines.
Authors: Kaisheng Liang, Xuelong Dai, Yanjie Li, Dong Wang, Bin Xiao
Abstract: Deep neural networks (DNNs) exhibit vulnerability to adversarial examples that can transfer across different DNN models. A particularly challenging problem is developing transferable targeted attacks that can mislead DNN models into predicting specific target classes. While various methods have been proposed to enhance attack transferability, they often incur substantial computational costs while yielding limited improvements. Recent clean feature mixup methods use random clean features to perturb the feature space but lack optimization for disrupting adversarial examples, overlooking the advantages of attack-specific perturbations. In this paper, we propose Feature Tuning Mixup (FTM), a novel method that enhances targeted attack transferability by combining both random and optimized noises in the feature space. FTM introduces learnable feature perturbations and employs an efficient stochastic update strategy for optimization. These learnable perturbations facilitate the generation of more robust adversarial examples with improved transferability. We further demonstrate that attack performance can be enhanced through an ensemble of multiple FTM-perturbed surrogate models. Extensive experiments on the ImageNet-compatible dataset across various DNN models demonstrate that our method achieves significant improvements over state-of-the-art methods while maintaining low computational cost.
Authors: Pengcheng Xu, Boyuan Jiang, Xiaobin Hu, Donghao Luo, Qingdong He, Jiangning Zhang, Chengjie Wang, Yunsheng Wu, Charles Ling, Boyu Wang
Abstract: Leveraging the large generative prior of the flow transformer for tuning-free image editing requires authentic inversion to project the image into the model's domain and a flexible invariance control mechanism to preserve non-target contents. However, the prevailing diffusion inversion performs deficiently in flow-based models, and the invariance control cannot reconcile diverse rigid and non-rigid editing tasks. To address these, we systematically analyze the \textbf{inversion and invariance} control based on the flow transformer. Specifically, we unveil that the Euler inversion shares a similar structure to DDIM yet is more susceptible to the approximation error. Thus, we propose a two-stage inversion to first refine the velocity estimation and then compensate for the leftover error, which pivots closely to the model prior and benefits editing. Meanwhile, we propose the invariance control that manipulates the text features within the adaptive layer normalization, connecting the changes in the text prompt to image semantics. This mechanism can simultaneously preserve the non-target contents while allowing rigid and non-rigid manipulation, enabling a wide range of editing types such as visual text, quantity, facial expression, etc. Experiments on versatile scenarios validate that our framework achieves flexible and accurate editing, unlocking the potential of the flow transformer for versatile image editing.
Authors: Ruoyu Chen, Siyuan Liang, Jingzhi Li, Shiming Liu, Maosen Li, Zhen Huang, Hua Zhang, Xiaochun Cao
Abstract: Advances in multimodal pre-training have propelled object-level foundation models, such as Grounding DINO and Florence-2, in tasks like visual grounding and object detection. However, interpreting these models' decisions has grown increasingly challenging. Existing interpretable attribution methods for object-level task interpretation have notable limitations: (1) gradient-based methods lack precise localization due to visual-textual fusion in foundation models, and (2) perturbation-based methods produce noisy saliency maps, limiting fine-grained interpretability. To address these, we propose a Visual Precision Search method that generates accurate attribution maps with fewer regions. Our method bypasses internal model parameters to overcome attribution issues from multimodal fusion, dividing inputs into sparse sub-regions and using consistency and collaboration scores to accurately identify critical decision-making regions. We also conducted a theoretical analysis of the boundary guarantees and scope of applicability of our method. Experiments on RefCOCO, MS COCO, and LVIS show our approach enhances object-level task interpretability over SOTA for Grounding DINO and Florence-2 across various evaluation metrics, with faithfulness gains of 23.7%, 31.6%, and 20.1% on MS COCO, LVIS, and RefCOCO for Grounding DINO, and 102.9% and 66.9% on MS COCO and RefCOCO for Florence-2. Additionally, our method can interpret failures in visual grounding and object detection tasks, surpassing existing methods across multiple evaluation metrics. The code will be released at https://github.com/RuoyuChen10/VPS.
Authors: Chan Hee Song, Valts Blukis, Jonathan Tremblay, Stephen Tyree, Yu Su, Stan Birchfield
Abstract: Spatial understanding is a crucial capability that enables robots to perceive their surroundings, reason about their environment, and interact with it meaningfully. In modern robotics, these capabilities are increasingly provided by vision-language models. However, these models face significant challenges in spatial reasoning tasks, as their training data are based on general-purpose image datasets that often lack sophisticated spatial understanding. For example, datasets frequently do not capture reference frame comprehension, yet effective spatial reasoning requires understanding whether to reason from ego-, world-, or object-centric perspectives. To address this issue, we introduce RoboSpatial, a large-scale dataset for spatial understanding in robotics. It consists of real indoor and tabletop scenes, captured as 3D scans and egocentric images, and annotated with rich spatial information relevant to robotics. The dataset includes 1M images, 5k 3D scans, and 3M annotated spatial relationships, and the pairing of 2D egocentric images with 3D scans makes it both 2D- and 3D- ready. Our experiments show that models trained with RoboSpatial outperform baselines on downstream tasks such as spatial affordance prediction, spatial relationship prediction, and robotics manipulation.
Authors: Shenghe Zheng, Hongzhi Wang
Abstract: With the rapid growth of deep learning, there is an increasing availability of open-source models for various tasks. However, single fine-tuned models often fall short of meeting the diverse needs of users. Model merging has thus emerged as an efficient method to integrate the capabilities of existing models into a unified model. Nevertheless, existing model merging methods face challenging trade-offs between performance and deployment costs, primarily due to task interference. For the first time, we reveal that task interference is evident in the frequency domain of model parameters, yet current efforts only focus on spatial domain solutions, which are largely ineffective in addressing frequency domain interference. To mitigate the impact of frequency domain interference, we propose FR-Merging, an innovative method that effectively filters harmful frequency domain interference on the backbone with minimal computational overhead. Since performance loss is inevitable with cost-free methods, we propose a lightweight task-specific expert module that dynamically compensates for information loss during merging. This proposed framework, FREE-Merging (FR-Merging with experts), strikes a balanced trade-off between training cost, inference latency, storage requirements, and performance. We demonstrate the effectiveness of both FR-Merging and FREE-Merging on multiple tasks across CV, NLP, and Multi-Modal domains and show that they can be flexibly adapted to specific needs.
Authors: Srikar Yellapragada, Alexandros Graikos, Kostas Triaridis, Prateek Prasanna, Rajarsi R. Gupta, Joel Saltz, Dimitris Samaras
Abstract: Diffusion models have revolutionized image generation, yet several challenges restrict their application to large-image domains, such as digital pathology and satellite imagery. Given that it is infeasible to directly train a model on 'whole' images from domains with potential gigapixel sizes, diffusion-based generative methods have focused on synthesizing small, fixed-size patches extracted from these images. However, generating small patches has limited applicability since patch-based models fail to capture the global structures and wider context of large images, which can be crucial for synthesizing (semantically) accurate samples. To overcome this limitation, we present ZoomLDM, a diffusion model tailored for generating images across multiple scales. Central to our approach is a novel magnification-aware conditioning mechanism that utilizes self-supervised learning (SSL) embeddings and allows the diffusion model to synthesize images at different 'zoom' levels, i.e., fixed-size patches extracted from large images at varying scales. ZoomLDM synthesizes coherent histopathology images that remain contextually accurate and detailed at different zoom levels, achieving state-of-the-art image generation quality across all scales and excelling in the data-scarce setting of generating thumbnails of entire large images. The multi-scale nature of ZoomLDM unlocks additional capabilities in large image generation, enabling computationally tractable and globally coherent image synthesis up to $4096 \times 4096$ pixels and $4\times$ super-resolution. Additionally, multi-scale features extracted from ZoomLDM are highly effective in multiple instance learning experiments.
Authors: Gyeongjin Kang, Jisang Yoo, Jihyeon Park, Seungtae Nam, Sangpil Kim, Hyeonsoo Im, Sangheon Shin, Eunbyung Park
Abstract: We propose SelfSplat, a novel 3D Gaussian Splatting model designed to perform pose-free and 3D prior-free generalizable 3D reconstruction from unposed multi-view images. These settings are inherently ill-posed due to the lack of ground-truth data, learned geometric information, and the need to achieve accurate 3D reconstruction without finetuning, making it difficult for conventional methods to achieve high-quality results. Our model addresses these challenges by effectively integrating explicit 3D representations with self-supervised depth and pose estimation techniques, resulting in reciprocal improvements in both pose accuracy and 3D reconstruction quality. Furthermore, we incorporate a matching-aware pose estimation network and a depth refinement module to enhance geometry consistency across views, ensuring more accurate and stable 3D reconstructions. To present the performance of our method, we evaluated it on large-scale real-world datasets, including RealEstate10K, ACID, and DL3DV. SelfSplat achieves superior results over previous state-of-the-art methods in both appearance and geometry quality, also demonstrates strong cross-dataset generalization capabilities. Extensive ablation studies and analysis also validate the effectiveness of our proposed methods. Code and pretrained models are available at https://gynjn.github.io/selfsplat/
Authors: Claudia Cuttano, Gabriele Trivigno, Gabriele Rosi, Carlo Masone, Giuseppe Averta
Abstract: Referring Video Object Segmentation (RVOS) relies on natural language expressions to segment an object in a video clip. Existing methods restrict reasoning either to independent short clips, losing global context, or process the entire video offline, impairing their application in a streaming fashion. In this work, we aim to surpass these limitations and design an RVOS method capable of effectively operating in streaming-like scenarios while retaining contextual information from past frames. We build upon the Segment-Anything 2 (SAM2) model, that provides robust segmentation and tracking capabilities and is naturally suited for streaming processing. We make SAM2 wiser, by empowering it with natural language understanding and explicit temporal modeling at the feature extraction stage, without fine-tuning its weights, and without outsourcing modality interaction to external models. To this end, we introduce a novel adapter module that injects temporal information and multi-modal cues in the feature extraction process. We further reveal the phenomenon of tracking bias in SAM2 and propose a learnable module to adjust its tracking focus when the current frame features suggest a new object more aligned with the caption. Our proposed method, SAMWISE, achieves state-of-the-art across various benchmarks, by adding a negligible overhead of less than 5 M parameters. Code is available at https://github.com/ClaudiaCuttano/SAMWISE .
Authors: Mehdi Zayene, Jannik Endres, Albias Havolli, Charles Corbi\`ere, Salim Cherkaoui, Alexandre Kontouli, Alexandre Alahi
Abstract: Despite progress in stereo depth estimation, omnidirectional imaging remains underexplored, mainly due to the lack of appropriate data. We introduce Helvipad, a real-world dataset for omnidirectional stereo depth estimation, featuring 40K video frames from video sequences across diverse environments, including crowded indoor and outdoor scenes with various lighting conditions. Collected using two 360{\deg} cameras in a top-bottom setup and a LiDAR sensor, the dataset includes accurate depth and disparity labels by projecting 3D point clouds onto equirectangular images. Additionally, we provide an augmented training set with an increased label density by using depth completion. We benchmark leading stereo depth estimation models for both standard and omnidirectional images. The results show that while recent stereo methods perform decently, a challenge persists in accurately estimating depth in omnidirectional imaging. To address this, we introduce necessary adaptations to stereo models, leading to improved performance.
Authors: Weimin Qiu, Jieke Wang, Meng Tang
Abstract: Diffusion models achieved unprecedented fidelity and diversity for synthesizing image, video, 3D assets, etc. However, subject mixing is an unresolved issue for diffusion-based image synthesis, particularly for synthesizing multiple similar-looking subjects. We propose Self-Cross Diffusion Guidance to penalize the overlap between cross-attention maps and the aggregated self-attention map. Compared to previous methods based on self-attention or cross-attention alone, our guidance is more effective in eliminating subject mixing. What's more, our guidance addresses subject mixing for all relevant patches beyond the most discriminant one, e.g., the beak of a bird. For each subject, we aggregate self-attention maps of patches with higher cross-attention values. Thus, the aggregated self-attention map forms a region that the whole subject attends to. Our training-free method boosts the performance of both Unet-based and Transformer-based diffusion models such as the Stable Diffusion series. We also release a similar subjects dataset (SSD), a challenging benchmark, and utilize GPT-4o for automatic and reliable evaluation. Extensive qualitative and quantitative results demonstrate the effectiveness of our self-cross diffusion guidance.
Authors: Guangshun Wei, Yuan Feng, Long Ma, Chen Wang, Yuanfeng Zhou, Changjian Li
Abstract: This paper presents PCDreamer, a novel method for point cloud completion. Traditional methods typically extract features from partial point clouds to predict missing regions, but the large solution space often leads to unsatisfactory results. More recent approaches have started to use images as extra guidance, effectively improving performance, but obtaining paired data of images and partial point clouds is challenging in practice. To overcome these limitations, we harness the relatively view-consistent multi-view diffusion priors within large models, to generate novel views of the desired shape. The resulting image set encodes both global and local shape cues, which are especially beneficial for shape completion. To fully exploit the priors, we have designed a shape fusion module for producing an initial complete shape from multi-modality input (i.e.,, images and point clouds), and a follow-up shape consolidation module to obtain the final complete shape by discarding unreliable points introduced by the inconsistency from diffusion priors. Extensive experimental results demonstrate our superior performance, especially in recovering fine details.
Authors: Yichong Lu, Yichi Cai, Shangzhan Zhang, Hongyu Zhou, Haoji Hu, Huimin Yu, Andreas Geiger, Yiyi Liao
Abstract: Photorealistic 3D vehicle models with high controllability are essential for autonomous driving simulation and data augmentation. While handcrafted CAD models provide flexible controllability, free CAD libraries often lack the high-quality materials necessary for photorealistic rendering. Conversely, reconstructed 3D models offer high-fidelity rendering but lack controllability. In this work, we introduce UrbanCAD, a framework that generates highly controllable and photorealistic 3D vehicle digital twins from a single urban image, leveraging a large collection of free 3D CAD models and handcrafted materials. To achieve this, we propose a novel pipeline that follows a retrieval-optimization manner, adapting to observational data while preserving fine-grained expert-designed priors for both geometry and material. This enables vehicles' realistic 360-degree rendering, background insertion, material transfer, relighting, and component manipulation. Furthermore, given multi-view background perspective and fisheye images, we approximate environment lighting using fisheye images and reconstruct the background with 3DGS, enabling the photorealistic insertion of optimized CAD models into rendered novel view backgrounds. Experimental results demonstrate that UrbanCAD outperforms baselines in terms of photorealism. Additionally, we show that various perception models maintain their accuracy when evaluated on UrbanCAD with in-distribution configurations but degrade when applied to realistic out-of-distribution data generated by our method. This suggests that UrbanCAD is a significant advancement in creating photorealistic, safety-critical driving scenarios for downstream applications.
Authors: Alex Hanson, Allen Tu, Geng Lin, Vasu Singla, Matthias Zwicker, Tom Goldstein
Abstract: 3D Gaussian Splatting (3D-GS) is a recent 3D scene reconstruction technique that enables real-time rendering of novel views by modeling scenes as parametric point clouds of differentiable 3D Gaussians. However, its rendering speed and model size still present bottlenecks, especially in resource-constrained settings. In this paper, we identify and address two key inefficiencies in 3D-GS to substantially improve rendering speed. These improvements also yield the ancillary benefits of reduced model size and training time. First, we optimize the rendering pipeline to precisely localize Gaussians in the scene, boosting rendering speed without altering visual fidelity. Second, we introduce a novel pruning technique and integrate it into the training pipeline, significantly reducing model size and training time while further raising rendering speed. Our Speedy-Splat approach combines these techniques to accelerate average rendering speed by a drastic $\mathit{6.71\times}$ across scenes from the Mip-NeRF 360, Tanks & Temples, and Deep Blending datasets.
Authors: Shuling Zhao, Fa-Ting Hong, Xiaoshui Huang, Dan Xu
Abstract: Talking head video generation aims to generate a realistic talking head video that preserves the person's identity from a source image and the motion from a driving video. Despite the promising progress made in the field, it remains a challenging and critical problem to generate videos with accurate poses and fine-grained facial details simultaneously. Essentially, facial motion is often highly complex to model precisely, and the one-shot source face image cannot provide sufficient appearance guidance during generation due to dynamic pose changes. To tackle the problem, we propose to jointly learn motion and appearance codebooks and perform multi-scale codebook compensation to effectively refine both the facial motion conditions and appearance features for talking face image decoding. Specifically, the designed multi-scale motion and appearance codebooks are learned simultaneously in a unified framework to store representative global facial motion flow and appearance patterns. Then, we present a novel multi-scale motion and appearance compensation module, which utilizes a transformer-based codebook retrieval strategy to query complementary information from the two codebooks for joint motion and appearance compensation. The entire process produces motion flows of greater flexibility and appearance features with fewer distortions across different scales, resulting in a high-quality talking head video generation framework. Extensive experiments on various benchmarks validate the effectiveness of our approach and demonstrate superior generation results from both qualitative and quantitative perspectives when compared to state-of-the-art competitors.
Authors: Xin Xie, Dong Gong
Abstract: Text-to-image diffusion model alignment is critical for improving the alignment between the generated images and human preferences. While training-based methods are constrained by high computational costs and dataset requirements, training-free alignment methods remain underexplored and are often limited by inaccurate guidance. We propose a plug-and-play training-free alignment method, DyMO, for aligning the generated images and human preferences during inference. Apart from text-aware human preference scores, we introduce a semantic alignment objective for enhancing the semantic alignment in the early stages of diffusion, relying on the fact that the attention maps are effective reflections of the semantics in noisy images. We propose dynamic scheduling of multiple objectives and intermediate recurrent steps to reflect the requirements at different steps. Experiments with diverse pre-trained diffusion models and metrics demonstrate the effectiveness and robustness of the proposed method.
