Authors: Chaoyu Li, Sid Padmanabhuni, Maryam Cheema, Hasti Seifi, Pooyan Fazli
Abstract: Video descriptions are crucial for blind and low vision (BLV) users to access visual content. However, current artificial intelligence models for generating descriptions often fall short due to limitations in the quality of human annotations within training datasets, resulting in descriptions that do not fully meet BLV users' needs. To address this gap, we introduce VideoA11y, an approach that leverages multimodal large language models (MLLMs) and video accessibility guidelines to generate descriptions tailored for BLV individuals. Using this method, we have curated VideoA11y-40K, the largest and most comprehensive dataset of 40,000 videos described for BLV users. Rigorous experiments across 15 video categories, involving 347 sighted participants, 40 BLV participants, and seven professional describers, showed that VideoA11y descriptions outperform novice human annotations and are comparable to trained human annotations in clarity, accuracy, objectivity, descriptiveness, and user satisfaction. We evaluated models on VideoA11y-40K using both standard and custom metrics, demonstrating that MLLMs fine-tuned on this dataset produce high-quality accessible descriptions. Code and dataset are available at https://people-robots.github.io/VideoA11y.
Authors: MohammadHossein Rezaei, Yicheng Fu, Phil Cuvin, Caleb Ziems, Yanzhe Zhang, Hao Zhu, Diyi Yang
Abstract: Human activity is moderated by norms. When performing actions in the real world, humans not only follow norms, but also consider the trade-off between different norms However, machines are often trained without explicit supervision on norm understanding and reasoning, especially when the norms are grounded in a physical and social context. To improve and evaluate the normative reasoning capability of vision-language models (VLMs), we present EgoNormia $\|\epsilon\|$, consisting of 1,853 ego-centric videos of human interactions, each of which has two related questions evaluating both the prediction and justification of normative actions. The normative actions encompass seven categories: safety, privacy, proxemics, politeness, cooperation, coordination/proactivity, and communication/legibility. To compile this dataset at scale, we propose a novel pipeline leveraging video sampling, automatic answer generation, filtering, and human validation. Our work demonstrates that current state-of-the-art vision-language models lack robust norm understanding, scoring a maximum of 45% on EgoNormia (versus a human bench of 92%). Our analysis of performance in each dimension highlights the significant risks of safety, privacy, and the lack of collaboration and communication capability when applied to real-world agents. We additionally show that through a retrieval-based generation method, it is possible to use EgoNomia to enhance normative reasoning in VLMs.
Authors: Yixiong Chen, Shawn Xu, Andrew Sellergren, Yossi Matias, Avinatan Hassidim, Shravya Shetty, Daniel Golden, Alan Yuille, Lin Yang
Abstract: Vision-language models have proven to be of great benefit for medical image analysis since they learn rich semantics from both images and reports. Prior efforts have focused on better alignment of image and text representations to enhance image understanding. However, though explicit reference to a prior image is common in Chest X-Ray (CXR) reports, aligning progression descriptions with the semantics differences in image pairs remains under-explored. In this work, we propose two components to address this issue. (1) A CXR report processing pipeline to extract temporal structure. It processes reports with a large language model (LLM) to separate the description and comparison contexts, and extracts fine-grained annotations from reports. (2) A contrastive captioner model for CXR, namely CoCa-CXR, to learn how to both describe images and their temporal progressions. CoCa-CXR incorporates a novel regional cross-attention module to identify local differences between paired CXR images. Extensive experiments show the superiority of CoCa-CXR on both progression analysis and report generation compared to previous methods. Notably, on MS-CXR-T progression classification, CoCa-CXR obtains 65.0% average testing accuracy on five pulmonary conditions, outperforming the previous state-of-the-art (SOTA) model BioViL-T by 4.8%. It also achieves a RadGraph F1 of 24.2% on MIMIC-CXR, which is comparable to the Med-Gemini foundation model.
Authors: Kyle Fogarty, Jing Yang, Chayan Kumar Patodi, Aadi Bhanti, Steven Chacko, Cengiz Oztireli, Ujwal Bonde
Abstract: Accurate 3D foot reconstruction is crucial for personalized orthotics, digital healthcare, and virtual fittings. However, existing methods struggle with incomplete scans and anatomical variations, particularly in self-scanning scenarios where user mobility is limited, making it difficult to capture areas like the arch and heel. We present a novel end-to-end pipeline that refines Structure-from-Motion (SfM) reconstruction. It first resolves scan alignment ambiguities using SE(3) canonicalization with a viewpoint prediction module, then completes missing geometry through an attention-based network trained on synthetically augmented point clouds. Our approach achieves state-of-the-art performance on reconstruction metrics while preserving clinically validated anatomical fidelity. By combining synthetic training data with learned geometric priors, we enable robust foot reconstruction under real-world capture conditions, unlocking new opportunities for mobile-based 3D scanning in healthcare and retail.
Authors: Hu Wang, Ibrahim Almakky, Congbo Ma, Numan Saeed, Mohammad Yaqub
Abstract: Model merging is an effective strategy to merge multiple models for enhancing model performances, and more efficient than ensemble learning as it will not introduce extra computation into inference. However, limited research explores if the merging process can occur within one model and enhance the model's robustness, which is particularly critical in the medical image domain. In the paper, we are the first to propose in-model merging (InMerge), a novel approach that enhances the model's robustness by selectively merging similar convolutional kernels in the deep layers of a single convolutional neural network (CNN) during the training process for classification. We also analytically reveal important characteristics that affect how in-model merging should be performed, serving as an insightful reference for the community. We demonstrate the feasibility and effectiveness of this technique for different CNN architectures on 4 prevalent datasets. The proposed InMerge-trained model surpasses the typically-trained model by a substantial margin. The code will be made public.
Authors: Li-Wei Chen, Ombretta Strafforello, Anne-Sofie Maerten, Tinne Tuytelaars, Johan Wagemans
Abstract: Image aesthetic assessment (IAA) evaluates image aesthetics, a task complicated by image diversity and user subjectivity. Current approaches address this in two stages: Generic IAA (GIAA) models estimate mean aesthetic scores, while Personal IAA (PIAA) models adapt GIAA using transfer learning to incorporate user subjectivity. However, a theoretical understanding of transfer learning between GIAA and PIAA, particularly concerning the impact of group composition, group size, aesthetic differences between groups and individuals, and demographic correlations, is lacking. This work establishes a theoretical foundation for IAA, proposing a unified model that encodes individual characteristics in a distributional format for both individual and group assessments. We show that transferring from GIAA to PIAA involves extrapolation, while the reverse involves interpolation, which is generally more effective for machine learning. Experiments with varying group compositions, including sub-sampling by group size and disjoint demographics, reveal significant performance variation even for GIAA, indicating that mean scores do not fully eliminate individual subjectivity. Performance variations and Gini index analysis reveal education as the primary factor influencing aesthetic differences, followed by photography and art experience, with stronger individual subjectivity observed in artworks than in photos. Our model uniquely supports both GIAA and PIAA, enhancing generalization across demographics.
Authors: Ji-Hun Oh, Kianoush Falahkheirkhah, Rohit Bhargava
Abstract: Label-free chemical imaging holds significant promise for improving digital pathology workflows. However, data acquisition speed remains a limiting factor for smooth clinical transition. To address this gap, we propose an adaptive strategy: initially scan the low information (LI) content of the entire tissue quickly, identify regions with high aleatoric uncertainty (AU), and selectively re-image them at better quality to capture higher information (HI) details. The primary challenge lies in distinguishing between high-AU regions that can be mitigated through HI imaging and those that cannot. However, since existing uncertainty frameworks cannot separate such AU subcategories, we propose a fine-grained disentanglement method based on post-hoc latent space analysis to unmix resolvable from irresolvable high-AU regions. We apply our approach to efficiently image infrared spectroscopic data of breast tissues, achieving superior segmentation performance using the acquired HI data compared to a random baseline. This represents the first algorithmic study focused on fine-grained AU disentanglement within dynamic image spaces (LI-to-HI), with novel application to streamline histopathology.
Authors: Joana C. Costa, Tiago Roxo, Hugo Proen\c{c}a, Pedro R. M. In\'acio
Abstract: State-of-the-art defense mechanisms are typically evaluated in the context of white-box attacks, which is not realistic, as it assumes the attacker can access the gradients of the target network. To protect against this scenario, Adversarial Training (AT) and Adversarial Distillation (AD) include adversarial examples during the training phase, and Adversarial Purification uses a generative model to reconstruct all the images given to the classifier. This paper considers an even more realistic evaluation scenario: gray-box attacks, which assume that the attacker knows the architecture and the dataset used to train the target network, but cannot access its gradients. We provide empirical evidence that models are vulnerable to gray-box attacks and propose LISArD, a defense mechanism that does not increase computational and temporal costs but provides robustness against gray-box and white-box attacks without including AT. Our method approximates a cross-correlation matrix, created with the embeddings of perturbed and clean images, to a diagonal matrix while simultaneously conducting classification learning. Our results show that LISArD can effectively protect against gray-box attacks, can be used in multiple architectures, and carries over its resilience to the white-box scenario. Also, state-of-the-art AD models underperform greatly when removing AT and/or moving to gray-box settings, highlighting the lack of robustness from existing approaches to perform in various conditions (aside from white-box settings). All the source code is available at https://github.com/Joana-Cabral/LISArD.
Authors: Mohammad Abu Tami, Mohammed Elhenawy, Huthaifa I. Ashqar
Abstract: Traffic safety remains a vital concern in contemporary urban settings, intensified by the increase of vehicles and the complicated nature of road networks. Traditional safety-critical event detection systems predominantly rely on sensor-based approaches and conventional machine learning algorithms, necessitating extensive data collection and complex training processes to adhere to traffic safety regulations. This paper introduces HazardNet, a small-scale Vision Language Model designed to enhance traffic safety by leveraging the reasoning capabilities of advanced language and vision models. We built HazardNet by fine-tuning the pre-trained Qwen2-VL-2B model, chosen for its superior performance among open-source alternatives and its compact size of two billion parameters. This helps to facilitate deployment on edge devices with efficient inference throughput. In addition, we present HazardQA, a novel Vision Question Answering (VQA) dataset constructed specifically for training HazardNet on real-world scenarios involving safety-critical events. Our experimental results show that the fine-tuned HazardNet outperformed the base model up to an 89% improvement in F1-Score and has comparable results with improvement in some cases reach up to 6% when compared to larger models, such as GPT-4o. These advancements underscore the potential of HazardNet in providing real-time, reliable traffic safety event detection, thereby contributing to reduced accidents and improved traffic management in urban environments. Both HazardNet model and the HazardQA dataset are available at https://huggingface.co/Tami3/HazardNet and https://huggingface.co/datasets/Tami3/HazardQA, respectively.
URLs: https://huggingface.co/Tami3/HazardNet, https://huggingface.co/datasets/Tami3/HazardQA,
Authors: Sari Masri, Huthaifa I. Ashqar, Mohammed Elhenawy
Abstract: Traffic control in unsignalized urban intersections presents significant challenges due to the complexity, frequent conflicts, and blind spots. This study explores the capability of leveraging Multimodal Large Language Models (MLLMs), such as GPT-4o, to provide logical and visual reasoning by directly using birds-eye-view videos of four-legged intersections. In this proposed method, GPT-4o acts as intelligent system to detect conflicts and provide explanations and recommendations for the drivers. The fine-tuned model achieved an accuracy of 77.14%, while the manual evaluation of the true predicted values of the fine-tuned GPT-4o showed significant achievements of 89.9% accuracy for model-generated explanations and 92.3% for the recommended next actions. These results highlight the feasibility of using MLLMs for real-time traffic management using videos as inputs, offering scalable and actionable insights into intersections traffic management and operation. Code used in this study is available at https://github.com/sarimasri3/Traffic-Intersection-Conflict-Detection-using-images.git.
URLs: https://github.com/sarimasri3/Traffic-Intersection-Conflict-Detection-using-images.git.
Authors: MD Wahiduzzaman Khan, Mingshan Jia, Shaolin Zhang, En Yu, Kaska Musial-Gabrys
Abstract: Facial appearance editing is crucial for digital avatars, AR/VR, and personalized content creation, driving realistic user experiences. However, preserving identity with generative models is challenging, especially in scenarios with limited data availability. Traditional methods often require multiple images and still struggle with unnatural face shifts, inconsistent hair alignment, or excessive smoothing effects. To overcome these challenges, we introduce a novel diffusion-based framework, InstaFace, to generate realistic images while preserving identity using only a single image. Central to InstaFace, we introduce an efficient guidance network that harnesses 3D perspectives by integrating multiple 3DMM-based conditionals without introducing additional trainable parameters. Moreover, to ensure maximum identity retention as well as preservation of background, hair, and other contextual features like accessories, we introduce a novel module that utilizes feature embeddings from a facial recognition model and a pre-trained vision-language model. Quantitative evaluations demonstrate that our method outperforms several state-of-the-art approaches in terms of identity preservation, photorealism, and effective control of pose, expression, and lighting.
Authors: Vladimir Zaigrajew, Hubert Baniecki, Przemyslaw Biecek
Abstract: Sparse autoencoders (SAEs) are useful for detecting and steering interpretable features in neural networks, with particular potential for understanding complex multimodal representations. Given their ability to uncover interpretable features, SAEs are particularly valuable for analyzing large-scale vision-language models (e.g., CLIP and SigLIP), which are fundamental building blocks in modern systems yet remain challenging to interpret and control. However, current SAE methods are limited by optimizing both reconstruction quality and sparsity simultaneously, as they rely on either activation suppression or rigid sparsity constraints. To this end, we introduce Matryoshka SAE (MSAE), a new architecture that learns hierarchical representations at multiple granularities simultaneously, enabling a direct optimization of both metrics without compromise. MSAE establishes a new state-of-the-art Pareto frontier between reconstruction quality and sparsity for CLIP, achieving 0.99 cosine similarity and less than 0.1 fraction of variance unexplained while maintaining ~80% sparsity. Finally, we demonstrate the utility of MSAE as a tool for interpreting and controlling CLIP by extracting over 120 semantic concepts from its representation to perform concept-based similarity search and bias analysis in downstream tasks like CelebA.
Authors: Chi Ruan
Abstract: While open-vocabulary object detectors require predefined categories during inference, generative object detectors overcome this limitation by endowing the model with text generation capabilities. However, existing generative object detection methods directly append an autoregressive language model to an object detector to generate texts for each detected object. This straightforward design leads to structural redundancy and increased processing time. In this paper, we propose a Real-Time GENerative Detection Transformer (RTGen), a real-time generative object detector with a succinct encoder-decoder architecture. Specifically, we introduce a novel Region-Language Decoder (RL-Decoder), which innovatively integrates a non-autoregressive language model into the detection decoder, enabling concurrent processing of object and text information. With these efficient designs, RTGen achieves a remarkable inference speed of 60.41 FPS. Moreover, RTGen obtains 18.6 mAP on the LVIS dataset, outperforming the previous SOTA method by 3.5 mAP.
Authors: Yifei Qian, Zhongliang Guo, Bowen Deng, Chun Tong Lei, Shuai Zhao, Chun Pong Lau, Xiaopeng Hong, Michael P. Pound
Abstract: Zero-shot object counting aims to count instances of arbitrary object categories specified by text descriptions. Existing methods typically rely on vision-language models like CLIP, but often exhibit limited sensitivity to text prompts. We present T2ICount, a diffusion-based framework that leverages rich prior knowledge and fine-grained visual understanding from pretrained diffusion models. While one-step denoising ensures efficiency, it leads to weakened text sensitivity. To address this challenge, we propose a Hierarchical Semantic Correction Module that progressively refines text-image feature alignment, and a Representational Regional Coherence Loss that provides reliable supervision signals by leveraging the cross-attention maps extracted from the denosing U-Net. Furthermore, we observe that current benchmarks mainly focus on majority objects in images, potentially masking models' text sensitivity. To address this, we contribute a challenging re-annotated subset of FSC147 for better evaluation of text-guided counting ability. Extensive experiments demonstrate that our method achieves superior performance across different benchmarks. Code is available at https://github.com/cha15yq/T2ICount.
Authors: Yuqian Chen, Leo Zekelman, Yui Lo, Suheyla Cetin-Karayumak, Tengfei Xue, Yogesh Rathi, Nikos Makris, Fan Zhang, Weidong Cai, Lauren J. O'Donnell
Abstract: Tractography parcellation classifies streamlines reconstructed from diffusion MRI into anatomically defined fiber tracts for clinical and research applications. However, clinical scans often have incomplete fields of view (FOV) where brain regions are partially imaged, leading to partial or truncated fiber tracts. To address this challenge, we introduce TractCloud-FOV, a deep learning framework that robustly parcellates tractography under conditions of incomplete FOV. We propose a novel training strategy, FOV-Cut Augmentation (FOV-CA), in which we synthetically cut tractograms to simulate a spectrum of real-world inferior FOV cutoff scenarios. This data augmentation approach enriches the training set with realistic truncated streamlines, enabling the model to achieve superior generalization. We evaluate the proposed TractCloud-FOV on both synthetically cut tractography and two real-life datasets with incomplete FOV. TractCloud-FOV significantly outperforms several state-of-the-art methods on all testing datasets in terms of streamline classification accuracy, generalization ability, tract anatomical depiction, and computational efficiency. Overall, TractCloud-FOV achieves efficient and consistent tractography parcellation in diffusion MRI with incomplete FOV.
Authors: Pengyu Zhang, Xieyuanli Chen, Yuwei Chen, Beizhen Bi, Zhuo Xu, Tian Jin, Xiaotao Huang, Liang Shen
Abstract: Ground penetrating radar (GPR) based localization has gained significant recognition in robotics due to its ability to detect stable subsurface features, offering advantages in environments where traditional sensors like cameras and LiDAR may struggle. However, existing methods are primarily focused on small-scale place recognition (PR), leaving the challenges of PR in large-scale maps unaddressed. These challenges include the inherent sparsity of underground features and the variability in underground dielectric constants, which complicate robust localization. In this work, we investigate the geometric relationship between GPR echo sequences and underground scenes, leveraging the robustness of directional features to inform our network design. We introduce learnable Gabor filters for the precise extraction of directional responses, coupled with a direction-aware attention mechanism for effective geometric encoding. To further enhance performance, we incorporate a shift-invariant unit and a multi-scale aggregation strategy to better accommodate variations in di-electric constants. Experiments conducted on public datasets demonstrate that our proposed EDENet not only surpasses existing solutions in terms of PR performance but also offers advantages in model size and computational efficiency.
Authors: Yu Pan, Bingrong Dai, Jiahao Chen, Lin Wang, Yi Du, Jiao Liu
Abstract: In recent years, Diffusion Models (DMs) have demonstrated significant advances in the field of image generation. However, according to current research, DMs are vulnerable to backdoor attacks, which allow attackers to control the model's output by inputting data containing covert triggers, such as a specific patch or phrase. Existing defense strategies are well equipped to thwart such attacks through backdoor detection and trigger inversion because previous attack methods are constrained by limited input spaces and triggers defined by low-dimensional features. To bridge these gaps, we propose Gungnir, a novel method that enables attackers to activate the backdoor in DMs through hidden style triggers within input images. Our approach proposes using stylistic features as triggers for the first time and implements backdoor attacks successfully in image2image tasks by utilizing Reconstructing-Adversarial Noise (RAN) and Short-Term-Timesteps-Retention (STTR) of DMs. Meanwhile, experiments demonstrate that our method can easily bypass existing defense methods. Among existing DM main backdoor defense frameworks, our approach achieves a 0\% backdoor detection rate (BDR). Our codes are available at https://github.com/paoche11/Gungnir.
Authors: Rishi Mukherjee, Sakshi Singh, Jack McWilliams, Junaed Sattar
Abstract: We introduce COU: Common Objects Underwater, an instance-segmented image dataset of commonly found man-made objects in multiple aquatic and marine environments. COU contains approximately 10K segmented images, annotated from images collected during a number of underwater robot field trials in diverse locations. COU has been created to address the lack of datasets with robust class coverage curated for underwater instance segmentation, which is particularly useful for training light-weight, real-time capable detectors for Autonomous Underwater Vehicles (AUVs). In addition, COU addresses the lack of diversity in object classes since the commonly available underwater image datasets focus only on marine life. Currently, COU contains images from both closed-water (pool) and open-water (lakes and oceans) environments, of 24 different classes of objects including marine debris, dive tools, and AUVs. To assess the efficacy of COU in training underwater object detectors, we use three state-of-the-art models to evaluate its performance and accuracy, using a combination of standard accuracy and efficiency metrics. The improved performance of COU-trained detectors over those solely trained on terrestrial data demonstrates the clear advantage of training with annotated underwater images. We make COU available for broad use under open-source licenses.
Authors: Shaobo Wang, Yicun Yang, Zhiyuan Liu, Chenghao Sun, Xuming Hu, Conghui He, Linfeng Zhang
Abstract: Dataset distillation has emerged as a powerful approach for reducing data requirements in deep learning. Among various methods, distribution matching-based approaches stand out for their balance of computational efficiency and strong performance. However, existing distance metrics used in distribution matching often fail to accurately capture distributional differences, leading to unreliable measures of discrepancy. In this paper, we reformulate dataset distillation as a minmax optimization problem and introduce Neural Characteristic Function Discrepancy (NCFD), a comprehensive and theoretically grounded metric for measuring distributional differences. NCFD leverages the Characteristic Function (CF) to encapsulate full distributional information, employing a neural network to optimize the sampling strategy for the CF's frequency arguments, thereby maximizing the discrepancy to enhance distance estimation. Simultaneously, we minimize the difference between real and synthetic data under this optimized NCFD measure. Our approach, termed Neural Characteristic Function Matching (\mymethod{}), inherently aligns the phase and amplitude of neural features in the complex plane for both real and synthetic data, achieving a balance between realism and diversity in synthetic samples. Experiments demonstrate that our method achieves significant performance gains over state-of-the-art methods on both low- and high-resolution datasets. Notably, we achieve a 20.5\% accuracy boost on ImageSquawk. Our method also reduces GPU memory usage by over 300$\times$ and achieves 20$\times$ faster processing speeds compared to state-of-the-art methods. To the best of our knowledge, this is the first work to achieve lossless compression of CIFAR-100 on a single NVIDIA 2080 Ti GPU using only 2.3 GB of memory.
Authors: Ojonugwa Oluwafemi Ejiga Peter, Md Mahmudur Rahman, Fahmi Khalifa
Abstract: The MEDVQA-GI challenge addresses the integration of AI-driven text-to-image generative models in medical diagnostics, aiming to enhance diagnostic capabilities through synthetic image generation. Existing methods primarily focus on static image analysis and lack the dynamic generation of medical imagery from textual descriptions. This study intends to partially close this gap by introducing a novel approach based on fine-tuned generative models to generate dynamic, scalable, and precise images from textual descriptions. Particularly, our system integrates fine-tuned Stable Diffusion and DreamBooth models, as well as Low-Rank Adaptation (LORA), to generate high-fidelity medical images. The problem is around two sub-tasks namely: image synthesis (IS) and optimal prompt production (OPG). The former creates medical images via verbal prompts, whereas the latter provides prompts that produce high-quality images in specified categories. The study emphasizes the limitations of traditional medical image generation methods, such as hand sketching, constrained datasets, static procedures, and generic models. Our evaluation measures showed that Stable Diffusion surpasses CLIP and DreamBooth + LORA in terms of producing high-quality, diversified images. Specifically, Stable Diffusion had the lowest Fr\'echet Inception Distance (FID) scores (0.099 for single center, 0.064 for multi-center, and 0.067 for combined), indicating higher image quality. Furthermore, it had the highest average Inception Score (2.327 across all datasets), indicating exceptional diversity and quality. This advances the field of AI-powered medical diagnosis. Future research will concentrate on model refining, dataset augmentation, and ethical considerations for efficiently implementing these advances into clinical practice
Authors: Xiang Xiang, Zhuo Xu, Yao Deng, Qinhao Zhou, Yifan Liang, Ke Chen, Qingfang Zheng, Yaowei Wang, Xilin Chen, Wen Gao
Abstract: In open-world remote sensing, deployed models must continuously adapt to a steady influx of new data, which often exhibits various shifts compared to what the model encountered during the training phase. To effectively handle the new data, models are required to detect semantic shifts, adapt to covariate shifts, and continuously update themselves. These challenges give rise to a variety of open-world tasks. However, existing open-world remote sensing studies typically train and test within a single dataset to simulate open-world conditions. Currently, there is a lack of large-scale benchmarks capable of evaluating multiple open-world tasks. In this paper, we introduce OpenEarthSensing, a large-scale fine-grained benchmark for open-world remote sensing. OpenEarthSensing includes 189 scene and objects categories, covering the vast majority of potential semantic shifts that may occur in the real world. Additionally, OpenEarthSensing encompasses five data domains with significant covariate shifts, including two RGB satellite domians, one RGB aerial domian, one MS RGB domian, and one infrared domian. The various domains provide a more comprehensive testbed for evaluating the generalization performance of open-world models. We conduct the baseline evaluation of current mainstream open-world tasks and methods on OpenEarthSensing, demonstrating that it serves as a challenging benchmark for open-world remote sensing.
Authors: John J. Han, Jie Ying Wu
Abstract: The lack of labeled datasets in 3D vision for surgical scenes inhibits the development of robust 3D reconstruction algorithms in the medical domain. Despite the popularity of Neural Radiance Fields and 3D Gaussian Splatting in the general computer vision community, these systems have yet to find consistent success in surgical scenes due to challenges such as non-stationary lighting and non-Lambertian surfaces. As a result, the need for labeled surgical datasets continues to grow. In this work, we introduce a differentiable rendering framework for material and lighting estimation from endoscopic images and known geometry. Compared to previous approaches that model lighting and material jointly as radiance, we explicitly disentangle these scene properties for robust and photorealistic novel view synthesis. To disambiguate the training process, we formulate domain-specific properties inherent in surgical scenes. Specifically, we model the scene lighting as a simple spotlight and material properties as a bidirectional reflectance distribution function, parameterized by a neural network. By grounding color predictions in the rendering equation, we can generate photorealistic images at arbitrary camera poses. We evaluate our method with various sequences from the Colonoscopy 3D Video Dataset and show that our method produces competitive novel view synthesis results compared with other approaches. Furthermore, we demonstrate that synthetic data can be used to develop 3D vision algorithms by finetuning a depth estimation model with our rendered outputs. Overall, we see that the depth estimation performance is on par with fine-tuning with the original real images.
Authors: Shanshan Wan, Yingmei Wei, Lai Kang, Tianrui Shen, Haixuan Wang, Yee-Hong Yang
Abstract: Visual Place Recognition (VPR) is a major challenge for robotics and autonomous systems, with the goal of predicting the location of an image based solely on its visual features. State-of-the-art (SOTA) models extract global descriptors using the powerful foundation model DINOv2 as backbone. These models either explore the cross-image correlation or propose a time-consuming two-stage re-ranking strategy to achieve better performance. However, existing works only utilize the final output of DINOv2, and the current cross-image correlation causes unstable retrieval results. To produce both discriminative and constant global descriptors, this paper proposes stable cross-image correlation enhanced model for VPR called SciceVPR. This model explores the full potential of DINOv2 in providing useful feature representations that implicitly encode valuable contextual knowledge. Specifically, SciceVPR first uses a multi-layer feature fusion module to capture increasingly detailed task-relevant channel and spatial information from the multi-layer output of DINOv2. Secondly, SciceVPR considers the invariant correlation between images within a batch as valuable knowledge to be distilled into the proposed self-enhanced encoder. In this way, SciceVPR can acquire fairly robust global features regardless of domain shifts (e.g., changes in illumination, weather and viewpoint between pictures taken in the same place). Experimental results demonstrate that the base variant, SciceVPR-B, outperforms SOTA one-stage methods with single input on multiple datasets with varying domain conditions. The large variant, SciceVPR-L, performs on par with SOTA two-stage models, scoring over 3% higher in Recall@1 compared to existing models on the challenging Tokyo24/7 dataset. Our code will be released at https://github.com/shuimushan/SciceVPR.
Authors: Aaryan Garg, Akash Kumar, Yogesh S Rawat
Abstract: In this work we study Weakly Supervised Spatio-Temporal Video Grounding (WSTVG), a challenging task of localizing subjects spatio-temporally in videos using only textual queries and no bounding box supervision. Inspired by recent advances in vision-language foundation models, we investigate their utility for WSTVG, leveraging their zero-shot grounding capabilities. However, we find that a simple adaptation lacks essential spatio-temporal grounding abilities. To bridge this gap, we introduce Tubelet Referral Grounding (TRG), which connects textual queries to tubelets to enable spatio-temporal predictions. Despite its promise, TRG struggles with compositional action understanding and dense scene scenarios. To address these limitations, we propose STPro, a novel progressive learning framework with two key modules: (1) Sub-Action Temporal Curriculum Learning (SA-TCL), which incrementally builds compositional action understanding, and (2) Congestion-Guided Spatial Curriculum Learning (CG-SCL), which adapts the model to complex scenes by spatially increasing task difficulty. STPro achieves state-of-the-art results on three benchmark datasets, with improvements of 1.0% on VidSTG-Declarative and 3.0% on HCSTVG-v1.
Authors: Hanbang Liang, Zhen Wang, Weihui Deng
Abstract: Diffusion models have demonstrated their powerful image generation capabilities, effectively fitting highly complex image distributions. These models can serve as strong priors for image restoration. Existing methods often utilize techniques like ControlNet to sample high quality images with low quality images from these priors. However, ControlNet typically involves copying a large part of the original network, resulting in a significantly large number of parameters as the prior scales up. In this paper, we propose a relatively lightweight Adapter that leverages the powerful generative capabilities of pretrained priors to achieve photo-realistic image restoration. The Adapters can be adapt to both denoising UNet and DiT, and performs excellent.
Authors: Dongki Jung, Jaehoon Choi, Yonghan Lee, Somi Jeong, Taejae Lee, Dinesh Manocha, Suyong Yeon
Abstract: We introduce the first learning-based dense matching algorithm, termed Equirectangular Projection-Oriented Dense Kernelized Feature Matching (EDM), specifically designed for omnidirectional images. Equirectangular projection (ERP) images, with their large fields of view, are particularly suited for dense matching techniques that aim to establish comprehensive correspondences across images. However, ERP images are subject to significant distortions, which we address by leveraging the spherical camera model and geodesic flow refinement in the dense matching method. To further mitigate these distortions, we propose spherical positional embeddings based on 3D Cartesian coordinates of the feature grid. Additionally, our method incorporates bidirectional transformations between spherical and Cartesian coordinate systems during refinement, utilizing a unit sphere to improve matching performance. We demonstrate that our proposed method achieves notable performance enhancements, with improvements of +26.72 and +42.62 in AUC@5{\deg} on the Matterport3D and Stanford2D3D datasets.
Authors: Dacheng Li, Yunhao Fang, Yukang Chen, Shuo Yang, Shiyi Cao, Justin Wong, Michael Luo, Xiaolong Wang, Hongxu Yin, Joseph E. Gonzalez, Ion Stoica, Song Han, Yao Lu
Abstract: Video generation models have rapidly progressed, positioning themselves as video world models capable of supporting decision-making applications like robotics and autonomous driving. However, current benchmarks fail to rigorously evaluate these claims, focusing only on general video quality, ignoring important factors to world models such as physics adherence. To bridge this gap, we propose WorldModelBench, a benchmark designed to evaluate the world modeling capabilities of video generation models in application-driven domains. WorldModelBench offers two key advantages: (1) Against to nuanced world modeling violations: By incorporating instruction-following and physics-adherence dimensions, WorldModelBench detects subtle violations, such as irregular changes in object size that breach the mass conservation law - issues overlooked by prior benchmarks. (2) Aligned with large-scale human preferences: We crowd-source 67K human labels to accurately measure 14 frontier models. Using our high-quality human labels, we further fine-tune an accurate judger to automate the evaluation procedure, achieving 8.6% higher average accuracy in predicting world modeling violations than GPT-4o with 2B parameters. In addition, we demonstrate that training to align human annotations by maximizing the rewards from the judger noticeably improve the world modeling capability. The website is available at https://worldmodelbench-team.github.io.
Authors: Ke Sun, Shen Chen, Taiping Yao, Ziyin Zhou, Jiayi Ji, Xiaoshuai Sun, Chia-Wen Lin, Rongrong Ji
Abstract: Face manipulation techniques have achieved significant advances, presenting serious challenges to security and social trust. Recent works demonstrate that leveraging multimodal models can enhance the generalization and interpretability of face forgery detection. However, existing annotation approaches, whether through human labeling or direct Multimodal Large Language Model (MLLM) generation, often suffer from hallucination issues, leading to inaccurate text descriptions, especially for high-quality forgeries. To address this, we propose Face Forgery Text Generator (FFTG), a novel annotation pipeline that generates accurate text descriptions by leveraging forgery masks for initial region and type identification, followed by a comprehensive prompting strategy to guide MLLMs in reducing hallucination. We validate our approach through fine-tuning both CLIP with a three-branch training framework combining unimodal and multimodal objectives, and MLLMs with our structured annotations. Experimental results demonstrate that our method not only achieves more accurate annotations with higher region identification accuracy, but also leads to improvements in model performance across various forgery detection benchmarks. Our Codes are available in https://github.com/skJack/VLFFD.git.
Authors: Kiranmayee Janardhan, Christy Bobby Thomas
Abstract: Glioma, a prevalent and heterogeneous tumor originating from the glial cells, can be differentiated as Low Grade Glioma (LGG) and High Grade Glioma (HGG) according to World Health Organization's norms. Classifying gliomas is essential for treatment protocols that depend extensively on subtype differentiation. For non-invasive glioma evaluation, Magnetic Resonance Imaging (MRI) offers vital information about the morphology and location of the the tumor. The versatility of MRI allows the classification of gliomas as LGG and HGG based on their texture, perfusion, and diffusion characteristics, and further for improving the diagnosis and providing tailored treatments. Nevertheless, the precise classification is complicated by tumor heterogeneity and overlapping radiomic characteristics. Thus, in this work, wavelet based novel fusion algorithm were implemented on multi-sequence T1, T1-contrast enhanced (T1CE), T2 and Fluid Attenuated Inversion Recovery (FLAIR) MRI images to compute the radiomics features. Furthermore, principal component analysis is applied to reduce the feature space and XGBoost, Support Vector Machine, and Random Forest Classifier are used for the classification. The result shows that the SVM algorithm performs comparatively well with an accuracy of 90.17%, precision of 91.04% and recall of 96.19%, F1-score of 93.53%, and AUC of 94.60% when implemented on BraTS 2018 dataset and with an accuracy of 91.34%, precision of 93.05% and recall of 96.13%, F1-score of 94.53%, and AUC of 93.71% for BraTS 2018 dataset. Thus, the proposed algorithm could be potentially implemented for the computer-aided diagnosis and grading system for gliomas.
Authors: Yuan Li, Cheng Lin, Yuan Liu, Xiaoxiao Long, Chenxu Zhang, Ningna Wang, Xin Li, Wenping Wang, Xiaohu Guo
Abstract: Diffusion-based 3D generation has made remarkable progress in recent years. However, existing 3D generative models often produce overly dense and unstructured meshes, which stand in stark contrast to the compact, structured, and sharply-edged Computer-Aided Design (CAD) models crafted by human designers. To address this gap, we introduce CADDreamer, a novel approach for generating boundary representations (B-rep) of CAD objects from a single image. CADDreamer employs a primitive-aware multi-view diffusion model that captures both local geometric details and high-level structural semantics during the generation process. By encoding primitive semantics into the color domain, the method leverages the strong priors of pre-trained diffusion models to align with well-defined primitives. This enables the inference of multi-view normal maps and semantic maps from a single image, facilitating the reconstruction of a mesh with primitive labels. Furthermore, we introduce geometric optimization techniques and topology-preserving extraction methods to mitigate noise and distortion in the generated primitives. These enhancements result in a complete and seamless B-rep of the CAD model. Experimental results demonstrate that our method effectively recovers high-quality CAD objects from single-view images. Compared to existing 3D generation techniques, the B-rep models produced by CADDreamer are compact in representation, clear in structure, sharp in edges, and watertight in topology.
Authors: Xiwen Liang, Min Lin, Weiqi Ruan, Rongtao Xu, Yuecheng Liu, Jiaqi Chen, Bingqian Lin, Yuzheng Zhuang, Xiaodan Liang
Abstract: Existing methods for vision-language task planning excel in short-horizon tasks but often fall short in complex, long-horizon planning within dynamic environments. These challenges primarily arise from the difficulty of effectively training models to produce high-quality reasoning processes for long-horizon tasks. To address this, we propose Structured Preference Optimization (SPO), which aims to enhance reasoning and action selection in long-horizon task planning through structured preference evaluation and optimized training strategies. Specifically, SPO introduces: 1) Preference-Based Scoring and Optimization, which systematically evaluates reasoning chains based on task relevance, visual grounding, and historical consistency; and 2) Curriculum-Guided Training, where the model progressively adapts from simple to complex tasks, improving its generalization ability in long-horizon scenarios and enhancing reasoning robustness. To advance research in vision-language long-horizon task planning, we introduce ExtendaBench, a comprehensive benchmark covering 1,509 tasks across VirtualHome and Habitat 2.0, categorized into ultra-short, short, medium, and long tasks. Experimental results demonstrate that SPO significantly improves reasoning quality and final decision accuracy, outperforming prior methods on long-horizon tasks and underscoring the effectiveness of preference-driven optimization in vision-language task planning. Specifically, SPO achieves a +5.98% GCR and +4.68% SR improvement in VirtualHome and a +3.30% GCR and +2.11% SR improvement in Habitat over the best-performing baselines.
Authors: Weijia Zhang, Fei Xie, Weidong Cai, Chao Ma
Abstract: Knowledge distillation (KD) aims to transfer the knowledge of a more capable yet cumbersome teacher model to a lightweight student model. In recent years, relation-based KD methods have fallen behind, as their instance-matching counterparts dominate in performance. In this paper, we revive relational KD by identifying and tackling several key issues in relation-based methods, including their susceptibility to overfitting and spurious responses. Specifically, we transfer novelly constructed affinity graphs that compactly encapsulate a wealth of beneficial inter-sample, inter-class, and inter-view correlations by exploiting virtual views and relations as a new kind of knowledge. As a result, the student has access to richer guidance signals and stronger regularisation throughout the distillation process. To further mitigate the adverse impact of spurious responses, we prune the affinity graphs by dynamically detaching redundant and unreliable edges. Extensive experiments on CIFAR-100 and ImageNet datasets demonstrate the superior performance of the proposed virtual relation matching (VRM) method over a range of models, architectures, and set-ups. For instance, VRM for the first time hits 74.0% accuracy for ResNet50-to-MobileNetV2 distillation on ImageNet, and improves DeiT-T by 14.44% on CIFAR-100 with a ResNet56 teacher. Thorough analyses are also conducted to gauge the soundness, properties, and complexity of our designs. Code and models will be released.
Authors: Yueyang Li, Lei Chen, Wenhao Dong, Shengyu Gong, Zijian Kang, Boyang Wei, Weiming Zeng, Hongjie Yan, Lingbin Bian, Wai Ting Siok, Nizhuan Wang
Abstract: Developing interpretable models for diagnosing neurodevelopmental disorders (NDDs) is highly valuable yet challenging, primarily due to the complexity of encoding, decoding and integrating imaging and non-imaging data. Many existing machine learning models struggle to provide comprehensive interpretability, often failing to extract meaningful biomarkers from imaging data, such as functional magnetic resonance imaging (fMRI), or lacking mechanisms to explain the significance of non-imaging data. In this paper, we propose the Interpretable Information Bottleneck Heterogeneous Graph Neural Network (I2B-HGNN), a novel framework designed to learn from fine-grained local patterns to comprehensive global multi-modal interactions. This framework comprises two key modules. The first module, the Information Bottleneck Graph Transformer (IBGraphFormer) for local patterns, integrates global modeling with brain connectomic-constrained graph neural networks to identify biomarkers through information bottleneck-guided pooling. The second module, the Information Bottleneck Heterogeneous Graph Attention Network (IB-HGAN) for global multi-modal interactions, facilitates interpretable multi-modal fusion of imaging and non-imaging data using heterogeneous graph neural networks. The results of the experiments demonstrate that I2B-HGNN excels in diagnosing NDDs with high accuracy, providing interpretable biomarker identification and effective analysis of non-imaging data.
Authors: Masoumeh Chapariniya, Hossein Ranjbar, Teodora Vukovic, Sarah Ebling, Volker Dellwo
Abstract: In the age of AI-driven generative technologies, traditional biometric recognition systems face unprecedented challenges, particularly from sophisticated deepfake and face reenactment techniques. In this study, we propose a Two-Stream Spatial-Temporal Transformer Framework for person identification using upper body keypoints visible during online conversations, which we term conversational keypoints. Our framework processes both spatial relationships between keypoints and their temporal evolution through two specialized branches: a Spatial Transformer (STR) that learns distinctive structural patterns in keypoint configurations, and a Temporal Transformer (TTR) that captures sequential motion patterns. Using the state-of-the-art Sapiens pose estimator, we extract 133 keypoints (based on COCO-WholeBody format) representing facial features, head pose, and hand positions. The framework was evaluated on a dataset of 114 individuals engaged in natural conversations, achieving recognition accuracies of 80.12% for the spatial stream, 63.61% for the temporal stream. We then explored two fusion strategies: a shared loss function approach achieving 82.22% accuracy, and a feature-level fusion method that concatenates feature maps from both streams, significantly improving performance to 94.86%. By jointly modeling both static anatomical relationships and dynamic movement patterns, our approach learns comprehensive identity signatures that are more robust to spoofing than traditional appearance-based methods.
Authors: Xiao Wang, Jingyun Hua, Weihong Lin, Yuanxing Zhang, Fuzheng Zhang, Jianlong Wu, Di Zhang, Liqiang Nie
Abstract: Recent Multi-modal Large Language Models (MLLMs) have made great progress in video understanding. However, their performance on videos involving human actions is still limited by the lack of high-quality data. To address this, we introduce a two-stage data annotation pipeline. First, we design strategies to accumulate videos featuring clear human actions from the Internet. Second, videos are annotated in a standardized caption format that uses human attributes to distinguish individuals and chronologically details their actions and interactions. Through this pipeline, we curate two datasets, namely HAICTrain and HAICBench. \textbf{HAICTrain} comprises 126K video-caption pairs generated by Gemini-Pro and verified for training purposes. Meanwhile, \textbf{HAICBench} includes 500 manually annotated video-caption pairs and 1,400 QA pairs, for a comprehensive evaluation of human action understanding. Experimental results demonstrate that training with HAICTrain not only significantly enhances human understanding abilities across 4 benchmarks, but can also improve text-to-video generation results. Both the HAICTrain and HAICBench are released at https://huggingface.co/datasets/KuaishouHAIC/HAIC.
Authors: Bach-Thuan Bui, Huy-Hoang Bui, Yasuyuki Fujii, Dinh-Tuan Tran, Joo-Ho Lee
Abstract: In this paper, we present a new approach for improving 3D point and line mapping regression for camera re-localization. Previous methods typically rely on feature matching (FM) with stored descriptors or use a single network to encode both points and lines. While FM-based methods perform well in large-scale environments, they become computationally expensive with a growing number of mapping points and lines. Conversely, approaches that learn to encode mapping features within a single network reduce memory footprint but are prone to overfitting, as they may capture unnecessary correlations between points and lines. We propose that these features should be learned independently, each with a distinct focus, to achieve optimal accuracy. To this end, we introduce a new architecture that learns to prioritize each feature independently before combining them for localization. Experimental results demonstrate that our approach significantly enhances the 3D map point and line regression performance for camera re-localization. The implementation of our method will be publicly available at: https://github.com/ais-lab/pl2map/.
Authors: Jiawen Li, Jiali Hu, Qiehe Sun, Renao Yan, Minxi Ouyang, Tian Guan, Anjia Han, Chao He, Yonghong He
Abstract: The emergence of foundation models in computational pathology has transformed histopathological image analysis, with whole slide imaging (WSI) diagnosis being a core application. Traditionally, weakly supervised fine-tuning via multiple instance learning (MIL) has been the primary method for adapting foundation models to WSIs. However, in this work we present a key experimental finding: a simple nonlinear mapping strategy combining mean pooling and a multilayer perceptron, called SiMLP, can effectively adapt patch-level foundation models to slide-level tasks without complex MIL-based learning. Through extensive experiments across diverse downstream tasks, we demonstrate the superior performance of SiMLP with state-of-the-art methods. For instance, on a large-scale pan-cancer classification task, SiMLP surpasses popular MIL-based methods by 3.52%. Furthermore, SiMLP shows strong learning ability in few-shot classification and remaining highly competitive with slide-level foundation models pretrained on tens of thousands of slides. Finally, SiMLP exhibits remarkable robustness and transferability in lung cancer subtyping. Overall, our findings challenge the conventional MIL-based fine-tuning paradigm, demonstrating that a task-agnostic representation strategy alone can effectively adapt foundation models to WSI analysis. These insights offer a unique and meaningful perspective for future research in digital pathology, paving the way for more efficient and broadly applicable methodologies.
Authors: Fadeel Sher Khan, Joshua Ebenezer, Hamid Sheikh, Seok-Jun Lee
Abstract: Smartphone cameras have become ubiquitous imaging tools, yet their small sensors and compact optics often limit spatial resolution and introduce distortions. Combining information from multiple low-resolution (LR) frames to produce a high-resolution (HR) image has been explored to overcome the inherent limitations of smartphone cameras. Despite the promise of multi-frame super-resolution (MFSR), current approaches are hindered by datasets that fail to capture the characteristic noise and motion patterns found in real-world handheld burst images. In this work, we address this gap by introducing a novel synthetic data engine that uses multi-exposure static images to synthesize LR-HR training pairs while preserving sensor-specific noise characteristics and image motion found during handheld burst photography. We also propose MFSR-GAN: a multi-scale RAW-to-RGB network for MFSR. Compared to prior approaches, MFSR-GAN emphasizes a "base frame" throughout its architecture to mitigate artifacts. Experimental results on both synthetic and real data demonstrates that MFSR-GAN trained with our synthetic engine yields sharper, more realistic reconstructions than existing methods for real-world MFSR.
Authors: Zelong Sun, Dong Jing, Zhiwu Lu
Abstract: Zero-Shot Composed Image Retrieval (ZS-CIR) aims to retrieve target images by integrating information from a composed query (reference image and modification text) without training samples. Existing methods primarily combine caption models and large language models (LLMs) to generate target captions based on composed queries but face various issues such as incompatibility, visual information loss, and insufficient reasoning. In this work, we propose CoTMR, a training-free framework crafted for ZS-CIR with novel Chain-of-thought (CoT) and Multi-scale Reasoning. Instead of relying on caption models for modality transformation, CoTMR employs the Large Vision-Language Model (LVLM) to achieve unified understanding and reasoning for composed queries. To enhance the reasoning reliability, we devise CIRCoT, which guides the LVLM through a step-by-step inference process using predefined subtasks. Considering that existing approaches focus solely on global-level reasoning, our CoTMR incorporates multi-scale reasoning to achieve more comprehensive inference via fine-grained predictions about the presence or absence of key elements at the object scale. Further, we design a Multi-Grained Scoring (MGS) mechanism, which integrates CLIP similarity scores of the above reasoning outputs with candidate images to realize precise retrieval. Extensive experiments demonstrate that our CoTMR not only drastically outperforms previous methods across four prominent benchmarks but also offers appealing interpretability.
Authors: Anh Tien Nguyen, Keunho Byeon, Kyungeun Kim, Jin Tae Kwak
Abstract: Recent advances in vision-language models (VLMs) have shown remarkable potential in bridging visual and textual modalities. In computational pathology, domain-specific VLMs, which are pre-trained on extensive histopathology image-text datasets, have succeeded in various downstream tasks. However, existing research has primarily focused on the pre-training process and direct applications of VLMs on the patch level, leaving their great potential for whole slide image (WSI) applications unexplored. In this study, we hypothesize that pre-trained VLMs inherently capture informative and interpretable WSI representations through quantitative feature extraction. To validate this hypothesis, we introduce Vision and Language Embeddings for Explainable WSI Representation (VLEER), a novel method designed to leverage VLMs for WSI representation. We systematically evaluate VLEER on three pathological WSI datasets, proving its better performance in WSI analysis compared to conventional vision features. More importantly, VLEER offers the unique advantage of interpretability, enabling direct human-readable insights into the results by leveraging the textual modality for detailed pathology annotations, providing clear reasoning for WSI-level pathology downstream tasks.
Authors: Shaoming Li, Qing Cai, Songqi Kong, Runqing Tan, Heng Tong, Shiji Qiu, Yongguo Jiang, Zhi Liu
Abstract: Reconstructing 3D shapes from a single image plays an important role in computer vision. Many methods have been proposed and achieve impressive performance. However, existing methods mainly focus on extracting semantic information from images and then simply concatenating it with 3D point clouds without further exploring the concatenated semantics. As a result, these entangled semantic features significantly hinder the reconstruction performance. In this paper, we propose a novel single-image 3D reconstruction method called Mining Effective Semantic Cues for 3D Reconstruction from a Single Image (MESC-3D), which can actively mine effective semantic cues from entangled features. Specifically, we design an Effective Semantic Mining Module to establish connections between point clouds and image semantic attributes, enabling the point clouds to autonomously select the necessary information. Furthermore, to address the potential insufficiencies in semantic information from a single image, such as occlusions, inspired by the human ability to represent 3D objects using prior knowledge drawn from daily experiences, we introduce a 3D Semantic Prior Learning Module. This module incorporates semantic understanding of spatial structures, enabling the model to interpret and reconstruct 3D objects with greater accuracy and realism, closely mirroring human perception of complex 3D environments. Extensive evaluations show that our method achieves significant improvements in reconstruction quality and robustness compared to prior works. Additionally, further experiments validate the strong generalization capabilities and excels in zero-shot preformance on unseen classes. Code is available at https://github.com/QINGQINGLE/MESC-3D.
Authors: Chunlin Zhong, Shuang Hao, Junhua Wu, Xiaona Chang, Jiwei Jiang, Xiu Nie, He Tang, Xiang Bai
Abstract: With the rapid development of computational pathology, many AI-assisted diagnostic tasks have emerged. Cellular nuclei segmentation can segment various types of cells for downstream analysis, but it relies on predefined categories and lacks flexibility. Moreover, pathology visual question answering can perform image-level understanding but lacks region-level detection capability. To address this, we propose a new benchmark called Pathology Visual Grounding (PathVG), which aims to detect regions based on expressions with different attributes. To evaluate PathVG, we create a new dataset named RefPath which contains 27,610 images with 33,500 language-grounded boxes. Compared to visual grounding in other domains, PathVG presents pathological images at multi-scale and contains expressions with pathological knowledge. In the experimental study, we found that the biggest challenge was the implicit information underlying the pathological expressions. Based on this, we proposed Pathology Knowledge-enhanced Network (PKNet) as the baseline model for PathVG. PKNet leverages the knowledge-enhancement capabilities of Large Language Models (LLMs) to convert pathological terms with implicit information into explicit visual features, and fuses knowledge features with expression features through the designed Knowledge Fusion Module (KFM). The proposed method achieves state-of-the-art performance on the PathVG benchmark.
Authors: Bj\"orn Braun, Rayan Armani, Manuel Meier, Max Moebus, Christian Holz
Abstract: Egocentric vision systems aim to understand the spatial surroundings and the wearer's behavior inside it, including motions, activities, and interaction with objects. Since a person's attention and situational responses are influenced by their physiological state, egocentric systems must also detect this state for better context awareness. In this paper, we propose egoPPG, a novel task for egocentric vision systems to extract a person's heart rate (HR) as a key indicator of the wearer's physiological state from the system's built-in sensors (e.g., eye tracking videos). We then propose EgoPulseFormer, a method that solely takes eye-tracking video as input to estimate a person's photoplethysmogram (PPG) from areas around the eyes to track HR values-without requiring additional or dedicated hardware. We demonstrate the downstream benefit of EgoPulseFormer on EgoExo4D, where we find that augmenting existing models with tracked HR values improves proficiency estimation by 14%. To train and validate EgoPulseFormer, we collected a dataset of 13+ hours of eye-tracking videos from Project Aria and contact-based blood volume pulse signals as well as an electrocardiogram (ECG) for ground-truth HR values. 25 participants performed diverse everyday activities such as office work, cooking, dancing, and exercising, which induced significant natural motion and HR variation (44-164 bpm). Our model robustly estimates HR (MAE=8.82 bpm) and captures patterns (r=0.81). Our results show how egocentric systems may unify environmental and physiological tracking to better understand user actions and internal states.
Authors: Hu Gao, Depeng Dang
Abstract: Image deblurring aims to restore high-quality images from blurred ones. While existing deblurring methods have made significant progress, most overlook the fact that the degradation degree varies across different regions. In this paper, we propose AIBNet, a network that adaptively identifies the blurred regions, enabling differential restoration of these regions. Specifically, we design a spatial feature differential handling block (SFDHBlock), with the core being the spatial domain feature enhancement module (SFEM). Through the feature difference operation, SFEM not only helps the model focus on the key information in the blurred regions but also eliminates the interference of implicit noise. Additionally, based on the fact that the difference between sharp and blurred images primarily lies in the high-frequency components, we propose a high-frequency feature selection block (HFSBlock). The HFSBlock first uses learnable filters to extract high-frequency features and then selectively retains the most important ones. To fully leverage the decoder's potential, we use a pre-trained model as the encoder and incorporate the above modules only in the decoder. Finally, to alleviate the resource burden during training, we introduce a progressive training strategy. Extensive experiments demonstrate that our AIBNet achieves superior performance in image deblurring.
Authors: Jiaming Chu, Lei Jin, Tao Wang, Junliang Xing, Jian Zhao
Abstract: The rapid development of image generation and editing algorithms in recent years has enabled ordinary user to produce realistic images. However, the current AI painting ecosystem predominantly relies on text-driven diffusion models (T2I), which pose challenges in accurately capturing user requirements. Furthermore, achieving compatibility with other modalities incurs substantial training costs. To this end, we introduce DiffBrush, which is compatible with T2I models and allows users to draw and edit images. By manipulating and adapting the internal representation of the diffusion model, DiffBrush guides the model-generated images to converge towards the user's hand-drawn sketches for user's specific needs without additional training. DiffBrush achieves control over the color, semantic, and instance of objects in images by continuously guiding the latent and instance-level attention map during the denoising process of the diffusion model. Besides, we propose a latent regeneration, which refines the randomly sampled noise in the diffusion model, obtaining a better image generation layout. Finally, users only need to roughly draw the mask of the instance (acceptable colors) on the canvas, DiffBrush can naturally generate the corresponding instance at the corresponding location.
Authors: Haonan An, Guang Hua, Zhengru Fang, Guowen Xu, Susanto Rahardja, Yuguang Fang
Abstract: The intellectual property of deep image-to-image models can be protected by the so-called box-free watermarking. It uses an encoder and a decoder, respectively, to embed into and extract from the model's output images invisible copyright marks. Prior works have improved watermark robustness, focusing on the design of better watermark encoders. In this paper, we reveal an overlooked vulnerability of the unprotected watermark decoder which is jointly trained with the encoder and can be exploited to train a watermark removal network. To defend against such an attack, we propose the decoder gradient shield (DGS) as a protection layer in the decoder API to prevent gradient-based watermark removal with a closed-form solution. The fundamental idea is inspired by the classical adversarial attack, but is utilized for the first time as a defensive mechanism in the box-free model watermarking. We then demonstrate that DGS can reorient and rescale the gradient directions of watermarked queries and stop the watermark remover's training loss from converging to the level without DGS, while retaining decoder output image quality. Experimental results verify the effectiveness of proposed method. Code of paper will be made available upon acceptance.
Authors: Utku Ozbulak, Seyed Amir Mousavi, Francesca Tozzi, Nikdokht Rashidian, Wouter Willaert, Wesley De Neve, Joris Vankerschaver
Abstract: Real-time video segmentation is a promising feature for AI-assisted surgery, providing intraoperative guidance by identifying surgical tools and anatomical structures. However, deploying state-of-the-art segmentation models, such as SAM2, in real-time settings is computationally demanding, which makes it essential to balance frame rate and segmentation performance. In this study, we investigate the impact of frame rate on zero-shot surgical video segmentation, evaluating SAM2's effectiveness across multiple frame sampling rates for cholecystectomy procedures. Surprisingly, our findings indicate that in conventional evaluation settings, frame rates as low as a single frame per second can outperform 25 FPS, as fewer frames smooth out segmentation inconsistencies. However, when assessed in a real-time streaming scenario, higher frame rates yield superior temporal coherence and stability, particularly for dynamic objects such as surgical graspers. Finally, we investigate human perception of real-time surgical video segmentation among professionals who work closely with such data and find that respondents consistently prefer high FPS segmentation mask overlays, reinforcing the importance of real-time evaluation in AI-assisted surgery.
Authors: Xue Yang, Tao Chen, Lei Guo, Wenbo Jiang, Ji Guo, Yongming Li, Jiaming He
Abstract: Reference-based image super-resolution (RefSR) represents a promising advancement in super-resolution (SR). In contrast to single-image super-resolution (SISR), RefSR leverages an additional reference image to help recover high-frequency details, yet its vulnerability to backdoor attacks has not been explored. To fill this research gap, we propose a novel attack framework called BadRefSR, which embeds backdoors in the RefSR model by adding triggers to the reference images and training with a mixed loss function. Extensive experiments across various backdoor attack settings demonstrate the effectiveness of BadRefSR. The compromised RefSR network performs normally on clean input images, while outputting attacker-specified target images on triggered input images. Our study aims to alert researchers to the potential backdoor risks in RefSR. Codes are available at https://github.com/xuefusiji/BadRefSR.
Authors: Zhengxuan Zhang, Yin Wu, Yuyu Luo, Nan Tang
Abstract: Visual Question Answering (VQA) focuses on providing answers to natural language questions by utilizing information from images. Although cutting-edge multimodal large language models (MLLMs) such as GPT-4o achieve strong performance on VQA tasks, they frequently fall short in accessing domain-specific or the latest knowledge. To mitigate this issue, retrieval-augmented generation (RAG) leveraging external knowledge bases (KBs), referred to as KB-VQA, emerges as a promising approach. Nevertheless, conventional unimodal retrieval techniques, which translate images into textual descriptions, often result in the loss of critical visual details. This study presents fine-grained knowledge units, which merge textual snippets with entity images stored in vector databases. Furthermore, we introduce a knowledge unit retrieval-augmented generation framework (KU-RAG) that integrates fine-grained retrieval with MLLMs. The proposed KU-RAG framework ensures precise retrieval of relevant knowledge and enhances reasoning capabilities through a knowledge correction chain. Experimental findings demonstrate that our approach significantly boosts the performance of leading KB-VQA methods, achieving improvements of up to 10%.