Authors: Tom\'a\v{s} Sou\v{c}ek, Prajwal Gatti, Michael Wray, Ivan Laptev, Dima Damen, Josef Sivic
Abstract: The goal of this work is to generate step-by-step visual instructions in the form of a sequence of images, given an input image that provides the scene context and the sequence of textual instructions. This is a challenging problem as it requires generating multi-step image sequences to achieve a complex goal while being grounded in a specific environment. Part of the challenge stems from the lack of large-scale training data for this problem. The contribution of this work is thus three-fold. First, we introduce an automatic approach for collecting large step-by-step visual instruction training data from instructional videos. We apply this approach to one million videos and create a large-scale, high-quality dataset of 0.6M sequences of image-text pairs. Second, we develop and train ShowHowTo, a video diffusion model capable of generating step-by-step visual instructions consistent with the provided input image. Third, we evaluate the generated image sequences across three dimensions of accuracy (step, scene, and task) and show our model achieves state-of-the-art results on all of them. Our code, dataset, and trained models are publicly available.
Authors: Yu Yuan, Xijun Wang, Yichen Sheng, Prateek Chennuri, Xingguang Zhang, Stanley Chan
Abstract: Image generation today can produce somewhat realistic images from text prompts. However, if one asks the generator to synthesize a specific camera setting such as creating different fields of view using a 24mm lens versus a 70mm lens, the generator will not be able to interpret and generate scene-consistent images. This limitation not only hinders the adoption of generative tools in professional photography but also highlights the broader challenge of aligning data-driven models with real-world physical settings. In this paper, we introduce Generative Photography, a framework that allows controlling camera intrinsic settings during content generation. The core innovation of this work are the concepts of Dimensionality Lifting and Differential Camera Intrinsics Learning, enabling smooth and consistent transitions across different camera settings. Experimental results show that our method produces significantly more scene-consistent photorealistic images than state-of-the-art models such as Stable Diffusion 3 and FLUX. Our code and additional results are available at https://generative-photography.github.io/project.
Authors: Zhaofeng Hu, Sifan Zhou, Shibo Zhao, Zhihang Yuan, Ci-Jyun Liang
Abstract: 3D single object tracking is essential in autonomous driving and robotics. Existing methods often struggle with sparse and incomplete point cloud scenarios. To address these limitations, we propose a Multimodal-guided Virtual Cues Projection (MVCP) scheme that generates virtual cues to enrich sparse point clouds. Additionally, we introduce an enhanced tracker MVCTrack based on the generated virtual cues. Specifically, the MVCP scheme seamlessly integrates RGB sensors into LiDAR-based systems, leveraging a set of 2D detections to create dense 3D virtual cues that significantly improve the sparsity of point clouds. These virtual cues can naturally integrate with existing LiDAR-based 3D trackers, yielding substantial performance gains. Extensive experiments demonstrate that our method achieves competitive performance on the NuScenes dataset.
Authors: Kiyohiro Nakayama, Jan Ackermann, Timur Levent Kesdogan, Yang Zheng, Maria Korosteleva, Olga Sorkine-Hornung, Leonidas J. Guibas, Guandao Yang, Gordon Wetzstein
Abstract: Apparel is essential to human life, offering protection, mirroring cultural identities, and showcasing personal style. Yet, the creation of garments remains a time-consuming process, largely due to the manual work involved in designing them. To simplify this process, we introduce AIpparel, a multimodal foundation model for generating and editing sewing patterns. Our model fine-tunes state-of-the-art large multimodal models (LMMs) on a custom-curated large-scale dataset of over 120,000 unique garments, each with multimodal annotations including text, images, and sewing patterns. Additionally, we propose a novel tokenization scheme that concisely encodes these complex sewing patterns so that LLMs can learn to predict them efficiently. AIpparel achieves state-of-the-art performance in single-modal tasks, including text-to-garment and image-to-garment prediction, and enables novel multimodal garment generation applications such as interactive garment editing. The project website is at https://georgenakayama.github.io/AIpparel/.
Authors: Hamid Gadirov, Qi Wu, David Bauer, Kwan-Liu Ma, Jos Roerdink, Steffen Frey
Abstract: We present HyperFLINT (Hypernetwork-based FLow estimation and temporal INTerpolation), a novel deep learning-based approach for estimating flow fields, temporally interpolating scalar fields, and facilitating parameter space exploration in spatio-temporal scientific ensemble data. This work addresses the critical need to explicitly incorporate ensemble parameters into the learning process, as traditional methods often neglect these, limiting their ability to adapt to diverse simulation settings and provide meaningful insights into the data dynamics. HyperFLINT introduces a hypernetwork to account for simulation parameters, enabling it to generate accurate interpolations and flow fields for each timestep by dynamically adapting to varying conditions, thereby outperforming existing parameter-agnostic approaches. The architecture features modular neural blocks with convolutional and deconvolutional layers, supported by a hypernetwork that generates weights for the main network, allowing the model to better capture intricate simulation dynamics. A series of experiments demonstrates HyperFLINT's significantly improved performance in flow field estimation and temporal interpolation, as well as its potential in enabling parameter space exploration, offering valuable insights into complex scientific ensembles.
Authors: Thomas Walker, Salvatore Esposito, Daniel Rebain, Amir Vaxman, Arno Onken, Changjian Li, Oisin Mac Aodha
Abstract: Reconstructing complex structures from planar cross-sections is a challenging problem, with wide-reaching applications in medical imaging, manufacturing, and topography. Out-of-the-box point cloud reconstruction methods can often fail due to the data sparsity between slicing planes, while current bespoke methods struggle to reconstruct thin geometric structures and preserve topological continuity. This is important for medical applications where thin vessel structures are present in CT and MRI scans. This paper introduces CrossSDF, a novel approach for extracting a 3D signed distance field from 2D signed distances generated from planar contours. Our approach makes the training of neural SDFs contour-aware by using losses designed for the case where geometry is known within 2D slices. Our results demonstrate a significant improvement over existing methods, effectively reconstructing thin structures and producing accurate 3D models without the interpolation artifacts or over-smoothing of prior approaches.
Authors: Yuto Matsubara, Ko Nishino
Abstract: We introduce a novel method for human shape and pose recovery that can fully leverage multiple static views. We target fixed-multiview people monitoring, including elderly care and safety monitoring, in which calibrated cameras can be installed at the corners of a room or an open space but whose configuration may vary depending on the environment. Our key idea is to formulate it as neural optimization. We achieve this with HeatFormer, a neural optimizer that iteratively refines the SMPL parameters given multiview images, which is fundamentally agonistic to the configuration of views. HeatFormer realizes this SMPL parameter estimation as heat map generation and alignment with a novel transformer encoder and decoder. We demonstrate the effectiveness of HeatFormer including its accuracy, robustness to occlusion, and generalizability through an extensive set of experiments. We believe HeatFormer can serve a key role in passive human behavior modeling.
Authors: Jaihyun Lew, Soohyuk Jang, Jaehoon Lee, Seungryong Yoo, Eunji Kim, Saehyung Lee, Jisoo Mok, Siwon Kim, Sungroh Yoon
Abstract: Transformers, a groundbreaking architecture proposed for Natural Language Processing (NLP), have also achieved remarkable success in Computer Vision. A cornerstone of their success lies in the attention mechanism, which models relationships among tokens. While the tokenization process in NLP inherently ensures that a single token does not contain multiple semantics, the tokenization of Vision Transformer (ViT) utilizes tokens from uniformly partitioned square image patches, which may result in an arbitrary mixing of visual concepts in a token. In this work, we propose to substitute the grid-based tokenization in ViT with superpixel tokenization, which employs superpixels to generate a token that encapsulates a sole visual concept. Unfortunately, the diverse shapes, sizes, and locations of superpixels make integrating superpixels into ViT tokenization rather challenging. Our tokenization pipeline, comprised of pre-aggregate extraction and superpixel-aware aggregation, overcomes the challenges that arise in superpixel tokenization. Extensive experiments demonstrate that our approach, which exhibits strong compatibility with existing frameworks, enhances the accuracy and robustness of ViT on various downstream tasks.
Authors: Wanyue Zhang, Rishabh Dabral, Vladislav Golyanik, Vasileios Choutas, Eduardo Alvarado, Thabo Beeler, Marc Habermann, Christian Theobalt
Abstract: We present BimArt, a novel generative approach for synthesizing 3D bimanual hand interactions with articulated objects. Unlike prior works, we do not rely on a reference grasp, a coarse hand trajectory, or separate modes for grasping and articulating. To achieve this, we first generate distance-based contact maps conditioned on the object trajectory with an articulation-aware feature representation, revealing rich bimanual patterns for manipulation. The learned contact prior is then used to guide our hand motion generator, producing diverse and realistic bimanual motions for object movement and articulation. Our work offers key insights into feature representation and contact prior for articulated objects, demonstrating their effectiveness in taming the complex, high-dimensional space of bimanual hand-object interactions. Through comprehensive quantitative experiments, we demonstrate a clear step towards simplified and high-quality hand-object animations that surpass the state of the art in motion quality and diversity. Project page: https://vcai.mpi-inf.mpg.de/projects/bimart/.
Authors: Leigang Qu, Haochuan Li, Wenjie Wang, Xiang Liu, Juncheng Li, Liqiang Nie, Tat-Seng Chua
Abstract: Large Multimodal Models (LMMs) have demonstrated impressive capabilities in multimodal understanding and generation, pushing forward advancements in text-to-image generation. However, achieving accurate text-image alignment for LMMs, particularly in compositional scenarios, remains challenging. Existing approaches, such as layout planning for multi-step generation and learning from human feedback or AI feedback, depend heavily on prompt engineering, costly human annotations, and continual upgrading, limiting flexibility and scalability. In this work, we introduce a model-agnostic iterative self-improvement framework (SILMM) that can enable LMMs to provide helpful and scalable self-feedback and optimize text-image alignment via Direct Preference Optimization (DPO). DPO can readily applied to LMMs that use discrete visual tokens as intermediate image representations; while it is less suitable for LMMs with continuous visual features, as obtaining generation probabilities is challenging. To adapt SILMM to LMMs with continuous features, we propose a diversity mechanism to obtain diverse representations and a kernel-based continuous DPO for alignment. Extensive experiments on three compositional text-to-image generation benchmarks validate the effectiveness and superiority of SILMM, showing improvements exceeding 30% on T2I-CompBench++ and around 20% on DPG-Bench.
Authors: Jan Rathjens, Shirin Reyhanian, David Kappel, Laurenz Wiskott
Abstract: Understanding the mechanisms underlying deep neural networks in computer vision remains a fundamental challenge. While many previous approaches have focused on visualizing intermediate representations within deep neural networks, particularly convolutional neural networks, these techniques have yet to be thoroughly explored in transformer-based vision models. In this study, we apply a modular approach of training inverse models to reconstruct input images from intermediate layers within a Detection Transformer and a Vision Transformer, showing that this approach is efficient and feasible. Through qualitative and quantitative evaluations of reconstructed images, we generate insights into the underlying mechanisms of these architectures, highlighting their similarities and differences in terms of contextual shape and preservation of image details, inter-layer correlation, and robustness to color perturbations. Our analysis illustrates how these properties emerge within the models, contributing to a deeper understanding of transformer-based vision models. The code for reproducing our experiments is available at github.com/wiskott-lab/inverse-tvm.
Authors: Linke Ouyang, Yuan Qu, Hongbin Zhou, Jiawei Zhu, Rui Zhang, Qunshu Lin, Bin Wang, Zhiyuan Zhao, Man Jiang, Xiaomeng Zhao, Jin Shi, Fan Wu, Pei Chu, Minghao Liu, Zhenxiang Li, Chao Xu, Bo Zhang, Botian Shi, Zhongying Tu, Conghui He
Abstract: Document content extraction is a critical task in computer vision, underpinning the data needs of large language models (LLMs) and retrieval-augmented generation (RAG) systems. Despite recent progress, current document parsing methods have not been fairly and comprehensively evaluated due to the narrow coverage of document types and the simplified, unrealistic evaluation procedures in existing benchmarks. To address these gaps, we introduce OmniDocBench, a novel benchmark featuring high-quality annotations across nine document sources, including academic papers, textbooks, and more challenging cases such as handwritten notes and densely typeset newspapers. OmniDocBench supports flexible, multi-level evaluations--ranging from an end-to-end assessment to the task-specific and attribute--based analysis using 19 layout categories and 15 attribute labels. We conduct a thorough evaluation of both pipeline-based methods and end-to-end vision-language models, revealing their strengths and weaknesses across different document types. OmniDocBench sets a new standard for the fair, diverse, and fine-grained evaluation in document parsing. Dataset and code are available at https://github.com/opendatalab/OmniDocBench.
Authors: Jingxi Chen, Brandon Y. Feng, Haoming Cai, Tianfu Wang, Levi Burner, Dehao Yuan, Cornelia Fermuller, Christopher A. Metzler, Yiannis Aloimonos
Abstract: Video Frame Interpolation aims to recover realistic missing frames between observed frames, generating a high-frame-rate video from a low-frame-rate video. However, without additional guidance, the large motion between frames makes this problem ill-posed. Event-based Video Frame Interpolation (EVFI) addresses this challenge by using sparse, high-temporal-resolution event measurements as motion guidance. This guidance allows EVFI methods to significantly outperform frame-only methods. However, to date, EVFI methods have relied on a limited set of paired event-frame training data, severely limiting their performance and generalization capabilities. In this work, we overcome the limited data challenge by adapting pre-trained video diffusion models trained on internet-scale datasets to EVFI. We experimentally validate our approach on real-world EVFI datasets, including a new one that we introduce. Our method outperforms existing methods and generalizes across cameras far better than existing approaches.
Authors: Briac Toussaint, Diego Thomas, Jean-S\'ebastien Franco
Abstract: SDF-based differential rendering frameworks have achieved state-of-the-art multiview 3D shape reconstruction. In this work, we re-examine this family of approaches by minimally reformulating its core appearance model in a way that simultaneously yields faster computation and increased performance. To this goal, we exhibit a physically-inspired minimal radiance parametrization decoupling angular and spatial contributions, by encoding them with a small number of features stored in two respective volumetric grids of different resolutions. Requiring as little as four parameters per voxel, and a tiny MLP call inside a single fully fused kernel, our approach allows to enhance performance with both surface and image (PSNR) metrics, while providing a significant training speedup and real-time rendering. We show this performance to be consistently achieved on real data over two widely different and popular application fields, generic object and human subject shape reconstruction, using four representative and challenging datasets.
Authors: Yan Xia, Yunxiang Lu, Rui Song, Oussema Dhaouadi, Jo\~ao F. Henriques, Daniel Cremers
Abstract: We tackle the problem of localizing traffic cameras within a 3D reference map and propose a novel image-to-point cloud registration (I2P) method, TrafficLoc, in a coarse-tofine matching fashion. To overcome the lack of large-scale real-world intersection datasets, we first introduce Carla Intersection, a new simulated dataset with 75 urban and rural intersections in Carla. We find that current I2P methods struggle with cross-modal matching under large viewpoint differences, especially at traffic intersections. TrafficLoc thus employs a novel Geometry-guided Attention Loss (GAL) to focus only on the corresponding geometric regions under different viewpoints during 2D-3D feature fusion. To address feature inconsistency in paired image patch-point groups, we further propose Inter-intra Contrastive Learning (ICL) to enhance separating 2D patch/3D group features within each intra-modality and introduce Dense Training Alignment (DTA) with soft-argmax for improving position regression. Extensive experiments show our TrafficLoc greatly improves the performance over the SOTA I2P methods (up to 86%) on Carla Intersection and generalizes well to real-world data. TrafficLoc also achieves new SOTA performance on KITTI and NuScenes datasets, demonstrating the superiority across both in-vehicle and traffic cameras. Our project page is publicly available at https://tum-luk.github.io/projects/trafficloc/.
Authors: Ximing Xing, Juncheng Hu, Guotao Liang, Jing Zhang, Dong Xu, Qian Yu
Abstract: The unprecedented advancements in Large Language Models (LLMs) have profoundly impacted natural language processing but have yet to fully embrace the realm of scalable vector graphics (SVG) generation. While LLMs encode partial knowledge of SVG data from web pages during training, recent findings suggest that semantically ambiguous and tokenized representations within LLMs may result in hallucinations in vector primitive predictions. Additionally, LLM training typically lacks modeling and understanding of the rendering sequence of vector paths, which can lead to occlusion between output vector primitives. In this paper, we present LLM4SVG, an initial yet substantial step toward bridging this gap by enabling LLMs to better understand and generate vector graphics. LLM4SVG facilitates a deeper understanding of SVG components through learnable semantic tokens, which precisely encode these tokens and their corresponding properties to generate semantically aligned SVG outputs. Using a series of learnable semantic tokens, a structured dataset for instruction following is developed to support comprehension and generation across two primary tasks. Our method introduces a modular architecture to existing large language models, integrating semantic tags, vector instruction encoders, fine-tuned commands, and powerful LLMs to tightly combine geometric, appearance, and language information. To overcome the scarcity of SVG-text instruction data, we developed an automated data generation pipeline that collected our SVGX-SFT Dataset, consisting of high-quality human-designed SVGs and 580k SVG instruction following data specifically crafted for LLM training, which facilitated the adoption of the supervised fine-tuning strategy popular in LLM development.
Authors: Bingwen Hu, Heng Liu, Zhedong Zheng, Ping Liu
Abstract: Convolutional Neural Networks (CNNs) have significantly advanced Image Super-Resolution (SR), yet most CNN-based methods rely solely on pixel-based transformations, often leading to artifacts and blurring, particularly under severe downsampling rates (\eg, 8$\times$ or 16$\times$). The recently developed text-guided SR approaches leverage textual descriptions to enhance their detail restoration capabilities but frequently struggle with effectively performing alignment, resulting in semantic inconsistencies. To address these challenges, we propose a multi-modal semantic enhancement framework that integrates textual semantics with visual features, effectively mitigating semantic mismatches and detail losses in highly degraded low-resolution (LR) images. Our method enables realistic, high-quality SR to be performed at large upscaling factors, with a maximum scaling ratio of 16$\times$. The framework integrates both text and image inputs using the prompt predictor, the Text-Image Fusion Block (TIFBlock), and the Iterative Refinement Module, leveraging Contrastive Language-Image Pretraining (CLIP) features to guide a progressive enhancement process with fine-grained alignment. This synergy produces high-resolution outputs with sharp textures and strong semantic coherence, even at substantial scaling factors. Extensive comparative experiments and ablation studies validate the effectiveness of our approach. Furthermore, by leveraging textual semantics, our method offers a degree of super-resolution editability, allowing for controlled enhancements while preserving semantic consistency.