Authors: Sabina Jangirova, Branislava Jankovic, Waseem Ullah, Latif U. Khan, Mohsen Guizani
Abstract: Wildfire catastrophes cause significant environmental degradation, human losses, and financial damage. To mitigate these severe impacts, early fire detection and warning systems are crucial. Current systems rely primarily on fixed CCTV cameras with a limited field of view, restricting their effectiveness in large outdoor environments. The fusion of intelligent fire detection with remote sensing improves coverage and mobility, enabling monitoring in remote and challenging areas. Existing approaches predominantly utilize convolutional neural networks and vision transformer models. While these architectures provide high accuracy in fire detection, their computational complexity limits real-time performance on edge devices such as UAVs. In our work, we present a lightweight fire detection model based on MobileViT-S, compressed through the distillation of knowledge from a stronger teacher model. The ablation study highlights the impact of a teacher model and the chosen distillation technique on the model's performance improvement. We generate activation map visualizations using Grad-CAM to confirm the model's ability to focus on relevant fire regions. The high accuracy and efficiency of the proposed model make it well-suited for deployment on satellites, UAVs, and IoT devices for effective fire detection. Experiments on common fire benchmarks demonstrate that our model suppresses the state-of-the-art model by 0.44%, 2.00% while maintaining a compact model size. Our model delivers the highest processing speed among existing works, achieving real-time performance on resource-constrained devices.
Authors: Fuyun Wang, Tong Zhang, Yuanzhi Wang, Yide Qiu, Xin Liu, Xu Guo, Zhen Cui
Abstract: In Open-set Supervised Anomaly Detection (OSAD), the existing methods typically generate pseudo anomalies to compensate for the scarcity of observed anomaly samples, while overlooking critical priors of normal samples, leading to less effective discriminative boundaries. To address this issue, we propose a Distribution Prototype Diffusion Learning (DPDL) method aimed at enclosing normal samples within a compact and discriminative distribution space. Specifically, we construct multiple learnable Gaussian prototypes to create a latent representation space for abundant and diverse normal samples and learn a Schr\"odinger bridge to facilitate a diffusive transition toward these prototypes for normal samples while steering anomaly samples away. Moreover, to enhance inter-sample separation, we design a dispersion feature learning way in hyperspherical space, which benefits the identification of out-of-distribution anomalies. Experimental results demonstrate the effectiveness and superiority of our proposed DPDL, achieving state-of-the-art performance on 9 public datasets.
Authors: Maximilian Rokuss, Yannick Kirchhoff, Seval Akbal, Balint Kovacs, Saikat Roy, Constantin Ulrich, Tassilo Wald, Lukas T. Rotkopf, Heinz-Peter Schlemmer, Klaus Maier-Hein
Abstract: In this work, we present LesionLocator, a framework for zero-shot longitudinal lesion tracking and segmentation in 3D medical imaging, establishing the first end-to-end model capable of 4D tracking with dense spatial prompts. Our model leverages an extensive dataset of 23,262 annotated medical scans, as well as synthesized longitudinal data across diverse lesion types. The diversity and scale of our dataset significantly enhances model generalizability to real-world medical imaging challenges and addresses key limitations in longitudinal data availability. LesionLocator outperforms all existing promptable models in lesion segmentation by nearly 10 dice points, reaching human-level performance, and achieves state-of-the-art results in lesion tracking, with superior lesion retrieval and segmentation accuracy. LesionLocator not only sets a new benchmark in universal promptable lesion segmentation and automated longitudinal lesion tracking but also provides the first open-access solution of its kind, releasing our synthetic 4D dataset and model to the community, empowering future advancements in medical imaging. Code is available at: www.github.com/MIC-DKFZ/LesionLocator
Authors: Woo Kyoung Han, Byeonghun Lee, Hyunmin Cho, Sunghoon Im, Kyong Hwan Jin
Abstract: We quantify the upper bound on the size of the implicit neural representation (INR) model from a digital perspective. The upper bound of the model size increases exponentially as the required bit-precision increases. To this end, we present a bit-plane decomposition method that makes INR predict bit-planes, producing the same effect as reducing the upper bound of the model size. We validate our hypothesis that reducing the upper bound leads to faster convergence with constant model size. Our method achieves lossless representation in 2D image and audio fitting, even for high bit-depth signals, such as 16-bit, which was previously unachievable. We pioneered the presence of bit bias, which INR prioritizes as the most significant bit (MSB). We expand the application of the INR task to bit depth expansion, lossless image compression, and extreme network quantization. Our source code is available at https://github.com/WooKyoungHan/LosslessINR
Authors: Mengzhu Li, Quanxing Zha, Hongjun Wu
Abstract: Dynamic Facial Expression Recognition (DFER) facilitates the understanding of psychological intentions through non-verbal communication. Existing methods struggle to manage irrelevant information, such as background noise and redundant semantics, which impacts both efficiency and effectiveness. In this work, we propose a novel supervised temporal soft masked autoencoder network for DFER, namely AdaTosk, which integrates a parallel supervised classification branch with the self-supervised reconstruction branch. The self-supervised reconstruction branch applies random binary hard mask to generate diverse training samples, encouraging meaningful feature representations in visible tokens. Meanwhile the classification branch employs an adaptive temporal soft mask to flexibly mask visible tokens based on their temporal significance. Its two key components, respectively of, class-agnostic and class-semantic soft masks, serve to enhance critical expression moments and reduce semantic redundancy over time. Extensive experiments conducted on widely-used benchmarks demonstrate that our AdaTosk remarkably reduces computational costs compared with current state-of-the-art methods while still maintaining competitive performance.
Authors: Junchao Zhu, Ruining Deng, Tianyuan Yao, Juming Xiong, Chongyu Qu, Junlin Guo, Siqi Lu, Yucheng Tang, Daguang Xu, Mengmeng Yin, Yu Wang, Shilin Zhao, Yaohong Wang, Haichun Yang, Yuankai Huo
Abstract: The rapid development of spatial transcriptomics (ST) offers new opportunities to explore the gene expression patterns within the spatial microenvironment. Current research integrates pathological images to infer gene expression, addressing the high costs and time-consuming processes to generate spatial transcriptomics data. However, as spatial transcriptomics resolution continues to improve, existing methods remain primarily focused on gene expression prediction at low-resolution spot levels. These methods face significant challenges, especially the information bottleneck, when they are applied to high-resolution HD data. To bridge this gap, this paper introduces MagNet, a multi-level attention graph network designed for accurate prediction of high-resolution HD data. MagNet employs cross-attention layers to integrate features from multi-resolution image patches hierarchically and utilizes a GAT-Transformer module to aggregate neighborhood information. By integrating multilevel features, MagNet overcomes the limitations posed by low-resolution inputs in predicting high-resolution gene expression. We systematically evaluated MagNet and existing ST prediction models on both a private spatial transcriptomics dataset and a public dataset at three different resolution levels. The results demonstrate that MagNet achieves state-of-the-art performance at both spot level and high-resolution bin levels, providing a novel methodology and benchmark for future research and applications in high-resolution HD-level spatial transcriptomics. Code is available at https://github.com/Junchao-Zhu/MagNet.
Authors: Chanhui Lee, Yeonghwan Song, Jeany Son
Abstract: Data-free Universal Adversarial Perturbation (UAP) is an image-agnostic adversarial attack that deceives deep neural networks using a single perturbation generated solely from random noise, without any data priors. However, traditional data-free UAP methods often suffer from limited transferability due to the absence of semantic information in random noise. To address this, we propose a novel data-free universal attack approach that generates a pseudo-semantic prior recursively from the UAPs, enriching semantic contents within the data-free UAP framework. Our method is based on the observation that UAPs inherently contain latent semantic information, enabling the generated UAP to act as an alternative data prior, by capturing a diverse range of semantics through region sampling. We further introduce a sample reweighting technique to emphasize hard examples by focusing on samples that are less affected by the UAP. By leveraging the semantic information from the pseudo-semantic prior, we also incorporate input transformations, typically ineffective in data-free UAPs due to the lack of semantic content in random priors, to boost black-box transferability. Comprehensive experiments on ImageNet show that our method achieves state-of-the-art performance in average fooling rate by a substantial margin, significantly improves attack transferability across various CNN architectures compared to existing data-free UAP methods, and even surpasses data-dependent UAP methods.
Authors: Jingru Fu, Yuqi Zheng, Neel Dey, Daniel Ferreira, Rodrigo Moreno
Abstract: Simulating prospective magnetic resonance imaging (MRI) scans from a given individual brain image is challenging, as it requires accounting for canonical changes in aging and/or disease progression while also considering the individual brain's current status and unique characteristics. While current deep generative models can produce high-resolution anatomically accurate templates for population-wide studies, their ability to predict future aging trajectories for individuals remains limited, particularly in capturing subject-specific neuroanatomical variations over time. In this study, we introduce Individualized Brain Synthesis (InBrainSyn), a framework for synthesizing high-resolution subject-specific longitudinal MRI scans that simulate neurodegeneration in both Alzheimer's disease (AD) and normal aging. InBrainSyn uses a parallel transport algorithm to adapt the population-level aging trajectories learned by a generative deep template network, enabling individualized aging synthesis. As InBrainSyn uses diffeomorphic transformations to simulate aging, the synthesized images are topologically consistent with the original anatomy by design. We evaluated InBrainSyn both quantitatively and qualitatively on AD and healthy control cohorts from the Open Access Series of Imaging Studies - version 3 dataset. Experimentally, InBrainSyn can also model neuroanatomical transitions between normal aging and AD. An evaluation of an external set supports its generalizability. Overall, with only a single baseline scan, InBrainSyn synthesizes realistic 3D spatiotemporal T1w MRI scans, producing personalized longitudinal aging trajectories. The code for InBrainSyn is available at: https://github.com/Fjr9516/InBrainSyn.
Authors: Emanuele Vivoli, Lorenzo Capineri, Marco Bertini
Abstract: The detection and removal of landmines is a complex and risky task that requires advanced remote sensing techniques to reduce the risk for the professionals involved in this task. In this paper, we propose a novel synthetic dataset for buried landmine detection to provide researchers with a valuable resource to observe, measure, locate, and address issues in landmine detection. The dataset consists of 41,800 microwave holographic images (2D) and their holographic inverted scans (3D) of different types of buried objects, including landmines, clutter, and pottery objects, and is collected by means of a microwave holography sensor. We evaluate the performance of several state-of-the-art deep learning models trained on our synthetic dataset for various classification tasks. While the results do not yield yet high performances, showing the difficulty of the proposed task, we believe that our dataset has significant potential to drive progress in the field of landmine detection thanks to the accuracy and resolution obtainable using holographic radars. To the best of our knowledge, our dataset is the first of its kind and will help drive further research on computer vision methods to automatize mine detection, with the overall goal of reducing the risks and the costs of the demining process.
Authors: Ziyi Zhang, Zhen Sun, Zongmin Zhang, Jihui Guo, Xinlei He
Abstract: Large Vision-Language Models (LVLMs) have become powerful and widely adopted in some practical applications. However, recent research has revealed their vulnerability to multimodal jailbreak attacks, whereby the model can be induced to generate harmful content, leading to safety risks. Although most LVLMs have undergone safety alignment, recent research shows that the visual modality is still vulnerable to jailbreak attacks. In our work, we discover that by using flowcharts with partially harmful information, LVLMs can be induced to provide additional harmful details. Based on this, we propose a jailbreak attack method based on auto-generated flowcharts, FC-Attack. Specifically, FC-Attack first fine-tunes a pre-trained LLM to create a step-description generator based on benign datasets. The generator is then used to produce step descriptions corresponding to a harmful query, which are transformed into flowcharts in 3 different shapes (vertical, horizontal, and S-shaped) as visual prompts. These flowcharts are then combined with a benign textual prompt to execute a jailbreak attack on LVLMs. Our evaluations using the Advbench dataset show that FC-Attack achieves over 90% attack success rates on Gemini-1.5, Llaval-Next, Qwen2-VL, and InternVL-2.5 models, outperforming existing LVLM jailbreak methods. Additionally, we investigate factors affecting the attack performance, including the number of steps and the font styles in the flowcharts. Our evaluation shows that FC-Attack can improve the jailbreak performance from 4% to 28% in Claude-3.5 by changing the font style. To mitigate the attack, we explore several defenses and find that AdaShield can largely reduce the jailbreak performance but with the cost of utility drop.
Authors: Chahine-Nicolas Zede, Laurent Carrafa, Val\'erie Gouet-Brunet
Abstract: Retrieval in 3D point clouds is a challenging task that consists in retrieving the most similar point clouds to a given query within a reference of 3D points. Current methods focus on comparing descriptors of point clouds in order to identify similar ones. Due to the complexity of this latter step, here we focus on the acceleration of the retrieval by adapting the Differentiable Search Index (DSI), a transformer-based approach initially designed for text information retrieval, for 3D point clouds retrieval. Our approach generates 1D identifiers based on the point descriptors, enabling direct retrieval in constant time. To adapt DSI to 3D data, we integrate Vision Transformers to map descriptors to these identifiers while incorporating positional and semantic encoding. The approach is evaluated for place recognition on a public benchmark comparing its retrieval capabilities against state-of-the-art methods, in terms of quality and speed of returned point clouds.
Authors: Christopher Wewer, Bart Pogodzinski, Bernt Schiele, Jan Eric Lenssen
Abstract: We introduce Spatial Reasoning Models (SRMs), a framework to perform reasoning over sets of continuous variables via denoising generative models. SRMs infer continuous representations on a set of unobserved variables, given observations on observed variables. Current generative models on spatial domains, such as diffusion and flow matching models, often collapse to hallucination in case of complex distributions. To measure this, we introduce a set of benchmark tasks that test the quality of complex reasoning in generative models and can quantify hallucination. The SRM framework allows to report key findings about importance of sequentialization in generation, the associated order, as well as the sampling strategies during training. It demonstrates, for the first time, that order of generation can successfully be predicted by the denoising network itself. Using these findings, we can increase the accuracy of specific reasoning tasks from <1% to >50%.
Authors: Sabine Muzellec, Andrea Alamia, Thomas Serre, Rufin VanRullen
Abstract: Neural synchrony is hypothesized to play a crucial role in how the brain organizes visual scenes into structured representations, enabling the robust encoding of multiple objects within a scene. However, current deep learning models often struggle with object binding, limiting their ability to represent multiple objects effectively. Inspired by neuroscience, we investigate whether synchrony-based mechanisms can enhance object encoding in artificial models trained for visual categorization. Specifically, we combine complex-valued representations with Kuramoto dynamics to promote phase alignment, facilitating the grouping of features belonging to the same object. We evaluate two architectures employing synchrony: a feedforward model and a recurrent model with feedback connections to refine phase synchronization using top-down information. Both models outperform their real-valued counterparts and complex-valued models without Kuramoto synchronization on tasks involving multi-object images, such as overlapping handwritten digits, noisy inputs, and out-of-distribution transformations. Our findings highlight the potential of synchrony-driven mechanisms to enhance deep learning models, improving their performance, robustness, and generalization in complex visual categorization tasks.
Authors: Yifei Xia, Suhan Ling, Fangcheng Fu, Yujie Wang, Huixia Li, Xuefeng Xiao, Bin Cui
Abstract: Generating high-fidelity long videos with Diffusion Transformers (DiTs) is often hindered by significant latency, primarily due to the computational demands of attention mechanisms. For instance, generating an 8-second 720p video (110K tokens) with HunyuanVideo takes about 600 PFLOPs, with around 500 PFLOPs consumed by attention computations. To address this issue, we propose AdaSpa, the first Dynamic Pattern and Online Precise Search sparse attention method. Firstly, to realize the Dynamic Pattern, we introduce a blockified pattern to efficiently capture the hierarchical sparsity inherent in DiTs. This is based on our observation that sparse characteristics of DiTs exhibit hierarchical and blockified structures between and within different modalities. This blockified approach significantly reduces the complexity of attention computation while maintaining high fidelity in the generated videos. Secondly, to enable Online Precise Search, we propose the Fused LSE-Cached Search with Head-adaptive Hierarchical Block Sparse Attention. This method is motivated by our finding that DiTs' sparse pattern and LSE vary w.r.t. inputs, layers, and heads, but remain invariant across denoising steps. By leveraging this invariance across denoising steps, it adapts to the dynamic nature of DiTs and allows for precise, real-time identification of sparse indices with minimal overhead. AdaSpa is implemented as an adaptive, plug-and-play solution and can be integrated seamlessly with existing DiTs, requiring neither additional fine-tuning nor a dataset-dependent profiling. Extensive experiments validate that AdaSpa delivers substantial acceleration across various models while preserving video quality, establishing itself as a robust and scalable approach to efficient video generation.
Authors: Jing-Yuan Chang
Abstract: Badminton, known for having the fastest ball speeds among all sports, presents significant challenges to the field of computer vision, including player identification, court line detection, shuttlecock trajectory tracking, and player stroke-type classification. In this paper, we introduce a novel video segmentation strategy to extract frames of each player's racket swing in a badminton broadcast match. These segmented frames are then processed by two existing models: one for Human Pose Estimation to obtain player skeletal joints, and the other for shuttlecock trajectory detection to extract shuttlecock trajectories. Leveraging these joints, trajectories, and player positions as inputs, we propose Badminton Stroke-type Transformer (BST) to classify player stroke-types in singles. To the best of our knowledge, experimental results demonstrate that our method outperforms the previous state-of-the-art on the largest publicly available badminton video dataset, ShuttleSet, which shows that effectively leveraging ball trajectory is likely to be a trend for racket sports action recognition.
Authors: Jingqiu Zhou, Lue Fan, Linjiang Huang, Xiaoyu Shi, Si Liu, Zhaoxiang Zhang, Hongsheng Li
Abstract: Driving scene reconstruction and rendering have advanced significantly using the 3D Gaussian Splatting. However, most prior research has focused on the rendering quality along a pre-recorded vehicle path and struggles to generalize to out-of-path viewpoints, which is caused by the lack of high-quality supervision in those out-of-path views. To address this issue, we introduce an Inverse View Warping technique to create compact and high-quality images as supervision for the reconstruction of the out-of-path views, enabling high-quality rendering results for those views. For accurate and robust inverse view warping, a depth bootstrap strategy is proposed to obtain on-the-fly dense depth maps during the optimization process, overcoming the sparsity and incompleteness of LiDAR depth data. Our method achieves superior in-path and out-of-path reconstruction and rendering performance on the widely used Waymo Open dataset. In addition, a simulator-based benchmark is proposed to obtain the out-of-path ground truth and quantitatively evaluate the performance of out-of-path rendering, where our method outperforms previous methods by a significant margin.
Authors: Yunfan Lu, Xiaogang Xu, Hao Lu, Yanlin Qian, Pengteng Li, Huizai Yao, Bin Yang, Junyi Li, Qianyi Cai, Weiyu Guo, Hui Xiong
Abstract: Event cameras, with a high dynamic range exceeding $120dB$, significantly outperform traditional embedded cameras, robustly recording detailed changing information under various lighting conditions, including both low- and high-light situations. However, recent research on utilizing event data has primarily focused on low-light image enhancement, neglecting image enhancement and brightness adjustment across a broader range of lighting conditions, such as normal or high illumination. Based on this, we propose a novel research question: how to employ events to enhance and adaptively adjust the brightness of images captured under broad lighting conditions? To investigate this question, we first collected a new dataset, SEE-600K, consisting of 610,126 images and corresponding events across 202 scenarios, each featuring an average of four lighting conditions with over a 1000-fold variation in illumination. Subsequently, we propose a framework that effectively utilizes events to smoothly adjust image brightness through the use of prompts. Our framework captures color through sensor patterns, uses cross-attention to model events as a brightness dictionary, and adjusts the image's dynamic range to form a broad light-range representation (BLR), which is then decoded at the pixel level based on the brightness prompt. Experimental results demonstrate that our method not only performs well on the low-light enhancement dataset but also shows robust performance on broader light-range image enhancement using the SEE-600K dataset. Additionally, our approach enables pixel-level brightness adjustment, providing flexibility for post-processing and inspiring more imaging applications. The dataset and source code are publicly available at:https://github.com/yunfanLu/SEE.
Authors: Jiuyang Dong, Junjun Jiang, Kui Jiang, Jiahan Li, Yongbing Zhang
Abstract: Although multi-instance learning (MIL) has succeeded in pathological image classification, it faces the challenge of high inference costs due to processing numerous patches from gigapixel whole slide images (WSIs). To address this, we propose HDMIL, a hierarchical distillation multi-instance learning framework that achieves fast and accurate classification by eliminating irrelevant patches. HDMIL consists of two key components: the dynamic multi-instance network (DMIN) and the lightweight instance pre-screening network (LIPN). DMIN operates on high-resolution WSIs, while LIPN operates on the corresponding low-resolution counterparts. During training, DMIN are trained for WSI classification while generating attention-score-based masks that indicate irrelevant patches. These masks then guide the training of LIPN to predict the relevance of each low-resolution patch. During testing, LIPN first determines the useful regions within low-resolution WSIs, which indirectly enables us to eliminate irrelevant regions in high-resolution WSIs, thereby reducing inference time without causing performance degradation. In addition, we further design the first Chebyshev-polynomials-based Kolmogorov-Arnold classifier in computational pathology, which enhances the performance of HDMIL through learnable activation layers. Extensive experiments on three public datasets demonstrate that HDMIL outperforms previous state-of-the-art methods, e.g., achieving improvements of 3.13% in AUC while reducing inference time by 28.6% on the Camelyon16 dataset.
Authors: Zineb Sordo, Eric Chagnon, Daniela Ushizima
Abstract: This review surveys the state-of-the-art in text-to-image and image-to-image generation within the scope of generative AI. We provide a comparative analysis of three prominent architectures: Variational Autoencoders, Generative Adversarial Networks and Diffusion Models. For each, we elucidate core concepts, architectural innovations, and practical strengths and limitations, particularly for scientific image understanding. Finally, we discuss critical open challenges and potential future research directions in this rapidly evolving field.
Authors: Yuheng Jia, Wesley Armour
Abstract: Visible-Infrared Person Re-Identification (VI-ReID) plays a crucial role in applications such as search and rescue, infrastructure protection, and nighttime surveillance. However, it faces significant challenges due to modality discrepancies, varying illumination, and frequent occlusions. To overcome these obstacles, we propose \textbf{AMINet}, an Adaptive Modality Interaction Network. AMINet employs multi-granularity feature extraction to capture comprehensive identity attributes from both full-body and upper-body images, improving robustness against occlusions and background clutter. The model integrates an interactive feature fusion strategy for deep intra-modal and cross-modal alignment, enhancing generalization and effectively bridging the RGB-IR modality gap. Furthermore, AMINet utilizes phase congruency for robust, illumination-invariant feature extraction and incorporates an adaptive multi-scale kernel MMD to align feature distributions across varying scales. Extensive experiments on benchmark datasets demonstrate the effectiveness of our approach, achieving a Rank-1 accuracy of $74.75\%$ on SYSU-MM01, surpassing the baseline by $7.93\%$ and outperforming the current state-of-the-art by $3.95\%$.
Authors: Martina Finocchiaro, Ronja Stern, Abraham George Smith, Jens Petersen, Kenny Erleben, Melanie Ganz
Abstract: High-resolution colon segmentation is crucial for clinical and research applications, such as digital twins and personalized medicine. However, the leading open-source abdominal segmentation tool, TotalSegmentator, struggles with accuracy for the colon, which has a complex and variable shape, requiring time-intensive labeling. Here, we present the first fully automatic high-resolution colon segmentation method. To develop it, we first created a high resolution colon dataset using a pipeline that combines region growing with interactive machine learning to efficiently and accurately label the colon on CT colonography (CTC) images. Based on the generated dataset consisting of 435 labeled CTC images we trained an nnU-Net model for fully automatic colon segmentation. Our fully automatic model achieved an average symmetric surface distance of 0.2 mm (vs. 4.0 mm from TotalSegmentator) and a 95th percentile Hausdorff distance of 1.0 mm (vs. 18 mm from TotalSegmentator). Our segmentation accuracy substantially surpasses TotalSegmentator. We share our trained model and pipeline code, providing the first and only open-source tool for high-resolution colon segmentation. Additionally, we created a large-scale dataset of publicly available high-resolution colon labels.
Authors: Zihan Huang, Xinyu Shi, Zecheng Hao, Tong Bu, Jianhao Ding, Zhaofei Yu, Tiejun Huang
Abstract: Spiking neural networks (SNNs) show great potential due to their energy efficiency, fast processing capabilities, and robustness. There are two main approaches to constructing SNNs. Direct training methods require much memory, while conversion methods offer a simpler and more efficient option. However, current conversion methods mainly focus on converting convolutional neural networks (CNNs) to SNNs. Converting Transformers to SNN is challenging because of the presence of non-linear modules. In this paper, we propose an Expectation Compensation Module to preserve the accuracy of the conversion. The core idea is to use information from the previous T time-steps to calculate the expected output at time-step T. We also propose a Multi-Threshold Neuron and the corresponding Parallel Parameter normalization to address the challenge of large time steps needed for high accuracy, aiming to reduce network latency and power consumption. Our experimental results demonstrate that our approach achieves state-of-the-art performance. For example, we achieve a top-1 accuracy of 88.60\% with only a 1\% loss in accuracy using 4 time steps while consuming only 35\% of the original power of the Transformer. To our knowledge, this is the first successful Artificial Neural Network (ANN) to SNN conversion for Spiking Transformers that achieves high accuracy, low latency, and low power consumption on complex datasets. The source codes of the proposed method are available at https://github.com/h-z-h-cell/Transformer-to-SNN-ECMT.
URLs: https://github.com/h-z-h-cell/Transformer-to-SNN-ECMT.
Authors: Otto Brookes, Maksim Kukushkin, Majid Mirmehdi, Colleen Stephens, Paula Dieguez, Thurston C. Hicks, Sorrel Jones, Kevin Lee, Maureen S. McCarthy, Amelia Meier, Emmanuelle Normand, Erin G. Wessling, Roman M. Wittig, Kevin Langergraber, Klaus Zuberb\"uhler, Lukas Boesch, Thomas Schmid, Mimi Arandjelovic, Hjalmar K\"uhl, Tilo Burghardt
Abstract: Computer vision analysis of camera trap video footage is essential for wildlife conservation, as captured behaviours offer some of the earliest indicators of changes in population health. Recently, several high-impact animal behaviour datasets and methods have been introduced to encourage their use; however, the role of behaviour-correlated background information and its significant effect on out-of-distribution generalisation remain unexplored. In response, we present the PanAf-FGBG dataset, featuring 20 hours of wild chimpanzee behaviours, recorded at over 350 individual camera locations. Uniquely, it pairs every video with a chimpanzee (referred to as a foreground video) with a corresponding background video (with no chimpanzee) from the same camera location. We present two views of the dataset: one with overlapping camera locations and one with disjoint locations. This setup enables, for the first time, direct evaluation of in-distribution and out-of-distribution conditions, and for the impact of backgrounds on behaviour recognition models to be quantified. All clips come with rich behavioural annotations and metadata including unique camera IDs and detailed textual scene descriptions. Additionally, we establish several baselines and present a highly effective latent-space normalisation technique that boosts out-of-distribution performance by +5.42% mAP for convolutional and +3.75% mAP for transformer-based models. Finally, we provide an in-depth analysis on the role of backgrounds in out-of-distribution behaviour recognition, including the so far unexplored impact of background durations (i.e., the count of background frames within foreground videos).
Authors: Maria Koshkina, James H. Elder
Abstract: In team sports analytics, long-term player tracking remains a challenging task due to player appearance similarity, occlusion, and dynamic motion patterns. Accurately re-identifying players and reconnecting tracklets after extended absences from the field of view or prolonged occlusions is crucial for robust analysis. We introduce SportsSUSHI, a hierarchical graph-based approach that leverages domain-specific features, including jersey numbers, team IDs, and field coordinates, to enhance tracking accuracy. SportsSUSHI achieves high performance on the SoccerNet dataset and a newly proposed hockey tracking dataset. Our hockey dataset, recorded using a stationary camera capturing the entire playing surface, contains long sequences and annotations for team IDs and jersey numbers, making it well-suited for evaluating long-term tracking capabilities. The inclusion of domain-specific features in our approach significantly improves association accuracy, as demonstrated in our experiments. The dataset and code are available at https://github.com/mkoshkina/sports-SUSHI.
Authors: Alberto Mario Ceballos-Arroyo, Jisoo Kim, Chu-Hsuan Lin, Lei Qin, Geoffrey S. Young, Huaizu Jiang
Abstract: Intracranial aneurysms are a major cause of morbidity and mortality worldwide, and detecting them manually is a complex, time-consuming task. Albeit automated solutions are desirable, the limited availability of training data makes it difficult to develop such solutions using typical supervised learning frameworks. In this work, we propose a novel pre-training strategy using more widely available unannotated head CT scan data to pre-train a 3D Vision Transformer model prior to fine-tuning for the aneurysm detection task. Specifically, we modify masked auto-encoder (MAE) pre-training in the following ways: we use a factorized self-attention mechanism to make 3D attention computationally viable, we restrict the masked patches to areas near arteries to focus on areas where aneurysms are likely to occur, and we reconstruct not only CT scan intensity values but also artery distance maps, which describe the distance between each voxel and the closest artery, thereby enhancing the backbone's learned representations. Compared with SOTA aneurysm detection models, our approach gains +4-8% absolute Sensitivity at a false positive rate of 0.5. Code and weights will be released.
Authors: Nita Mulliqi (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Anders Blilie (Department of Pathology, Stavanger University Hospital, Stavanger, Norway,Faculty of Health Sciences, University of Stavanger, Stavanger, Norway), Xiaoyi Ji (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Kelvin Szolnoky (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Henrik Olsson (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Sol Erika Boman (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden,Department of Molecular Medicine,Surgery, Karolinska Institutet, Stockholm, Sweden), Matteo Titus (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Geraldine Martinez Gonzalez (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Julia Anna Mielcarz (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Masi Valkonen (Institute of Biomedicine, University of Turku, Turku, Finland), Einar Gudlaugsson (Department of Pathology, Stavanger University Hospital, Stavanger, Norway), Svein R. Kjosavik (The General Practice,Care Coordination Research Group, Stavanger University Hospital, Norway,Department of Global Public Health,Primary Care, Faculty of Medicine, University of Bergen, Norway), Jos\'e Asenjo (Department of Pathology, Synlab, Madrid, Spain), Marcello Gambacorta (Department of Pathology, Synlab, Brescia, Italy), Paolo Libretti (Department of Pathology, Synlab, Brescia, Italy), Marcin Braun (Department of Pathology, Chair of Oncology, Medical University of Lodz, Lodz, Poland), Radzislaw Kordek (Department of Pathology, Chair of Oncology, Medical University of Lodz, Lodz, Poland), Roman {\L}owicki (1st Department of Urology, Medical University of Lodz, Lodz, Poland), Kristina Hotakainen (Department of Clinical Chemistry,Hematology, University of Helsinki, Helsinki, Finland,Laboratory Services, Mehil\"ainen Oy, Helsinki, Finland), P\"aivi V\"are (Department of Pathology, Mehil\"ainen L\"ansi-Pohja Hospital, Kemi, Finland), Bodil Ginnerup Pedersen (Department of Radiology, Aarhus University Hospital, Aarhus, Denmark,Department of Clinical Medicine, Aarhus University, Aarhus, Denmark), Karina Dalsgaard S{\o}rensen (Department of Clinical Medicine, Aarhus University, Aarhus, Denmark,Department of Molecular Medicine, Aarhus University Hospital, Aarhus, Denmark), Benedicte Parm Ulh{\o}i (Department of Pathology, Aarhus University Hospital, Aarhus, Denmark), Pekka Ruusuvuori (Institute of Biomedicine, University of Turku, Turku, Finland,InFLAMES Research Flagship, University of Turku, Turku, Finland,Faculty of Medicine,Health Technology, Tampere University, Tampere, Finland), Brett Delahunt (Malaghan Institute of Medical Research, Wellington, New Zealand,Department of Oncology,Pathology, Karolinska Institutet, Stockholm, Sweden), Hemamali Samaratunga (Aquesta Uropathology,University of Queensland, QLD, Brisbane, Australia), Toyonori Tsuzuki (Department of Surgical Pathology, School of Medicine, Aichi Medical University, Nagoya, Japan), Emilius A. M. Janssen (Department of Pathology, Stavanger University Hospital, Stavanger, Norway,Department of Chemistry, Bioscience,Environmental Engineering, University of Stavanger, Stavanger, Norway,Institute for Biomedicine,Glycomics, Griffith University, Queensland, Australia), Lars Egevad (Department of Oncology,Pathology, Karolinska Institutet, Stockholm, Sweden), Martin Eklund (Department of Medical Epidemiology,Biostatistics, Karolinska Institutet, Stockholm, Sweden), Kimmo Kartasalo (Department of Medical Epidemiology,Biostatistics, SciLifeLab, Karolinska Institutet, Stockholm, Sweden)
Abstract: The role of artificial intelligence (AI) in pathology has evolved from aiding diagnostics to uncovering predictive morphological patterns in whole slide images (WSIs). Recently, foundation models (FMs) leveraging self-supervised pre-training have been widely advocated as a universal solution for diverse downstream tasks. However, open questions remain about their clinical applicability and generalization advantages over end-to-end learning using task-specific (TS) models. Here, we focused on AI with clinical-grade performance for prostate cancer diagnosis and Gleason grading. We present the largest validation of AI for this task, using over 100,000 core needle biopsies from 7,342 patients across 15 sites in 11 countries. We compared two FMs with a fully end-to-end TS model in a multiple instance learning framework. Our findings challenge assumptions that FMs universally outperform TS models. While FMs demonstrated utility in data-scarce scenarios, their performance converged with - and was in some cases surpassed by - TS models when sufficient labeled training data were available. Notably, extensive task-specific training markedly reduced clinically significant misgrading, misdiagnosis of challenging morphologies, and variability across different WSI scanners. Additionally, FMs used up to 35 times more energy than the TS model, raising concerns about their sustainability. Our results underscore that while FMs offer clear advantages for rapid prototyping and research, their role as a universal solution for clinically applicable medical AI remains uncertain. For high-stakes clinical applications, rigorous validation and consideration of task-specific training remain critically important. We advocate for integrating the strengths of FMs and end-to-end learning to achieve robust and resource-efficient AI pathology solutions fit for clinical use.
Authors: Xi Tang, Jihao Qiu, Lingxi Xie, Yunjie Tian, Jianbin Jiao, Qixiang Ye
Abstract: Multimodal large language models (MLLMs) have enabled open-world visual understanding by injecting visual input as extra tokens into large language models (LLMs) as contexts. However, when the visual input changes from a single image to a long video, the above paradigm encounters difficulty because the vast amount of video tokens has significantly exceeded the maximal capacity of MLLMs. Therefore, existing video-based MLLMs are mostly established upon sampling a small portion of tokens from input data, which can cause key information to be lost and thus produce incorrect answers. This paper presents a simple yet effective algorithm named Adaptive Keyframe Sampling (AKS). It inserts a plug-and-play module known as keyframe selection, which aims to maximize the useful information with a fixed number of video tokens. We formulate keyframe selection as an optimization involving (1) the relevance between the keyframes and the prompt, and (2) the coverage of the keyframes over the video, and present an adaptive algorithm to approximate the best solution. Experiments on two long video understanding benchmarks validate that Adaptive Keyframe Sampling improves video QA accuracy (beyond strong baselines) upon selecting informative keyframes. Our study reveals the importance of information pre-filtering in video-based MLLMs. Code is available at https://github.com/ncTimTang/AKS.
Authors: Sherlon Almeida da Silva, Davi Geiger, Luiz Velho, Moacir Antonelli Ponti
Abstract: We innovate in stereo vision by explicitly providing analytical 3D surface models as viewed by a cyclopean eye model that incorporate depth discontinuities and occlusions. This geometrical foundation combined with learned stereo features allows our system to benefit from the strengths of both approaches. We also invoke a prior monocular model of surfaces to fill in occlusion regions or texture-less regions where data matching is not sufficient. Our results already are on par with the state-of-the-art purely data-driven methods and are of much better visual quality, emphasizing the importance of the 3D geometrical model to capture critical visual information. Such qualitative improvements may find applicability in virtual reality, for a better human experience, as well as in robotics, for reducing critical errors. Our approach aims to demonstrate that understanding and modeling geometrical properties of 3D surfaces is beneficial to computer vision research.
Authors: Xueyun Tian, Wei Li, Bingbing Xu, Yige Yuan, Yuanzhuo Wang, Huawei Shen
Abstract: Despite significant progress in diffusion-based image generation, subject-driven generation and instruction-based editing remain challenging. Existing methods typically treat them separately, struggling with limited high-quality data and poor generalization. However, both tasks require capturing complex visual variations while maintaining consistency between inputs and outputs. Therefore, we propose MIGE, a unified framework that standardizes task representations using multimodal instructions. It treats subject-driven generation as creation on a blank canvas and instruction-based editing as modification of an existing image, establishing a shared input-output formulation. MIGE introduces a novel multimodal encoder that maps free-form multimodal instructions into a unified vision-language space, integrating visual and semantic features through a feature fusion mechanism.This unification enables joint training of both tasks, providing two key advantages: (1) Cross-Task Enhancement: By leveraging shared visual and semantic representations, joint training improves instruction adherence and visual consistency in both subject-driven generation and instruction-based editing. (2) Generalization: Learning in a unified format facilitates cross-task knowledge transfer, enabling MIGE to generalize to novel compositional tasks, including instruction-based subject-driven editing. Experiments show that MIGE excels in both subject-driven generation and instruction-based editing while setting a state-of-the-art in the new task of instruction-based subject-driven editing. Code and model have been publicly available at https://github.com/Eureka-Maggie/MIGE.
Authors: Abhishek Jha, Tinne Tuytelaars, Yuki M. Asano
Abstract: Following the success in NLP, the best vision models are now in the billion parameter ranges. Adapting these large models to a target distribution has become computationally and economically prohibitive. Addressing this challenge, we introduce UpStep, an Unsupervised Parameter-efficient Source-free post-pretraining approach, designed to efficiently adapt a base model from a source domain to a target domain: i) we design a self-supervised training scheme to adapt a pretrained model on an unlabeled target domain in a setting where source domain data is unavailable. Such source-free setting comes with the risk of catastrophic forgetting, hence, ii) we propose center vector regularization (CVR), a set of auxiliary operations that minimize catastrophic forgetting and additionally reduces the computational cost by skipping backpropagation in 50\% of the training iterations. Finally iii) we perform this adaptation process in a parameter-efficient way by adapting the pretrained model through low-rank adaptation methods, resulting in a fraction of parameters to optimize. We utilize various general backbone architectures, both supervised and unsupervised, trained on Imagenet as our base model and adapt them to a diverse set of eight target domains demonstrating the adaptability and generalizability of our proposed approach.
Authors: Zhiyu Tan, Junyan Wang, Hao Yang, Luozheng Qin, Hesen Chen, Qiang Zhou, Hao Li
Abstract: Text-to-video generation has demonstrated promising progress with the advent of diffusion models, yet existing approaches are limited by dataset quality and computational resources. To address these limitations, this paper presents a comprehensive approach that advances both data curation and model design. We introduce CFC-VIDS-1M, a high-quality video dataset constructed through a systematic coarse-to-fine curation pipeline. The pipeline first evaluates video quality across multiple dimensions, followed by a fine-grained stage that leverages vision-language models to enhance text-video alignment and semantic richness. Building upon the curated dataset's emphasis on visual quality and temporal coherence, we develop RACCOON, a transformer-based architecture with decoupled spatial-temporal attention mechanisms. The model is trained through a progressive four-stage strategy designed to efficiently handle the complexities of video generation. Extensive experiments demonstrate that our integrated approach of high-quality data curation and efficient training strategy generates visually appealing and temporally coherent videos while maintaining computational efficiency. We will release our dataset, code, and models.
Authors: L. Degeorge, A. Ghosh, N. Dufour, D. Picard, V. Kalogeiton
Abstract: Recent text-to-image (T2I) generation models have achieved remarkable results by training on billion-scale datasets, following a `bigger is better' paradigm that prioritizes data quantity over quality. We challenge this established paradigm by demonstrating that strategic data augmentation of small, well-curated datasets can match or outperform models trained on massive web-scraped collections. Using only ImageNet enhanced with well-designed text and image augmentations, we achieve a +2 overall score over SD-XL on GenEval and +5 on DPGBench while using just 1/10th the parameters and 1/1000th the training images. Our results suggest that strategic data augmentation, rather than massive datasets, could offer a more sustainable path forward for T2I generation.
Authors: Jingtao Zhan, Jiahao Zhao, Jiayu Li, Yiqun Liu, Bo Zhang, Qingyao Ai, Jiaxin Mao, Hongning Wang, Min Zhang, Shaoping Ma
Abstract: How does intelligence emerge? We propose that intelligence is not a sudden gift or random occurrence, but rather a necessary trait for species to survive through Natural Selection. If a species passes the test of Natural Selection, it demonstrates the intelligence to survive in nature. Extending this perspective, we introduce Intelligence Test, a method to quantify the intelligence of any subject on any task. Like how species evolve by trial and error, Intelligence Test quantifies intelligence by the number of failed attempts before success. Fewer failures correspond to higher intelligence. When the expectation and variance of failure counts are both finite, it signals the achievement of an autonomous level of intelligence. Using Intelligence Test, we comprehensively evaluate existing AI systems. Our results show that while AI systems achieve a level of autonomy in simple tasks, they are still far from autonomous in more complex tasks, such as vision, search, recommendation, and language. While scaling model size might help, this would come at an astronomical cost. Projections suggest that achieving general autonomy would require unimaginable $10^{26}$ parameters. Even if Moore's Law continuously holds, such a parameter scale would take $70$ years. This staggering cost highlights the complexity of human tasks and the inadequacies of current AI. To further understand this phenomenon, we conduct a theoretical analysis. Our simulations suggest that human tasks possess a criticality property. As a result, autonomy requires a deep understanding of the task's underlying mechanisms. Current AI, however, does not fully grasp these mechanisms and instead relies on superficial mimicry, making it difficult to reach an autonomous level. We believe Intelligence Test can not only guide the future development of AI but also offer profound insights into the intelligence of humans ourselves.
Authors: Shaharukh Khan, Ayush Tarun, Ali Faraz, Palash Kamble, Vivek Dahiya, Praveen Pokala, Ashish Kulkarni, Chandra Khatri, Abhinav Ravi, Shubham Agarwal
Abstract: In this work, we provide the system description of our submission as part of the English to Lowres Multimodal Translation Task at the Workshop on Asian Translation (WAT2024). We introduce Chitranuvad, a multimodal model that effectively integrates Multilingual LLM and a vision module for Multimodal Translation. Our method uses a ViT image encoder to extract visual representations as visual token embeddings which are projected to the LLM space by an adapter layer and generates translation in an autoregressive fashion. We participated in all the three tracks (Image Captioning, Text only and Multimodal translation tasks) for Indic languages (ie. English translation to Hindi, Bengali and Malyalam) and achieved SOTA results for Hindi in all of them on the Challenge set while remaining competitive for the other languages in the shared task.
Authors: Julius Broomfield, Kartik Sharma, Srijan Kumar
Abstract: Large language models (LLMs) have recently demonstrated remarkable advancements in embodying diverse personas, enhancing their effectiveness as conversational agents and virtual assistants. Consequently, LLMs have made significant strides in processing and integrating multimodal information. However, even though human personas can be expressed in both text and image, the extent to which the modality of a persona impacts the embodiment by the LLM remains largely unexplored. In this paper, we investigate how do different modalities influence the expressiveness of personas in multimodal LLMs. To this end, we create a novel modality-parallel dataset of 40 diverse personas varying in age, gender, occupation, and location. This consists of four modalities to equivalently represent a persona: image-only, text-only, a combination of image and small text, and typographical images, where text is visually stylized to convey persona-related attributes. We then create a systematic evaluation framework with 60 questions and corresponding metrics to assess how well LLMs embody each persona across its attributes and scenarios. Comprehensive experiments on $5$ multimodal LLMs show that personas represented by detailed text show more linguistic habits, while typographical images often show more consistency with the persona. Our results reveal that LLMs often overlook persona-specific details conveyed through images, highlighting underlying limitations and paving the way for future research to bridge this gap. We release the data and code at https://github.com/claws-lab/persona-modality .
Authors: Zhuohang Li, Chao Yan, Nicholas J. Jackson, Wendi Cui, Bo Li, Jiaxin Zhang, Bradley A. Malin
Abstract: Advancements in Large Vision-Language Models (LVLMs) have demonstrated promising performance in a variety of vision-language tasks involving image-conditioned free-form text generation. However, growing concerns about hallucinations in LVLMs, where the generated text is inconsistent with the visual context, are becoming a major impediment to deploying these models in applications that demand guaranteed reliability. In this paper, we introduce a framework to address this challenge, ConfLVLM, which is grounded on conformal prediction to achieve finite-sample distribution-free statistical guarantees on the factuality of LVLM output. This framework treats an LVLM as a hypothesis generator, where each generated text detail (or claim) is considered an individual hypothesis. It then applies a statistical hypothesis testing procedure to verify each claim using efficient heuristic uncertainty measures to filter out unreliable claims before returning any responses to users. We conduct extensive experiments covering three representative application domains, including general scene understanding, medical radiology report generation, and document understanding. Remarkably, ConfLVLM reduces the error rate of claims generated by LLaVa-1.5 for scene descriptions from 87.8\% to 10.0\% by filtering out erroneous claims with a 95.3\% true positive rate. Our results further demonstrate that ConfLVLM is highly flexible, and can be applied to any black-box LVLMs paired with any uncertainty measure for any image-conditioned free-form text generation task while providing a rigorous guarantee on controlling the risk of hallucination.
Authors: Sajjad Saleem, Muhammad Imran Sharif
Abstract: The lungs are the essential organs of respiration, and this system is significant in the carbon dioxide and exchange between oxygen that occurs in human life. However, several lung diseases, which include pneumonia, tuberculosis, COVID-19, and lung cancer, are serious healthiness challenges and demand early and precise diagnostics. The methodological study has proposed a new deep learning framework called NASNet-ViT, which effectively incorporates the convolution capability of NASNet with the global attention mechanism capability of Vision Transformer ViT. The proposed model will classify the lung conditions into five classes: Lung cancer, COVID-19, pneumonia, TB, and normal. A sophisticated multi-faceted preprocessing strategy called MixProcessing has been used to improve diagnostic accuracy. This preprocessing combines wavelet transform, adaptive histogram equalization, and morphological filtering techniques. The NASNet-ViT model performs at state of the art, achieving an accuracy of 98.9%, sensitivity of 0.99, an F1-score of 0.989, and specificity of 0.987, outperforming other state of the art architectures such as MixNet-LD, D-ResNet, MobileNet, and ResNet50. The model's efficiency is further emphasized by its compact size, 25.6 MB, and a low computational time of 12.4 seconds, hence suitable for real-time, clinically constrained environments. These results reflect the high-quality capability of NASNet-ViT in extracting meaningful features and recognizing various types of lung diseases with very high accuracy. This work contributes to medical image analysis by providing a robust and scalable solution for diagnostics in lung diseases.
Authors: Vicente Balmaseda, Bokun Wang, Ching-Long Lin, Tianbao Yang
Abstract: In self-supervised contrastive learning, negative pairs are typically constructed using an anchor image and a sample drawn from the entire dataset, excluding the anchor. However, this approach can result in the creation of negative pairs with similar semantics, referred to as "false negatives", leading to their embeddings being falsely pushed apart. To address this issue, we introduce GloFND, an optimization-based approach that automatically learns on the fly the threshold for each anchor data to identify its false negatives during training. In contrast to previous methods for false negative discovery, our approach globally detects false negatives across the entire dataset rather than locally within the mini-batch. Moreover, its per-iteration computation cost remains independent of the dataset size. Experimental results on image and image-text data demonstrate the effectiveness of the proposed method. Our implementation is available at https://github.com/vibalcam/GloFND .
Authors: Zhiqiang Shen, Peng Cao, Jinzhu Yang, Osmar R. Zaiane, Zhaolin Chen
Abstract: Due to the domain shifts between training and testing medical images, learned segmentation models often experience significant performance degradation during deployment. In this paper, we first decompose an image into its style code and content map and reveal that domain shifts in medical images involve: \textbf{style shifts} (\emph{i.e.}, differences in image appearance) and \textbf{content shifts} (\emph{i.e.}, variations in anatomical structures), the latter of which has been largely overlooked. To this end, we propose \textbf{StyCona}, a \textbf{sty}le \textbf{con}tent decomposition-based data \textbf{a}ugmentation method that innovatively augments both image style and content within the rank-one space, for domain generalizable medical image segmentation. StyCona is a simple yet effective plug-and-play module that substantially improves model generalization without requiring additional training parameters or modifications to the segmentation model architecture. Experiments on cross-sequence, cross-center, and cross-modality medical image segmentation settings with increasingly severe domain shifts, demonstrate the effectiveness of StyCona and its superiority over state-of-the-arts. The code is available at https://github.com/Senyh/StyCona.
Authors: Yuepeng Hu, Zhengyuan Jiang, Neil Zhenqiang Gong
Abstract: Text-to-image models can generate harmful images when presented with unsafe prompts, posing significant safety and societal risks. Alignment methods aim to modify these models to ensure they generate only non-harmful images, even when exposed to unsafe prompts. A typical text-to-image model comprises two main components: 1) a text encoder and 2) a diffusion module. Existing alignment methods mainly focus on modifying the diffusion module to prevent harmful image generation. However, this often significantly impacts the model's behavior for safe prompts, causing substantial quality degradation of generated images. In this work, we propose SafeText, a novel alignment method that fine-tunes the text encoder rather than the diffusion module. By adjusting the text encoder, SafeText significantly alters the embedding vectors for unsafe prompts, while minimally affecting those for safe prompts. As a result, the diffusion module generates non-harmful images for unsafe prompts while preserving the quality of images for safe prompts. We evaluate SafeText on multiple datasets of safe and unsafe prompts, including those generated through jailbreak attacks. Our results show that SafeText effectively prevents harmful image generation with minor impact on the images for safe prompts, and SafeText outperforms six existing alignment methods. We will publish our code and data after paper acceptance.
Authors: Changyeon Kim, Minho Heo, Doohyun Lee, Jinwoo Shin, Honglak Lee, Joseph J. Lim, Kimin Lee
Abstract: Reinforcement Learning (RL) agents have demonstrated their potential across various robotic tasks. However, they still heavily rely on human-engineered reward functions, requiring extensive trial-and-error and access to target behavior information, often unavailable in real-world settings. This paper introduces REDS: REward learning from Demonstration with Segmentations, a novel reward learning framework that leverages action-free videos with minimal supervision. Specifically, REDS employs video demonstrations segmented into subtasks from diverse sources and treats these segments as ground-truth rewards. We train a dense reward function conditioned on video segments and their corresponding subtasks to ensure alignment with ground-truth reward signals by minimizing the Equivalent-Policy Invariant Comparison distance. Additionally, we employ contrastive learning objectives to align video representations with subtasks, ensuring precise subtask inference during online interactions. Our experiments show that REDS significantly outperforms baseline methods on complex robotic manipulation tasks in Meta-World and more challenging real-world tasks, such as furniture assembly in FurnitureBench, with minimal human intervention. Moreover, REDS facilitates generalization to unseen tasks and robot embodiments, highlighting its potential for scalable deployment in diverse environments.
Authors: Yichi Zhang, Bohao Lv, Le Xue, Wenbo Zhang, Yuchen Liu, Yu Fu, Yuan Cheng, Yuan Qi
Abstract: Deep learning-based medical image segmentation typically requires large amount of labeled data for training, making it less applicable in clinical settings due to high annotation cost. Semi-supervised learning (SSL) has emerged as an appealing strategy due to its less dependence on acquiring abundant annotations from experts compared to fully supervised methods. Beyond existing model-centric advancements of SSL by designing novel regularization strategies, we anticipate a paradigmatic shift due to the emergence of promptable segmentation foundation models with universal segmentation capabilities using positional prompts represented by Segment Anything Model (SAM). In this paper, we present SemiSAM+, a foundation model-driven SSL framework to efficiently learn from limited labeled data for medical image segmentation. SemiSAM+ consists of one or multiple promptable foundation models as generalist models, and a trainable task-specific segmentation model as specialist model. For a given new segmentation task, the training is based on the specialist-generalist collaborative learning procedure, where the trainable specialist model delivers positional prompts to interact with the frozen generalist models to acquire pseudo-labels, and then the generalist model output provides the specialist model with informative and efficient supervision which benefits the automatic segmentation and prompt generation in turn. Extensive experiments on two public datasets and one in-house clinical dataset demonstrate that SemiSAM+ achieves significant performance improvement, especially under extremely limited annotation scenarios, and shows strong efficiency as a plug-and-play strategy that can be easily adapted to different specialist and generalist models.
Authors: Zhaoyang Jia, Bin Li, Jiahao Li, Wenxuan Xie, Linfeng Qi, Houqiang Li, Yan Lu
Abstract: We introduce a practical real-time neural video codec (NVC) designed to deliver high compression ratio, low latency and broad versatility. In practice, the coding speed of NVCs depends on 1) computational costs, and 2) non-computational operational costs, such as memory I/O and the number of function calls. While most efficient NVCs prioritize reducing computational cost, we identify operational cost as the primary bottleneck to achieving higher coding speed. Leveraging this insight, we introduce a set of efficiency-driven design improvements focused on minimizing operational costs. Specifically, we employ implicit temporal modeling to eliminate complex explicit motion modules, and use single low-resolution latent representations rather than progressive downsampling. These innovations significantly accelerate NVC without sacrificing compression quality. Additionally, we implement model integerization for consistent cross-device coding and a module-bank-based rate control scheme to improve practical adaptability. Experiments show our proposed DCVC-RT achieves an impressive average encoding/decoding speed at 125.2/112.8 fps (frames per second) for 1080p video, while saving an average of 21% in bitrate compared to H.266/VTM. The code is available at https://github.com/microsoft/DCVC.
Authors: Qiao Yan, Yuchen Yuan, Xiaowei Hu, Yihan Wang, Jiaqi Xu, Jinpeng Li, Chi-Wing Fu, Pheng-Ann Heng
Abstract: The increasing use of vision-language models (VLMs) in healthcare applications presents great challenges related to hallucinations, in which the models may generate seemingly plausible results that are in fact incorrect. Such hallucinations can jeopardize clinical decision making, potentially harming the diagnosis and treatments. In this work, we propose MedHallTune, a large-scale benchmark designed specifically to evaluate and mitigate hallucinations in medical VLMs. Comprising over 100,000 images and 1,000,000 instruction pairs, MedHallTune includes both hallucination and non-hallucination samples, each with ground-truth annotations. We conduct a comprehensive evaluation of current medical and general VLMs using MedHallTune, assessing their performance across key metrics, including clinical accuracy, relevance, detail level, and risk level. The experimental results show that fine-tuning with MedHallTune successfully improves the ability of several existing models to manage hallucinations and boost their zero-shot performance on downstream visual-question-answering (VQA) tasks, making them more reliable for practical medical applications. Our work contributes to the development of more trustworthy VLMs. Codes and dataset will be available at \href{https://github.com/russellyq/MedHallTune}{MedHallTune}.
Authors: Tao Chen, Chenhui Wang, Zhihao Chen, Hongming Shan
Abstract: While deep learning has significantly advanced medical image segmentation, most existing methods still struggle with handling complex anatomical regions. Cascaded or deep supervision-based approaches attempt to address this challenge through multi-scale feature learning but fail to establish sufficient inter-scale dependencies, as each scale relies solely on the features of the immediate predecessor. To this end, we propose the AutoRegressive Segmentation framework via next-scale mask prediction, termed AR-Seg, which progressively predicts the next-scale mask by explicitly modeling dependencies across all previous scales within a unified architecture. AR-Seg introduces three innovations: (1) a multi-scale mask autoencoder that quantizes the mask into multi-scale token maps to capture hierarchical anatomical structures, (2) a next-scale autoregressive mechanism that progressively predicts next-scale masks to enable sufficient inter-scale dependencies, and (3) a consensus-aggregation strategy that combines multiple sampled results to generate a more accurate mask, further improving segmentation robustness. Extensive experimental results on two benchmark datasets with different modalities demonstrate that AR-Seg outperforms state-of-the-art methods while explicitly visualizing the intermediate coarse-to-fine segmentation process.
Authors: Yongqi Tian, Xueyu Sun, Haoyuan He, Linji Hao, Ning Ding, Caigui Jiang
Abstract: Hand-object interaction(HOI) is the fundamental link between human and environment, yet its dexterous and complex pose significantly challenges for gesture control. Despite significant advances in AI and robotics, enabling machines to understand and simulate hand-object interactions, capturing the semantics of functional grasping tasks remains a considerable challenge. While previous work can generate stable and correct 3D grasps, they are still far from achieving functional grasps due to unconsidered grasp semantics. To address this challenge, we propose an innovative two-stage framework, Functional Grasp Synthesis Net (FGS-Net), for generating 3D HOI driven by functional text. This framework consists of a text-guided 3D model generator, Functional Grasp Generator (FGG), and a pose optimization strategy, Functional Grasp Refiner (FGR). FGG generates 3D models of hands and objects based on text input, while FGR fine-tunes the poses using Object Pose Approximator and energy functions to ensure the relative position between the hand and object aligns with human intent and remains physically plausible. Extensive experiments demonstrate that our approach achieves precise and high-quality HOI generation without requiring additional 3D annotation data.
Authors: Rongchang Lu, Bingcheng Liao, Haowen Hou, Jiahang Lv, Xin Hai
Abstract: Magnetic Resonance Imaging (MRI) Super-Resolution (SR) addresses the challenges such as long scan times and expensive equipment by enhancing image resolution from low-quality inputs acquired in shorter scan times in clinical settings. However, current SR techniques still have problems such as limited ability to capture both local and global static patterns effectively and efficiently. To address these limitations, we propose Delta-WKV, a novel MRI super-resolution model that combines Meta-in-Context Learning (MiCL) with the Delta rule to better recognize both local and global patterns in MRI images. This approach allows Delta-WKV to adjust weights dynamically during inference, improving pattern recognition with fewer parameters and less computational effort, without using state-space modeling. Additionally, inspired by Receptance Weighted Key Value (RWKV), Delta-WKV uses a quad-directional scanning mechanism with time-mixing and channel-mixing structures to capture long-range dependencies while maintaining high-frequency details. Tests on the IXI and fastMRI datasets show that Delta-WKV outperforms existing methods, improving PSNR by 0.06 dB and SSIM by 0.001, while reducing training and inference times by over 15\%. These results demonstrate its efficiency and potential for clinical use with large datasets and high-resolution imaging.