Authors: Samuel Teodoro, Agus Gunawan, Soo Ye Kim, Jihyong Oh, Munchurl Kim
Abstract: Recent AI-based video editing has enabled users to edit videos through simple text prompts, significantly simplifying the editing process. However, recent zero-shot video editing techniques primarily focus on global or single-object edits, which can lead to unintended changes in other parts of the video. When multiple objects require localized edits, existing methods face challenges, such as unfaithful editing, editing leakage, and lack of suitable evaluation datasets and metrics. To overcome these limitations, we propose $\textbf{P}$robability $\textbf{R}$edistribution for $\textbf{I}$nstance-aware $\textbf{M}$ulti-object Video $\textbf{Edit}$ing ($\textbf{PRIMEdit}$). PRIMEdit is a zero-shot framework that introduces two key modules: (i) Instance-centric Probability Redistribution (IPR) to ensure precise localization and faithful editing and (ii) Disentangled Multi-instance Sampling (DMS) to prevent editing leakage. Additionally, we present our new MIVE Dataset for video editing featuring diverse video scenarios, and introduce the Cross-Instance Accuracy (CIA) Score to evaluate editing leakage in multi-instance video editing tasks. Our extensive qualitative, quantitative, and user study evaluations demonstrate that PRIMEdit significantly outperforms recent state-of-the-art methods in terms of editing faithfulness, accuracy, and leakage prevention, setting a new benchmark for multi-instance video editing.
Authors: Sebastian Koch, Johanna Wald, Mirco Colosi, Narunas Vaskevicius, Pedro Hermosilla, Federico Tombari, Timo Ropinski
Abstract: Neural radiance fields are an emerging 3D scene representation and recently even been extended to learn features for scene understanding by distilling open-vocabulary features from vision-language models. However, current method primarily focus on object-centric representations, supporting object segmentation or detection, while understanding semantic relationships between objects remains largely unexplored. To address this gap, we propose RelationField, the first method to extract inter-object relationships directly from neural radiance fields. RelationField represents relationships between objects as pairs of rays within a neural radiance field, effectively extending its formulation to include implicit relationship queries. To teach RelationField complex, open-vocabulary relationships, relationship knowledge is distilled from multi-modal LLMs. To evaluate RelationField, we solve open-vocabulary 3D scene graph generation tasks and relationship-guided instance segmentation, achieving state-of-the-art performance in both tasks. See the project website at https://relationfield.github.io.
Authors: Yiyu Zhuang, Jiaxi Lv, Hao Wen, Qing Shuai, Ailing Zeng, Hao Zhu, Shifeng Chen, Yujiu Yang, Xun Cao, Wei Liu
Abstract: Creating a high-fidelity, animatable 3D full-body avatar from a single image is a challenging task due to the diverse appearance and poses of humans and the limited availability of high-quality training data. To achieve fast and high-quality human reconstruction, this work rethinks the task from the perspectives of dataset, model, and representation. First, we introduce a large-scale HUman-centric GEnerated dataset, HuGe100K, consisting of 100K diverse, photorealistic sets of human images. Each set contains 24-view frames in specific human poses, generated using a pose-controllable image-to-multi-view model. Next, leveraging the diversity in views, poses, and appearances within HuGe100K, we develop a scalable feed-forward transformer model to predict a 3D human Gaussian representation in a uniform space from a given human image. This model is trained to disentangle human pose, body shape, clothing geometry, and texture. The estimated Gaussians can be animated without post-processing. We conduct comprehensive experiments to validate the effectiveness of the proposed dataset and method. Our model demonstrates the ability to efficiently reconstruct photorealistic humans at 1K resolution from a single input image using a single GPU instantly. Additionally, it seamlessly supports various applications, as well as shape and texture editing tasks. Project page: https://yiyuzhuang.github.io/IDOL/.
Authors: Jen-Hao Rick Chang, Yuyang Wang, Miguel Angel Bautista Martin, Jiatao Gu, Xiaoming Zhao, Josh Susskind, Oncel Tuzel
Abstract: We introduce a latent 3D representation that models 3D surfaces as probability density functions in 3D, i.e., p(x,y,z), with flow-matching. Our representation is specifically designed for consumption by machine learning models, offering continuity and compactness by construction while requiring only point clouds and minimal data preprocessing. Despite being a data-driven method, our use of flow matching in the 3D space enables interesting geometry properties, including the capabilities to perform zero-shot estimation of surface normal and deformation field. We evaluate with several machine learning tasks, including 3D-CLIP, unconditional generative models, single-image conditioned generative model, and intersection-point estimation. Across all experiments, our models achieve competitive performance to existing baselines, while requiring less preprocessing and auxiliary information from training data.
Authors: Quan Dao, Hao Phung, Trung Dao, Dimitris Metaxas, Anh Tran
Abstract: Flow matching has emerged as a promising framework for training generative models, demonstrating impressive empirical performance while offering relative ease of training compared to diffusion-based models. However, this method still requires numerous function evaluations in the sampling process. To address these limitations, we introduce a self-corrected flow distillation method that effectively integrates consistency models and adversarial training within the flow-matching framework. This work is a pioneer in achieving consistent generation quality in both few-step and one-step sampling. Our extensive experiments validate the effectiveness of our method, yielding superior results both quantitatively and qualitatively on CelebA-HQ and zero-shot benchmarks on the COCO dataset. Our implementation is released at https://github.com/VinAIResearch/SCFlow
Authors: Azim Ospanov, Mohammad Jalali, Farzan Farnia
Abstract: The use of CLIP embeddings to assess the alignment of samples produced by text-to-image generative models has been extensively explored in the literature. While the widely adopted CLIPScore, derived from the cosine similarity of text and image embeddings, effectively measures the relevance of a generated image, it does not quantify the diversity of images generated by a text-to-image model. In this work, we extend the application of CLIP embeddings to quantify and interpret the intrinsic diversity of text-to-image models, which is responsible for generating diverse images from similar text prompts. To achieve this, we propose a decomposition of the CLIP-based kernel covariance matrix of image data into text-based and non-text-based components. Using the Schur complement of the joint image-text kernel covariance matrix, we perform this decomposition and define the matrix-based entropy of the decomposed component as the \textit{Schur Complement Entropy (SCE)} score, a measure of the intrinsic diversity of a text-to-image model based on data collected with varying text prompts. Additionally, we demonstrate the use of the Schur complement-based decomposition to nullify the influence of a given prompt in the CLIP embedding of an image, enabling focus or defocus of embeddings on specific objects or properties for downstream tasks. We present several numerical results that apply our Schur complement-based approach to evaluate text-to-image models and modify CLIP image embeddings. The codebase is available at https://github.com/aziksh-ospanov/CLIP-DISSECTION
Authors: Wenkun He, Yun Liu, Ruitao Liu, Li Yi
Abstract: Synthesizing realistic human-object interaction motions is a critical problem in VR/AR and human animation. Unlike the commonly studied scenarios involving a single human or hand interacting with one object, we address a more generic multi-body setting with arbitrary numbers of humans, hands, and objects. This complexity introduces significant challenges in synchronizing motions due to the high correlations and mutual influences among bodies. To address these challenges, we introduce SyncDiff, a novel method for multi-body interaction synthesis using a synchronized motion diffusion strategy. SyncDiff employs a single diffusion model to capture the joint distribution of multi-body motions. To enhance motion fidelity, we propose a frequency-domain motion decomposition scheme. Additionally, we introduce a new set of alignment scores to emphasize the synchronization of different body motions. SyncDiff jointly optimizes both data sample likelihood and alignment likelihood through an explicit synchronization strategy. Extensive experiments across four datasets with various multi-body configurations demonstrate the superiority of SyncDiff over existing state-of-the-art motion synthesis methods.
Authors: Yuqian Yuan, Hang Zhang, Wentong Li, Zesen Cheng, Boqiang Zhang, Long Li, Xin Li, Deli Zhao, Wenqiao Zhang, Yueting Zhuang, Jianke Zhu, Lidong Bing
Abstract: Video Large Language Models (Video LLMs) have recently exhibited remarkable capabilities in general video understanding. However, they mainly focus on holistic comprehension and struggle with capturing fine-grained spatial and temporal details. Besides, the lack of high-quality object-level video instruction data and a comprehensive benchmark further hinders their advancements. To tackle these challenges, we introduce the VideoRefer Suite to empower Video LLM for finer-level spatial-temporal video understanding, i.e., enabling perception and reasoning on any objects throughout the video. Specially, we thoroughly develop VideoRefer Suite across three essential aspects: dataset, model, and benchmark. Firstly, we introduce a multi-agent data engine to meticulously curate a large-scale, high-quality object-level video instruction dataset, termed VideoRefer-700K. Next, we present the VideoRefer model, which equips a versatile spatial-temporal object encoder to capture precise regional and sequential representations. Finally, we meticulously create a VideoRefer-Bench to comprehensively assess the spatial-temporal understanding capability of a Video LLM, evaluating it across various aspects. Extensive experiments and analyses demonstrate that our VideoRefer model not only achieves promising performance on video referring benchmarks but also facilitates general video understanding capabilities.
Authors: Huabin Liu, Filip Ilievski, Cees G. M. Snoek
Abstract: This paper proposes the first video-grounded entailment tree reasoning method for commonsense video question answering (VQA). Despite the remarkable progress of large visual-language models (VLMs), there are growing concerns that they learn spurious correlations between videos and likely answers, reinforced by their black-box nature and remaining benchmarking biases. Our method explicitly grounds VQA tasks to video fragments in four steps: entailment tree construction, video-language entailment verification, tree reasoning, and dynamic tree expansion. A vital benefit of the method is its generalizability to current video and image-based VLMs across reasoning types. To support fair evaluation, we devise a de-biasing procedure based on large-language models that rewrites VQA benchmark answer sets to enforce model reasoning. Systematic experiments on existing and de-biased benchmarks highlight the impact of our method components across benchmarks, VLMs, and reasoning types.
Authors: Xueyi Ke, Satoshi Tsutsui, Yayun Zhang, Bihan Wen
Abstract: Infants develop complex visual understanding rapidly, even preceding the acquisition of linguistic skills. As computer vision seeks to replicate the human vision system, understanding infant visual development may offer valuable insights. In this paper, we present an interdisciplinary study exploring this question: can a computational model that imitates the infant learning process develop broader visual concepts that extend beyond the vocabulary it has heard, similar to how infants naturally learn? To investigate this, we analyze a recently published model in Science by Vong et al., which is trained on longitudinal, egocentric images of a single child paired with transcribed parental speech. We perform neuron labeling to identify visual concept neurons hidden in the model's internal representations. We then demonstrate that these neurons can recognize objects beyond the model's original vocabulary. Furthermore, we compare the differences in representation between infant models and those in modern computer vision models, such as CLIP and ImageNet pre-trained model. Ultimately, our work bridges cognitive science and computer vision by analyzing the internal representations of a computational model trained on an infant visual and linguistic inputs. Our code is available at https://github.com/Kexueyi/discover_infant_vis.
Authors: Dominick Reilly, Manish Kumar Govind, Le Xue, Srijan Das
Abstract: Large Vision Language Models (LVLMs) have demonstrated impressive capabilities in video understanding, yet their adoption for Activities of Daily Living (ADL) remains limited by their inability to capture fine-grained interactions and spatial relationships. To address this, we aim to leverage the complementary nature of egocentric views to enhance LVLM's understanding of exocentric ADL videos. Consequently, we propose ego2exo knowledge distillation to learn ego-augmented exp representations. While effective, this approach requires paired ego-exo videos, which are impractical to collect at scale. To address this, we propose Skeleton-guided Synthetic Ego Generation (SK-EGO), which leverages human skeleton motion to generate synthetic ego views from exocentric videos. To enhance the ego representation of LVLMs trained on synthetic data, we develop a domain-agnostic bootstrapped ego2exo strategy that effectively transfers knowledge from real ego-exo pairs to synthetic ego-exo pairs, while mitigating domain misalignment. We find that the exo representations of our ego-augmented LVLMs successfully learn to extract ego-perspective cues, demonstrated through comprehensive evaluation on six ADL benchmarks and our proposed Ego-in-Exo PerceptionMCQ benchmark designed specifically to assess egocentric understanding from exocentric videos. Code, models, and data will be open-sourced at https://github.com/dominickrei/EgoExo4ADL.
Authors: Shangjin Zhai, Zhichao Ye, Jialin Liu, Weijian Xie, Jiaqi Hu, Zhen Peng, Hua Xue, Danpeng Chen, Xiaomeng Wang, Lei Yang, Nan Wang, Haomin Liu, Guofeng Zhang
Abstract: Recent advances in large reconstruction and generative models have significantly improved scene reconstruction and novel view generation. However, due to compute limitations, each inference with these large models is confined to a small area, making long-range consistent scene generation challenging. To address this, we propose StarGen, a novel framework that employs a pre-trained video diffusion model in an autoregressive manner for long-range scene generation. The generation of each video clip is conditioned on the 3D warping of spatially adjacent images and the temporally overlapping image from previously generated clips, improving spatiotemporal consistency in long-range scene generation with precise pose control. The spatiotemporal condition is compatible with various input conditions, facilitating diverse tasks, including sparse view interpolation, perpetual view generation, and layout-conditioned city generation. Quantitative and qualitative evaluations demonstrate StarGen's superior scalability, fidelity, and pose accuracy compared to state-of-the-art methods. Project page: https://zju3dv.github.io/StarGen.
Authors: Nirit Alkalay, Roy Orfaig, Ben-Zion Bobrovsky
Abstract: 4D panoptic LiDAR segmentation is essential for scene understanding in autonomous driving and robotics, combining semantic and instance segmentation with temporal consistency. Current methods, like 4D-PLS and 4D-STOP, use a tracking-by-detection methodology, employing deep learning networks to perform semantic and instance segmentation on each frame. To maintain temporal consistency, large-size instances detected in the current frame are compared and associated with instances within a temporal window that includes the current and preceding frames. However, their reliance on short-term instance detection, lack of motion estimation, and exclusion of small-sized instances lead to frequent identity switches and reduced tracking performance. We address these issues with the NextStop1 tracker, which integrates Kalman filter-based motion estimation, data association, and lifespan management, along with a tracklet state concept to improve prioritization. Evaluated using the LiDAR Segmentation and Tracking Quality (LSTQ) metric on the SemanticKITTI validation set, NextStop demonstrated enhanced tracking performance, particularly for small-sized objects like people and bicyclists, with fewer ID switches, earlier tracking initiation, and improved reliability in complex environments. The source code is available at https://github.com/AIROTAU/NextStop
Authors: Jiwen Yu, Yiran Qin, Xintao Wang, Pengfei Wan, Di Zhang, Xihui Liu
Abstract: Generative videos have the potential to revolutionize game development by autonomously creating new content. In this paper, we present GameFactory, a framework for action-controlled scene-generalizable game video generation. We first address the fundamental challenge of action controllability by introducing GF-Minecraft, a action-annotated game video dataset without human bias, and developing a action control module that enables precise control over both keyboard and mouse inputs. We further extend to support autoregressive generation for unlimited-length interactive videos. More importantly, GameFactory tackles the critical challenge of scene-generalizable action control, which most existing methods fail to address. To enable the creation of entirely new and diverse games beyond fixed styles and scenes, we leverage the open-domain generative priors from pre-trained video diffusion models. To bridge the domain gap between open-domain priors and small-scale game datasets, we propose a multi-phase training strategy with a domain adapter that decouples game style learning from action control. This decoupling ensures that action control learning is no longer bound to specific game styles, thereby achieving scene-generalizable action control. Experimental results demonstrate that GameFactory effectively generates open-domain action-controllable game videos, representing a significant step forward in AI-driven game generation. Our dataset and project page are publicly available at https://yujiwen.github.io/gamefactory/.
Authors: Ryan Burgert, Yuancheng Xu, Wenqi Xian, Oliver Pilarski, Pascal Clausen, Mingming He, Li Ma, Yitong Deng, Lingxiao Li, Mohsen Mousavi, Michael Ryoo, Paul Debevec, Ning Yu
Abstract: Generative modeling aims to transform random noise into structured outputs. In this work, we enhance video diffusion models by allowing motion control via structured latent noise sampling. This is achieved by just a change in data: we pre-process training videos to yield structured noise. Consequently, our method is agnostic to diffusion model design, requiring no changes to model architectures or training pipelines. Specifically, we propose a novel noise warping algorithm, fast enough to run in real time, that replaces random temporal Gaussianity with correlated warped noise derived from optical flow fields, while preserving the spatial Gaussianity. The efficiency of our algorithm enables us to fine-tune modern video diffusion base models using warped noise with minimal overhead, and provide a one-stop solution for a wide range of user-friendly motion control: local object motion control, global camera movement control, and motion transfer. The harmonization between temporal coherence and spatial Gaussianity in our warped noise leads to effective motion control while maintaining per-frame pixel quality. Extensive experiments and user studies demonstrate the advantages of our method, making it a robust and scalable approach for controlling motion in video diffusion models. Video results are available on our webpage: https://eyeline-research.github.io/Go-with-the-Flow. Source code and model checkpoints are available on GitHub: https://github.com/Eyeline-Research/Go-with-the-Flow.
URLs: https://eyeline-research.github.io/Go-with-the-Flow., https://github.com/Eyeline-Research/Go-with-the-Flow.
Authors: Eunjin Kim, Hyeonjin Kim, Kyong Hwan Jin, Jaejun Yoo
Abstract: While prior methods in Continuous Spatial-Temporal Video Super-Resolution (C-STVSR) employ Implicit Neural Representation (INR) for continuous encoding, they often struggle to capture the complexity of video data, relying on simple coordinate concatenation and pre-trained optical flow networks for motion representation. Interestingly, we find that adding position encoding, contrary to common observations, does not improve--and even degrades--performance. This issue becomes particularly pronounced when combined with pre-trained optical flow networks, which can limit the model's flexibility. To address these issues, we propose BF-STVSR, a C-STVSR framework with two key modules tailored to better represent spatial and temporal characteristics of video: 1) B-spline Mapper for smooth temporal interpolation, and 2) Fourier Mapper for capturing dominant spatial frequencies. Our approach achieves state-of-the-art in various metrics, including PSNR and SSIM, showing enhanced spatial details and natural temporal consistency. Our code is available https://github.com/Eunjnnn/bfstvsr.
Authors: Kazi Hasan Ibn Arif, Sajib Acharjee Dip, Khizar Hussain, Lang Zhang, Chris Thomas
Abstract: Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities in understanding and describing visual content, achieving state-of-the-art performance across various vision-language tasks. However, these models often generate descriptions containing objects or details that are absent in the input image, a phenomenon commonly known as hallucination. Our work investigates the key reasons behind this issue by analyzing the pattern of self-attention in transformer layers. We find that hallucinations often arise from the progressive weakening of attention weight to visual tokens in the deeper layers of the LLM. Some previous works naively boost the attention of all visual tokens to mitigate this issue, resulting in suboptimal hallucination reduction. To address this, we identify two critical sets of visual tokens that facilitate the transfer of visual information from the vision encoder to the LLM. Local tokens encode grounded information about objects present in an image, while summary tokens capture the overall aggregated representation of the image. Importantly, these two sets of tokens require different levels of weight enhancement. To this end, we propose \textbf{PAINT} (\textbf{P}aying \textbf{A}ttention to \textbf{IN}formed \textbf{T}okens), a plug-and-play framework that intervenes in the self-attention mechanism of the LLM, selectively boosting the attention weights of local and summary tokens with experimentally learned margins. Evaluation on the MSCOCO image captioning dataset demonstrate that our approach reduces hallucination rates by up to 62.3\% compared to baseline models while maintaining accuracy.
Authors: Jinghan You, Shanglin Li, Yuanrui Sun, Jiangchuan Wei, Mingyu Guo, Chao Feng, Jiao Ran
Abstract: Vision Transformers (ViTs) have revolutionized large-scale visual modeling, yet remain underexplored in face recognition (FR) where CNNs still dominate. We identify a critical bottleneck: CNN-inspired training paradigms fail to unlock ViT's potential, leading to suboptimal performance and convergence instability.To address this challenge, we propose LVFace, a ViT-based FR model that integrates Progressive Cluster Optimization (PCO) to achieve superior results. Specifically, PCO sequentially applies negative class sub-sampling (NCS) for robust and fast feature alignment from random initialization, feature expectation penalties for centroid stabilization, performing cluster boundary refinement through full-batch training without NCS constraints. LVFace establishes a new state-of-the-art face recognition baseline, surpassing leading approaches such as UniFace and TopoFR across multiple benchmarks. Extensive experiments demonstrate that LVFace delivers consistent performance gains, while exhibiting scalability to large-scale datasets and compatibility with mainstream VLMs and LLMs. Notably, LVFace secured 1st place in the ICCV 2021 Masked Face Recognition (MFR)-Ongoing Challenge (March 2025), proving its efficacy in real-world scenarios.
Authors: Peiqing Yang, Shangchen Zhou, Jixin Zhao, Qingyi Tao, Chen Change Loy
Abstract: Auxiliary-free human video matting methods, which rely solely on input frames, often struggle with complex or ambiguous backgrounds. To address this, we propose MatAnyone, a robust framework tailored for target-assigned video matting. Specifically, building on a memory-based paradigm, we introduce a consistent memory propagation module via region-adaptive memory fusion, which adaptively integrates memory from the previous frame. This ensures semantic stability in core regions while preserving fine-grained details along object boundaries. For robust training, we present a larger, high-quality, and diverse dataset for video matting. Additionally, we incorporate a novel training strategy that efficiently leverages large-scale segmentation data, boosting matting stability. With this new network design, dataset, and training strategy, MatAnyone delivers robust and accurate video matting results in diverse real-world scenarios, outperforming existing methods.
Authors: Tao Ren, Zishi Zhang, Zehao Li, Jingyang Jiang, Shentao Qin, Guanghao Li, Yan Li, Yi Zheng, Xinping Li, Min Zhan, Yijie Peng
Abstract: The probabilistic diffusion model (DM), generating content by inferencing through a recursive chain structure, has emerged as a powerful framework for visual generation. After pre-training on enormous unlabeled data, the model needs to be properly aligned to meet requirements for downstream applications. How to efficiently align the foundation DM is a crucial task. Contemporary methods are either based on Reinforcement Learning (RL) or truncated Backpropagation (BP). However, RL and truncated BP suffer from low sample efficiency and biased gradient estimation respectively, resulting in limited improvement or, even worse, complete training failure. To overcome the challenges, we propose the Recursive Likelihood Ratio (RLR) optimizer, a zeroth-order informed fine-tuning paradigm for DM. The zeroth-order gradient estimator enables the computation graph rearrangement within the recursive diffusive chain, making the RLR's gradient estimator an unbiased one with the lower variance than other methods. We provide theoretical guarantees for the performance of the RLR. Extensive experiments are conducted on image and video generation tasks to validate the superiority of the RLR. Furthermore, we propose a novel prompt technique that is natural for the RLR to achieve a synergistic effect.
Authors: Quan Dao, Khanh Doan, Di Liu, Trung Le, Dimitris Metaxas
Abstract: Consistency models are a new family of generative models capable of producing high-quality samples in either a single step or multiple steps. Recently, consistency models have demonstrated impressive performance, achieving results on par with diffusion models in the pixel space. However, the success of scaling consistency training to large-scale datasets, particularly for text-to-image and video generation tasks, is determined by performance in the latent space. In this work, we analyze the statistical differences between pixel and latent spaces, discovering that latent data often contains highly impulsive outliers, which significantly degrade the performance of iCT in the latent space. To address this, we replace Pseudo-Huber losses with Cauchy losses, effectively mitigating the impact of outliers. Additionally, we introduce a diffusion loss at early timesteps and employ optimal transport (OT) coupling to further enhance performance. Lastly, we introduce the adaptive scaling-$c$ scheduler to manage the robust training process and adopt Non-scaling LayerNorm in the architecture to better capture the statistics of the features and reduce outlier impact. With these strategies, we successfully train latent consistency models capable of high-quality sampling with one or two steps, significantly narrowing the performance gap between latent consistency and diffusion models. The implementation is released here: https://github.com/quandao10/sLCT/
Authors: Joohyun Kwon, Hanbyel Cho, Junmo Kim
Abstract: Recent 4D dynamic scene editing methods require editing thousands of 2D images used for dynamic scene synthesis and updating the entire scene with additional training loops, resulting in several hours of processing to edit a single dynamic scene. Therefore, these methods are not scalable with respect to the temporal dimension of the dynamic scene (i.e., the number of timesteps). In this work, we propose Instruct-4DGS, an efficient dynamic scene editing method that is more scalable in terms of temporal dimension. To achieve computational efficiency, we leverage a 4D Gaussian representation that models a 4D dynamic scene by combining static 3D Gaussians with a Hexplane-based deformation field, which captures dynamic information. We then perform editing solely on the static 3D Gaussians, which is the minimal but sufficient component required for visual editing. To resolve the misalignment between the edited 3D Gaussians and the deformation field, which may arise from the editing process, we introduce a refinement stage using a score distillation mechanism. Extensive editing results demonstrate that Instruct-4DGS is efficient, reducing editing time by more than half compared to existing methods while achieving high-quality edits that better follow user instructions.
Authors: Yihua Cheng, Hengfei Wang, Zhongqun Zhang, Yang Yue, Bo Eun Kim, Feng Lu, Hyung Jin Chang
Abstract: 3D and 2D gaze estimation share the fundamental objective of capturing eye movements but are traditionally treated as two distinct research domains. In this paper, we introduce a novel cross-task few-shot 2D gaze estimation approach, aiming to adapt a pre-trained 3D gaze estimation network for 2D gaze prediction on unseen devices using only a few training images. This task is highly challenging due to the domain gap between 3D and 2D gaze, unknown screen poses, and limited training data. To address these challenges, we propose a novel framework that bridges the gap between 3D and 2D gaze. Our framework contains a physics-based differentiable projection module with learnable parameters to model screen poses and project 3D gaze into 2D gaze. The framework is fully differentiable and can integrate into existing 3D gaze networks without modifying their original architecture. Additionally, we introduce a dynamic pseudo-labelling strategy for flipped images, which is particularly challenging for 2D labels due to unknown screen poses. To overcome this, we reverse the projection process by converting 2D labels to 3D space, where flipping is performed. Notably, this 3D space is not aligned with the camera coordinate system, so we learn a dynamic transformation matrix to compensate for this misalignment. We evaluate our method on MPIIGaze, EVE, and GazeCapture datasets, collected respectively on laptops, desktop computers, and mobile devices. The superior performance highlights the effectiveness of our approach, and demonstrates its strong potential for real-world applications.
Authors: Toby Perrett, Ahmad Darkhalil, Saptarshi Sinha, Omar Emara, Sam Pollard, Kranti Parida, Kaiting Liu, Prajwal Gatti, Siddhant Bansal, Kevin Flanagan, Jacob Chalk, Zhifan Zhu, Rhodri Guerrier, Fahd Abdelazim, Bin Zhu, Davide Moltisanti, Michael Wray, Hazel Doughty, Dima Damen
Abstract: We present a validation dataset of newly-collected kitchen-based egocentric videos, manually annotated with highly detailed and interconnected ground-truth labels covering: recipe steps, fine-grained actions, ingredients with nutritional values, moving objects, and audio annotations. Importantly, all annotations are grounded in 3D through digital twinning of the scene, fixtures, object locations, and primed with gaze. Footage is collected from unscripted recordings in diverse home environments, making HDEPIC the first dataset collected in-the-wild but with detailed annotations matching those in controlled lab environments. We show the potential of our highly-detailed annotations through a challenging VQA benchmark of 26K questions assessing the capability to recognise recipes, ingredients, nutrition, fine-grained actions, 3D perception, object motion, and gaze direction. The powerful long-context Gemini Pro only achieves 38.5% on this benchmark, showcasing its difficulty and highlighting shortcomings in current VLMs. We additionally assess action recognition, sound recognition, and long-term video-object segmentation on HD-EPIC. HD-EPIC is 41 hours of video in 9 kitchens with digital twins of 413 kitchen fixtures, capturing 69 recipes, 59K fine-grained actions, 51K audio events, 20K object movements and 37K object masks lifted to 3D. On average, we have 263 annotations per minute of our unscripted videos.
Authors: Chung-Ho Wu, Yang-Jung Chen, Ying-Huan Chen, Jie-Ying Lee, Bo-Hsu Ke, Chun-Wei Tuan Mu, Yi-Chuan Huang, Chin-Yang Lin, Min-Hung Chen, Yen-Yu Lin, Yu-Lun Liu
Abstract: Three-dimensional scene inpainting is crucial for applications from virtual reality to architectural visualization, yet existing methods struggle with view consistency and geometric accuracy in 360{\deg} unbounded scenes. We present AuraFusion360, a novel reference-based method that enables high-quality object removal and hole filling in 3D scenes represented by Gaussian Splatting. Our approach introduces (1) depth-aware unseen mask generation for accurate occlusion identification, (2) Adaptive Guided Depth Diffusion, a zero-shot method for accurate initial point placement without requiring additional training, and (3) SDEdit-based detail enhancement for multi-view coherence. We also introduce 360-USID, the first comprehensive dataset for 360{\deg} unbounded scene inpainting with ground truth. Extensive experiments demonstrate that AuraFusion360 significantly outperforms existing methods, achieving superior perceptual quality while maintaining geometric accuracy across dramatic viewpoint changes.
Authors: Wonjoon Jin, Qi Dai, Chong Luo, Seung-Hwan Baek, Sunghyun Cho
Abstract: We present FloVD, a novel video diffusion model for camera-controllable video generation. FloVD leverages optical flow to represent the motions of the camera and moving objects. This approach offers two key benefits. Since optical flow can be directly estimated from videos, our approach allows for the use of arbitrary training videos without ground-truth camera parameters. Moreover, as background optical flow encodes 3D correlation across different viewpoints, our method enables detailed camera control by leveraging the background motion. To synthesize natural object motion while supporting detailed camera control, our framework adopts a two-stage video synthesis pipeline consisting of optical flow generation and flow-conditioned video synthesis. Extensive experiments demonstrate the superiority of our method over previous approaches in terms of accurate camera control and natural object motion synthesis.
Authors: Pengfei Zhou, Xukun Shen, Yong Hu
Abstract: Recent advances in Score Distillation Sampling (SDS) have improved 3D human generation from textual descriptions. However, existing methods still face challenges in accurately aligning 3D models with long and complex textual inputs. To address this challenge, we propose a novel framework that introduces contrastive preferences, where human-level preference models, guided by both positive and negative prompts, assist SDS for improved alignment. Specifically, we design a preference optimization module that integrates multiple models to comprehensively capture the full range of textual features. Furthermore, we introduce a negation preference module to mitigate over-optimization of irrelevant details by leveraging static-dynamic negation prompts, effectively preventing ``reward hacking". Extensive experiments demonstrate that our method achieves state-of-the-art results, significantly enhancing texture realism and visual alignment with textual descriptions, particularly for long and complex inputs.
Authors: Insu Han, Zeliang Zhang, Zhiyuan Wang, Yifan Zhu, Susan Liang, Jiani Liu, Haiting Lin, Mingjie Zhao, Chenliang Xu, Kun Wan, Wentian Zhao
Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable performance across diverse applications. However, their computational overhead during deployment remains a critical bottleneck. While Key-Value (KV) caching effectively trades memory for computation to enhance inference efficiency, the growing memory footprint from extensive KV caches significantly reduces throughput and restricts prolonged deployment on memory-constrained GPU devices. To address this challenge, we propose CalibQuant, a simple yet highly effective visual quantization strategy that drastically reduces both memory and computational overhead. Specifically, CalibQuant introduces an extreme 1-bit quantization scheme, complemented by novel post-scaling and calibration techniques tailored to the intrinsic patterns of KV caches, thereby ensuring high efficiency without compromising model performance. Leveraging Triton for runtime optimization, we achieve a 10x throughput increase on InternVL models. Our method is designed to be plug-and-play, seamlessly integrating with various existing MLLMs without requiring architectural changes. Extensive experiments confirm that our approach significantly reduces memory usage while maintaining computational efficiency and preserving multimodal capabilities. Codes are available at https://github.com/insuhan/calibquant.
Authors: Sanghyun Yi, Qingfeng Liu, Mostafa El-Khamy
Abstract: Diffusion Transformers for video generation have gained significant research interest since the impressive performance of SORA. Efficient deployment of such generative-AI models on GPUs has been demonstrated with dynamic quantization. However, resource-constrained devices cannot support dynamic quantization, and need static quantization of the models for their efficient deployment on AI processors. In this paper, we propose a novel method for the post-training quantization of OpenSora\cite{opensora}, a Video Diffusion Transformer, without relying on dynamic quantization techniques. Our approach employs static quantization, achieving video quality comparable to FP16 and dynamically quantized ViDiT-Q methods, as measured by CLIP, and VQA metrics. In particular, we utilize per-step calibration data to adequately provide a post-training statically quantized model for each time step, incorporating channel-wise quantization for weights and tensor-wise quantization for activations. By further applying the smooth-quantization technique, we can obtain high-quality video outputs with the statically quantized models. Extensive experimental results demonstrate that static quantization can be a viable alternative to dynamic quantization for video diffusion transformers, offering a more efficient approach without sacrificing performance.
Authors: Chen-Wei Chang, Cheng-De Fan, Chia-Che Chang, Yi-Chen Lo, Yu-Chee Tseng, Jiun-Long Huang, Yu-Lun Liu
Abstract: Color constancy methods often struggle to generalize across different camera sensors due to varying spectral sensitivities. We present GCC, which leverages diffusion models to inpaint color checkers into images for illumination estimation. Our key innovations include (1) a single-step deterministic inference approach that inpaints color checkers reflecting scene illumination, (2) a Laplacian decomposition technique that preserves checker structure while allowing illumination-dependent color adaptation, and (3) a mask-based data augmentation strategy for handling imprecise color checker annotations. By harnessing rich priors from pre-trained diffusion models, GCC demonstrates strong robustness in challenging cross-camera scenarios. These results highlight our method's effective generalization capability across different camera characteristics without requiring sensor-specific training, making it a versatile and practical solution for real-world applications.
Authors: Xin Ye, Burhaneddin Yaman, Sheng Cheng, Feng Tao, Abhirup Mallik, Liu Ren
Abstract: Bird's-eye-view (BEV) representations play a crucial role in autonomous driving tasks. Despite recent advancements in BEV generation, inherent noise, stemming from sensor limitations and the learning process, remains largely unaddressed, resulting in suboptimal BEV representations that adversely impact the performance of downstream tasks. To address this, we propose BEVDiffuser, a novel diffusion model that effectively denoises BEV feature maps using the ground-truth object layout as guidance. BEVDiffuser can be operated in a plug-and-play manner during training time to enhance existing BEV models without requiring any architectural modifications. Extensive experiments on the challenging nuScenes dataset demonstrate BEVDiffuser's exceptional denoising and generation capabilities, which enable significant enhancement to existing BEV models, as evidenced by notable improvements of 12.3\% in mAP and 10.1\% in NDS achieved for 3D object detection without introducing additional computational complexity. Moreover, substantial improvements in long-tail object detection and under challenging weather and lighting conditions further validate BEVDiffuser's effectiveness in denoising and enhancing BEV representations.
Authors: SeonHwa Kim, Jiwon Kim, Soobin Park, Donghoon Ahn, Jiwon Kang, Seungryong Kim, Kyong Hwan Jin, Eunju Cha
Abstract: Score distillation sampling (SDS) demonstrates a powerful capability for text-conditioned 2D image and 3D object generation by distilling the knowledge from learned score functions. However, SDS often suffers from blurriness caused by noisy gradients. When SDS meets the image editing, such degradations can be reduced by adjusting bias shifts using reference pairs, but the de-biasing techniques are still corrupted by erroneous gradients. To this end, we introduce Identity-preserving Distillation Sampling (IDS), which compensates for the gradient leading to undesired changes in the results. Based on the analysis that these errors come from the text-conditioned scores, a new regularization technique, called fixed-point iterative regularization (FPR), is proposed to modify the score itself, driving the preservation of the identity even including poses and structures. Thanks to a self-correction by FPR, the proposed method provides clear and unambiguous representations corresponding to the given prompts in image-to-image editing and editable neural radiance field (NeRF). The structural consistency between the source and the edited data is obviously maintained compared to other state-of-the-art methods.