Authors: Yuxiang Chen, Haocheng Xi, Jun Zhu, Jianfei Chen
Abstract: Pre-training Transformers in FP4 precision is becoming a promising approach to gain substantial speedup, but it comes with a considerable loss of accuracy. Microscaling (MX) data format provides a fine-grained per-group quantization method to improve the representation ability of the FP4 format and is supported by the next-generation Blackwell GPU architecture. However, training with MXFP4 data format still results in significant degradation and there is a lack of systematic research on the reason. In this work, we propose a novel training method TetraJet for a more accurate FP4 training. We comprehensively evaluate all of the quantizers involved in the training, and identify the weight oscillation problem in the forward pass as the main source of the degradation in MXFP4 training. Therefore, we introduce two novel methods, EMA Quantizer (Q-EMA) and Adaptive Ramping Optimizer (Q-Ramping), to resolve the oscillation problem. Extensive experiments on Vision Transformers demonstrate that TetraJet consistently outperforms the existing 4-bit training methods, and Q-EMA & Q-Ramping can provide additional enhancement by effectively reducing oscillation. We decreased the accuracy degradation by more than $50\%$ compared to the baseline, and can even achieve competitive performance compared to full precision training. The codes are available at https://github.com/thu-ml/TetraJet-MXFP4Training
Authors: Haozhong Sun, Zhongsen Li, Chenlin Du, Haokun Li, Yajie Wang, Huijun Chen
Abstract: Quantitative magnetic resonance imaging (qMRI) requires multi-phase acqui-sition, often relying on reduced data sampling and reconstruction algorithms to accelerate scans, which inherently poses an ill-posed inverse problem. While many studies focus on measuring uncertainty during this process, few explore how to leverage it to enhance reconstruction performance. In this paper, we in-troduce PUQ, a novel approach that pioneers the use of uncertainty infor-mation for qMRI reconstruction. PUQ employs a two-stage reconstruction and parameter fitting framework, where phase-wise uncertainty is estimated during reconstruction and utilized in the fitting stage. This design allows uncertainty to reflect the reliability of different phases and guide information integration during parameter fitting. We evaluated PUQ on in vivo T1 and T2 mapping datasets from healthy subjects. Compared to existing qMRI reconstruction methods, PUQ achieved the state-of-the-art performance in parameter map-pings, demonstrating the effectiveness of uncertainty guidance. Our code is available at https://anonymous.4open.science/r/PUQ-75B2/.
Authors: Silin Chen, Kangjian Di, Yichu Xu, Han-Jia Ye, Wenhan Luo, Ningmu Zou
Abstract: Unsupervised image anomaly detection (UAD) has become a critical process in industrial and medical applications, but it faces growing challenges due to increasing concerns over data privacy. The limited class diversity inherent to one-class classification tasks, combined with distribution biases caused by variations in products across and within clients, poses significant challenges for preserving data privacy with federated UAD. Thus, this article proposes an efficient federated learning method with dynamic memory and memory-reduce for unsupervised image anomaly detection, called FedDyMem. Considering all client data belongs to a single class (i.e., normal sample) in UAD and the distribution of intra-class features demonstrates significant skewness, FedDyMem facilitates knowledge sharing between the client and server through the client's dynamic memory bank instead of model parameters. In the local clients, a memory generator and a metric loss are employed to improve the consistency of the feature distribution for normal samples, leveraging the local model to update the memory bank dynamically. For efficient communication, a memory-reduce method based on weighted averages is proposed to significantly decrease the scale of memory banks. On the server, global memory is constructed and distributed to individual clients through k-means aggregation. Experiments conducted on six industrial and medical datasets, comprising a mixture of six products or health screening types derived from eleven public datasets, demonstrate the effectiveness of FedDyMem.
Authors: Nesryne Mejri, Enjie Ghorbel, Anis Kacem, Pavel Chernakov, Niki Foteinopoulou, Djamila Aouada
Abstract: This paper introduces the first fully unsupervised domain adaptation (UDA) framework for unsupervised anomaly detection (UAD). The performance of UAD techniques degrades significantly in the presence of a domain shift, difficult to avoid in a real-world setting. While UDA has contributed to solving this issue in binary and multi-class classification, such a strategy is ill-posed in UAD. This might be explained by the unsupervised nature of the two tasks, namely, domain adaptation and anomaly detection. Herein, we first formulate this problem that we call the two-fold unsupervised curse. Then, we propose a pioneering solution to this curse, considered intractable so far, by assuming that anomalies are rare. Specifically, we leverage clustering techniques to identify a dominant cluster in the target feature space. Posed as the normal cluster, the latter is aligned with the source normal features. Concretely, given a one-class source set and an unlabeled target set composed mostly of normal data and some anomalies, we fit the source features within a hypersphere while jointly aligning them with the features of the dominant cluster from the target set. The paper provides extensive experiments and analysis on common adaptation benchmarks for anomaly detection, demonstrating the relevance of both the newly introduced paradigm and the proposed approach. The code will be made publicly available.
Authors: Ganzhao Yuan
Abstract: This paper proposes {\sf AAPG-SPIDER}, an Adaptive Accelerated Proximal Gradient (AAPG) method with variance reduction for minimizing composite nonconvex finite-sum functions. It integrates three acceleration techniques: adaptive stepsizes, Nesterov's extrapolation, and the recursive stochastic path-integrated estimator SPIDER. While targeting stochastic finite-sum problems, {\sf AAPG-SPIDER} simplifies to {\sf AAPG} in the full-batch, non-stochastic setting, which is also of independent interest. To our knowledge, {\sf AAPG-SPIDER} and {\sf AAPG} are the first learning-rate-free methods to achieve optimal iteration complexity for this class of \textit{composite} minimization problems. Specifically, {\sf AAPG} achieves the optimal iteration complexity of $\mathcal{O}(N \epsilon^{-2})$, while {\sf AAPG-SPIDER} achieves $\mathcal{O}(N + \sqrt{N} \epsilon^{-2})$ for finding $\epsilon$-approximate stationary points, where $N$ is the number of component functions. Under the Kurdyka-Lojasiewicz (KL) assumption, we establish non-ergodic convergence rates for both methods. Preliminary experiments on sparse phase retrieval and linear eigenvalue problems demonstrate the superior performance of {\sf AAPG-SPIDER} and {\sf AAPG} compared to existing methods.
Authors: Youngjin Yoo, Bogdan Georgescu, Yanbo Zhang, Sasa Grbic, Han Liu, Gabriela D. Aldea, Thomas J. Re, Jyotipriya Das, Poikavila Ullaskrishnan, Eva Eibenberger, Andrei Chekkoury, Uttam K. Bodanapally, Savvas Nicolaou, Pina C. Sanelli, Thomas J. Schroeppel, Yvonne W. Lui, Eli Gibson
Abstract: Recent advancements in AI and medical imaging offer transformative potential in emergency head CT interpretation for reducing assessment times and improving accuracy in the face of an increasing request of such scans and a global shortage in radiologists. This study introduces a 3D foundation model for detecting diverse neuro-trauma findings with high accuracy and efficiency. Using large language models (LLMs) for automatic labeling, we generated comprehensive multi-label annotations for critical conditions. Our approach involved pretraining neural networks for hemorrhage subtype segmentation and brain anatomy parcellation, which were integrated into a pretrained comprehensive neuro-trauma detection network through multimodal fine-tuning. Performance evaluation against expert annotations and comparison with CT-CLIP demonstrated strong triage accuracy across major neuro-trauma findings, such as hemorrhage and midline shift, as well as less frequent critical conditions such as cerebral edema and arterial hyperdensity. The integration of neuro-specific features significantly enhanced diagnostic capabilities, achieving an average AUC of 0.861 for 16 neuro-trauma conditions. This work advances foundation models in medical imaging, serving as a benchmark for future AI-assisted neuro-trauma diagnostics in emergency radiology.
Authors: Marina D'Amato, Jeroen van der Laak, Francesco Ciompi
Abstract: Accurate tumor detection in digital pathology whole-slide images (WSIs) is crucial for cancer diagnosis and treatment planning. Multiple Instance Learning (MIL) has emerged as a widely used approach for weakly-supervised tumor detection with large-scale data without the need for manual annotations. However, traditional MIL methods often depend on classification tasks that require tumor-free cases as negative examples, which are challenging to obtain in real-world clinical workflows, especially for surgical resection specimens. We address this limitation by reformulating tumor detection as a regression task, estimating tumor percentages from WSIs, a clinically available target across multiple cancer types. In this paper, we provide an analysis of the proposed weakly-supervised regression framework by applying it to multiple organs, specimen types and clinical scenarios. We characterize the robustness of our framework to tumor percentage as a noisy regression target, and introduce a novel concept of amplification technique to improve tumor detection sensitivity when learning from small tumor regions. Finally, we provide interpretable insights into the model's predictions by analyzing visual attention and logit maps. Our code is available at https://github.com/DIAGNijmegen/tumor-percentage-mil-regression.
URLs: https://github.com/DIAGNijmegen/tumor-percentage-mil-regression.
Authors: Eli Verwimp, Guy Hacohen, Tinne Tuytelaars
Abstract: Continual learning aims to enable models to adapt to new datasets without losing performance on previously learned data, often assuming that prior data is no longer available. However, in many practical scenarios, both old and new data are accessible. In such cases, good performance on both datasets is typically achieved by abandoning the model trained on the previous data and re-training a new model from scratch on both datasets. This training from scratch is computationally expensive. In contrast, methods that leverage the previously trained model and old data are worthy of investigation, as they could significantly reduce computational costs. Our evaluation framework quantifies the computational savings of such methods while maintaining or exceeding the performance of training from scratch. We identify key optimization aspects -- initialization, regularization, data selection, and hyper-parameters -- that can each contribute to reducing computational costs. For each aspect, we propose effective first-step methods that already yield substantial computational savings. By combining these methods, we achieve up to 2.7x reductions in computation time across various computer vision tasks, highlighting the potential for further advancements in this area.
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: Shashwat Gupta, Sarthak Gupta, Akshan Agrawal, Mahim Naaz, Rajanikanth Yadav, Priyanka Bagade
Abstract: Comb Sign is an important imaging biomarker to detect multiple gastrointestinal diseases. It shows up as increased blood flow along the intestinal wall indicating potential abnormality, which helps doctors diagnose inflammatory conditions. Despite its clinical significance, current detection methods are manual, time-intensive, and prone to subjective interpretation due to the need for multi-planar image-orientation. To the best of our knowledge, we are the first to propose a fully automated technique for the detection of Comb Sign from CTE scans. Our novel approach is based on developing a probabilistic map that shows areas of pathological hypervascularity by identifying fine vascular bifurcations and wall enhancement via processing through stepwise algorithmic modules. These modules include utilising deep learning segmentation model, a Gaussian Mixture Model (GMM), vessel extraction using vesselness filter, iterative probabilistic enhancement of vesselness via neighborhood maximization and a distance-based weighting scheme over the vessels. Experimental results demonstrate that our pipeline effectively identifies Comb Sign, offering an objective, accurate, and reliable tool to enhance diagnostic accuracy in Crohn's disease and related hypervascular conditions where Comb Sign is considered as one of the important biomarkers.
Authors: Tatiana A. Bubba, Matteo Santacesaria, Andrea Sebastiani
Abstract: Deep learning has emerged as a powerful tool for solving inverse problems in imaging, including computed tomography (CT). However, most approaches require paired training data with ground truth images, which can be difficult to obtain, e.g., in medical applications. We present TomoSelfDEQ, a self-supervised Deep Equilibrium (DEQ) framework for sparse-angle CT reconstruction that trains directly on undersampled measurements. We establish theoretical guarantees showing that, under suitable assumptions, our self-supervised updates match those of fully-supervised training with a loss including the (possibly non-unitary) forward operator like the CT forward map. Numerical experiments on sparse-angle CT data confirm this finding, also demonstrating that TomoSelfDEQ outperforms existing self-supervised methods, achieving state-of-the-art results with as few as 16 projection angles.
Authors: Komal Kumar, Tajamul Ashraf, Omkar Thawakar, Rao Muhammad Anwer, Hisham Cholakkal, Mubarak Shah, Ming-Hsuan Yang, Phillip H. S. Torr, Salman Khan, Fahad Shahbaz Khan
Abstract: Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement learning, and test-time scaling have emerged as critical strategies for optimizing LLMs performance, ensuring robustness, and improving adaptability across various real-world tasks. This survey provides a systematic exploration of post-training methodologies, analyzing their role in refining LLMs beyond pretraining, addressing key challenges such as catastrophic forgetting, reward hacking, and inference-time trade-offs. We highlight emerging directions in model alignment, scalable adaptation, and inference-time reasoning, and outline future research directions. We also provide a public repository to continually track developments in this fast-evolving field: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training.
URLs: https://github.com/mbzuai-oryx/Awesome-LLM-Post-training.
Authors: Hengyuan Ma, Xiatian Zhu, Jianfeng Feng, Li Zhang
Abstract: Score-based generative models (SGMs) have recently emerged as a promising class of generative models. However, a fundamental limitation is that their sampling process is slow due to a need for many (e.g., 2000) iterations of sequential computations. An intuitive acceleration method is to reduce the sampling iterations which however causes severe performance degradation. We assault this problem to the ill-conditioned issues of the Langevin dynamics and reverse diffusion in the sampling process. Under this insight, we propose a novel preconditioned diffusion sampling (PDS) method that leverages matrix preconditioning to alleviate the aforementioned problem. PDS alters the sampling process of a vanilla SGM at marginal extra computation cost and without model retraining. Theoretically, we prove that PDS preserves the output distribution of the SGM, with no risk of inducing systematical bias to the original sampling process. We further theoretically reveal a relation between the parameter of PDS and the sampling iterations, easing the parameter estimation under varying sampling iterations. Extensive experiments on various image datasets with a variety of resolutions and diversity validate that our PDS consistently accelerates off-the-shelf SGMs whilst maintaining the synthesis quality. In particular, PDS can accelerate by up to 28x on more challenging high-resolution (1024x1024) image generation. Compared with the latest generative models (e.g., CLD-SGM and Analytic-DDIM), PDS can achieve the best sampling quality on CIFAR-10 at an FID score of 1.99. Our code is publicly available to foster any further research https://github.com/fudan-zvg/PDS.
Authors: Finlay G. C. Hudson, William A. P. Smith
Abstract: In this paper, we introduce a novel method for zero-shot, cross-domain image retrieval. Our key contribution is a test-time Iterative Cluster-free Re-ranking process that leverages gallery-gallery feature information to establish semantic links between query and gallery images. This enables the retrieval of relevant images even when they do not exhibit similar visual features but share underlying semantic concepts. This can be combined with any pre-existing cross-domain feature extraction backbone to improve retrieval performance. However, when combined with a carefully chosen Vision Transformer backbone and combination of zero-shot retrieval losses, our approach yields state-of-the-art results on the Sketchy, TU-Berlin and QuickDraw sketch-based retrieval benchmarks. We show that our re-ranking also improves performance with other backbones and outperforms other re-ranking methods applied with our backbone. Importantly, unlike many previous methods, none of the components in our approach are engineered specifically towards the sketch-based image retrieval task - it can be generally applied to any cross-domain, zero-shot retrieval task. We therefore also present new results on zero-shot cartoon-to-photo and art-to-product retrieval using the Office-Home dataset. Project page: finlay-hudson.github.io/icfrr, code available at: github.com/finlay-hudson/ICFRR
Authors: Aneeq Zia, Max Berniker, Rogerio Garcia Nespolo, Conor Perreault, Kiran Bhattacharyya, Xi Liu, Ziheng Wang, Satoshi Kondo, Satoshi Kasai, Kousuke Hirasawa, Bo Liu, David Austin, Yiheng Wang, Michal Futrega, Jean-Francois Puget, Zhenqiang Li, Yoichi Sato, Ryo Fujii, Ryo Hachiuma, Mana Masuda, Hideo Saito, An Wang, Mengya Xu, Mobarakol Islam, Long Bai, Winnie Pang, Hongliang Ren, Chinedu Nwoye, Luca Sestini, Nicolas Padoy, Maximilian Nielsen, Samuel Sch\"uttler, Thilo Sentker, H\"umeyra Husseini, Ivo Baltruschat, R\"udiger Schmitz, Ren\'e Werner, Aleksandr Matsun, Mugariya Farooq, Numan Saaed, Jose Renato Restom Viera, Mohammad Yaqub, Neil Getty, Fangfang Xia, Zixuan Zhao, Xiaotian Duan, Xing Yao, Ange Lou, Hao Yang, Jintong Han, Jack Noble, Jie Ying Wu, Tamer Abdulbaki Alshirbaji, Nour Aldeen Jalal, Herag Arabian, Ning Ding, Knut Moeller, Weiliang Chen, Quan He, Muhammad Bilal, Taofeek Akinosho, Adnan Qayyum, Massimo Caputo, Hunaid Vohra, Michael Loizou, Anuoluwapo Ajayi, Ilhem Berrou, Faatihah Niyi-Odumosu, Charlie Budd, Oluwatosin Alabi, Tom Vercauteren, Ruoxi Zhao, Ayberk Acar, John Han, Jumanh Atoum, Yinhong Qin, Jie Ying Wu, Surong Hua, Lu Ping, Wenming Wu, Rongfeng Wei, Jinlin Wu, You Pang, Zhen Chen, Tim Jaspers, Amine Yamlahi, Piotr Kalinowski, Dominik Michael, Tim R\"a dsch, Marco H\"ubner, Danail Stoyanov, Stefanie Speidel, Lena Maier-Hein, Anthony Jarc
Abstract: Robotic assisted (RA) surgery promises to transform surgical intervention. Intuitive Surgical is committed to fostering these changes and the machine learning models and algorithms that will enable them. With these goals in mind we have invited the surgical data science community to participate in a yearly competition hosted through the Medical Imaging Computing and Computer Assisted Interventions (MICCAI) conference. With varying changes from year to year, we have challenged the community to solve difficult machine learning problems in the context of advanced RA applications. Here we document the results of these challenges, focusing on surgical tool localization (SurgToolLoc). The publicly released dataset that accompanies these challenges is detailed in a separate paper arXiv:2501.09209 [1].
Authors: Chunhui Zhang, Xin Sun, Yiqian Yang, Li Liu, Qiong Liu, Xi Zhou, Yanfeng Wang
Abstract: Current mainstream vision-language (VL) tracking framework consists of three parts, \ie a visual feature extractor, a language feature extractor, and a fusion model. To pursue better performance, a natural modus operandi for VL tracking is employing customized and heavier unimodal encoders, and multi-modal fusion models. Albeit effective, existing VL trackers separate feature extraction and feature integration, resulting in extracted features that lack semantic guidance and have limited target-aware capability in complex scenarios, \eg similar distractors and extreme illumination. In this work, inspired by the recent success of exploring foundation models with unified architecture for both natural language and computer vision tasks, we propose an All-in-One framework, which learns joint feature extraction and interaction by adopting a unified transformer backbone. Specifically, we mix raw vision and language signals to generate language-injected vision tokens, which we then concatenate before feeding into the unified backbone architecture. This approach achieves feature integration in a unified backbone, removing the need for carefully-designed fusion modules and resulting in a more effective and efficient VL tracking framework. To further improve the learning efficiency, we introduce a multi-modal alignment module based on cross-modal and intra-modal contrastive objectives, providing more reasonable representations for the unified All-in-One transformer backbone. Extensive experiments on five benchmarks, \ie OTB99-L, TNL2K, LaSOT, LaSOT$_{\rm Ext}$ and WebUAV-3M, demonstrate the superiority of the proposed tracker against existing state-of-the-arts on VL tracking. Codes will be made publicly available at https://github.com/983632847/All-in-One.
Authors: Yiming Zhou, Yuexing Peng, Junchuan Yu, Daqing Ge, Wei Xiang
Abstract: As a natural disaster, landslide often brings tremendous losses to human lives, so it urgently demands reliable detection of landslide risks. When detecting relic landslides that present important information for landslide risk warning, problems such as visual blur and small-sized dataset cause great challenges when using remote sensing images. To extract accurate semantic features, a hyper-pixel-wise contrastive learning augmented segmentation network (HPCL-Net) is proposed, which augments the local salient feature extraction from boundaries of landslides through HPCL and fuses heterogeneous information in the semantic space from high-resolution remote sensing images and digital elevation model data. For full utilization of precious samples, a global hyper-pixel-wise sample pair queues-based contrastive learning method is developed, which includes the construction of global queues that store hyper-pixel-wise samples and the updating scheme of a momentum encoder, reliably enhancing the extraction ability of semantic features. The proposed HPCL-Net is evaluated on the Loess Plateau relic landslide dataset and experimental results verify that the proposed HPCL-Net greatly outperforms existing models, where the mIoU is increased from 0.620 to 0.651, the Landslide IoU is improved from 0.334 to 0.394 and the F1score is enhanced from 0.501 to 0.565.
Authors: Shilin Xu, Haobo Yuan, Qingyu Shi, Lu Qi, Jingbo Wang, Yibo Yang, Yining Li, Kai Chen, Yunhai Tong, Bernard Ghanem, Xiangtai Li, Ming-Hsuan Yang
Abstract: Recent segmentation methods, which adopt large-scale data training and transformer architecture, aim to create one foundation model that can perform multiple tasks. However, most of these methods rely on heavy encoder and decoder frameworks, hindering their performance in real-time scenarios. To explore real-time segmentation, recent advancements primarily focus on semantic segmentation within specific environments, such as autonomous driving. However, they often overlook the generalization ability of these models across diverse scenarios. Therefore, to fill this gap, this work explores a novel real-time segmentation setting called real-time multi-purpose segmentation. It contains three fundamental sub-tasks: interactive segmentation, panoptic segmentation, and video instance segmentation. Unlike previous methods, which use a specific design for each task, we aim to use only a single end-to-end model to accomplish all these tasks in real-time. To meet real-time requirements and balance multi-task learning, we present a novel dynamic convolution-based method, Real-Time Multi-Purpose SAM (RMP-SAM). It contains an efficient encoder and an efficient decoupled adapter to perform prompt-driven decoding. Moreover, we further explore different training strategies and one new adapter design to boost co-training performance further. We benchmark several strong baselines by extending existing works to support our multi-purpose segmentation. Extensive experiments demonstrate that RMP-SAM is effective and generalizes well on proposed benchmarks and other specific semantic tasks. Our implementation of RMP-SAM achieves the optimal balance between accuracy and speed for these tasks.Our code and model are available at https://github.com/xushilin1/RAP-SAM/.
Authors: Michail Mamalakis, Antonios Mamalakis, Ingrid Agartz, Lynn Egeland M{\o}rch-Johnsen, Graham Murray, John Suckling, Pietro Lio
Abstract: The accelerated progress of artificial intelligence (AI) has popularized deep learning models across various domains, yet their inherent opacity poses challenges, particularly in critical fields like healthcare, medicine, and the geosciences. Explainable AI (XAI) has emerged to shed light on these 'black box' models, aiding in deciphering their decision-making processes. However, different XAI methods often produce significantly different explanations, leading to high inter-method variability that increases uncertainty and undermines trust in deep networks' predictions. In this study, we address this challenge by introducing a novel framework designed to enhance the explainability of deep networks through a dual focus on maximizing both accuracy and comprehensibility in the explanations. Our framework integrates outputs from multiple established XAI methods and leverages a non-linear neural network model, termed the 'explanation optimizer,' to construct a unified, optimal explanation. The optimizer evaluates explanations using two key metrics: faithfulness (accuracy in reflecting the network's decisions) and complexity (comprehensibility). By balancing these, it provides accurate and accessible explanations, addressing a key XAI limitation. Experiments on multi-class and binary classification in 2D object and 3D neuroscience imaging confirm its efficacy. Our optimizer achieved faithfulness scores 155% and 63% higher than the best XAI methods in 3D and 2D tasks, respectively, while also reducing complexity for better understanding. These results demonstrate that optimal explanations based on specific quality criteria are achievable, offering a solution to the issue of inter-method variability in the current XAI literature and supporting more trustworthy deep network predictions
Authors: Pritam Sarkar, Sayna Ebrahimi, Ali Etemad, Ahmad Beirami, Sercan \"O. Ar{\i}k, Tomas Pfister
Abstract: Despite their significant advancements, Multimodal Large Language Models (MLLMs) often generate factually inaccurate information, referred to as hallucination. In this work, we address object hallucinations in MLLMs, where information is generated about an object not present in the input image. We introduce Data-augmented Phrase-level Alignment (DPA), a novel loss which can be applied to instruction-tuned off-the-shelf MLLMs to mitigate hallucinations, while preserving their general vision-language capabilities. To fine-tune MLLMs with DPA, we first generate a set of `hallucinated' and `correct' response pairs through generative data augmentation by selectively altering the ground-truth information of the correct responses at a phrase level. The DPA loss is then used to train MLLMs to reduce the likelihood of hallucinated phrases compared to the correct ones. Our thorough evaluation on various benchmarks confirms the effectiveness of DPA in mitigating hallucination while retaining the out-of-the-box performance of the MLLMs on general tasks. For instance, MLLMs finetuned with DPA, which we refer to as Hallucination Attenuated Language and Vision Assistant (HALVA), improve F1 by up to 13.4% on hallucination visual question-answering and reduce the hallucination rate by up to 4.2% on image description tasks.
Authors: Haolin Li, Jinyang Liu, Mario Sznaier, Octavia Camps
Abstract: Photo-realistic image rendering from scene 3D reconstruction is a fundamental problem in 3D computer vision. This domain has seen considerable advancements owing to the advent of recent neural rendering techniques. These techniques predominantly aim to focus on learning volumetric representations of 3D scenes and refining these representations via loss functions derived from their rendering. Among these, 3D Gaussian Splatting (3D-GS) has emerged as a preferred method, surpassing Neural Radiance Fields' (NeRFs) quality and rendering speed. 3D-GS uses parameterized 3D Gaussians to model both spatial locations and color information, combined with a tile-based fast rendering technique. Despite its superior performance, using 3D Gaussian kernels has inherent limitations in accurately representing discontinuous functions, notably at edges and corners corresponding to shape discontinuities, and across varying textures due to color discontinuities. In this paper, we introduce 3D Half-Gaussian (\textbf{3D-HGS}) kernels, which can be used as a plug-and-play kernel, to address this issue. Our experiments demonstrate their capability to improve the performance of current 3D-GS related methods and achieve state-of-the-art rendering quality performance on various datasets without compromising their rendering speed.
Authors: Drew Linsley, Peisen Zhou, Alekh Karkada Ashok, Akash Nagaraj, Gaurav Gaonkar, Francis E Lewis, Zygmunt Pizlo, Thomas Serre
Abstract: Visual perspective taking (VPT) is the ability to perceive and reason about the perspectives of others. It is an essential feature of human intelligence, which develops over the first decade of life and requires an ability to process the 3D structure of visual scenes. A growing number of reports have indicated that deep neural networks (DNNs) become capable of analyzing 3D scenes after training on large image datasets. We investigated if this emergent ability for 3D analysis in DNNs is sufficient for VPT with the 3D perception challenge (3D-PC): a novel benchmark for 3D perception in humans and DNNs. The 3D-PC is comprised of three 3D-analysis tasks posed within natural scene images: 1. a simple test of object depth order, 2. a basic VPT task (VPT-basic), and 3. another version of VPT (VPT-Strategy) designed to limit the effectiveness of "shortcut" visual strategies. We tested human participants (N=33) and linearly probed or text-prompted over 300 DNNs on the challenge and found that nearly all of the DNNs approached or exceeded human accuracy in analyzing object depth order. Surprisingly, DNN accuracy on this task correlated with their object recognition performance. In contrast, there was an extraordinary gap between DNNs and humans on VPT-basic. Humans were nearly perfect, whereas most DNNs were near chance. Fine-tuning DNNs on VPT-basic brought them close to human performance, but they, unlike humans, dropped back to chance when tested on VPT-Strategy. Our challenge demonstrates that the training routines and architectures of today's DNNs are well-suited for learning basic 3D properties of scenes and objects but are ill-suited for reasoning about these properties as humans do. We release our 3D-PC datasets and code to help bridge this gap in 3D perception between humans and machines.
Authors: Kailas Dayanandan, Nikhil Kumar, Anand Sinha, Brejesh Lall
Abstract: The dual thinking framework considers fast, intuitive, and slower logical processing. The perception of dual thinking in vision requires images where inferences from intuitive and logical processing differ, and the latter is under-explored in current studies. We introduce a novel adversarial dataset to provide evidence for the dual thinking framework in human vision, which also facilitates the study of the qualitative behavior of deep learning models. Our psychophysical studies show the presence of multiple inferences in rapid succession, and analysis of errors shows that the early stopping of visual processing can result in missing relevant information. MLLMs (Multi-modal Large Language Models) and VLMs (Vision Language Models) have made significant progress in correcting errors in intuitive processing in human vision and showed enhanced performance on images requiring logical processing. However, their improvements in logical processing have not kept pace with their advancements in intuitive processing. In contrast, segmentation models exhibit errors similar to those seen in intuitive human processing and lack understanding of sub-structures, as indicated by errors related to sub-components in identified instances. As AI (Artificial Intelligence)-based systems find increasing applications in safety-critical domains like autonomous driving, the integration of logical processing capabilities becomes essential. This not only enhances performance but also addresses the limitations of scaling-based approaches while ensuring robustness and reliability in real-world environments.