Authors: Zezeng Li, Xiaoyu Du, Na Lei, Liming Chen, Weimin Wang
Abstract: Adversarial attacks exploit the vulnerability of deep models against adversarial samples. Existing point cloud attackers are tailored to specific models, iteratively optimizing perturbations based on gradients in either a white-box or black-box setting. Despite their promising attack performance, they often struggle to produce transferable adversarial samples due to overfitting the specific parameters of surrogate models. To overcome this issue, we shift our focus to the data distribution itself and introduce a novel approach named NoPain, which employs optimal transport (OT) to identify the inherent singular boundaries of the data manifold for cross-network point cloud attacks. Specifically, we first calculate the OT mapping from noise to the target feature space, then identify singular boundaries by locating non-differentiable positions. Finally, we sample along singular boundaries to generate adversarial point clouds. Once the singular boundaries are determined, NoPain can efficiently produce adversarial samples without the need of iterative updates or guidance from the surrogate classifiers. Extensive experiments demonstrate that the proposed end-to-end method outperforms baseline approaches in terms of both transferability and efficiency, while also maintaining notable advantages even against defense strategies. Code and model are available at https://github.com/cognaclee/nopain
Authors: Yufan Zhang, Yu Ji, Yu Guo, Jinwei Ye
Abstract: We present a snapshot imaging technique for recovering 3D surrounding views of miniature scenes. Due to their intricacy, miniature scenes with objects sized in millimeters are difficult to reconstruct, yet miniatures are common in life and their 3D digitalization is desirable. We design a catadioptric imaging system with a single camera and eight pairs of planar mirrors for snapshot 3D reconstruction from a dollhouse perspective. We place paired mirrors on nested pyramid surfaces for capturing surrounding multi-view images in a single shot. Our mirror design is customizable based on the size of the scene for optimized view coverage. We use the 3D Gaussian Splatting (3DGS) representation for scene reconstruction and novel view synthesis. We overcome the challenge posed by our sparse view input by integrating visual hull-derived depth constraint. Our method demonstrates state-of-the-art performance on a variety of synthetic and real miniature scenes.
Authors: Guanyao Wu, Haoyu Liu, Hongming Fu, Yichuan Peng, Jinyuan Liu, Xin Fan, Risheng Liu
Abstract: Multi-modality image fusion, particularly infrared and visible, plays a crucial role in integrating diverse modalities to enhance scene understanding. Although early research prioritized visual quality, preserving fine details and adapting to downstream tasks remains challenging. Recent approaches attempt task-specific design but rarely achieve "The Best of Both Worlds" due to inconsistent optimization goals. To address these issues, we propose a novel method that leverages the semantic knowledge from the Segment Anything Model (SAM) to Grow the quality of fusion results and Enable downstream task adaptability, namely SAGE. Specifically, we design a Semantic Persistent Attention (SPA) Module that efficiently maintains source information via the persistent repository while extracting high-level semantic priors from SAM. More importantly, to eliminate the impractical dependence on SAM during inference, we introduce a bi-level optimization-driven distillation mechanism with triplet losses, which allow the student network to effectively extract knowledge. Extensive experiments show that our method achieves a balance between high-quality visual results and downstream task adaptability while maintaining practical deployment efficiency. The code is available at https://github.com/RollingPlain/SAGE_IVIF.
Authors: Nuria Alina Chandra, Ryan Murtfeldt, Lin Qiu, Arnab Karmakar, Hannah Lee, Emmanuel Tanumihardja, Kevin Farhat, Ben Caffee, Sejin Paik, Changyeon Lee, Jongwook Choi, Aerin Kim, Oren Etzioni
Abstract: In the age of increasingly realistic generative AI, robust deepfake detection is essential for mitigating fraud and disinformation. While many deepfake detectors report high accuracy on academic datasets, we show that these academic benchmarks are out of date and not representative of real-world deepfakes. We introduce Deepfake-Eval-2024, a new deepfake detection benchmark consisting of in-the-wild deepfakes collected from social media and deepfake detection platform users in 2024. Deepfake-Eval-2024 consists of 45 hours of videos, 56.5 hours of audio, and 1,975 images, encompassing the latest manipulation technologies. The benchmark contains diverse media content from 88 different websites in 52 different languages. We find that the performance of open-source state-of-the-art deepfake detection models drops precipitously when evaluated on Deepfake-Eval-2024, with AUC decreasing by 50% for video, 48% for audio, and 45% for image models compared to previous benchmarks. We also evaluate commercial deepfake detection models and models finetuned on Deepfake-Eval-2024, and find that they have superior performance to off-the-shelf open-source models, but do not yet reach the accuracy of deepfake forensic analysts. The dataset is available at https://github.com/nuriachandra/Deepfake-Eval-2024.
Authors: Alaa Mazouz, Sumanta Chaudhuri, Marco Cagnanzzo, Mihai Mitrea, Enzo Tartaglione, Attilio Fiandrotti
Abstract: Learnable Image Compression (LIC) has shown the potential to outperform standardized video codecs in RD efficiency, prompting the research for hardware-friendly implementations. Most existing LIC hardware implementations prioritize latency to RD-efficiency and through an extensive exploration of the hardware design space. We present a novel design paradigm where the burden of tuning the design for a specific hardware platform is shifted towards model dimensioning and without compromising on RD-efficiency. First, we design a framework for distilling a leaner student LIC model from a reference teacher: by tuning a single model hyperparameters, we can meet the constraints of different hardware platforms without a complex hardware design exploration. Second, we propose a hardware-friendly implementation of the Generalized Divisive Normalization - GDN activation that preserves RD efficiency even post parameter quantization. Third, we design a pipelined FPGA configuration which takes full advantage of available FPGA resources by leveraging parallel processing and optimizing resource allocation. Our experiments with a state of the art LIC model show that we outperform all existing FPGA implementations while performing very close to the original model.
Authors: Ali Bahri, Moslem Yazdanpanah, Mehrdad Noori, Sahar Dastani, Milad Cheraghalikhani, David Osowiechi, Gustavo Adolfo Vargas Hakim, Farzad Beizaee, Ismail Ben Ayed, Christian Desrosiers
Abstract: State space models have shown significant promise in Natural Language Processing (NLP) and, more recently, computer vision. This paper introduces a new methodology leveraging Mamba and Masked Autoencoder networks for point cloud data in both supervised and self-supervised learning. We propose three key contributions to enhance Mamba's capability in processing complex point cloud structures. First, we exploit the spectrum of a graph Laplacian to capture patch connectivity, defining an isometry-invariant traversal order that is robust to viewpoints and better captures shape manifolds than traditional 3D grid-based traversals. Second, we adapt segmentation via a recursive patch partitioning strategy informed by Laplacian spectral components, allowing finer integration and segment analysis. Third, we address token placement in Masked Autoencoder for Mamba by restoring tokens to their original positions, which preserves essential order and improves learning. Extensive experiments demonstrate the improvements of our approach in classification, segmentation, and few-shot tasks over state-of-the-art baselines.
Authors: Ruslan Gokhman, Jialu Li, Youshan Zhang
Abstract: Automating teaching presents unique challenges, as replicating human interaction and adaptability is complex. Automated systems cannot often provide nuanced, real-time feedback that aligns with students' individual learning paces or comprehension levels, which can hinder effective support for diverse needs. This is especially challenging in fields where abstract concepts require adaptive explanations. In this paper, we propose a vision language retrieval augmented generation (named VL-RAG) system that has the potential to bridge this gap by delivering contextually relevant, visually enriched responses that can enhance comprehension. By leveraging a database of tailored answers and images, the VL-RAG system can dynamically retrieve information aligned with specific questions, creating a more interactive and engaging experience that fosters deeper understanding and active student participation. It allows students to explore concepts visually and verbally, promoting deeper understanding and reducing the need for constant human oversight while maintaining flexibility to expand across different subjects and course material.
Authors: Weixi Zheng, Aoling Huang, Jingping Yuan, Haoyu Zhao, Zhou Zhao, Yongchao Xu, Thierry G\'eraud
Abstract: In histopathology, intelligent diagnosis of Whole Slide Images (WSIs) is essential for automating and objectifying diagnoses, reducing the workload of pathologists. However, diagnostic models often face the challenge of forgetting previously learned data during incremental training on datasets from different sources. To address this issue, we propose a new framework PaGMIL to mitigate catastrophic forgetting in breast cancer WSI classification. Our framework introduces two key components into the common MIL model architecture. First, it leverages microscopic pathological prior to select more accurate and diverse representative patches for MIL. Secondly, it trains separate classification heads for each task and uses macroscopic pathological prior knowledge, treating the thumbnail as a prompt guide (PG) to select the appropriate classification head. We evaluate the continual learning performance of PaGMIL across several public breast cancer datasets. PaGMIL achieves a better balance between the performance of the current task and the retention of previous tasks, outperforming other continual learning methods. Our code will be open-sourced upon acceptance.
Authors: Xianjie Liu, Keren Fu, Qijun Zhao
Abstract: Dichotomous Image Segmentation (DIS) is a high-precision object segmentation task for high-resolution natural images. The current mainstream methods focus on the optimization of local details but overlook the fundamental challenge of modeling the integrity of objects. We have found that the depth integrity-prior implicit in the the pseudo-depth maps generated by Depth Anything Model v2 and the local detail features of image patches can jointly address the above dilemmas. Based on the above findings, we have designed a novel Patch-Depth Fusion Network (PDFNet) for high-precision dichotomous image segmentation. The core of PDFNet consists of three aspects. Firstly, the object perception is enhanced through multi-modal input fusion. By utilizing the patch fine-grained strategy, coupled with patch selection and enhancement, the sensitivity to details is improved. Secondly, by leveraging the depth integrity-prior distributed in the depth maps, we propose an integrity-prior loss to enhance the uniformity of the segmentation results in the depth maps. Finally, we utilize the features of the shared encoder and, through a simple depth refinement decoder, improve the ability of the shared encoder to capture subtle depth-related information in the images. Experiments on the DIS-5K dataset show that PDFNet significantly outperforms state-of-the-art non-diffusion methods. Due to the incorporation of the depth integrity-prior, PDFNet achieves or even surpassing the performance of the latest diffusion-based methods while using less than 11% of the parameters of diffusion-based methods. The source code at https://github.com/Tennine2077/PDFNet
Authors: Shiu-hong Kao, Yu-Wing Tai, Chi-Keung Tang
Abstract: Reasoning segmentation is a challenging vision-language task that aims to output the segmentation mask with respect to a complex, implicit, and even non-visual query text. Previous works incorporated multimodal Large Language Models (MLLMs) with segmentation models to approach the difficult problem. However, their segmentation quality often falls short in complex cases, particularly when dealing with out-of-domain objects with intricate structures, blurry boundaries, occlusions, or high similarity with surroundings. In this paper, we introduce ThinkFirst, a training-free reasoning segmentation framework that leverages GPT's chain of thought to address these challenging cases. Our approach allows GPT-4o or other powerful MLLMs to generate a detailed, chain-of-thought description of an image. This summarized description is then passed to a language-instructed segmentation assistant to aid the segmentation process. Our framework allows users to easily interact with the segmentation agent using multimodal inputs, such as easy text and image scribbles, for successive refinement or communication. We evaluate the performance of ThinkFirst on diverse objects. Extensive experiments show that, this zero-shot-CoT approach significantly improves the vanilla reasoning segmentation agent, both qualitatively and quantitatively, while being less sensitive or critical to user-supplied prompts after Thinking First.
Authors: Junwei Luo, Yingying Zhang, Xue Yang, Kang Wu, Qi Zhu, Lei Liang, Jingdong Chen, Yansheng Li
Abstract: Efficient vision-language understanding of large Remote Sensing Images (RSIs) is meaningful but challenging. Current Large Vision-Language Models (LVLMs) typically employ limited pre-defined grids to process images, leading to information loss when handling gigapixel RSIs. Conversely, using unlimited grids significantly increases computational costs. To preserve image details while reducing computational complexity, we propose a text-guided token pruning method with Dynamic Image Pyramid (DIP) integration. Our method introduces: (i) a Region Focus Module (RFM) that leverages text-aware region localization capability to identify critical vision tokens, and (ii) a coarse-to-fine image tile selection and vision token pruning strategy based on DIP, which is guided by RFM outputs and avoids directly processing the entire large imagery. Additionally, existing benchmarks for evaluating LVLMs' perception ability on large RSI suffer from limited question diversity and constrained image sizes. We construct a new benchmark named LRS-VQA, which contains 7,333 QA pairs across 8 categories, with image length up to 27,328 pixels. Our method outperforms existing high-resolution strategies on four datasets using the same data. Moreover, compared to existing token reduction methods, our approach demonstrates higher efficiency under high-resolution settings. Dataset and code are in https://github.com/VisionXLab/LRS-VQA.
Authors: Quoc-Tien Nguyen, Hong-Hai Nguyen, Van-Thong Huynh
Abstract: In this study, we present an approach for efficient spatiotemporal feature extraction using MobileNetV4 and a multi-scale 3D MLP-Mixer-based temporal aggregation module. MobileNetV4, with its Universal Inverted Bottleneck (UIB) blocks, serves as the backbone for extracting hierarchical feature representations from input image sequences, ensuring both computational efficiency and rich semantic encoding. To capture temporal dependencies, we introduce a three-level MLP-Mixer module, which processes spatial features at multiple resolutions while maintaining structural integrity. Experimental results on the ABAW 8th competition demonstrate the effectiveness of our approach, showing promising performance in affective behavior analysis. By integrating an efficient vision backbone with a structured temporal modeling mechanism, the proposed framework achieves a balance between computational efficiency and predictive accuracy, making it well-suited for real-time applications in mobile and embedded computing environments.
Authors: Jun Yu, Lingsi Zhu, Yanjun Chi, Yunxiang Zhang, Yang Zheng, Yongqi Wang, Xilong Lu
Abstract: Emotional Mimicry Intensity (EMI) estimation plays a pivotal role in understanding human social behavior and advancing human-computer interaction. The core challenges lie in dynamic correlation modeling and robust fusion of multimodal temporal signals. To address the limitations of existing methods--insufficient exploitation of cross-modal synergies, sensitivity to noise, and constrained fine-grained alignment capabilities--this paper proposes a dual-stage cross-modal alignment framework. Stage 1 develops vision-text and audio-text contrastive learning networks based on a CLIP architecture, achieving preliminary feature-space alignment through modality-decoupled pre-training. Stage 2 introduces a temporal-aware dynamic fusion module integrating Temporal Convolutional Networks (TCN) and gated bidirectional LSTM to capture macro-evolution patterns of facial expressions and local dynamics of acoustic features, respectively. A novel quality-guided fusion strategy further enables differentiable weight allocation for modality compensation under occlusion and noise. Experiments on the Hume-Vidmimic2 dataset demonstrate superior performance with an average Pearson correlation coefficient of 0.51 across six emotion dimensions on the validate set. Remarkably, our method achieved 0.68 on the test set, securing runner-up in the EMI Challenge Track of the 8th ABAW (Affective Behavior Analysis in the Wild) Competition, offering a novel pathway for fine-grained emotion analysis in open environments.
Authors: Byung Hyun Lee, Sungjin Lim, Se Young Chun
Abstract: Fine-tuning based concept erasing has demonstrated promising results in preventing generation of harmful contents from text-to-image diffusion models by removing target concepts while preserving remaining concepts. To maintain the generation capability of diffusion models after concept erasure, it is necessary to remove only the image region containing the target concept when it locally appears in an image, leaving other regions intact. However, prior arts often compromise fidelity of the other image regions in order to erase the localized target concept appearing in a specific area, thereby reducing the overall performance of image generation. To address these limitations, we first introduce a framework called localized concept erasure, which allows for the deletion of only the specific area containing the target concept in the image while preserving the other regions. As a solution for the localized concept erasure, we propose a training-free approach, dubbed Gated Low-rank adaptation for Concept Erasure (GLoCE), that injects a lightweight module into the diffusion model. GLoCE consists of low-rank matrices and a simple gate, determined only by several generation steps for concepts without training. By directly applying GLoCE to image embeddings and designing the gate to activate only for target concepts, GLoCE can selectively remove only the region of the target concepts, even when target and remaining concepts coexist within an image. Extensive experiments demonstrated GLoCE not only improves the image fidelity to text prompts after erasing the localized target concepts, but also outperforms prior arts in efficacy, specificity, and robustness by large margin and can be extended to mass concept erasure.
Authors: Sumin In, Youngdong Jang, Utae Jeong, MinHyuk Jang, Hyeongcheol Park, Eunbyung Park, Sangpil Kim
Abstract: 3D Gaussian Splatting (3DGS) enables rapid differentiable rendering for 3D reconstruction and novel view synthesis, leading to its widespread commercial use. Consequently, copyright protection via watermarking has become critical. However, because 3DGS relies on millions of Gaussians, which require gigabytes of storage, efficient transfer and storage require compression. Existing 3DGS watermarking methods are vulnerable to quantization-based compression, often resulting in the loss of the embedded watermark. To address this challenge, we propose a novel watermarking method that ensures watermark robustness after model compression while maintaining high rendering quality. In detail, we incorporate a quantization distortion layer that simulates compression during training, preserving the watermark under quantization-based compression. Also, we propose a learnable watermark embedding feature that embeds the watermark into the anchor feature, ensuring structural consistency and seamless integration into the 3D scene. Furthermore, we present a frequency-aware anchor growing mechanism to enhance image quality in high-frequency regions by effectively identifying Guassians within these regions. Experimental results confirm that our method preserves the watermark and maintains superior image quality under high compression, validating it as a promising approach for a secure 3DGS model.
Authors: Harshal Kausadikar, Tanvi Kale, Onkar Susladkar, Sparsh Mittal
Abstract: In medieval India, the Marathi language was written using the Modi script. The texts written in Modi script include extensive knowledge about medieval sciences, medicines, land records and authentic evidence about Indian history. Around 40 million documents are in poor condition and have not yet been transliterated. Furthermore, only a few experts in this domain can transliterate this script into English or Devanagari. Most of the past research predominantly focuses on individual character recognition. A system that can transliterate Modi script documents to Devanagari script is needed. We propose the MoDeTrans dataset, comprising 2,043 images of Modi script documents accompanied by their corresponding textual transliterations in Devanagari. We further introduce MoScNet (\textbf{Mo}di \textbf{Sc}ript \textbf{Net}work), a novel Vision-Language Model (VLM) framework for transliterating Modi script images into Devanagari text. MoScNet leverages Knowledge Distillation, where a student model learns from a teacher model to enhance transliteration performance. The final student model of MoScNet has better performance than the teacher model while having 163$\times$ lower parameters. Our work is the first to perform direct transliteration from the handwritten Modi script to the Devanagari script. MoScNet also shows competitive results on the optical character recognition (OCR) task.
Authors: Zhenglin Zhou, Fan Ma, Hehe Fan, Tat-Seng Chua
Abstract: Animatable head avatar generation typically requires extensive data for training. To reduce the data requirements, a natural solution is to leverage existing data-free static avatar generation methods, such as pre-trained diffusion models with score distillation sampling (SDS), which align avatars with pseudo ground-truth outputs from the diffusion model. However, directly distilling 4D avatars from video diffusion often leads to over-smooth results due to spatial and temporal inconsistencies in the generated video. To address this issue, we propose Zero-1-to-A, a robust method that synthesizes a spatial and temporal consistency dataset for 4D avatar reconstruction using the video diffusion model. Specifically, Zero-1-to-A iteratively constructs video datasets and optimizes animatable avatars in a progressive manner, ensuring that avatar quality increases smoothly and consistently throughout the learning process. This progressive learning involves two stages: (1) Spatial Consistency Learning fixes expressions and learns from front-to-side views, and (2) Temporal Consistency Learning fixes views and learns from relaxed to exaggerated expressions, generating 4D avatars in a simple-to-complex manner. Extensive experiments demonstrate that Zero-1-to-A improves fidelity, animation quality, and rendering speed compared to existing diffusion-based methods, providing a solution for lifelike avatar creation. Code is publicly available at: https://github.com/ZhenglinZhou/Zero-1-to-A.
Authors: Zichen Liu, Kunlun Xu, Bing Su, Xu Zou, Yuxin Peng, Jiahuan Zhou
Abstract: Pre-trained on tremendous image-text pairs, vision-language models like CLIP have demonstrated promising zero-shot generalization across numerous image-based tasks. However, extending these capabilities to video tasks remains challenging due to limited labeled video data and high training costs. Recent video prompting methods attempt to adapt CLIP for video tasks by introducing learnable prompts, but they typically rely on a single static prompt for all video sequences, overlooking the diverse temporal dynamics and spatial variations that exist across frames. This limitation significantly hinders the model's ability to capture essential temporal information for effective video understanding. To address this, we propose an integrated Spatial-TempOral dynamic Prompting (STOP) model which consists of two complementary modules, the intra-frame spatial prompting and inter-frame temporal prompting. Our intra-frame spatial prompts are designed to adaptively highlight discriminative regions within each frame by leveraging intra-frame attention and temporal variation, allowing the model to focus on areas with substantial temporal dynamics and capture fine-grained spatial details. Additionally, to highlight the varying importance of frames for video understanding, we further introduce inter-frame temporal prompts, dynamically inserting prompts between frames with high temporal variance as measured by frame similarity. This enables the model to prioritize key frames and enhances its capacity to understand temporal dependencies across sequences. Extensive experiments on various video benchmarks demonstrate that STOP consistently achieves superior performance against state-of-the-art methods. The code is available at https://github.com/zhoujiahuan1991/CVPR2025-STOP.
Authors: Yike Yuan, Ziyu Wang, Zihao Huang, Defa Zhu, Xun Zhou, Jingyi Yu, Qiyang Min
Abstract: Diffusion models have emerged as mainstream framework in visual generation. Building upon this success, the integration of Mixture of Experts (MoE) methods has shown promise in enhancing model scalability and performance. In this paper, we introduce Race-DiT, a novel MoE model for diffusion transformers with a flexible routing strategy, Expert Race. By allowing tokens and experts to compete together and select the top candidates, the model learns to dynamically assign experts to critical tokens. Additionally, we propose per-layer regularization to address challenges in shallow layer learning, and router similarity loss to prevent mode collapse, ensuring better expert utilization. Extensive experiments on ImageNet validate the effectiveness of our approach, showcasing significant performance gains while promising scaling properties.
Authors: Tim Seizinger, Florin-Alexandru Vasluianu, Marcos V. Conde, Zongwei Wu, Radu Timofte
Abstract: Bokeh rendering methods play a key role in creating the visually appealing, softly blurred backgrounds seen in professional photography. While recent learning-based approaches show promising results, generating realistic Bokeh with variable strength remains challenging. Existing methods require additional inputs and suffer from unrealistic Bokeh reproduction due to reliance on synthetic data. In this work, we propose Bokehlicious, a highly efficient network that provides intuitive control over Bokeh strength through an Aperture-Aware Attention mechanism, mimicking the physical lens aperture. To further address the lack of high-quality real-world data, we present RealBokeh, a novel dataset featuring 23,000 high-resolution (24-MP) images captured by professional photographers, covering diverse scenes with varied aperture and focal length settings. Evaluations on both our new RealBokeh and established Bokeh rendering benchmarks show that Bokehlicious consistently outperforms SOTA methods while significantly reducing computational cost and exhibiting strong zero-shot generalization. Our method and dataset further extend to defocus deblurring, achieving competitive results on the RealDOF benchmark. Our code and data can be found at https://github.com/TimSeizinger/Bokehlicious
Authors: Chun-Han Yao, Yiming Xie, Vikram Voleti, Huaizu Jiang, Varun Jampani
Abstract: We present Stable Video 4D 2.0 (SV4D 2.0), a multi-view video diffusion model for dynamic 3D asset generation. Compared to its predecessor SV4D, SV4D 2.0 is more robust to occlusions and large motion, generalizes better to real-world videos, and produces higher-quality outputs in terms of detail sharpness and spatio-temporal consistency. We achieve this by introducing key improvements in multiple aspects: 1) network architecture: eliminating the dependency of reference multi-views and designing blending mechanism for 3D and frame attention, 2) data: enhancing quality and quantity of training data, 3) training strategy: adopting progressive 3D-4D training for better generalization, and 4) 4D optimization: handling 3D inconsistency and large motion via 2-stage refinement and progressive frame sampling. Extensive experiments demonstrate significant performance gain by SV4D 2.0 both visually and quantitatively, achieving better detail (-14\% LPIPS) and 4D consistency (-44\% FV4D) in novel-view video synthesis and 4D optimization (-12\% LPIPS and -24\% FV4D) compared to SV4D. Project page: https://sv4d20.github.io.
Authors: Yingying Fan, Quanwei Yang, Kaisiyuan Wang, Hang Zhou, Yingying Li, Haocheng Feng, Errui Ding, Yu Wu, Jingdong Wang
Abstract: Current digital human studies focusing on lip-syncing and body movement are no longer sufficient to meet the growing industrial demand, while human video generation techniques that support interacting with real-world environments (e.g., objects) have not been well investigated. Despite human hand synthesis already being an intricate problem, generating objects in contact with hands and their interactions presents an even more challenging task, especially when the objects exhibit obvious variations in size and shape. To tackle these issues, we present a novel video Reenactment framework focusing on Human-Object Interaction (HOI) via an adaptive Layout-instructed Diffusion model (Re-HOLD). Our key insight is to employ specialized layout representation for hands and objects, respectively. Such representations enable effective disentanglement of hand modeling and object adaptation to diverse motion sequences. To further improve the generation quality of HOI, we design an interactive textural enhancement module for both hands and objects by introducing two independent memory banks. We also propose a layout adjustment strategy for the cross-object reenactment scenario to adaptively adjust unreasonable layouts caused by diverse object sizes during inference. Comprehensive qualitative and quantitative evaluations demonstrate that our proposed framework significantly outperforms existing methods. Project page: https://fyycs.github.io/Re-HOLD.
Authors: Duanrui Yu, Jing You, Xin Pei, Anqi Qu, Dingyu Wang, Shaocheng Jia
Abstract: Collaborative perception allows real-time inter-agent information exchange and thus offers invaluable opportunities to enhance the perception capabilities of individual agents. However, limited communication bandwidth in practical scenarios restricts the inter-agent data transmission volume, consequently resulting in performance declines in collaborative perception systems. This implies a trade-off between perception performance and communication cost. To address this issue, we propose Which2comm, a novel multi-agent 3D object detection framework leveraging object-level sparse features. By integrating semantic information of objects into 3D object detection boxes, we introduce semantic detection boxes (SemDBs). Innovatively transmitting these information-rich object-level sparse features among agents not only significantly reduces the demanding communication volume, but also improves 3D object detection performance. Specifically, a fully sparse network is constructed to extract SemDBs from individual agents; a temporal fusion approach with a relative temporal encoding mechanism is utilized to obtain the comprehensive spatiotemporal features. Extensive experiments on the V2XSet and OPV2V datasets demonstrate that Which2comm consistently outperforms other state-of-the-art methods on both perception performance and communication cost, exhibiting better robustness to real-world latency. These results present that for multi-agent collaborative 3D object detection, transmitting only object-level sparse features is sufficient to achieve high-precision and robust performance.
Authors: Zhengqing Gao, Dongting Hu, Jia-Wang Bian, Huan Fu, Yan Li, Tongliang Liu, Mingming Gong, Kun Zhang
Abstract: 3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis but is limited by the substantial number of Gaussian primitives required, posing challenges for deployment on lightweight devices. Recent methods address this issue by compressing the storage size of densified Gaussians, yet fail to preserve rendering quality and efficiency. To overcome these limitations, we propose ProtoGS to learn Gaussian prototypes to represent Gaussian primitives, significantly reducing the total Gaussian amount without sacrificing visual quality. Our method directly uses Gaussian prototypes to enable efficient rendering and leverage the resulting reconstruction loss to guide prototype learning. To further optimize memory efficiency during training, we incorporate structure-from-motion (SfM) points as anchor points to group Gaussian primitives. Gaussian prototypes are derived within each group by clustering of K-means, and both the anchor points and the prototypes are optimized jointly. Our experiments on real-world and synthetic datasets prove that we outperform existing methods, achieving a substantial reduction in the number of Gaussians, and enabling high rendering speed while maintaining or even enhancing rendering fidelity.
Authors: Hyungyu Choi, Young Kyun Jang, Chanho Eom
Abstract: Vision-language models like CLIP have shown impressive capabilities in aligning images and text, but they often struggle with lengthy and detailed text descriptions because of their training focus on short and concise captions. We present GOAL (Global-local Object Alignment Learning), a novel fine-tuning method that enhances CLIP's ability to handle lengthy text by leveraging both global and local semantic alignments between image and lengthy text. Our approach consists of two key components: Local Image-Sentence Matching (LISM), which identifies corresponding pairs between image segments and descriptive sentences, and Token Similarity-based Learning (TSL), which efficiently propagates local element attention through these matched pairs. Evaluating GOAL on three new benchmarks for image-lengthy text retrieval, we demonstrate significant improvements over baseline CLIP fine-tuning, establishing a simple yet effective approach for adapting CLIP to detailed textual descriptions. Through extensive experiments, we show that our method's focus on local semantic alignment alongside global context leads to more nuanced and representative embeddings, particularly beneficial for tasks requiring fine-grained understanding of lengthy text descriptions.
Authors: Yuzhi Li, Haojun Xu, Feng Tian
Abstract: With the rising popularity of short video platforms, the demand for video production has increased substantially. However, high-quality video creation continues to rely heavily on professional editing skills and a nuanced understanding of visual language. To address this challenge, the Shot Sequence Ordering (SSO) task in AI-assisted video editing has emerged as a pivotal approach for enhancing video storytelling and the overall viewing experience. Nevertheless, the progress in this field has been impeded by a lack of publicly available benchmark datasets. In response, this paper introduces two novel benchmark datasets, AVE-Order and ActivityNet-Order. Additionally, we employ the Kendall Tau distance as an evaluation metric for the SSO task and propose the Kendall Tau Distance-Cross Entropy Loss. We further introduce the concept of Cinematology Embedding, which incorporates movie metadata and shot labels as prior knowledge into the SSO model, and constructs the AVE-Meta dataset to validate the method's effectiveness. Experimental results indicate that the proposed loss function and method substantially enhance SSO task accuracy. All datasets are publicly accessible at https://github.com/litchiar/ShotSeqBench.
Authors: Kelly O. Marshall, Omid Poursaeed, Sergiu Oprea, Amit Kumar, Anushrut Jignasu, Chinmay Hegde, Yilei Li, Rakesh Ranjan
Abstract: 3D indoor scene generation is an important problem for the design of digital and real-world environments. To automate this process, a scene generation model should be able to not only generate plausible scene layouts, but also take into consideration visual features and style preferences. Existing methods for this task exhibit very limited control over these attributes, only allowing text inputs in the form of simple object-level descriptions or pairwise spatial relationships. Our proposed method Decorum enables users to control the scene generation process with natural language by adopting language-based representations at each stage. This enables us to harness recent advancements in Large Language Models (LLMs) to model language-to-language mappings. In addition, we show that using a text-based representation allows us to select furniture for our scenes using a novel object retrieval method based on multimodal LLMs. Evaluations on the benchmark 3D-FRONT dataset show that our methods achieve improvements over existing work in text-conditioned scene synthesis and object retrieval.
Authors: Yiheng Zhong, Zihong Luo, Chengzhi Liu, Feilong Tang, Zelin Peng, Ming Hu, Yingzhen Hu, Jionglong Su, Zongyuan Geand, Imran Razzak
Abstract: Segment Anything Model (SAM) demonstrates powerful zero-shot capabilities; however, its accuracy and robustness significantly decrease when applied to medical image segmentation. Existing methods address this issue through modality fusion, integrating textual and image information to provide more detailed priors. In this study, we argue that the granularity of text and the domain gap affect the accuracy of the priors. Furthermore, the discrepancy between high-level abstract semantics and pixel-level boundary details in images can introduce noise into the fusion process. To address this, we propose Prior-Guided SAM (PG-SAM), which employs a fine-grained modality prior aligner to leverage specialized medical knowledge for better modality alignment. The core of our method lies in efficiently addressing the domain gap with fine-grained text from a medical LLM. Meanwhile, it also enhances the priors' quality after modality alignment, ensuring more accurate segmentation. In addition, our decoder enhances the model's expressive capabilities through multi-level feature fusion and iterative mask optimizer operations, supporting unprompted learning. We also propose a unified pipeline that effectively supplies high-quality semantic information to SAM. Extensive experiments on the Synapse dataset demonstrate that the proposed PG-SAM achieves state-of-the-art performance. Our anonymous code is released at https://github.com/logan-0623/PG-SAM.
Authors: Sherry X. Chen, Misha Sra, Pradeep Sen
Abstract: Although natural language instructions offer an intuitive way to guide automated image editing, deep-learning models often struggle to achieve high-quality results, largely due to the difficulty of creating large, high-quality training datasets. To do this, previous approaches have typically relied on text-to-image (T2I) generative models to produce pairs of original and edited images that simulate the input/output of an instruction-guided image-editing model. However, these image pairs often fail to align with the specified edit instructions due to the limitations of T2I models, which negatively impacts models trained on such datasets. To address this, we present Instruct-CLIP (I-CLIP), a selfsupervised method that learns the semantic changes between original and edited images to refine and better align the instructions in existing datasets. Furthermore, we adapt Instruct-CLIP to handle noisy latent images and diffusion timesteps so that it can be used to train latent diffusion models (LDMs) and efficiently enforce alignment between the edit instruction and the image changes in latent space at any step of the diffusion pipeline. We use Instruct-CLIP to correct the InstructPix2Pix dataset and get over 120K refined samples we then use to fine-tune their model, guided by our novel I-CLIP-based loss function. The resulting model can produce edits that are more aligned with the given instructions. Our code and dataset are available at https://github.com/SherryXTChen/Instruct-CLIP.git.
Authors: Wencheng Zhu, Yuexin Wang, Hongxuan Li, Pengfei Zhu, Qinghua Hu
Abstract: Vision-language models bridge visual and linguistic understanding and have proven to be powerful for video recognition tasks. Existing approaches primarily rely on parameter-efficient fine-tuning of image-text pre-trained models, yet they often suffer from limited interpretability and poor generalization due to inadequate temporal modeling. To address these, we propose a simple yet effective video-to-text discretization framework. Our method repurposes the frozen text encoder to construct a visual codebook from video class labels due to the many-to-one contrastive alignment between visual and textual embeddings in multimodal pretraining. This codebook effectively transforms temporal visual data into textual tokens via feature lookups and offers interpretable video representations through explicit video modeling. Then, to enhance robustness against irrelevant or noisy frames, we introduce a confidence-aware fusion module that dynamically weights keyframes by assessing their semantic relevance via the codebook. Furthermore, our method incorporates learnable text prompts to conduct adaptive codebook updates. Extensive experiments on HMDB-51, UCF-101, SSv2, and Kinetics-400 have validated the superiority of our approach, achieving more competitive improvements over state-of-the-art methods. The code will be publicly available at https://github.com/isxinxin/VTD-CLIP.
Authors: Zhichao Sun, Huazhang Hu, Yidong Ma, Gang Liu, Nemo Chen, Xu Tang, Yao Hu, Yongchao Xu
Abstract: With the exponential growth of data, traditional object detection methods are increasingly struggling to handle vast vocabulary object detection tasks effectively. We analyze two key limitations of classification-based detectors: positive gradient dilution, where rare positive categories receive insufficient learning signals, and hard negative gradient dilution, where discriminative gradients are overwhelmed by numerous easy negatives. To address these challenges, we propose CQ-DINO, a category query-based object detection framework that reformulates classification as a contrastive task between object queries and learnable category queries. Our method introduces image-guided query selection, which reduces the negative space by adaptively retrieving top-K relevant categories per image via cross-attention, thereby rebalancing gradient distributions and facilitating implicit hard example mining. Furthermore, CQ-DINO flexibly integrates explicit hierarchical category relationships in structured datasets (e.g., V3Det) or learns implicit category correlations via self-attention in generic datasets (e.g., COCO). Experiments demonstrate that CQ-DINO achieves superior performance on the challenging V3Det benchmark (surpassing previous methods by 2.1% AP) while maintaining competitiveness in COCO. Our work provides a scalable solution for real-world detection systems requiring wide category coverage. The dataset and code will be publicly at https://github.com/RedAIGC/CQ-DINO.