Authors: Isaac Lin, Tianye Wang, Shang Gao, Shiming Tang, Tai Sing Lee
Abstract: Convolutional neural networks (CNNs) have been shown to be state-of-the-art models for visual cortical neurons. Cortical neurons in the primary visual cortex are sensitive to contextual information mediated by extensive horizontal and feedback connections. Standard CNNs integrate global contextual information to model contextual modulation via two mechanisms: successive convolutions and a fully connected readout layer. In this paper, we find that self-attention (SA), an implementation of non-local network mechanisms, can improve neural response predictions over parameter-matched CNNs in two key metrics: tuning curve correlation and peak tuning. We introduce peak tuning as a metric to evaluate a model's ability to capture a neuron's top feature preference. We factorize networks to assess each context mechanism, revealing that information in the local receptive field is most important for modeling overall tuning, but surround information is critically necessary for characterizing the tuning peak. We find that self-attention can replace posterior spatial-integration convolutions when learned incrementally, and is further enhanced in the presence of a fully connected readout layer, suggesting that the two context mechanisms are complementary. Finally, we find that decomposing receptive field learning and contextual modulation learning in an incremental manner may be an effective and robust mechanism for learning surround-center interactions.
Authors: Yingyan Li, Lue Fan, Jiawei He, Yuqi Wang, Yuntao Chen, Zhaoxiang Zhang, Tieniu Tan
Abstract: In autonomous driving, end-to-end planners directly utilize raw sensor data, enabling them to extract richer scene features and reduce information loss compared to traditional planners. This raises a crucial research question: how can we develop better scene feature representations to fully leverage sensor data in end-to-end driving? Self-supervised learning methods show great success in learning rich feature representations in NLP and computer vision. Inspired by this, we propose a novel self-supervised learning approach using the LAtent World model (LAW) for end-to-end driving. LAW predicts future scene features based on current features and ego trajectories. This self-supervised task can be seamlessly integrated into perception-free and perception-based frameworks, improving scene feature learning and optimizing trajectory prediction. LAW achieves state-of-the-art performance across multiple benchmarks, including real-world open-loop benchmark nuScenes, NAVSIM, and simulator-based closed-loop benchmark CARLA. The code is released at https://github.com/BraveGroup/LAW.
Authors: Moritz Kappel, Florian Hahlbohm, Timon Scholz, Susana Castillo, Christian Theobalt, Martin Eisemann, Vladislav Golyanik, Marcus Magnor
Abstract: Dynamic reconstruction and spatiotemporal novel-view synthesis of non-rigidly deforming scenes recently gained increased attention. While existing work achieves impressive quality and performance on multi-view or teleporting camera setups, most methods fail to efficiently and faithfully recover motion and appearance from casual monocular captures. This paper contributes to the field by introducing a new method for dynamic novel view synthesis from monocular video, such as casual smartphone captures. Our approach represents the scene as a $\textit{dynamic neural point cloud}$, an implicit time-conditioned point distribution that encodes local geometry and appearance in separate hash-encoded neural feature grids for static and dynamic regions. By sampling a discrete point cloud from our model, we can efficiently render high-quality novel views using a fast differentiable rasterizer and neural rendering network. Similar to recent work, we leverage advances in neural scene analysis by incorporating data-driven priors like monocular depth estimation and object segmentation to resolve motion and depth ambiguities originating from the monocular captures. In addition to guiding the optimization process, we show that these priors can be exploited to explicitly initialize our scene representation to drastically improve optimization speed and final image quality. As evidenced by our experimental evaluation, our dynamic point cloud model not only enables fast optimization and real-time frame rates for interactive applications, but also achieves competitive image quality on monocular benchmark sequences. Our code and data are available online: https://moritzkappel.github.io/projects/dnpc/.
Authors: Yue Jiang, Jiawei Chen, Dingkang Yang, Mingcheng Li, Shunli Wang, Tong Wu, Ke Li, Lihua Zhang
Abstract: Automatic medical report generation (MRG), which possesses significant research value as it can aid radiologists in clinical diagnosis and report composition, has garnered increasing attention. Despite recent progress, generating accurate reports remains arduous due to the requirement for precise clinical comprehension and disease diagnosis inference. Furthermore, owing to the limited accessibility of medical data and the imbalanced distribution of diseases, the underrepresentation of rare diseases in training data makes large-scale medical visual language models (LVLMs) prone to hallucinations, such as omissions or fabrications, severely undermining diagnostic performance and further intensifying the challenges for MRG in practice. In this study, to effectively mitigate hallucinations in medical report generation, we propose a chain-of-medical-thought approach (CoMT), which intends to imitate the cognitive process of human doctors by decomposing diagnostic procedures. The radiological features with different importance are structured into fine-grained medical thought chains to enhance the inferential ability during diagnosis, thereby alleviating hallucination problems and enhancing the diagnostic accuracy of MRG. The code and dataset have been released at https://github.com/FRENKIE-CHIANG/CoMT.
Authors: Jianzhu Guo, Dingyun Zhang, Xiaoqiang Liu, Zhizhou Zhong, Yuan Zhang, Pengfei Wan, Di Zhang
Abstract: Portrait Animation aims to synthesize a lifelike video from a single source image, using it as an appearance reference, with motion (i.e., facial expressions and head pose) derived from a driving video, audio, text, or generation. Instead of following mainstream diffusion-based methods, we explore and extend the potential of the implicit-keypoint-based framework, which effectively balances computational efficiency and controllability. Building upon this, we develop a video-driven portrait animation framework named LivePortrait with a focus on better generalization, controllability, and efficiency for practical usage. To enhance the generation quality and generalization ability, we scale up the training data to about 69 million high-quality frames, adopt a mixed image-video training strategy, upgrade the network architecture, and design better motion transformation and optimization objectives. Additionally, we discover that compact implicit keypoints can effectively represent a kind of blendshapes and meticulously propose a stitching and two retargeting modules, which utilize a small MLP with negligible computational overhead, to enhance the controllability. Experimental results demonstrate the efficacy of our framework even compared to diffusion-based methods. The generation speed remarkably reaches 12.8ms on an RTX 4090 GPU with PyTorch. The inference code and models are available at https://github.com/KwaiVGI/LivePortrait
Authors: Amir Mohammad Karimi Mamaghan, Samuele Papa, Karl Henrik Johansson, Stefan Bauer, Andrea Dittadi
Abstract: Object-centric (OC) representations, which model visual scenes as compositions of discrete objects, have the potential to be used in various downstream tasks to achieve systematic compositional generalization and facilitate reasoning. However, these claims have yet to be thoroughly validated empirically. Recently, foundation models have demonstrated unparalleled capabilities across diverse domains, from language to computer vision, positioning them as a potential cornerstone of future research for a wide range of computational tasks. In this paper, we conduct an extensive empirical study on representation learning for downstream Visual Question Answering (VQA), which requires an accurate compositional understanding of the scene. We thoroughly investigate the benefits and trade-offs of OC models and alternative approaches including large pre-trained foundation models on both synthetic and real-world data, ultimately identifying a promising path to leverage the strengths of both paradigms. The extensiveness of our study, encompassing over 600 downstream VQA models and 15 different types of upstream representations, also provides several additional insights that we believe will be of interest to the community at large.
Authors: Yiming Xie, Chun-Han Yao, Vikram Voleti, Huaizu Jiang, Varun Jampani
Abstract: We present Stable Video 4D (SV4D), a latent video diffusion model for multi-frame and multi-view consistent dynamic 3D content generation. Unlike previous methods that rely on separately trained generative models for video generation and novel view synthesis, we design a unified diffusion model to generate novel view videos of dynamic 3D objects. Specifically, given a monocular reference video, SV4D generates novel views for each video frame that are temporally consistent. We then use the generated novel view videos to optimize an implicit 4D representation (dynamic NeRF) efficiently, without the need for cumbersome SDS-based optimization used in most prior works. To train our unified novel view video generation model, we curate a dynamic 3D object dataset from the existing Objaverse dataset. Extensive experimental results on multiple datasets and user studies demonstrate SV4D's state-of-the-art performance on novel-view video synthesis as well as 4D generation compared to prior works.
Authors: Yabiao Wang, Shuo Wang, Jiangning Zhang, Ke Fan, Jiafu Wu, Zhucun Xue, Yong Liu
Abstract: Human-human motion generation is essential for understanding humans as social beings. Current methods fall into two main categories: single-person-based methods and separate modeling-based methods. To delve into this field, we abstract the overall generation process into a general framework MetaMotion, which consists of two phases: temporal modeling and interaction mixing. For temporal modeling, the single-person-based methods concatenate two people into a single one directly, while the separate modeling-based methods skip the modeling of interaction sequences. The inadequate modeling described above resulted in sub-optimal performance and redundant model parameters. In this paper, we introduce TIMotion (Temporal and Interactive Modeling), an efficient and effective framework for human-human motion generation. Specifically, we first propose Causal Interactive Injection to model two separate sequences as a causal sequence leveraging the temporal and causal properties. Then we present Role-Evolving Scanning to adjust to the change in the active and passive roles throughout the interaction. Finally, to generate smoother and more rational motion, we design Localized Pattern Amplification to capture short-term motion patterns. Extensive experiments on InterHuman and InterX demonstrate that our method achieves superior performance. The project code will be released upon acceptance. Project page: https://aigc-explorer.github.io/TIMotion-page/
Authors: Qianlong Xiang, Miao Zhang, Yuzhang Shang, Jianlong Wu, Yan Yan, Liqiang Nie
Abstract: Diffusion models (DMs) have demonstrated exceptional generative capabilities across various domains, including image, video, and so on. A key factor contributing to their effectiveness is the high quantity and quality of data used during training. However, mainstream DMs now consume increasingly large amounts of data. For example, training a Stable Diffusion model requires billions of image-text pairs. This enormous data requirement poses significant challenges for training large DMs due to high data acquisition costs and storage expenses. To alleviate this data burden, we propose a novel scenario: using existing DMs as data sources to train new DMs with any architecture. We refer to this scenario as Data-Free Knowledge Distillation for Diffusion Models (DKDM), where the generative ability of DMs is transferred to new ones in a data-free manner. To tackle this challenge, we make two main contributions. First, we introduce a DKDM objective that enables the training of new DMs via distillation, without requiring access to the data. Second, we develop a dynamic iterative distillation method that efficiently extracts time-domain knowledge from existing DMs, enabling direct retrieval of training data without the need for a prolonged generative process. To the best of our knowledge, we are the first to explore this scenario. Experimental results demonstrate that our data-free approach not only achieves competitive generative performance but also, in some instances, outperforms models trained with the entire dataset.
Authors: Jaewoo Kim, Uehwan Kim
Abstract: While current state-of-the-art Scene Change Detection (SCD) approaches achieve impressive results in well-trained research data, they become unreliable under unseen environments and different temporal conditions; in-domain performance drops from 77.6\% to 8.0\% in a previously unseen environment and to 4.6\% under a different temporal condition -- calling for generalizable SCD and benchmark. In this work, we propose the Generalizable Scene Change Detection Framework (GeSCF), which addresses unseen domain performance and temporal consistency -- to meet the growing demand for anything SCD. Our method leverages the pre-trained Segment Anything Model (SAM) in a zero-shot manner. For this, we design Initial Pseudo-mask Generation and Geometric-Semantic Mask Matching -- seamlessly turning user-guided prompt and single-image based segmentation into scene change detection for a pair of inputs without guidance. Furthermore, we define the Generalizable Scene Change Detection (GeSCD) benchmark along with novel metrics and an evaluation protocol to facilitate SCD research in generalizability. In the process, we introduce the ChangeVPR dataset, a collection of challenging image pairs with diverse environmental scenarios -- including urban, suburban, and rural settings. Extensive experiments across various datasets demonstrate that GeSCF achieves an average performance gain of 19.2\% on existing SCD datasets and 30.0\% on the ChangeVPR dataset, nearly doubling the prior art performance. We believe our work can lay a solid foundation for robust and generalizable SCD research.
Authors: Weimin Lyu, Jiachen Yao, Saumya Gupta, Lu Pang, Tao Sun, Lingjie Yi, Lijie Hu, Haibin Ling, Chao Chen
Abstract: The emergence of Vision-Language Models (VLMs) represents a significant advancement in integrating computer vision with Large Language Models (LLMs) to generate detailed text descriptions from visual inputs. Despite their growing importance, the security of VLMs, particularly against backdoor attacks, is under explored. Moreover, prior works often assume attackers have access to the original training data, which is often unrealistic. In this paper, we address a more practical and challenging scenario where attackers must rely solely on Out-Of-Distribution (OOD) data. We introduce VLOOD (Backdooring Vision-Language Models with Out-of-Distribution Data), a novel approach with two key contributions: (1) demonstrating backdoor attacks on VLMs in complex image-to-text tasks while minimizing degradation of the original semantics under poisoned inputs, and (2) proposing innovative techniques for backdoor injection without requiring any access to the original training data. Our evaluation on image captioning and visual question answering (VQA) tasks confirms the effectiveness of VLOOD, revealing a critical security vulnerability in VLMs and laying the foundation for future research on securing multimodal models against sophisticated threats.
Authors: Dexuan Ding, Lei Wang, Liyun Zhu, Tom Gedeon, Piotr Koniusz
Abstract: In computer vision tasks, features often come from diverse representations, domains (e.g., indoor and outdoor), and modalities (e.g., text, images, and videos). Effectively fusing these features is essential for robust performance, especially with the availability of powerful pre-trained models like vision-language models. However, common fusion methods, such as concatenation, element-wise operations, and non-linear techniques, often fail to capture structural relationships, deep feature interactions, and suffer from inefficiency or misalignment of features across domains or modalities. In this paper, we shift from high-dimensional feature space to a lower-dimensional, interpretable graph space by constructing relationship graphs that encode feature relationships at different levels, e.g., clip, frame, patch, token, etc. To capture deeper interactions, we expand graphs through iterative graph relationship updates and introduce a learnable graph fusion operator to integrate these expanded relationships for more effective fusion. Our approach is relationship-centric, operates in a homogeneous space, and is mathematically principled, resembling element-wise relationship score aggregation via multilinear polynomials. We demonstrate the effectiveness of our graph-based fusion method on video anomaly detection, showing strong performance across multi-representational, multi-modal, and multi-domain feature fusion tasks.
Authors: Sihyun Yu, Sangkyung Kwak, Huiwon Jang, Jongheon Jeong, Jonathan Huang, Jinwoo Shin, Saining Xie
Abstract: Recent studies have shown that the denoising process in (generative) diffusion models can induce meaningful (discriminative) representations inside the model, though the quality of these representations still lags behind those learned through recent self-supervised learning methods. We argue that one main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations. Moreover, training can be made easier by incorporating high-quality external visual representations, rather than relying solely on the diffusion models to learn them independently. We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders. The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs. For instance, our method can speed up SiT training by over 17.5$\times$, matching the performance (without classifier-free guidance) of a SiT-XL model trained for 7M steps in less than 400K steps. In terms of final generation quality, our approach achieves state-of-the-art results of FID=1.42 using classifier-free guidance with the guidance interval.
Authors: Qian Yu, Peng-Tao Jiang, Hao Zhang, Jinwei Chen, Bo Li, Lihe Zhang, Huchuan Lu
Abstract: In the realm of high-resolution (HR), fine-grained image segmentation, the primary challenge is balancing broad contextual awareness with the precision required for detailed object delineation, capturing intricate details and the finest edges of objects. Diffusion models, trained on vast datasets comprising billions of image-text pairs, such as SD V2.1, have revolutionized text-to-image synthesis by delivering exceptional quality, fine detail resolution, and strong contextual awareness, making them an attractive solution for high-resolution image segmentation. To this end, we propose DiffDIS, a diffusion-driven segmentation model that taps into the potential of the pre-trained U-Net within diffusion models, specifically designed for high-resolution, fine-grained object segmentation. By leveraging the robust generalization capabilities and rich, versatile image representation prior of the SD models, coupled with a task-specific stable one-step denoising approach, we significantly reduce the inference time while preserving high-fidelity, detailed generation. Additionally, we introduce an auxiliary edge generation task to not only enhance the preservation of fine details of the object boundaries, but reconcile the probabilistic nature of diffusion with the deterministic demands of segmentation. With these refined strategies in place, DiffDIS serves as a rapid object mask generation model, specifically optimized for generating detailed binary maps at high resolutions, while demonstrating impressive accuracy and swift processing. Experiments on the DIS5K dataset demonstrate the superiority of DiffDIS, achieving state-of-the-art results through a streamlined inference process. The source code will be publicly available at https://github.com/qianyu-dlut/DiffDIS.
Authors: Linfeng Xu, Fanman Meng, Qingbo Wu, Lili Pan, Heqian Qiu, Lanxiao Wang, Kailong Chen, Kanglei Geng, Yilei Qian, Haojie Wang, Shuchang Zhou, Shimou Ling, Zejia Liu, Nanlin Chen, Yingjie Xu, Shaoxu Cheng, Bowen Tan, Ziyong Xu, Hongliang Li
Abstract: The application of activity recognition in the ``AI + Education" field is gaining increasing attention. However, current work mainly focuses on the recognition of activities in manually captured videos and a limited number of activity types, with little attention given to recognizing activities in surveillance images from real classrooms. Activity recognition in classroom surveillance images faces multiple challenges, such as class imbalance and high activity similarity. To address this gap, we constructed a novel multimodal dataset focused on classroom surveillance image activity recognition called ARIC (Activity Recognition In Classroom). The ARIC dataset has advantages of multiple perspectives, 32 activity categories, three modalities, and real-world classroom scenarios. In addition to the general activity recognition tasks, we also provide settings for continual learning and few-shot continual learning. We hope that the ARIC dataset can act as a facilitator for future analysis and research for open teaching scenarios. You can download preliminary data from https://ivipclab.github.io/publication_ARIC/ARIC.
Authors: Shicheng Xu, Liang Pang, Yunchang Zhu, Huawei Shen, Xueqi Cheng
Abstract: Vision-language alignment in Large Vision-Language Models (LVLMs) successfully enables LLMs to understand visual input. However, we find that existing vision-language alignment methods fail to transfer the existing safety mechanism for text in LLMs to vision, which leads to vulnerabilities in toxic image. To explore the cause of this problem, we give the insightful explanation of where and how the safety mechanism of LVLMs operates and conduct comparative analysis between text and vision. We find that the hidden states at the specific transformer layers play a crucial role in the successful activation of safety mechanism, while the vision-language alignment at hidden states level in current methods is insufficient. This results in a semantic shift for input images compared to text in hidden states, therefore misleads the safety mechanism. To address this, we propose a novel Text-Guided vision-language Alignment method (TGA) for LVLMs. TGA retrieves the texts related to input vision and uses them to guide the projection of vision into the hidden states space in LLMs. Experiments show that TGA not only successfully transfers the safety mechanism for text in basic LLMs to vision in vision-language alignment for LVLMs without any safety fine-tuning on the visual modality but also maintains the general performance on various vision tasks (Safe and Good).
Authors: Zhangheng Li, Keen You, Haotian Zhang, Di Feng, Harsh Agrawal, Xiujun Li, Mohana Prasad Sathya Moorthy, Jeff Nichols, Yinfei Yang, Zhe Gan
Abstract: Building a generalist model for user interface (UI) understanding is challenging due to various foundational issues, such as platform diversity, resolution variation, and data limitation. In this paper, we introduce Ferret-UI 2, a multimodal large language model (MLLM) designed for universal UI understanding across a wide range of platforms, including iPhone, Android, iPad, Webpage, and AppleTV. Building on the foundation of Ferret-UI, Ferret-UI 2 introduces three key innovations: support for multiple platform types, high-resolution perception through adaptive scaling, and advanced task training data generation powered by GPT-4o with set-of-mark visual prompting. These advancements enable Ferret-UI 2 to perform complex, user-centered interactions, making it highly versatile and adaptable for the expanding diversity of platform ecosystems. Extensive empirical experiments on referring, grounding, user-centric advanced tasks (comprising 9 subtasks $\times$ 5 platforms), GUIDE next-action prediction dataset, and GUI-World multi-platform benchmark demonstrate that Ferret-UI 2 significantly outperforms Ferret-UI, and also shows strong cross-platform transfer capabilities.
Authors: Tuan Duc Ngo, Peiye Zhuang, Chuang Gan, Evangelos Kalogerakis, Sergey Tulyakov, Hsin-Ying Lee, Chaoyang Wang
Abstract: Tracking dense 3D motion from monocular videos remains challenging, particularly when aiming for pixel-level precision over long sequences. We introduce DELTA, a novel method that efficiently tracks every pixel in 3D space, enabling accurate motion estimation across entire videos. Our approach leverages a joint global-local attention mechanism for reduced-resolution tracking, followed by a transformer-based upsampler to achieve high-resolution predictions. Unlike existing methods, which are limited by computational inefficiency or sparse tracking, DELTA delivers dense 3D tracking at scale, running over 8x faster than previous methods while achieving state-of-the-art accuracy. Furthermore, we explore the impact of depth representation on tracking performance and identify log-depth as the optimal choice. Extensive experiments demonstrate the superiority of DELTA on multiple benchmarks, achieving new state-of-the-art results in both 2D and 3D dense tracking tasks. Our method provides a robust solution for applications requiring fine-grained, long-term motion tracking in 3D space.
Authors: Sara Oblak, Despoina Paschalidou, Sanja Fidler, Matan Atzmon
Abstract: Reconstructing a dynamic scene from image inputs is a fundamental computer vision task with many downstream applications. Despite recent advancements, existing approaches still struggle to achieve high-quality reconstructions from unseen viewpoints and timestamps. This work introduces the ReMatching framework, designed to improve reconstruction quality by incorporating deformation priors into dynamic reconstruction models. Our approach advocates for velocity-field based priors, for which we suggest a matching procedure that can seamlessly supplement existing dynamic reconstruction pipelines. The framework is highly adaptable and can be applied to various dynamic representations. Moreover, it supports integrating multiple types of model priors and enables combining simpler ones to create more complex classes. Our evaluations on popular benchmarks involving both synthetic and real-world dynamic scenes demonstrate that augmenting current state-of-the-art methods with our approach leads to a clear improvement in reconstruction accuracy.
Authors: Liang Mi, Weijun Wang, Wenming Tu, Qingfeng He, Rui Kong, Xinyu Fang, Yazhu Dong, Yikang Zhang, Yunchun Li, Meng Li, Haipeng Dai, Guihai Chen, Yunxin Liu
Abstract: Large Multimodal Models (LMMs) have shown significant progress in various complex vision tasks with the solid linguistic and reasoning capacity inherited from large language models (LMMs). Low-rank adaptation (LoRA) offers a promising method to integrate external knowledge into LMMs, compensating for their limitations on domain-specific tasks. However, the existing LoRA model serving is excessively computationally expensive and causes extremely high latency. In this paper, we present an end-to-end solution that empowers diverse vision tasks and enriches vision applications with LoRA LMMs. Our system, VaLoRA, enables accurate and efficient vision tasks by 1) an accuracy-aware LoRA adapter generation approach that generates LoRA adapters rich in domain-specific knowledge to meet application-specific accuracy requirements, 2) an adaptive-tiling LoRA adapters batching operator that efficiently computes concurrent heterogeneous LoRA adapters, and 3) a flexible LoRA adapter orchestration mechanism that manages application requests and LoRA adapters to achieve the lowest average response latency. We prototype VaLoRA on five popular vision tasks on three LMMs. Experiment results reveal that VaLoRA improves 24-62% of the accuracy compared to the original LMMs and reduces 20-89% of the latency compared to the state-of-the-art LoRA model serving systems.
Authors: Abhisek Ray, Ayush Raj, Maheshkumar H. Kolekar
Abstract: Extracting multiscale contextual information and higher-order correlations among skeleton sequences using Graph Convolutional Networks (GCNs) alone is inadequate for effective action classification. Hypergraph convolution addresses the above issues but cannot harness the long-range dependencies. The transformer proves to be effective in capturing these dependencies and making complex contextual features accessible. We propose an Autoregressive Adaptive HyperGraph Transformer (AutoregAd-HGformer) model for in-phase (autoregressive and discrete) and out-phase (adaptive) hypergraph generation. The vector quantized in-phase hypergraph equipped with powerful autoregressive learned priors produces a more robust and informative representation suitable for hyperedge formation. The out-phase hypergraph generator provides a model-agnostic hyperedge learning technique to align the attributes with input skeleton embedding. The hybrid (supervised and unsupervised) learning in AutoregAd-HGformer explores the action-dependent feature along spatial, temporal, and channel dimensions. The extensive experimental results and ablation study indicate the superiority of our model over state-of-the-art hypergraph architectures on the NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.
Authors: Wanquan Feng, Jiawei Liu, Pengqi Tu, Tianhao Qi, Mingzhen Sun, Tianxiang Ma, Songtao Zhao, Siyu Zhou, Qian He
Abstract: Video generation technologies are developing rapidly and have broad potential applications. Among these technologies, camera control is crucial for generating professional-quality videos that accurately meet user expectations. However, existing camera control methods still suffer from several limitations, including control precision and the neglect of the control for subject motion dynamics. In this work, we propose I2VControl-Camera, a novel camera control method that significantly enhances controllability while providing adjustability over the strength of subject motion. To improve control precision, we employ point trajectory in the camera coordinate system instead of only extrinsic matrix information as our control signal. To accurately control and adjust the strength of subject motion, we explicitly model the higher-order components of the video trajectory expansion, not merely the linear terms, and design an operator that effectively represents the motion strength. We use an adapter architecture that is independent of the base model structure. Experiments on static and dynamic scenes show that our framework outperformances previous methods both quantitatively and qualitatively. The project page is: https://wanquanf.github.io/I2VControlCamera .
Authors: Ranjan Sapkota, Achyut Paudel, Manoj Karkee
Abstract: Currently, deep learning-based instance segmentation for various applications (e.g., Agriculture) is predominantly performed using a labor-intensive process involving extensive field data collection using sophisticated sensors, followed by careful manual annotation of images, presenting significant logistical and financial challenges to researchers and organizations. The process also slows down the model development and training process. In this study, we presented a novel method for deep learning-based instance segmentation of apples in commercial orchards that eliminates the need for labor-intensive field data collection and manual annotation. Utilizing a Large Language Model (LLM), we synthetically generated orchard images and automatically annotated them using the Segment Anything Model (SAM) integrated with a YOLO11 base model. This method significantly reduces reliance on physical sensors and manual data processing, presenting a major advancement in "Agricultural AI". The synthetic, auto-annotated dataset was used to train the YOLO11 model for Apple instance segmentation, which was then validated on real orchard images. The results showed that the automatically generated annotations achieved a Dice Coefficient of 0.9513 and an IoU of 0.9303, validating the accuracy and overlap of the mask annotations. All YOLO11 configurations, trained solely on these synthetic datasets with automated annotations, accurately recognized and delineated apples, highlighting the method's efficacy. Specifically, the YOLO11m-seg configuration achieved a mask precision of 0.902 and a mask mAP@50 of 0.833 on test images collected from a commercial orchard. Additionally, the YOLO11l-seg configuration outperformed other models in validation on 40 LLM-generated images, achieving the highest mask precision and mAP@50 metrics. Keywords: YOLO, SAM, SAMv2, YOLO11, YOLOv11, Segment Anything, YOLO-SAM
Authors: Shuchen Weng, Haojie Zheng, Peixuan Zhan, Yuchen Hong, Han Jiang, Si Li, Boxin Shi
Abstract: We introduce VIRES, a video instance repainting method with sketch and text guidance, enabling video instance repainting, replacement, generation, and removal. Existing approaches struggle with temporal consistency and accurate alignment with the provided sketch sequence. VIRES leverages the generative priors of text-to-video models to maintain temporal consistency and produce visually pleasing results. We propose the Sequential ControlNet with the standardized self-scaling, which effectively extracts structure layouts and adaptively captures high-contrast sketch details. We further augment the diffusion transformer backbone with the sketch attention to interpret and inject fine-grained sketch semantics. A sketch-aware encoder ensures that repainted results are aligned with the provided sketch sequence. Additionally, we contribute the VireSet, a dataset with detailed annotations tailored for training and evaluating video instance editing methods. Experimental results demonstrate the effectiveness of VIRES, which outperforms state-of-the-art methods in visual quality, temporal consistency, condition alignment, and human ratings. Project page:https://suimuc.github.io/suimu.github.io/projects/VIRES/
URLs: https://suimuc.github.io/suimu.github.io/projects/VIRES/
Authors: Youngjae Cho, Gwangyeol Kim, Sirojbek Safarov, Seongdeok Bang, Jaewoo Park
Abstract: Detecting anomalies in industrial settings is challenging due to the scarcity of labeled anomalous data. Generative models can mitigate this issue by synthesizing realistic defect samples, but existing approaches often fail to model the crucial interplay between defects and their background. This oversight leads to unrealistic anomalies, especially in scenarios where contextual consistency is essential (i.e., logical anomaly). To address this, we propose a novel background-aware defect generation framework, where the background influences defect denoising without affecting the background itself by ensuring realistic synthesis while preserving structural integrity. Our method leverages a disentanglement loss to separate the background' s denoising process from the defect, enabling controlled defect synthesis through DDIM Inversion. We theoretically demonstrate that our approach maintains background fidelity while generating contextually accurate defects. Extensive experiments on MVTec AD and MVTec Loco benchmarks validate our mehtod's superiority over existing techniques in both defect generation quality and anomaly detection performance.