Authors: Jinho Jeong, Sangmin Han, Jinwoo Kim, Seon Joo Kim
Abstract: In this paper, we propose LSRNA, a novel framework for higher-resolution (exceeding 1K) image generation using diffusion models by leveraging super-resolution directly in the latent space. Existing diffusion models struggle with scaling beyond their training resolutions, often leading to structural distortions or content repetition. Reference-based methods address the issues by upsampling a low-resolution reference to guide higher-resolution generation. However, they face significant challenges: upsampling in latent space often causes manifold deviation, which degrades output quality. On the other hand, upsampling in RGB space tends to produce overly smoothed outputs. To overcome these limitations, LSRNA combines Latent space Super-Resolution (LSR) for manifold alignment and Region-wise Noise Addition (RNA) to enhance high-frequency details. Our extensive experiments demonstrate that integrating LSRNA outperforms state-of-the-art reference-based methods across various resolutions and metrics, while showing the critical role of latent space upsampling in preserving detail and sharpness. The code is available at https://github.com/3587jjh/LSRNA.
Authors: Luchao Wang, Qian Ren, Kaimin Liao, Hua Wang, Zhi Chen, Yaohua Tang
Abstract: Recent years have witnessed remarkable success of 3D Gaussian Splatting (3DGS) in novel view synthesis, surpassing prior differentiable rendering methods in both quality and efficiency. However, its training process suffers from coupled opacity-color optimization that frequently converges to local minima, producing floater artifacts that degrade visual fidelity. We present StableGS, a framework that eliminates floaters through cross-view depth consistency constraints while introducing a dual-opacity GS model to decouple geometry and material properties of translucent objects. To further enhance reconstruction quality in weakly-textured regions, we integrate DUSt3R depth estimation, significantly improving geometric stability. Our method fundamentally addresses 3DGS training instabilities, outperforming existing state-of-the-art methods across open-source datasets.
Authors: Wei Deng, Mengshi Qi, Huadong Ma
Abstract: Large Vision-Language Models (VLMs), such as GPT-4, have achieved remarkable success across various fields. However, there are few studies on 3D indoor scene generation with VLMs. This paper considers this task as a planning problem subject to spatial and layout common sense constraints. To solve the problem with a VLM, we propose a new global-local tree search algorithm. Globally, the method places each object sequentially and explores multiple placements during each placement process, where the problem space is represented as a tree. To reduce the depth of the tree, we decompose the scene structure hierarchically, i.e. room level, region level, floor object level, and supported object level. The algorithm independently generates the floor objects in different regions and supported objects placed on different floor objects. Locally, we also decompose the sub-task, the placement of each object, into multiple steps. The algorithm searches the tree of problem space. To leverage the VLM model to produce positions of objects, we discretize the top-down view space as a dense grid and fill each cell with diverse emojis to make to cells distinct. We prompt the VLM with the emoji grid and the VLM produces a reasonable location for the object by describing the position with the name of emojis. The quantitative and qualitative experimental results illustrate our approach generates more plausible 3D scenes than state-of-the-art approaches. Our source code is available at https://github.com/dw-dengwei/TreeSearchGen .
Authors: Xiaoyu Zhang, Weihong Pan, Chong Bao, Xiyu Zhang, Xiaojun Xiang, Hanqing Jiang, Hujun Bao
Abstract: Humans perceive and comprehend their surroundings through information spanning multiple frequencies. In immersive scenes, people naturally scan their environment to grasp its overall structure while examining fine details of objects that capture their attention. However, current NeRF frameworks primarily focus on modeling either high-frequency local views or the broad structure of scenes with low-frequency information, which is limited to balancing both. We introduce FA-NeRF, a novel frequency-aware framework for view synthesis that simultaneously captures the overall scene structure and high-definition details within a single NeRF model. To achieve this, we propose a 3D frequency quantification method that analyzes the scene's frequency distribution, enabling frequency-aware rendering. Our framework incorporates a frequency grid for fast convergence and querying, a frequency-aware feature re-weighting strategy to balance features across different frequency contents. Extensive experiments show that our method significantly outperforms existing approaches in modeling entire scenes while preserving fine details. Project page: https://coscatter.github.io/LookCloser/
Authors: Soulaimene Turki, Daniel Panangian, Houda Chaabouni-Chouayakh, Ksenia Bittner
Abstract: Three-dimensional urban reconstruction of buildings from single-view images has attracted significant attention over the past two decades. However, recent methods primarily focus on rooftops from aerial images, often overlooking essential geometrical details. Additionally, there is a notable lack of datasets containing complete 3D point clouds for entire buildings, along with challenges in obtaining reliable camera pose information for aerial images. This paper addresses these challenges by presenting a novel methodology, AIM2PC , which utilizes our generated dataset that includes complete 3D point clouds and determined camera poses. Our approach takes features from a single aerial image as input and concatenates them with essential additional conditions, such as binary masks and Sobel edge maps, to enable more edge-aware reconstruction. By incorporating a point cloud diffusion model based on Centered denoising Diffusion Probabilistic Models (CDPM), we project these concatenated features onto the partially denoised point cloud using our camera poses at each diffusion step. The proposed method is able to reconstruct the complete 3D building point cloud, including wall information and demonstrates superior performance compared to existing baseline techniques. To allow further comparisons with our methodology the dataset has been made available at https://github.com/Soulaimene/AIM2PCDataset
Authors: Takashi Isobe, He Cui, Dong Zhou, Mengmeng Ge, Dong Li, Emad Barsoum
Abstract: Text-to-Video (T2V) generation has attracted significant attention for its ability to synthesize realistic videos from textual descriptions. However, existing models struggle to balance computational efficiency and high visual quality, particularly on resource-limited devices, e.g.,iGPUs and mobile phones. Most prior work prioritizes visual fidelity while overlooking the need for smaller, more efficient models suitable for real-world deployment. To address this challenge, we propose a lightweight T2V framework, termed Hummingbird, which prunes existing models and enhances visual quality through visual feedback learning. Our approach reduces the size of the U-Net from 1.4 billion to 0.7 billion parameters, significantly improving efficiency while preserving high-quality video generation. Additionally, we introduce a novel data processing pipeline that leverages Large Language Models (LLMs) and Video Quality Assessment (VQA) models to enhance the quality of both text prompts and video data. To support user-driven training and style customization, we publicly release the full training code, including data processing and model training. Extensive experiments show that our method achieves a 31X speedup compared to state-of-the-art models such as VideoCrafter2, while also attaining the highest overall score on VBench. Moreover, our method supports the generation of videos with up to 26 frames, addressing the limitations of existing U-Net-based methods in long video generation. Notably, the entire training process requires only four GPUs, yet delivers performance competitive with existing leading methods. Hummingbird presents a practical and efficient solution for T2V generation, combining high performance, scalability, and flexibility for real-world applications.
Authors: Juncen Guo, Xiaoguang Zhu, Lianlong Sun, Liangyu Teng, Di Li, Yang Liu, Liang Song
Abstract: Class-incremental Learning (CIL) enables models to continuously learn new class knowledge while memorizing previous classes, facilitating their adaptation and evolution in dynamic environments. Traditional CIL methods are mainly based on visual features, which limits their ability to handle complex scenarios. In contrast, Vision-Language Models (VLMs) show promising potential to promote CIL by integrating pretrained knowledge with textual features. However, previous methods make it difficult to overcome catastrophic forgetting while preserving the generalization capabilities of VLMs. To tackle these challenges, we propose Feature Calibration enhanced Parameter Synthesis (FCPS) in this paper. Specifically, our FCPS employs a specific parameter adjustment mechanism to iteratively refine the proportion of original visual features participating in the final class determination, ensuring the model's foundational generalization capabilities. Meanwhile, parameter integration across different tasks achieves a balance between learning new class knowledge and retaining old knowledge. Experimental results on popular benchmarks (e.g., CIFAR100 and ImageNet100) validate the superiority of the proposed method.
Authors: Taeyeop Lee, Bowen Wen, Minjun Kang, Gyuree Kang, In So Kweon, Kuk-Jin Yoon
Abstract: We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, Toyota-Light, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d
Authors: Linwei Chen, Lin Gu, Liang Li, Chenggang Yan, Ying Fu
Abstract: While Dynamic Convolution (DY-Conv) has shown promising performance by enabling adaptive weight selection through multiple parallel weights combined with an attention mechanism, the frequency response of these weights tends to exhibit high similarity, resulting in high parameter costs but limited adaptability. In this work, we introduce Frequency Dynamic Convolution (FDConv), a novel approach that mitigates these limitations by learning a fixed parameter budget in the Fourier domain. FDConv divides this budget into frequency-based groups with disjoint Fourier indices, enabling the construction of frequency-diverse weights without increasing the parameter cost. To further enhance adaptability, we propose Kernel Spatial Modulation (KSM) and Frequency Band Modulation (FBM). KSM dynamically adjusts the frequency response of each filter at the spatial level, while FBM decomposes weights into distinct frequency bands in the frequency domain and modulates them dynamically based on local content. Extensive experiments on object detection, segmentation, and classification validate the effectiveness of FDConv. We demonstrate that when applied to ResNet-50, FDConv achieves superior performance with a modest increase of +3.6M parameters, outperforming previous methods that require substantial increases in parameter budgets (e.g., CondConv +90M, KW +76.5M). Moreover, FDConv seamlessly integrates into a variety of architectures, including ConvNeXt, Swin-Transformer, offering a flexible and efficient solution for modern vision tasks. The code is made publicly available at https://github.com/Linwei-Chen/FDConv.
Authors: Ruichuan An, Sihan Yang, Ming Lu, Renrui Zhang, Kai Zeng, Yulin Luo, Jiajun Cao, Hao Liang, Ying Chen, Qi She, Shanghang Zhang, Wentao Zhang
Abstract: Current vision-language models (VLMs) show exceptional abilities across diverse tasks, such as visual question answering. To enhance user experience, recent studies investigate VLM personalization to understand user-provided concepts. However, they mainly focus on single-concept personalization, neglecting the existence and interplay of multiple concepts, which limits real-world applicability. This paper proposes the first multi-concept personalization paradigm, MC-LLaVA. Specifically, MC-LLaVA employs a multi-concept instruction tuning strategy, effectively integrating multiple concepts in a single training step. To reduce the costs related to joint training, we propose a personalized textual prompt that uses visual token information to initialize concept tokens. Additionally, we introduce a personalized visual prompt during inference, aggregating location confidence maps for enhanced recognition and grounding capabilities. To advance multi-concept personalization research, we further contribute a high-quality instruction tuning dataset. We carefully collect images with multiple characters and objects from movies and manually generate question-answer samples for multi-concept scenarios, featuring superior diversity. Comprehensive qualitative and quantitative experiments demonstrate that MC-LLaVA can achieve impressive multi-concept personalized responses, paving the way for VLMs to become better user-specific assistants. The code and dataset will be publicly available at https://github.com/arctanxarc/MC-LLaVA}.
Authors: Zunnan Xu, Zhentao Yu, Zixiang Zhou, Jun Zhou, Xiaoyu Jin, Fa-Ting Hong, Xiaozhong Ji, Junwei Zhu, Chengfei Cai, Shiyu Tang, Qin Lin, Xiu Li, Qinglin Lu
Abstract: We introduce HunyuanPortrait, a diffusion-based condition control method that employs implicit representations for highly controllable and lifelike portrait animation. Given a single portrait image as an appearance reference and video clips as driving templates, HunyuanPortrait can animate the character in the reference image by the facial expression and head pose of the driving videos. In our framework, we utilize pre-trained encoders to achieve the decoupling of portrait motion information and identity in videos. To do so, implicit representation is adopted to encode motion information and is employed as control signals in the animation phase. By leveraging the power of stable video diffusion as the main building block, we carefully design adapter layers to inject control signals into the denoising unet through attention mechanisms. These bring spatial richness of details and temporal consistency. HunyuanPortrait also exhibits strong generalization performance, which can effectively disentangle appearance and motion under different image styles. Our framework outperforms existing methods, demonstrating superior temporal consistency and controllability. Our project is available at https://kkakkkka.github.io/HunyuanPortrait.
Authors: Aether Team, Haoyi Zhu, Yifan Wang, Jianjun Zhou, Wenzheng Chang, Yang Zhou, Zizun Li, Junyi Chen, Chunhua Shen, Jiangmiao Pang, Tong He
Abstract: The integration of geometric reconstruction and generative modeling remains a critical challenge in developing AI systems capable of human-like spatial reasoning. This paper proposes Aether, a unified framework that enables geometry-aware reasoning in world models by jointly optimizing three core capabilities: (1) 4D dynamic reconstruction, (2) action-conditioned video prediction, and (3) goal-conditioned visual planning. Through task-interleaved feature learning, Aether achieves synergistic knowledge sharing across reconstruction, prediction, and planning objectives. Building upon video generation models, our framework demonstrates unprecedented synthetic-to-real generalization despite never observing real-world data during training. Furthermore, our approach achieves zero-shot generalization in both action following and reconstruction tasks, thanks to its intrinsic geometric modeling. Remarkably, even without real-world data, its reconstruction performance is comparable with or even better than that of domain-specific models. Additionally, Aether employs camera trajectories as geometry-informed action spaces, enabling effective action-conditioned prediction and visual planning. We hope our work inspires the community to explore new frontiers in physically-reasonable world modeling and its applications.
Authors: Bowen Hu, Weiheng Yao, Sibo Qiao, Hieu Pham, Shuqiang Wang, Michael Kwok-Po Ng
Abstract: In minimally-invasive brain surgeries with indirect and narrow operating environments, 3D brain reconstruction is crucial. However, as requirements of accuracy for some new minimally-invasive surgeries (such as brain-computer interface surgery) are higher and higher, the outputs of conventional 3D reconstruction, such as point cloud (PC), are facing the challenges that sample points are too sparse and the precision is insufficient. On the other hand, there is a scarcity of high-density point cloud datasets, which makes it challenging to train models for direct reconstruction of high-density brain point clouds. In this work, a novel model named stereoscopic-aware graph generative adversarial network (SG-GAN) with two stages is proposed to generate fine high-density PC conditioned on a single image. The Stage-I GAN sketches the primitive shape and basic structure of the organ based on the given image, yielding Stage-I point clouds. The Stage-II GAN takes the results from Stage-I and generates high-density point clouds with detailed features. The Stage-II GAN is capable of correcting defects and restoring the detailed features of the region of interest (ROI) through the up-sampling process. Furthermore, a parameter-free-attention-based free-transforming module is developed to learn the efficient features of input, while upholding a promising performance. Comparing with the existing methods, the SG-GAN model shows superior performance in terms of visual quality, objective measurements, and performance in classification, as demonstrated by comprehensive results measured by several evaluation metrics including PC-to-PC error and Chamfer distance.
Authors: Jacob Yeung, Andrew F. Luo, Gabriel Sarch, Margaret M. Henderson, Deva Ramanan, Michael J. Tarr
Abstract: While computer vision models have made incredible strides in static image recognition, they still do not match human performance in tasks that require the understanding of complex, dynamic motion. This is notably true for real-world scenarios where embodied agents face complex and motion-rich environments. Our approach, BrainNRDS (Brain-Neural Representations of Dynamic Stimuli), leverages state-of-the-art video diffusion models to decouple static image representation from motion generation, enabling us to utilize fMRI brain activity for a deeper understanding of human responses to dynamic visual stimuli. Conversely, we also demonstrate that information about the brain's representation of motion can enhance the prediction of optical flow in artificial systems. Our novel approach leads to four main findings: (1) Visual motion, represented as fine-grained, object-level resolution optical flow, can be decoded from brain activity generated by participants viewing video stimuli; (2) Video encoders outperform image-based models in predicting video-driven brain activity; (3) Brain-decoded motion signals enable realistic video reanimation based only on the initial frame of the video; and (4) We extend prior work to achieve full video decoding from video-driven brain activity. BrainNRDS advances our understanding of how the brain represents spatial and temporal information in dynamic visual scenes. Our findings demonstrate the potential of combining brain imaging with video diffusion models for developing more robust and biologically-inspired computer vision systems. We show additional decoding and encoding examples on this site: https://brain-nrds.github.io/.
Authors: Yunpeng Jiang, Paul Weng, Yutong Ban
Abstract: Data augmentation is widely applied and has shown its benefits in different machine learning tasks. However, as recently observed, it may have an unfair effect in multi-class classification. While data augmentation generally improves the overall performance (and therefore is beneficial for many classes), it can actually be detrimental for other classes, which can be problematic in some application domains. In this paper, to counteract this phenomenon, we propose CLAM, a CLAss-dependent Multiplicative-weights method. To derive it, we first formulate the training of a classifier as a non-linear optimization problem that aims at simultaneously maximizing the individual class performances and balancing them. By rewriting this optimization problem as an adversarial two-player game, we propose a novel multiplicative weight algorithm, for which we prove the convergence. Interestingly, our formulation also reveals that the class-dependent effects of data augmentation is not due to data augmentation only, but is in fact a general phenomenon. Our empirical results over five datasets demonstrate that the performance of learned classifiers is indeed more fairly distributed over classes, with only limited impact on the average accuracy.
Authors: Yan Song Hu, Nicolas Abboud, Muhammad Qasim Ali, Adam Srebrnjak Yang, Imad Elhajj, Daniel Asmar, Yuhao Chen, John S. Zelek
Abstract: Real-time SLAM with dense 3D mapping is computationally challenging, especially on resource-limited devices. The recent development of 3D Gaussian Splatting (3DGS) offers a promising approach for real-time dense 3D reconstruction. However, existing 3DGS-based SLAM systems struggle to balance hardware simplicity, speed, and map quality. Most systems excel in one or two of the aforementioned aspects but rarely achieve all. A key issue is the difficulty of initializing 3D Gaussians while concurrently conducting SLAM. To address these challenges, we present Monocular GSO (MGSO), a novel real-time SLAM system that integrates photometric SLAM with 3DGS. Photometric SLAM provides dense structured point clouds for 3DGS initialization, accelerating optimization and producing more efficient maps with fewer Gaussians. As a result, experiments show that our system generates reconstructions with a balance of quality, memory efficiency, and speed that outperforms the state-of-the-art. Furthermore, our system achieves all results using RGB inputs. We evaluate the Replica, TUM-RGBD, and EuRoC datasets against current live dense reconstruction systems. Not only do we surpass contemporary systems, but experiments also show that we maintain our performance on laptop hardware, making it a practical solution for robotics, A/R, and other real-time applications.
Authors: Zheda Mai, Ping Zhang, Cheng-Hao Tu, Hong-You Chen, Li Zhang, Wei-Lun Chao
Abstract: Parameter-efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyperparameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT.
Authors: Yang-Che Sun, Cheng Yu Yeo, Ernie Chu, Jun-Cheng Chen, Yu-Lun Liu
Abstract: In this work, we propose using a unified representation, termed Factorized Features, for low-level vision tasks, where we test on Single Image Super-Resolution (SISR) and Image Compression. Motivated by the shared principles between these tasks, they require recovering and preserving fine image details, whether by enhancing resolution for SISR or reconstructing compressed data for Image Compression. Unlike previous methods that mainly focus on network architecture, our proposed approach utilizes a basis-coefficient decomposition as well as an explicit formulation of frequencies to capture structural components and multi-scale visual features in images, which addresses the core challenges of both tasks. We replace the representation of prior models from simple feature maps with Factorized Features to validate the potential for broad generalizability. In addition, we further optimize the pipelines by leveraging the mergeable-basis property of our Factorized Features, which consolidates shared structures on multi-frame compression and super-resolution. Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4% in PSNR over the baseline in Super-Resolution (SR) and 9.35% BD-rate reduction in Image Compression compared to the previous SOTA.
Authors: Weixin Mao, Weiheng Zhong, Zhou Jiang, Dong Fang, Zhongyue Zhang, Zihan Lan, Haosheng Li, Fan Jia, Tiancai Wang, Haoqiang Fan, Osamu Yoshie
Abstract: Existing robot policies predominantly adopt the task-centric approach, requiring end-to-end task data collection. This results in limited generalization to new tasks and difficulties in pinpointing errors within long-horizon, multi-stage tasks. To address this, we propose RoboMatrix, a skill-centric hierarchical framework designed for scalable robot task planning and execution in open-world environments. RoboMatrix extracts general meta-skills from diverse complex tasks, enabling the completion of unseen tasks through skill composition. Its architecture consists of a high-level scheduling layer that utilizes large language models (LLMs) for task decomposition, an intermediate skill layer housing meta-skill models, and a low-level hardware layer for robot control. A key innovation of our work is the introduction of the first unified vision-language-action (VLA) model capable of seamlessly integrating both movement and manipulation within one model. This is achieved by combining vision and language prompts to generate discrete actions. Experimental results demonstrate that RoboMatrix achieves a 50% higher success rate than task-centric baselines when applied to unseen objects, scenes, and tasks. To advance open-world robotics research, we will open-source code, hardware designs, model weights, and datasets at https://github.com/WayneMao/RoboMatrix.
Authors: S\'ebastien Pi\'erard, Ana\"is Halin, Anthony Cioppa, Adrien Deli\`ege, Marc Van Droogenbroeck
Abstract: Ranking entities such as algorithms, devices, methods, or models based on their performances, while accounting for application-specific preferences, is a challenge. To address this challenge, we establish the foundations of a universal theory for performance-based ranking. First, we introduce a rigorous framework built on top of both the probability and order theories. Our new framework encompasses the elements necessary to (1) manipulate performances as mathematical objects, (2) express which performances are worse than or equivalent to others, (3) model tasks through a variable called satisfaction, (4) consider properties of the evaluation, (5) define scores, and (6) specify application-specific preferences through a variable called importance. On top of this framework, we propose the first axiomatic definition of performance orderings and performance-based rankings. Then, we introduce a universal parametric family of scores, called ranking scores, that can be used to establish rankings satisfying our axioms, while considering application-specific preferences. Finally, we show, in the case of two-class classification, that the family of ranking scores encompasses well-known performance scores, including the accuracy, the true positive rate (recall, sensitivity), the true negative rate (specificity), the positive predictive value (precision), and F1. However, we also show that some other scores commonly used to compare classifiers are unsuitable to derive performance orderings satisfying the axioms.
Authors: Byeonghyeon Lee, Youbin Kim, Yongjae Jo, Hyunsu Kim, Hyemi Park, Yangkyu Kim, Debabrata Mandal, Praneeth Chakravarthula, Inki Kim, Eunbyung Park
Abstract: Metalens is an emerging optical system with an irreplaceable merit in that it can be manufactured in ultra-thin and compact sizes, which shows great promise in various applications. Despite its advantage in miniaturization, its practicality is constrained by spatially varying aberrations and distortions, which significantly degrade the image quality. Several previous arts have attempted to address different types of aberrations, yet most of them are mainly designed for the traditional bulky lens and ineffective to remedy harsh aberrations of the metalens. While there have existed aberration correction methods specifically for metalens, they still fall short of restoration quality. In this work, we propose a novel aberration correction framework for metalens-captured images, harnessing Vision Transformers (ViT) that have the potential to restore metalens images with non-uniform aberrations. Specifically, we devise a Multiple Adaptive Filters Guidance (MAFG), where multiple Wiener filters enrich the degraded input images with various noise-detail balances and a cross-attention module reweights the features considering the different degrees of aberrations. In addition, we introduce a Spatial and Transposed self-Attention Fusion (STAF) module, which aggregates features from spatial self-attention and transposed self-attention modules to further ameliorate aberration correction. We conduct extensive experiments, including correcting aberrated images and videos, and clean 3D reconstruction. The proposed method outperforms the previous arts by a significant margin. We further fabricate a metalens and verify the practicality of our method by restoring the images captured with the manufactured metalens. Code and pre-trained models are available at https://benhenryl.github.io/Metalens-Transformer.
Authors: Jiong Lin, Lechen Zhang, Kwansoo Lee, Jialong Ning, Judah Goldfeder, Hod Lipson
Abstract: Robot description models are essential for simulation and control, yet their creation often requires significant manual effort. To streamline this modeling process, we introduce AutoURDF, an unsupervised approach for constructing description files for unseen robots from point cloud frames. Our method leverages a cluster-based point cloud registration model that tracks the 6-DoF transformations of point clusters. Through analyzing cluster movements, we hierarchically address the following challenges: (1) moving part segmentation, (2) body topology inference, and (3) joint parameter estimation. The complete pipeline produces robot description files that are fully compatible with existing simulators. We validate our method across a variety of robots, using both synthetic and real-world scan data. Results indicate that our approach outperforms previous methods in registration and body topology estimation accuracy, offering a scalable solution for automated robot modeling.
Authors: Meilong Xu, Saumya Gupta, Xiaoling Hu, Chen Li, Shahira Abousamra, Dimitris Samaras, Prateek Prasanna, Chao Chen
Abstract: Accurately modeling multi-class cell topology is crucial in digital pathology, as it provides critical insights into tissue structure and pathology. The synthetic generation of cell topology enables realistic simulations of complex tissue environments, enhances downstream tasks by augmenting training data, aligns more closely with pathologists' domain knowledge, and offers new opportunities for controlling and generalizing the tumor microenvironment. In this paper, we propose a novel approach that integrates topological constraints into a diffusion model to improve the generation of realistic, contextually accurate cell topologies. Our method refines the simulation of cell distributions and interactions, increasing the precision and interpretability of results in downstream tasks such as cell detection and classification. To assess the topological fidelity of generated layouts, we introduce a new metric, Topological Frechet Distance (TopoFD), which overcomes the limitations of traditional metrics like FID in evaluating topological structure. Experimental results demonstrate the effectiveness of our approach in generating multi-class cell layouts that capture intricate topological relationships. Code is available at https://github.com/Melon-Xu/TopoCellGen.
Authors: Jesse Hagenaars, Yilun Wu, Federico Paredes-Vall\'es, Stein Stroobants, Guido de Croon
Abstract: Event cameras provide low-latency perception for only milliwatts of power. This makes them highly suitable for resource-restricted, agile robots such as small flying drones. Self-supervised learning based on contrast maximization holds great potential for event-based robot vision, as it foregoes the need for high-frequency ground truth and allows for online learning in the robot's operational environment. However, online, on-board learning raises the major challenge of achieving sufficient computational efficiency for real-time learning, while maintaining competitive visual perception performance. In this work, we improve the time and memory efficiency of the contrast maximization pipeline, making on-device learning of low-latency monocular depth possible. We demonstrate that online learning on board a small drone yields more accurate depth estimates and more successful obstacle avoidance behavior compared to only pre-training. Benchmarking experiments show that the proposed pipeline is not only efficient, but also achieves state-of-the-art depth estimation performance among self-supervised approaches. Our work taps into the unused potential of online, on-device robot learning, promising smaller reality gaps and better performance.
Authors: Qi Wu, Janick Martinez Esturo, Ashkan Mirzaei, Nicolas Moenne-Loccoz, Zan Gojcic
Abstract: 3D Gaussian Splatting (3DGS) enables efficient reconstruction and high-fidelity real-time rendering of complex scenes on consumer hardware. However, due to its rasterization-based formulation, 3DGS is constrained to ideal pinhole cameras and lacks support for secondary lighting effects. Recent methods address these limitations by tracing the particles instead, but, this comes at the cost of significantly slower rendering. In this work, we propose 3D Gaussian Unscented Transform (3DGUT), replacing the EWA splatting formulation with the Unscented Transform that approximates the particles through sigma points, which can be projected exactly under any nonlinear projection function. This modification enables trivial support of distorted cameras with time dependent effects such as rolling shutter, while retaining the efficiency of rasterization. Additionally, we align our rendering formulation with that of tracing-based methods, enabling secondary ray tracing required to represent phenomena such as reflections and refraction within the same 3D representation. The source code is available at: https://github.com/nv-tlabs/3dgrut.
Authors: Uri Gadot, Assaf Shocher, Shie Mannor, Gal Chechik, Assaf Hallak
Abstract: Video encoders optimize compression for human perception by minimizing reconstruction error under bit-rate constraints. In many modern applications such as autonomous driving, an overwhelming majority of videos serve as input for AI systems performing tasks like object recognition or segmentation, rather than being watched by humans. It is therefore useful to optimize the encoder for a downstream task instead of for perceptual image quality. However, a major challenge is how to combine such downstream optimization with existing standard video encoders, which are highly efficient and popular. Here, we address this challenge by controlling the Quantization Parameters (QPs) at the macro-block level to optimize the downstream task. This granular control allows us to prioritize encoding for task-relevant regions within each frame. We formulate this optimization problem as a Reinforcement Learning (RL) task, where the agent learns to balance long-term implications of choosing QPs on both task performance and bit-rate constraints. Notably, our policy does not require the downstream task as an input during inference, making it suitable for streaming applications and edge devices such as vehicles. We demonstrate significant improvements in two tasks, car detection, and ROI (saliency) encoding. Our approach improves task performance for a given bit rate compared to traditional task agnostic encoding methods, paving the way for more efficient task-aware video compression.
Authors: Christian Tinauer, Maximilian Sackl, Rudolf Stollberger, Stefan Ropele, Christian Langkammer
Abstract: Backgrounds. Deep neural networks have demonstrated high accuracy in classifying Alzheimer's disease (AD). This study aims to enlighten the underlying black-box nature and reveal individual contributions of T1-weighted (T1w) gray-white matter texture, volumetric information and preprocessing on classification performance. Methods. We utilized T1w MRI data from the Alzheimer's Disease Neuroimaging Initiative to distinguish matched AD patients (990 MRIs) from healthy controls (990 MRIs). Preprocessing included skull stripping and binarization at varying thresholds to systematically eliminate texture information. A deep neural network was trained on these configurations, and the model performance was compared using McNemar tests with discrete Bonferroni-Holm correction. Layer-wise Relevance Propagation (LRP) and structural similarity metrics between heatmaps were applied to analyze learned features. Results. Classification performance metrics (accuracy, sensitivity, and specificity) were comparable across all configurations, indicating a negligible influence of T1w gray- and white signal texture. Models trained on binarized images demonstrated similar feature performance and relevance distributions, with volumetric features such as atrophy and skull-stripping features emerging as primary contributors. Conclusions. We revealed a previously undiscovered Clever Hans effect in a widely used AD MRI dataset. Deep neural networks classification predominantly rely on volumetric features, while eliminating gray-white matter T1w texture did not decrease the performance. This study clearly demonstrates an overestimation of the importance of gray-white matter contrasts, at least for widely used structural T1w images, and highlights potential misinterpretation of performance metrics.
Authors: Yuheng Ji, Huajie Tan, Jiayu Shi, Xiaoshuai Hao, Yuan Zhang, Hengyuan Zhang, Pengwei Wang, Mengdi Zhao, Yao Mu, Pengju An, Xinda Xue, Qinghang Su, Huaihai Lyu, Xiaolong Zheng, Jiaming Liu, Zhongyuan Wang, Shanghang Zhang
Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have shown remarkable capabilities across various multimodal contexts. However, their application in robotic scenarios, particularly for long-horizon manipulation tasks, reveals significant limitations. These limitations arise from the current MLLMs lacking three essential robotic brain capabilities: Planning Capability, which involves decomposing complex manipulation instructions into manageable sub-tasks; Affordance Perception, the ability to recognize and interpret the affordances of interactive objects; and Trajectory Prediction, the foresight to anticipate the complete manipulation trajectory necessary for successful execution. To enhance the robotic brain's core capabilities from abstract to concrete, we introduce ShareRobot, a high-quality heterogeneous dataset that labels multi-dimensional information such as task planning, object affordance, and end-effector trajectory. ShareRobot's diversity and accuracy have been meticulously refined by three human annotators. Building on this dataset, we developed RoboBrain, an MLLM-based model that combines robotic and general multi-modal data, utilizes a multi-stage training strategy, and incorporates long videos and high-resolution images to improve its robotic manipulation capabilities. Extensive experiments demonstrate that RoboBrain achieves state-of-the-art performance across various robotic tasks, highlighting its potential to advance robotic brain capabilities.
Authors: Haoqiang Kang, Enna Sachdeva, Piyush Gupta, Sangjae Bae, Kwonjoon Lee
Abstract: Vision-Language Models (VLMs) have recently shown promising advancements in sequential decision-making tasks through task-specific fine-tuning. However, common fine-tuning methods, such as Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) techniques like Proximal Policy Optimization (PPO), present notable limitations: SFT assumes Independent and Identically Distributed (IID) data, while PPO focuses on maximizing cumulative rewards. These limitations often restrict solution diversity and hinder generalization in multi-step reasoning tasks. To address these challenges, we introduce a novel framework, GFlowVLM, a framework that fine-tune VLMs using Generative Flow Networks (GFlowNets) to promote generation of diverse solutions for complex reasoning tasks. GFlowVLM models the environment as a non-Markovian decision process, allowing it to capture long-term dependencies essential for real-world applications. It takes observations and task descriptions as inputs to prompt chain-of-thought (CoT) reasoning which subsequently guides action selection. We use task based rewards to fine-tune VLM with GFlowNets. This approach enables VLMs to outperform prior fine-tuning methods, including SFT and RL. Empirical results demonstrate the effectiveness of GFlowVLM on complex tasks such as card games (NumberLine, BlackJack) and embodied planning tasks (ALFWorld), showing enhanced training efficiency, solution diversity, and stronger generalization capabilities across both in-distribution and out-of-distribution scenarios.
Authors: Huiyang Shao, Xin Xia, Yuhong Yang, Yuxi Ren, Xing Wang, Xuefeng Xiao
Abstract: Diffusion models have achieved remarkable success across various domains. However, their slow generation speed remains a critical challenge. Existing acceleration methods, while aiming to reduce steps, often compromise sample quality, controllability, or introduce training complexities. Therefore, we propose RayFlow, a novel diffusion framework that addresses these limitations. Unlike previous methods, RayFlow guides each sample along a unique path towards an instance-specific target distribution. This method minimizes sampling steps while preserving generation diversity and stability. Furthermore, we introduce Time Sampler, an importance sampling technique to enhance training efficiency by focusing on crucial timesteps. Extensive experiments demonstrate RayFlow's superiority in generating high-quality images with improved speed, control, and training efficiency compared to existing acceleration techniques.
Authors: Hong Zheng, Yucheng Chen, Nan Mu, Xiaoning Li
Abstract: Accurate segmentation of the ventricles from cardiac magnetic resonance images (CMRIs) is crucial for enhancing the diagnosis and analysis of heart conditions. Deep learning-based segmentation methods have recently garnered significant attention due to their impressive performance. However, these segmentation methods are typically good at partitioning regularly shaped organs, such as the left ventricle (LV) and the myocardium (MYO), whereas they perform poorly on irregularly shaped organs, such as the right ventricle (RV). In this study, we argue that this limitation of segmentation models stems from their insufficient generalization ability to address the distribution shift of segmentation targets across slices, cardiac phases, and disease conditions. To overcome this issue, we present a Multi-Disease-Aware Training Strategy (MTS) and restructure the introduced CMRI datasets into multi-disease datasets. Additionally, we propose a specialized data processing technique for preprocessing input images to support the MTS. To validate the effectiveness of our method, we performed control group experiments and cross-validation tests. The experimental results show that (1) network models trained using our proposed strategy achieved superior segmentation performance, particularly in RV segmentation, and (2) these networks exhibited robust performance even when applied to data from unknown diseases.
Authors: Meva Himmetoglu (for the Alzheimer's Disease Neuroimaging Initiative), Ilja Ciernik (for the Alzheimer's Disease Neuroimaging Initiative), Ender Konukoglu (for the Alzheimer's Disease Neuroimaging Initiative)
Abstract: Segmenting healthy tissue structures alongside lesions in brain Magnetic Resonance Images (MRI) remains a challenge for today's algorithms due to lesion-caused disruption of the anatomy and lack of jointly labeled training datasets, where both healthy tissues and lesions are labeled on the same images. In this paper, we propose a method that is robust to lesion-caused disruptions and can be trained from disparately labeled training sets, i.e., without requiring jointly labeled samples, to automatically segment both. In contrast to prior work, we decouple healthy tissue and lesion segmentation in two paths to leverage multi-sequence acquisitions and merge information with an attention mechanism. During inference, an image-specific adaptation reduces adverse influences of lesion regions on healthy tissue predictions. During training, the adaptation is taken into account through meta-learning and co-training is used to learn from disparately labeled training images. Our model shows an improved performance on several anatomical structures and lesions on a publicly available brain glioblastoma dataset compared to the state-of-the-art segmentation methods.