Authors: Rui Zhou, Jingbin Liu, Junbin Xie, Jianyu Zhang, Yingze Hu, Jiele Zhao
Abstract: Visual-inertial odometry (VIO) is widely used in various fields, such as robots, drones, and autonomous vehicles. However, real-world scenes often feature dynamic objects, compromising the accuracy of VIO. The diversity and partial occlusion of these objects present a tough challenge for existing dynamic VIO methods. To tackle this challenge, we introduce ADUGS-VINS, which integrates an enhanced SORT algorithm along with a promptable foundation model into VIO, thereby improving pose estimation accuracy in environments with diverse dynamic objects and frequent occlusions. We evaluated our proposed method using multiple public datasets representing various scenes, as well as in a real-world scenario involving diverse dynamic objects. The experimental results demonstrate that our proposed method performs impressively in multiple scenarios, outperforming other state-of-the-art methods. This highlights its remarkable generalization and adaptability in diverse dynamic environments, showcasing its potential to handle various dynamic objects in practical applications.
Authors: Eerik Alamikkotervo, Henrik Toikka, Kari Tammi, Risto Ojala
Abstract: Robust road segmentation in all road conditions is required for safe autonomous driving and advanced driver assistance systems. Supervised deep learning methods provide accurate road segmentation in the domain of their training data but cannot be trusted in out-of-distribution scenarios. Including the whole distribution in the trainset is challenging as each sample must be labeled by hand. Trajectory-based self-supervised methods offer a potential solution as they can learn from the traversed route without manual labels. However, existing trajectory-based methods use learning schemes that rely only on the camera or only on the lidar. In this paper, trajectory-based learning is implemented jointly with lidar and camera for increased performance. Our method outperforms recent standalone camera- and lidar-based methods when evaluated with a challenging winter driving dataset including countryside and suburb driving scenes. The source code is available at https://github.com/eerik98/lidar-camera-road-autolabeling.git
URLs: https://github.com/eerik98/lidar-camera-road-autolabeling.git
Authors: Jingyu Lin, Jiaqi Gu, Lubin Fan, Bojian Wu, Yujing Lou, Renjie Chen, Ligang Liu, Jieping Ye
Abstract: Generating high-quality novel view renderings of 3D Gaussian Splatting (3DGS) in scenes featuring transient objects is challenging. We propose a novel hybrid representation, termed as HybridGS, using 2D Gaussians for transient objects per image and maintaining traditional 3D Gaussians for the whole static scenes. Note that, the 3DGS itself is better suited for modeling static scenes that assume multi-view consistency, but the transient objects appear occasionally and do not adhere to the assumption, thus we model them as planar objects from a single view, represented with 2D Gaussians. Our novel representation decomposes the scene from the perspective of fundamental viewpoint consistency, making it more reasonable. Additionally, we present a novel multi-view regulated supervision method for 3DGS that leverages information from co-visible regions, further enhancing the distinctions between the transients and statics. Then, we propose a straightforward yet effective multi-stage training strategy to ensure robust training and high-quality view synthesis across various settings. Experiments on benchmark datasets show our state-of-the-art performance of novel view synthesis in both indoor and outdoor scenes, even in the presence of distracting elements.
Authors: Zun Wang, Jialu Li, Yicong Hong, Songze Li, Kunchang Li, Shoubin Yu, Yi Wang, Yu Qiao, Yali Wang, Mohit Bansal, Limin Wang
Abstract: Creating high-quality data for training robust language-instructed agents is a long-lasting challenge in embodied AI. In this paper, we introduce a Self-Refining Data Flywheel (SRDF) that generates high-quality and large-scale navigational instruction-trajectory pairs by iteratively refining the data pool through the collaboration between two models, the instruction generator and the navigator, without any human-in-the-loop annotation. Specifically, SRDF starts with using a base generator to create an initial data pool for training a base navigator, followed by applying the trained navigator to filter the data pool. This leads to higher-fidelity data to train a better generator, which can, in turn, produce higher-quality data for training the next-round navigator. Such a flywheel establishes a data self-refining process, yielding a continuously improved and highly effective dataset for large-scale language-guided navigation learning. Our experiments demonstrate that after several flywheel rounds, the navigator elevates the performance boundary from 70% to 78% SPL on the classic R2R test set, surpassing human performance (76%) for the first time. Meanwhile, this process results in a superior generator, evidenced by a SPICE increase from 23.5 to 26.2, better than all previous VLN instruction generation methods. Finally, we demonstrate the scalability of our method through increasing environment and instruction diversity, and the generalization ability of our pre-trained navigator across various downstream navigation tasks, surpassing state-of-the-art methods by a large margin in all cases.
Authors: Rick Akkerman, Haiwen Feng, Michael J. Black, Dimitrios Tzionas, Victoria Fern\'andez Abrevaya
Abstract: Predicting the dynamics of interacting objects is essential for both humans and intelligent systems. However, existing approaches are limited to simplified, toy settings and lack generalizability to complex, real-world environments. Recent advances in generative models have enabled the prediction of state transitions based on interventions, but focus on generating a single future state which neglects the continuous dynamics resulting from the interaction. To address this gap, we propose InterDyn, a novel framework that generates videos of interactive dynamics given an initial frame and a control signal encoding the motion of a driving object or actor. Our key insight is that large video generation models can act as both neural renderers and implicit physics simulators, having learned interactive dynamics from large-scale video data. To effectively harness this capability, we introduce an interactive control mechanism that conditions the video generation process on the motion of the driving entity. Qualitative results demonstrate that InterDyn generates plausible, temporally consistent videos of complex object interactions while generalizing to unseen objects. Quantitative evaluations show that InterDyn outperforms baselines that focus on static state transitions. This work highlights the potential of leveraging video generative models as implicit physics engines. Code and trained models will be released at: https://interdyn.is.tue.mpg.de/
Authors: Wenyu Zhang, Wei En Ng, Lixin Ma, Yuwen Wang, Jungqi Zhao, Allison Koenecke, Boyang Li, Lu Wang
Abstract: Current vision-language models may grasp basic spatial cues and simple directions (e.g. left, right, front, back), but struggle with the multi-dimensional spatial reasoning necessary for human-like understanding and real-world applications. To address this gap, we develop SPHERE (Spatial Perception and Hierarchical Evaluation of REasoning), a hierarchical evaluation framework supported by a new human-annotated dataset. SPHERE systematically probes models across increasing levels of complexity, from fundamental skills to multi-skill integration and high-level reasoning that combines spatial, visual, and logical understanding. Benchmark evaluation of state-of-the-art models reveals significant deficiencies, especially in reasoning about distance and proximity, understanding both egocentric and allocentric perspectives, and applying spatial logic in physical contexts. These findings expose critical blind spots in existing models and underscore the need for more advanced spatial reasoning techniques, driving the development of vision-language models that align more closely with human spatial cognition. The SPHERE benchmark is available at https://github.com/zwenyu/SPHERE-VLM.
Authors: Yushu Li, Yongyi Su, Adam Goodge, Kui Jia, Xun Xu
Abstract: Vision-language models (VLMs) have revolutionized machine learning by leveraging large pre-trained models to tackle various downstream tasks. Although label, training, and data efficiency have improved, many state-of-the-art VLMs still require task-specific hyperparameter tuning and fail to fully exploit test samples. To overcome these challenges, we propose a graph-based approach for label-efficient adaptation and inference. Our method dynamically constructs a graph over text prompts, few-shot examples, and test samples, using label propagation for inference without task-specific tuning. Unlike existing zero-shot label propagation techniques, our approach requires no additional unlabeled support set and effectively leverages the test sample manifold through dynamic graph expansion. We further introduce a context-aware feature re-weighting mechanism to improve task adaptation accuracy. Additionally, our method supports efficient graph expansion, enabling real-time inductive inference. Extensive evaluations on downstream tasks, such as fine-grained categorization and out-of-distribution generalization, demonstrate the effectiveness of our approach. The source code is available at https://github.com/Yushu-Li/ECALP.
Authors: Subin Erattakulangara, Karthika Kelat, Katie Burnham, Rachel Balbi, Sarah E. Gerard, David Meyer, Sajan Goud Lingala
Abstract: Accurate segmentation of the vocal tract from magnetic resonance imaging (MRI) data is essential for various voice and speech applications. Manual segmentation is time intensive and susceptible to errors. This study aimed to evaluate the efficacy of deep learning algorithms for automatic vocal tract segmentation from 3D MRI.
Authors: Tongxu Zhang, Bei Wang
Abstract: In recent years, point cloud upsampling has been widely applied in fields such as 3D reconstruction. Our study investigates the factors influencing point cloud upsampling on both global and local levels through representation learning. Specifically, the paper inputs global and local information of the same point cloud model object into two encoders to extract these features, fuses them, and then feeds the combined features into an upsampling decoder. The goal is to address issues of sparsity and noise in point clouds by leveraging prior knowledge from both global and local inputs. And the proposed framework can be applied to any state-of-the-art point cloud upsampling neural network. Experiments were conducted on a series of autoencoder-based models utilizing deep learning, yielding interpretability for both global and local inputs, and it has been proven in the results that our proposed framework can further improve the upsampling effect in previous SOTA works. At the same time, the Saliency Map reflects the differences between global and local feature inputs, as well as the effectiveness of training with both inputs in parallel.
Authors: Branislava Jankovic, Sabina Jangirova, Waseem Ullah, Latif U. Khan, Mohsen Guizani
Abstract: Dangerous surroundings and difficult-to-reach landscapes introduce significant complications for adequate disaster management and recuperation. These problems can be solved by engaging unmanned aerial vehicles (UAVs) provided with embedded platforms and optical sensors. In this work, we focus on enabling onboard aerial image processing to ensure proper and real-time disaster detection. Such a setting usually causes challenges due to the limited hardware resources of UAVs. However, privacy, connectivity, and latency issues can be avoided. We suggest a UAV-assisted edge framework for disaster detection, leveraging our proposed model optimized for onboard real-time aerial image classification. The optimization of the model is achieved using post-training quantization techniques. To address the limited number of disaster cases in existing benchmark datasets and therefore ensure real-world adoption of our model, we construct a novel dataset, DisasterEye, featuring disaster scenes captured by UAVs and individuals on-site. Experimental results reveal the efficacy of our model, reaching high accuracy with lowered inference latency and memory use on both traditional machines and resource-limited devices. This shows that the scalability and adaptability of our method make it a powerful solution for real-time disaster management on resource-constrained UAV platforms.
Authors: Junyeob Baek, Yi-Fu Wu, Gautam Singh, Sungjin Ahn
Abstract: Humans have an innate ability to decompose their perceptions of the world into objects and their attributes, such as colors, shapes, and movement patterns. This cognitive process enables us to imagine novel futures by recombining familiar concepts. However, replicating this ability in artificial intelligence systems has proven challenging, particularly when it comes to modeling videos into compositional concepts and generating unseen, recomposed futures without relying on auxiliary data, such as text, masks, or bounding boxes. In this paper, we propose Dreamweaver, a neural architecture designed to discover hierarchical and compositional representations from raw videos and generate compositional future simulations. Our approach leverages a novel Recurrent Block-Slot Unit (RBSU) to decompose videos into their constituent objects and attributes. In addition, Dreamweaver uses a multi-future-frame prediction objective to capture disentangled representations for dynamic concepts more effectively as well as static concepts. In experiments, we demonstrate our model outperforms current state-of-the-art baselines for world modeling when evaluated under the DCI framework across multiple datasets. Furthermore, we show how the modularized concept representations of our model enable compositional imagination, allowing the generation of novel videos by recombining attributes from previously seen objects. cun-bjy.github.io/dreamweaver-website
Authors: Alexandre Filiot, Nicolas Dop, Oussama Tchita, Auriane Riou, R\'emy Dubois, Thomas Peeters, Daria Valter, Marin Scalbert, Charlie Saillard, Genevi\`eve Robin, Antoine Olivier
Abstract: In recent years, the advent of foundation models (FM) for digital pathology has relied heavily on scaling the pre-training datasets and the model size, yielding large and powerful models. While it resulted in improving the performance on diverse downstream tasks, it also introduced increased computational cost and inference time. In this work, we explore the distillation of a large foundation model into a smaller one, reducing the number of parameters by several orders of magnitude. Leveraging distillation techniques, our distilled model, H0-mini, achieves nearly comparable performance to large FMs at a significantly reduced inference cost. It is evaluated on several public benchmarks, achieving 3rd place on the HEST benchmark and 5th place on the EVA benchmark. Additionally, a robustness analysis conducted on the PLISM dataset demonstrates that our distilled model reaches excellent robustness to variations in staining and scanning conditions, significantly outperforming other state-of-the art models. This opens new perspectives to design lightweight and robust models for digital pathology, without compromising on performance.
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: Shilong Zhang, Wenbo Li, Shoufa Chen, Chongjian Ge, Peize Sun, Yida Zhang, Yi Jiang, Zehuan Yuan, Binyue Peng, Ping Luo
Abstract: DiT diffusion models have achieved great success in text-to-video generation, leveraging their scalability in model capacity and data scale. High content and motion fidelity aligned with text prompts, however, often require large model parameters and a substantial number of function evaluations (NFEs). Realistic and visually appealing details are typically reflected in high resolution outputs, further amplifying computational demands especially for single stage DiT models. To address these challenges, we propose a novel two stage framework, FlashVideo, which strategically allocates model capacity and NFEs across stages to balance generation fidelity and quality. In the first stage, prompt fidelity is prioritized through a low resolution generation process utilizing large parameters and sufficient NFEs to enhance computational efficiency. The second stage establishes flow matching between low and high resolutions, effectively generating fine details with minimal NFEs. Quantitative and visual results demonstrate that FlashVideo achieves state-of-the-art high resolution video generation with superior computational efficiency. Additionally, the two-stage design enables users to preview the initial output and accordingly adjust the prompt before committing to full-resolution generation, thereby significantly reducing computational costs and wait times as well as enhancing commercial viability.
Authors: Krishna Sri Ipsit Mantri, Carola-Bibiane Sch\"onlieb, Bruno Ribeiro, Chaim Baskin, Moshe Eliasof
Abstract: Pre-trained Vision Transformers now serve as powerful tools for computer vision. Yet, efficiently adapting them for multiple tasks remains a challenge that arises from the need to modify the rich hidden representations encoded by the learned weight matrices, without inducing interference between tasks. Current parameter-efficient methods like LoRA, which apply low-rank updates, force tasks to compete within constrained subspaces, ultimately degrading performance. We introduce DiTASK a novel Diffeomorphic Multi-Task Fine-Tuning approach that maintains pre-trained representations by preserving weight matrix singular vectors, while enabling task-specific adaptations through neural diffeomorphic transformations of the singular values. By following this approach, DiTASK enables both shared and task-specific feature modulations with minimal added parameters. Our theoretical analysis shows that DITASK achieves full-rank updates during optimization, preserving the geometric structure of pre-trained features, and establishing a new paradigm for efficient multi-task learning (MTL). Our experiments on PASCAL MTL and NYUD show that DiTASK achieves state-of-the-art performance across four dense prediction tasks, using 75% fewer parameters than existing methods. Our code is available [here](https://github.com/ipsitmantri/DiTASK).
Authors: Jianyi Peng, Fan Lu, Bin Li, Yuan Huang, Sanqing Qu, Guang Chen
Abstract: Image-to-point cloud cross-modal Visual Place Recognition (VPR) is a challenging task where the query is an RGB image, and the database samples are LiDAR point clouds. Compared to single-modal VPR, this approach benefits from the widespread availability of RGB cameras and the robustness of point clouds in providing accurate spatial geometry and distance information. However, current methods rely on intermediate modalities that capture either the vertical or horizontal field of view, limiting their ability to fully exploit the complementary information from both sensors. In this work, we propose an innovative initial retrieval + re-rank method that effectively combines information from range (or RGB) images and Bird's Eye View (BEV) images. Our approach relies solely on a computationally efficient global descriptor similarity search process to achieve re-ranking. Additionally, we introduce a novel similarity label supervision technique to maximize the utility of limited training data. Specifically, we employ points average distance to approximate appearance similarity and incorporate an adaptive margin, based on similarity differences, into the vanilla triplet loss. Experimental results on the KITTI dataset demonstrate that our method significantly outperforms state-of-the-art approaches.
Authors: Ke Cao, Jing Wang, Ao Ma, Jiasong Feng, Zhanjie Zhang, Xuanhua He, Shanyuan Liu, Bo Cheng, Dawei Leng, Yuhui Yin, Jie Zhang
Abstract: The Diffusion Transformer plays a pivotal role in advancing text-to-image and text-to-video generation, owing primarily to its inherent scalability. However, existing controlled diffusion transformer methods incur significant parameter and computational overheads and suffer from inefficient resource allocation due to their failure to account for the varying relevance of control information across different transformer layers. To address this, we propose the Relevance-Guided Efficient Controllable Generation framework, RelaCtrl, enabling efficient and resource-optimized integration of control signals into the Diffusion Transformer. First, we evaluate the relevance of each layer in the Diffusion Transformer to the control information by assessing the "ControlNet Relevance Score"-i.e., the impact of skipping each control layer on both the quality of generation and the control effectiveness during inference. Based on the strength of the relevance, we then tailor the positioning, parameter scale, and modeling capacity of the control layers to reduce unnecessary parameters and redundant computations. Additionally, to further improve efficiency, we replace the self-attention and FFN in the commonly used copy block with the carefully designed Two-Dimensional Shuffle Mixer (TDSM), enabling efficient implementation of both the token mixer and channel mixer. Both qualitative and quantitative experimental results demonstrate that our approach achieves superior performance with only 15% of the parameters and computational complexity compared to PixArt-delta.
Authors: Xinhang Liu, Chi-Keung Tang, Yu-Wing Tai
Abstract: Constructing photorealistic virtual worlds has applications across various fields, but it often requires the extensive labor of highly trained professionals to operate conventional 3D modeling software. To democratize this process, we introduce WorldCraft, a system where large language model (LLM) agents leverage procedural generation to create indoor and outdoor scenes populated with objects, allowing users to control individual object attributes and the scene layout using intuitive natural language commands. In our framework, a coordinator agent manages the overall process and works with two specialized LLM agents to complete the scene creation: ForgeIt, which integrates an ever-growing manual through auto-verification to enable precise customization of individual objects, and ArrangeIt, which formulates hierarchical optimization problems to achieve a layout that balances ergonomic and aesthetic considerations. Additionally, our pipeline incorporates a trajectory control agent, allowing users to animate the scene and operate the camera through natural language interactions. Our system is also compatible with off-the-shelf deep 3D generators to enrich scene assets. Through evaluations and comparisons with state-of-the-art methods, we demonstrate the versatility of WorldCraft, ranging from single-object customization to intricate, large-scale interior and exterior scene designs. This system empowers non-professionals to bring their creative visions to life.
Authors: Chunyu Zhao, Wentao Mu, Xian Zhou, Wenbo Liu, Fei Yan, Tao Deng
Abstract: Driver attention recognition in driving scenarios is a popular direction in traffic scene perception technology. It aims to understand human driver attention to focus on specific targets/objects in the driving scene. However, traffic scenes contain not only a large amount of visual information but also semantic information related to driving tasks. Existing methods lack attention to the actual semantic information present in driving scenes. Additionally, the traffic scene is a complex and dynamic process that requires constant attention to objects related to the current driving task. Existing models, influenced by their foundational frameworks, tend to have large parameter counts and complex structures. Therefore, this paper proposes a real-time saliency Mamba network based on the latest Mamba framework. As shown in Figure 1, our model uses very few parameters (0.08M, only 0.09~11.16% of other models), while maintaining SOTA performance or achieving over 98% of the SOTA model's performance.
Authors: Rui Li, Xiaowei Zhao
Abstract: As a novel and challenging task, referring segmentation combines computer vision and natural language processing to localize and segment objects based on textual descriptions. While referring image segmentation (RIS) has been extensively studied in natural images, little attention has been given to aerial imagery, particularly from unmanned aerial vehicles (UAVs). The unique challenges of UAV imagery, including complex spatial scales, occlusions, and varying object orientations, render existing RIS approaches ineffective. A key limitation has been the lack of UAV-specific datasets, as manually annotating pixel-level masks and generating textual descriptions is labour-intensive and time-consuming. To address this gap, we design an automatic labelling pipeline that leverages pre-existing UAV segmentation datasets and Multimodal Large Language Models (MLLM) for generating textual descriptions. Furthermore, we propose Aerial Referring Transformer (AeroReformer), a novel framework for UAV referring image segmentation (UAV-RIS), featuring a Vision-Language Cross-Attention Module (VLCAM) for effective cross-modal understanding and a Rotation-Aware Multi-Scale Fusion (RAMSF) decoder to enhance segmentation accuracy in aerial scenes. Extensive experiments on two newly developed datasets demonstrate the superiority of AeroReformer over existing methods, establishing a new benchmark for UAV-RIS. The datasets and code will be publicly available at: https://github.com/lironui/AeroReformer.
Authors: Yaxuan Huang, Xili Dai, Jianan Wang, Xianbiao Qi, Yixing Yuan, Xiangyu Yue
Abstract: Room layout estimation from multiple-perspective images is poorly investigated due to the complexities that emerge from multi-view geometry, which requires muti-step solutions such as camera intrinsic and extrinsic estimation, image matching, and triangulation. However, in 3D reconstruction, the advancement of recent 3D foundation models such as DUSt3R has shifted the paradigm from the traditional multi-step structure-from-motion process to an end-to-end single-step approach. To this end, we introduce Plane-DUSt3R, a novel method for multi-view room layout estimation leveraging the 3D foundation model DUSt3R. Plane-DUSt3R incorporates the DUSt3R framework and fine-tunes on a room layout dataset (Structure3D) with a modified objective to estimate structural planes. By generating uniform and parsimonious results, Plane-DUSt3R enables room layout estimation with only a single post-processing step and 2D detection results. Unlike previous methods that rely on single-perspective or panorama image, Plane-DUSt3R extends the setting to handle multiple-perspective images. Moreover, it offers a streamlined, end-to-end solution that simplifies the process and reduces error accumulation. Experimental results demonstrate that Plane-DUSt3R not only outperforms state-of-the-art methods on the synthetic dataset but also proves robust and effective on in the wild data with different image styles such as cartoon.Our code is available at: https://github.com/justacar/Plane-DUSt3R
Authors: Liangtao Shi, Ting Liu, Xiantao Hu, Yue Hu, Quanjun Yin, Richang Hong
Abstract: Visual grounding aims to ground an image region through natural language, which heavily relies on cross-modal alignment. Most existing methods transfer visual/linguistic knowledge separately by fully fine-tuning uni-modal pre-trained models, followed by a simple stack of visual-language transformers for multimodal fusion. However, these approaches not only limit adequate interaction between visual and linguistic contexts, but also incur significant computational costs. Therefore, to address these issues, we explore a step-wise multimodal fusion and adaption framework, namely SwimVG. Specifically, SwimVG proposes step-wise multimodal prompts (Swip) and cross-modal interactive adapters (CIA) for visual grounding, replacing the cumbersome transformer stacks for multimodal fusion. Swip can improve {the} alignment between the vision and language representations step by step, in a token-level fusion manner. In addition, weight-level CIA further promotes multimodal fusion by cross-modal interaction. Swip and CIA are both parameter-efficient paradigms, and they fuse the cross-modal features from shallow to deep layers gradually. Experimental results on four widely-used benchmarks demonstrate that SwimVG achieves remarkable abilities and considerable benefits in terms of efficiency. Our code is available at https://github.com/liuting20/SwimVG.
Authors: Bingxuan Li, Yiwei Wang, Jiuxiang Gu, Kai-Wei Chang, Nanyun Peng
Abstract: Chart generation aims to generate code to produce charts satisfying the desired visual properties, e.g., texts, layout, color, and type. It has great potential to empower the automatic professional report generation in financial analysis, research presentation, education, and healthcare. In this work, we build a vision-language model (VLM) based multi-agent framework for effective automatic chart generation. Generating high-quality charts requires both strong visual design skills and precise coding capabilities that embed the desired visual properties into code. Such a complex multi-modal reasoning process is difficult for direct prompting of VLMs. To resolve these challenges, we propose METAL, a multi-agent framework that decomposes the task of chart generation into the iterative collaboration among specialized agents. METAL achieves 5.2% improvement over the current best result in the chart generation task. The METAL framework exhibits the phenomenon of test-time scaling: its performance increases monotonically as the logarithmic computational budget grows from 512 to 8192 tokens. In addition, we find that separating different modalities during the critique process of METAL boosts the self-correction capability of VLMs in the multimodal context.
Authors: Yunpeng Gao, Chenhui Li, Zhongrui You, Junli Liu, Zhen Li, Pengan Chen, Qizhi Chen, Zhonghan Tang, Liansheng Wang, Penghui Yang, Yiwen Tang, Yuhang Tang, Shuai Liang, Songyi Zhu, Ziqin Xiong, Yifei Su, Xinyi Ye, Jianan Li, Yan Ding, Dong Wang, Zhigang Wang, Bin Zhao, Xuelong Li
Abstract: Vision-Language Navigation (VLN) aims to guide agents through an environment by leveraging both language instructions and visual cues, playing a pivotal role in embodied AI. Indoor VLN has been extensively studied, whereas outdoor aerial VLN remains underexplored. The potential reason is that outdoor aerial view encompasses vast areas, making data collection more challenging, which results in a lack of benchmarks. To address this problem, we propose OpenFly, a platform comprising a versatile toolchain and large-scale benchmark for aerial VLN. Firstly, we develop a highly automated toolchain for data collection, enabling automatic point cloud acquisition, scene semantic segmentation, flight trajectory creation, and instruction generation. Secondly, based on the toolchain, we construct a large-scale aerial VLN dataset with 100k trajectories, covering diverse heights and lengths across 18 scenes. The corresponding visual data are generated using various rendering engines and advanced techniques, including Unreal Engine, GTA V, Google Earth, and 3D Gaussian Splatting (3D GS). All data exhibit high visual quality. Particularly, 3D GS supports real-to-sim rendering, further enhancing the realism of the dataset. Thirdly, we propose OpenFly-Agent, a keyframe-aware VLN model, which takes language instructions, current observations, and historical keyframes as input, and outputs flight actions directly. Extensive analyses and experiments are conducted, showcasing the superiority of our OpenFly platform and OpenFly-Agent. The toolchain, dataset, and codes will be open-sourced.
Authors: Guangfeng Jiang, Jun Liu, Yongxuan Lv, Yuzhi Wu, Xianfei Li, Wenlong Liao, Tao He, Pai Peng
Abstract: Outdoor LiDAR point cloud 3D instance segmentation is a crucial task in autonomous driving. However, it requires laborious human efforts to annotate the point cloud for training a segmentation model. To address this challenge, we propose a YoCo framework, which generates 3D pseudo labels using minimal coarse click annotations in the bird's eye view plane. It is a significant challenge to produce high-quality pseudo labels from sparse annotations. Our YoCo framework first leverages vision foundation models combined with geometric constraints from point clouds to enhance pseudo label generation. Second, a temporal and spatial-based label updating module is designed to generate reliable updated labels. It leverages predictions from adjacent frames and utilizes the inherent density variation of point clouds (dense near, sparse far). Finally, to further improve label quality, an IoU-guided enhancement module is proposed, replacing pseudo labels with high-confidence and high-IoU predictions. Experiments on the Waymo dataset demonstrate YoCo's effectiveness and generality, achieving state-of-the-art performance among weakly supervised methods and surpassing fully supervised Cylinder3D. Additionally, the YoCo is suitable for various networks, achieving performance comparable to fully supervised methods with minimal fine-tuning using only 0.8% of the fully labeled data, significantly reducing annotation costs.
Authors: Yimin Zhu, Linlin Xu
Abstract: Data augmentation effectively addresses the imbalanced-small sample data (ISSD) problem in hyperspectral image classification (HSIC). While most methodologies extend features in the latent space, few leverage text-driven generation to create realistic and diverse samples. Recently, text-guided diffusion models have gained significant attention due to their ability to generate highly diverse and high-quality images based on text prompts in natural image synthesis. Motivated by this, this paper proposes Txt2HSI-LDM(VAE), a novel language-informed hyperspectral image synthesis method to address the ISSD in HSIC. The proposed approach uses a denoising diffusion model, which iteratively removes Gaussian noise to generate hyperspectral samples conditioned on textual descriptions. First, to address the high-dimensionality of hyperspectral data, a universal variational autoencoder (VAE) is designed to map the data into a low-dimensional latent space, which provides stable features and reduces the inference complexity of diffusion model. Second, a semi-supervised diffusion model is designed to fully take advantage of unlabeled data. Random polygon spatial clipping (RPSC) and uncertainty estimation of latent feature (LF-UE) are used to simulate the varying degrees of mixing. Third, the VAE decodes HSI from latent space generated by the diffusion model with the language conditions as input. In our experiments, we fully evaluate synthetic samples' effectiveness from statistical characteristics and data distribution in 2D-PCA space. Additionally, visual-linguistic cross-attention is visualized on the pixel level to prove that our proposed model can capture the spatial layout and geometry of the generated data. Experiments demonstrate that the performance of the proposed Txt2HSI-LDM(VAE) surpasses the classical backbone models, state-of-the-art CNNs, and semi-supervised methods.
Authors: Jiaxing Li, Lin Jiang, Zeqi Ma, Kaihang Jiang, Xiaozhao Fang, Jie Wen
Abstract: Deep online cross-modal hashing has gained much attention from researchers recently, as its promising applications with low storage requirement, fast retrieval efficiency and cross modality adaptive, etc. However, there still exists some technical hurdles that hinder its applications, e.g., 1) how to extract the coexistent semantic relevance of cross-modal data, 2) how to achieve competitive performance when handling the real time data streams, 3) how to transfer the knowledge learned from offline to online training in a lightweight manner. To address these problems, this paper proposes a lightweight contrastive distilled hashing (LCDH) for cross-modal retrieval, by innovatively bridging the offline and online cross-modal hashing by similarity matrix approximation in a knowledge distillation framework. Specifically, in the teacher network, LCDH first extracts the cross-modal features by the contrastive language-image pre-training (CLIP), which are further fed into an attention module for representation enhancement after feature fusion. Then, the output of the attention module is fed into a FC layer to obtain hash codes for aligning the sizes of similarity matrices for online and offline training. In the student network, LCDH extracts the visual and textual features by lightweight models, and then the features are fed into a FC layer to generate binary codes. Finally, by approximating the similarity matrices, the performance of online hashing in the lightweight student network can be enhanced by the supervision of coexistent semantic relevance that is distilled from the teacher network. Experimental results on three widely used datasets demonstrate that LCDH outperforms some state-of-the-art methods.
Authors: Zijie Zhou, Zhangshuo Qi, Luqi Cheng, Guangming Xiong
Abstract: Robust and accurate localization is critical for autonomous driving. Traditional GNSS-based localization methods suffer from signal occlusion and multipath effects in urban environments. Meanwhile, methods relying on high-definition (HD) maps are constrained by the high costs associated with the construction and maintenance of HD maps. Standard-definition (SD) maps-based methods, on the other hand, often exhibit unsatisfactory performance or poor generalization ability due to overfitting. To address these challenges, we propose SegLocNet, a multimodal GNSS-free localization network that achieves precise localization using bird's-eye-view (BEV) semantic segmentation. SegLocNet employs a BEV segmentation network to generate semantic maps from multiple sensor inputs, followed by an exhaustive matching process to estimate the vehicle's ego pose. This approach avoids the limitations of regression-based pose estimation and maintains high interpretability and generalization. By introducing a unified map representation, our method can be applied to both HD and SD maps without any modifications to the network architecture, thereby balancing localization accuracy and area coverage. Extensive experiments on the nuScenes and Argoverse datasets demonstrate that our method outperforms the current state-of-the-art methods, and that our method can accurately estimate the ego pose in urban environments without relying on GNSS, while maintaining strong generalization ability. Our code and pre-trained model will be released publicly.
Authors: Xuzheng Yang, Junzhuo Liu, Peng Wang, Guoqing Wang, Yang Yang, Heng Tao Shen
Abstract: Referring Expression Comprehension (REC) is a foundational cross-modal task that evaluates the interplay of language understanding, image comprehension, and language-to-image grounding. To advance this field, we introduce a new REC dataset with two key features. First, it is designed with controllable difficulty levels, requiring fine-grained reasoning across object categories, attributes, and relationships. Second, it incorporates negative text and images generated through fine-grained editing, explicitly testing a model's ability to reject non-existent targets, an often-overlooked yet critical challenge in existing datasets. To address fine-grained compositional REC, we propose novel methods based on a Specialist-MLLM collaboration framework, leveraging the complementary strengths of them: Specialist Models handle simpler tasks efficiently, while MLLMs are better suited for complex reasoning. Based on this synergy, we introduce two collaborative strategies. The first, Slow-Fast Adaptation (SFA), employs a routing mechanism to adaptively delegate simple tasks to Specialist Models and complex tasks to MLLMs. Additionally, common error patterns in both models are mitigated through a target-refocus strategy. The second, Candidate Region Selection (CRS), generates multiple bounding box candidates based on Specialist Model and uses the advanced reasoning capabilities of MLLMs to identify the correct target. Extensive experiments on our dataset and other challenging compositional benchmarks validate the effectiveness of our approaches. The SFA strategy achieves a trade-off between localization accuracy and efficiency, and the CRS strategy greatly boosts the performance of both Specialist Models and MLLMs. We aim for this work to offer valuable insights into solving complex real-world tasks by strategically combining existing tools for maximum effectiveness, rather than reinventing them.
Authors: Arthur Pignet, John Klein, Genevieve Robin, Antoine Olivier
Abstract: Deploying digital pathology models across medical centers is challenging due to distribution shifts. Recent advances in domain generalization improve model transferability in terms of aggregated performance measured by the Area Under Curve (AUC). However, clinical regulations often require to control the transferability of other metrics, such as prescribed sensitivity levels. We introduce a novel approach to control the sensitivity of whole slide image (WSI) classification models, based on optimal transport and Multiple Instance Learning (MIL). Validated across multiple cohorts and tasks, our method enables robust sensitivity control with only a handful of calibration samples, providing a practical solution for reliable deployment of computational pathology systems.
Authors: Qingsen Yan, Yixu Feng, Cheng Zhang, Guansong Pang, Kangbiao Shi, Peng Wu, Wei Dong, Jinqiu Sun, Yanning Zhang
Abstract: Low-Light Image Enhancement (LLIE) is a crucial computer vision task that aims to restore detailed visual information from corrupted low-light images. Many existing LLIE methods are based on standard RGB (sRGB) space, which often produce color bias and brightness artifacts due to inherent high color sensitivity in sRGB. While converting the images using Hue, Saturation and Value (HSV) color space helps resolve the brightness issue, it introduces significant red and black noise artifacts. To address this issue, we propose a new color space for LLIE, namely Horizontal/Vertical-Intensity (HVI), defined by polarized HS maps and learnable intensity. The former enforces small distances for red coordinates to remove the red artifacts, while the latter compresses the low-light regions to remove the black artifacts. To fully leverage the chromatic and intensity information, a novel Color and Intensity Decoupling Network (CIDNet) is further introduced to learn accurate photometric mapping function under different lighting conditions in the HVI space. Comprehensive results from benchmark and ablation experiments show that the proposed HVI color space with CIDNet outperforms the state-of-the-art methods on 10 datasets. The code is available at https://github.com/Fediory/HVI-CIDNet.
Authors: Xuangeng Chu, Nabarun Goswami, Ziteng Cui, Hanqin Wang, Tatsuya Harada
Abstract: Speech-driven 3D facial animation aims to generate realistic lip movements and facial expressions for 3D head models from arbitrary audio clips. Although existing diffusion-based methods are capable of producing natural motions, their slow generation speed limits their application potential. In this paper, we introduce a novel autoregressive model that achieves real-time generation of highly synchronized lip movements and realistic head poses and eye blinks by learning a mapping from speech to a multi-scale motion codebook. Furthermore, our model can adapt to unseen speaking styles using sample motion sequences, enabling the creation of 3D talking avatars with unique personal styles beyond the identities seen during training. Extensive evaluations and user studies demonstrate that our method outperforms existing approaches in lip synchronization accuracy and perceived quality.
Authors: Jianyu Zhang, L\'eon Bottou
Abstract: It is impossible today to pretend that the practice of machine learning is always compatible with the idea that training and testing data follow the same distribution. Several authors have recently used ensemble techniques to show how scenarios involving multiple data distributions are best served by representations that are both richer than those obtained by regularizing for the best in-distribution performance, and richer than those obtained under the influence of the implicit sparsity bias of common stochastic gradient procedures. This contribution investigates the use of very high dropout rates instead of ensembles to obtain such rich representations. Although training a deep network from scratch using such dropout rates is virtually impossible, fine-tuning a large pre-trained model under such conditions is not only possible but also achieves out-of-distribution performances that exceed those of both ensembles and weight averaging methods such as model soups. This result has practical significance because the importance of the fine-tuning scenario has considerably grown in recent years. This result also provides interesting insights on the nature of rich representations and on the intrinsically linear nature of fine-tuning a large network using a comparatively small dataset.
Authors: Cheng Yang, Chufan Shi, Yaxin Liu, Bo Shui, Junjie Wang, Mohan Jing, Linran Xu, Xinyu Zhu, Siheng Li, Yuxiang Zhang, Gongye Liu, Xiaomei Nie, Deng Cai, Yujiu Yang
Abstract: We introduce a new benchmark, ChartMimic, aimed at assessing the visually-grounded code generation capabilities of large multimodal models (LMMs). ChartMimic utilizes information-intensive visual charts and textual instructions as inputs, requiring LMMs to generate the corresponding code for chart rendering. ChartMimic includes 4,800 human-curated (figure, instruction, code) triplets, which represent the authentic chart use cases found in scientific papers across various domains (e.g., Physics, Computer Science, Economics, etc). These charts span 18 regular types and 4 advanced types, diversifying into 201 subcategories. Furthermore, we propose multi-level evaluation metrics to provide an automatic and thorough assessment of the output code and the rendered charts. Unlike existing code generation benchmarks, ChartMimic places emphasis on evaluating LMMs' capacity to harmonize a blend of cognitive capabilities, encompassing visual understanding, code generation, and cross-modal reasoning. The evaluation of $3$ proprietary models and 14 open-weight models highlights the substantial challenges posed by ChartMimic. Even the advanced GPT-4o, InternVL2-Llama3-76B only achieved an average score across Direct Mimic and Customized Mimic tasks of 82.2 and 61.6, respectively, indicating significant room for improvement. We anticipate that ChartMimic will inspire the development of LMMs, advancing the pursuit of artificial general intelligence.
Authors: Abdul-Lateef Haji-Ali, Marcelo Pereyra, Luke Shaw, Konstantinos Zygalakis
Abstract: Generative diffusion models have recently emerged as a powerful strategy to perform stochastic sampling in Bayesian inverse problems, delivering remarkably accurate solutions for a wide range of challenging applications. However, diffusion models often require a large number of neural function evaluations per sample in order to deliver accurate posterior samples. As a result, using diffusion models as stochastic samplers for Monte Carlo integration in Bayesian computation can be highly computationally expensive, particularly in applications that require a substantial number of Monte Carlo samples for conducting uncertainty quantification analyses. This cost is especially high in large-scale inverse problems such as computational imaging, which rely on large neural networks that are expensive to evaluate. With quantitative imaging applications in mind, this paper presents a Multilevel Monte Carlo strategy that significantly reduces the cost of Bayesian computation with diffusion models. This is achieved by exploiting cost-accuracy trade-offs inherent to diffusion models to carefully couple models of different levels of accuracy in a manner that significantly reduces the overall cost of the calculation, without reducing the final accuracy. The proposed approach achieves a $4\times$-to-$8\times$ reduction in computational cost w.r.t. standard techniques across three benchmark imaging problems.
Authors: Xiao Yu, Baolin Peng, Vineeth Vajipey, Hao Cheng, Michel Galley, Jianfeng Gao, Zhou Yu
Abstract: Autonomous agents have demonstrated significant potential in automating complex multistep decision-making tasks. However, even state-of-the-art vision-language models (VLMs), such as GPT-4o, still fall short of human-level performance, particularly in intricate web environments and long-horizon tasks. To address these limitations, we present ExACT, an approach to combine test-time search and self-learning to build o1-like models for agentic applications. We first introduce Reflective Monte Carlo Tree Search (R-MCTS), a novel test time algorithm designed to enhance AI agents' ability to explore decision space on the fly. R-MCTS extends traditional MCTS by 1) incorporating contrastive reflection, allowing agents to learn from past interactions and dynamically improve their search efficiency; and 2) using multi-agent debate for reliable state evaluation. Next, we introduce Exploratory Learning, a novel learning strategy to teach agents to search at inference time without relying on any external search algorithms. On the challenging VisualWebArena benchmark, our GPT-4o based R-MCTS agent achieves a 6% to 30% relative improvement across various tasks compared to the previous state-of-the-art. Additionally, we show that the knowledge and experience gained from test-time search can be effectively transferred back to GPT-4o via fine-tuning. After Exploratory Learning, GPT-4o 1) demonstrates the ability to explore the environment, evaluate a state, and backtrack to viable ones when it detects that the current state cannot lead to success, and 2) matches 87% of R-MCTS's performance while using significantly less compute. Notably, our work demonstrates the compute scaling properties in both training - data collection with R-MCTS - and testing time. These results suggest a promising research direction to enhance VLMs' capabilities for agentic applications via test-time search and self-learning.
Authors: Pengcheng Chen, Wenhao Li, Nicole Gunderson, Jeremy Ruthberg, Randall Bly, Zhenglong Sun, Waleed M. Abuzeid, Eric J. Seibel
Abstract: In endoscopic sinus surgery (ESS), intraoperative CT (iCT) offers valuable intraoperative assessment but is constrained by slow deployment and radiation exposure, limiting its clinical utility. Endoscope-based monocular 3D reconstruction is a promising alternative; however, existing techniques often struggle to achieve the submillimeter precision required for dense reconstruction. In this work, we propose an iterative online learning approach that leverages Neural Radiance Fields (NeRF) as an intermediate representation, enabling monocular depth estimation and 3D reconstruction without relying on prior medical data. Our method attains a point-to-point accuracy below 0.5 mm, with a demonstrated theoretical depth accuracy of 0.125 $\pm$ 0.443 mm. We validate our approach across synthetic, phantom, and real endoscopic scenarios, confirming its accuracy and reliability. These results underscore the potential of our pipeline as an iCT alternative, meeting the demanding submillimeter accuracy standards required in ESS.
Authors: Yongyi Su, Yushu Li, Nanqing Liu, Kui Jia, Xulei Yang, Chuan-Sheng Foo, Xun Xu
Abstract: Test-time adaptation (TTA) updates the model weights during the inference stage using testing data to enhance generalization. However, this practice exposes TTA to adversarial risks. Existing studies have shown that when TTA is updated with crafted adversarial test samples, also known as test-time poisoned data, the performance on benign samples can deteriorate. Nonetheless, the perceived adversarial risk may be overstated if the poisoned data is generated under overly strong assumptions. In this work, we first review realistic assumptions for test-time data poisoning, including white-box versus grey-box attacks, access to benign data, attack order, and more. We then propose an effective and realistic attack method that better produces poisoned samples without access to benign samples, and derive an effective in-distribution attack objective. We also design two TTA-aware attack objectives. Our benchmarks of existing attack methods reveal that the TTA methods are more robust than previously believed. In addition, we analyze effective defense strategies to help develop adversarially robust TTA methods. The source code is available at https://github.com/Gorilla-Lab-SCUT/RTTDP.
Authors: Ke Wang, Nikolaos Dimitriadis, Alessandro Favero, Guillermo Ortiz-Jimenez, Francois Fleuret, Pascal Frossard
Abstract: Fine-tuning pre-trained models has become the standard approach to endow them with specialized knowledge, but it poses fundamental challenges. In particular, \textit{(i)} fine-tuning often leads to catastrophic forgetting, where improvements on a target domain degrade generalization on other tasks, and \textit{(ii)} merging fine-tuned checkpoints from disparate tasks can lead to significant performance loss. To address these challenges, we introduce LiNeS, Layer-increasing Network Scaling, a post-training editing technique designed to preserve pre-trained generalization while enhancing fine-tuned task performance. LiNeS scales parameter updates linearly based on their layer depth within the network, maintaining shallow layers close to their pre-trained values to preserve general features while allowing deeper layers to retain task-specific representations. In multi-task model merging scenarios, layer-wise scaling of merged parameters reduces negative task interference. LiNeS demonstrates significant improvements in both single-task and multi-task settings across various benchmarks in vision and natural language processing. It mitigates forgetting, enhances out-of-distribution generalization, integrates seamlessly with existing multi-task model merging baselines improving their performance across benchmarks and model sizes, and can boost generalization when merging LLM policies aligned with different rewards via RLHF. Our method is simple to implement, computationally efficient and complementary to many existing techniques. Our source code is available at https://github.com/wang-kee/LiNeS
Authors: Bingyu Yang, Qingyao Tian, Huai Liao, Xinyan Huang, Jinlin Wu, Jingdi Hu, Hongbin Liu
Abstract: Continuous and accurate segmentation of airways in chest CT images is essential for preoperative planning and real-time bronchoscopy navigation. Despite advances in deep learning for medical image segmentation, maintaining airway continuity remains a challenge, particularly due to intra-class imbalance between large and small branches and blurred CT scan details. To address these challenges, we propose a progressive curriculum learning pipeline and a Scale-Enhanced U-Net (SE-UNet) to enhance segmentation continuity. Specifically, our progressive curriculum learning pipeline consists of three stages: extracting main airways, identifying small airways, and repairing discontinuities. The cropping sampling strategy in each stage reduces feature interference between airways of different scales, effectively addressing the challenge of intra-class imbalance. In the third training stage, we present an Adaptive Topology-Responsive Loss (ATRL) to guide the network to focus on airway continuity. The progressive training pipeline shares the same SE-UNet, integrating multi-scale inputs and Detail Information Enhancers (DIEs) to enhance information flow and effectively capture the intricate details of small airways. Additionally, we propose a robust airway tree parsing method and hierarchical evaluation metrics to provide more clinically relevant and precise analysis. Experiments on both in-house and public datasets demonstrate that our method outperforms existing approaches, significantly improving the accuracy of small airways and the completeness of the airway tree. The code will be released upon publication.
Authors: Ajinkya Deshpande, Deep Gupta, Ankit Bhurane, Nisha Meshram, Sneha Singh, Petia Radeva
Abstract: Hepatocellular carcinoma (HCC) is a common type of liver cancer whose early-stage diagnosis is a common challenge, mainly due to the manual assessment of hematoxylin and eosin-stained whole slide images, which is a time-consuming process and may lead to variability in decision-making. For accurate detection of HCC, we propose a hybrid deep learning-based architecture that uses transfer learning to extract the features from pre-trained convolutional neural network (CNN) models and a classifier made up of a sequence of fully connected layers. This study uses a publicly available The Cancer Genome Atlas Hepatocellular Carcinoma (TCGA-LIHC)database (n=491) for model development and database of Kasturba Gandhi Medical College (KMC), India for validation. The pre-processing step involves patch extraction, colour normalization, and augmentation that results in 3920 patches for the TCGA dataset. The developed hybrid deep neural network consisting of a CNN-based pre-trained feature extractor and a customized artificial neural network-based classifier is trained using five-fold cross-validation. For this study, eight different state-of-the-art models are trained and tested as feature extractors for the proposed hybrid model. The proposed hybrid model with ResNet50-based feature extractor provided the sensitivity, specificity, F1-score, accuracy, and AUC of 100.00%, 100.00%, 100.00%, 100.00%, and 1.00, respectively on the TCGA database. On the KMC database, EfficientNetb3 resulted in the optimal choice of the feature extractor giving sensitivity, specificity, F1-score, accuracy, and AUC of 96.97, 98.85, 96.71, 96.71, and 0.99, respectively. The proposed hybrid models showed improvement in accuracy of 2% and 4% over the pre-trained models in TCGA-LIHC and KMC databases.
Authors: Marien Renaud, Arthur Leclaire, Nicolas Papadakis
Abstract: One key ingredient of image restoration is to define a realistic prior on clean images to complete the missing information in the observation. State-of-the-art restoration methods rely on a neural network to encode this prior. Moreover, typical image distributions are invariant to some set of transformations, such as rotations or flips. However, most deep architectures are not designed to represent an invariant image distribution. Recent works have proposed to overcome this difficulty by including equivariance properties within a Plug-and-Play paradigm. In this work, we propose a unified framework named Equivariant Regularization by Denoising (ERED) based on equivariant denoisers and stochastic optimization. We analyze the convergence of this algorithm and discuss its practical benefit.
Authors: Eliot Xing, Vernon Luk, Jean Oh
Abstract: Recent advances in GPU-based parallel simulation have enabled practitioners to collect large amounts of data and train complex control policies using deep reinforcement learning (RL), on commodity GPUs. However, such successes for RL in robotics have been limited to tasks sufficiently simulated by fast rigid-body dynamics. Simulation techniques for soft bodies are comparatively several orders of magnitude slower, thereby limiting the use of RL due to sample complexity requirements. To address this challenge, this paper presents both a novel RL algorithm and a simulation platform to enable scaling RL on tasks involving rigid bodies and deformables. We introduce Soft Analytic Policy Optimization (SAPO), a maximum entropy first-order model-based actor-critic RL algorithm, which uses first-order analytic gradients from differentiable simulation to train a stochastic actor to maximize expected return and entropy. Alongside our approach, we develop Rewarped, a parallel differentiable multiphysics simulation platform that supports simulating various materials beyond rigid bodies. We re-implement challenging manipulation and locomotion tasks in Rewarped, and show that SAPO outperforms baselines over a range of tasks that involve interaction between rigid bodies, articulations, and deformables. Additional details at https://rewarped.github.io/.
Authors: Arth Shukla, Stone Tao, Hao Su
Abstract: High-quality benchmarks are the foundation for embodied AI research, enabling significant advancements in long-horizon navigation, manipulation and rearrangement tasks. However, as frontier tasks in robotics get more advanced, they require faster simulation speed, more intricate test environments, and larger demonstration datasets. To this end, we present MS-HAB, a holistic benchmark for low-level manipulation and in-home object rearrangement. First, we provide a GPU-accelerated implementation of the Home Assistant Benchmark (HAB). We support realistic low-level control and achieve over 3x the speed of prior magical grasp implementations at a fraction of the GPU memory usage. Second, we train extensive reinforcement learning (RL) and imitation learning (IL) baselines for future work to compare against. Finally, we develop a rule-based trajectory filtering system to sample specific demonstrations from our RL policies which match predefined criteria for robot behavior and safety. Combining demonstration filtering with our fast environments enables efficient, controlled data generation at scale.
Authors: Eichi Takaya, Shinnosuke Yamamoto
Abstract: Annotation of medical images, such as MRI and CT scans, is crucial for evaluating treatment efficacy and planning radiotherapy. However, the extensive workload of medical professionals limits their ability to annotate large image datasets, posing a bottleneck for AI applications in medical imaging. To address this, we propose In-context Cascade Segmentation (ICS), a novel method that minimizes annotation requirements while achieving high segmentation accuracy for sequential medical images. ICS builds on the UniverSeg framework, which performs few-shot segmentation using support images without additional training. By iteratively adding the inference results of each slice to the support set, ICS propagates information forward and backward through the sequence, ensuring inter-slice consistency. We evaluate the proposed method on the HVSMR dataset, which includes segmentation tasks for eight cardiac regions. Experimental results demonstrate that ICS significantly improves segmentation performance in complex anatomical regions, particularly in maintaining boundary consistency across slices, compared to baseline methods. The study also highlights the impact of the number and position of initial support slices on segmentation accuracy. ICS offers a promising solution for reducing annotation burdens while delivering robust segmentation results, paving the way for its broader adoption in clinical and research applications.
Authors: Masanari Ohi, Masahiro Kaneko, Naoaki Okazaki, Nakamasa Inoue
Abstract: Vision-language models (VLMs) have shown impressive abilities across a range of multi-modal tasks. However, existing metrics for evaluating the quality of text generated by VLMs typically focus on an overall evaluation for a specific task, such as image captioning. While the overall evaluation is essential for any task, the criteria prioritized can differ depending on the task, making it challenging for current metrics to adapt to multi-task scenarios. To address this limitation, we propose HarmonicEval, a reference-free comprehensive evaluation metric that aggregates criterion-wise scores to produce the overall score in a bottom-up manner. Furthermore, we construct the Multi-task Multi-criteria Human Evaluation (MMHE) dataset, which comprises 18,000 expert human judgments across four multi-modal tasks. Our experiments demonstrate that HarmonicEval achieves higher correlations with human judgments than conventional metrics while providing numerical scores for each criterion.
Authors: Anh Bui, Trang Vu, Long Vuong, Trung Le, Paul Montague, Tamas Abraham, Junae Kim, Dinh Phung
Abstract: Concept erasure has emerged as a promising technique for mitigating the risk of harmful content generation in diffusion models by selectively unlearning undesirable concepts. The common principle of previous works to remove a specific concept is to map it to a fixed generic concept, such as a neutral concept or just an empty text prompt. In this paper, we demonstrate that this fixed-target strategy is suboptimal, as it fails to account for the impact of erasing one concept on the others. To address this limitation, we model the concept space as a graph and empirically analyze the effects of erasing one concept on the remaining concepts. Our analysis uncovers intriguing geometric properties of the concept space, where the influence of erasing a concept is confined to a local region. Building on this insight, we propose the Adaptive Guided Erasure (AGE) method, which \emph{dynamically} selects optimal target concepts tailored to each undesirable concept, minimizing unintended side effects. Experimental results show that AGE significantly outperforms state-of-the-art erasure methods on preserving unrelated concepts while maintaining effective erasure performance. Our code is published at {https://github.com/tuananhbui89/Adaptive-Guided-Erasure}.
URLs: https://github.com/tuananhbui89/Adaptive-Guided-Erasure
Authors: Xin Gao, Jian Pu
Abstract: Multi-View Representation Learning (MVRL) aims to derive a unified representation from multi-view data by leveraging shared and complementary information across views. However, when views are irregularly missing, the incomplete data can lead to representations that lack sufficiency and consistency. To address this, we propose Multi-View Permutation of Variational Auto-Encoders (MVP), which excavates invariant relationships between views in incomplete data. MVP establishes inter-view correspondences in the latent space of Variational Auto-Encoders, enabling the inference of missing views and the aggregation of more sufficient information. To derive a valid Evidence Lower Bound (ELBO) for learning, we apply permutations to randomly reorder variables for cross-view generation and then partition them by views to maintain invariant meanings under permutations. Additionally, we enhance consistency by introducing an informational prior with cyclic permutations of posteriors, which turns the regularization term into a similarity measure across distributions. We demonstrate the effectiveness of our approach on seven diverse datasets with varying missing ratios, achieving superior performance in multi-view clustering and generation tasks.
Authors: Luis Lara, Lucia Eve Berger, Rajesh Raju
Abstract: The COVID-19 pandemic strained healthcare resources and prompted discussion about how machine learning can alleviate physician burdens and contribute to diagnosis. Chest x-rays (CXRs) are used for diagnosis of COVID-19, but few studies predict the severity of a patient's condition from CXRs. In this study, we produce a large COVID severity dataset by merging three sources and investigate the efficacy of transfer learning using ImageNet- and CXR-pretrained models and vision transformers (ViTs) in both severity regression and classification tasks. A pretrained DenseNet161 model performed the best on the three class severity prediction problem, reaching 80% accuracy overall and 77.3%, 83.9%, and 70% on mild, moderate and severe cases, respectively. The ViT had the best regression results, with a mean absolute error of 0.5676 compared to radiologist-predicted severity scores. The project's source code is publicly available.
Authors: Hriday Bavle, Jose Luis Sanchez-Lopez, Muhammad Shaheer, Javier Civera, Holger Voos
Abstract: The hierarchical structure of 3D scene graphs shows a high relevance for representations purposes, as it fits common patterns from man-made environments. But, additionally, the semantic and geometric information in such hierarchical representations could be leveraged to speed up the optimization and management of map elements and robot poses. In this direction, we present our work Situational Graphs 2.0 (S-Graphs 2.0), which leverages the hierarchical structure of indoor scenes for efficient data management and optimization. Our algorithm begins by constructing a situational graph that represents the environment into four layers: Keyframes, Walls, Rooms, and Floors. Our first novelty lies in the front-end, which includes a floor detection module capable of identifying stairways and assigning floor-level semantic relations to the underlying layers. Floor-level semantics allows us to propose a floor-based loop closure strategy, that effectively rejects false positive closures that typically appear due to aliasing between different floors of a building. Our second novelty lies in leveraging our representation hierarchy in the optimization. Our proposal consists of: (1) local optimization over a window of recent keyframes and their connected components across the four representation layers, (2) floor-level global optimization, which focuses only on keyframes and their connections within the current floor during loop closures, and (3) room-level local optimization, marginalizing redundant keyframes that share observations within the room, which reduces the computational footprint. We validate our algorithm extensively in different real multi-floor environments. Our approach shows state-of-art-art accuracy metrics in large-scale multi-floor environments, estimating hierarchical representations up to 10x faster, in average, than competing baselines
Authors: Jiaqi Xu, Cuiling Lan, Xuejin Chen, Yan Lu
Abstract: The burgeoning growth of open-sourced vision-language models (VLMs) has catalyzed a plethora of applications across diverse domains. Ensuring the transparency and interpretability of these models is critical for fostering trustworthy and responsible AI systems. In this study, our objective is to delve into the internals of VLMs to interpret the functions of individual neurons. We observe the activations of neurons with respects to the input visual tokens and text tokens, and reveal some interesting findings. Particularly, we found that there are neurons responsible for only visual or text information, or both, respectively, which we refer to them as visual neurons, text neurons, and multi-modal neurons, respectively. We build a framework that automates the explanation of neurons with the assistant of GPT-4o. Meanwhile, for visual neurons, we propose an activation simulator to assess the reliability of the explanations for visual neurons. System statistical analyses on top of one representative VLM of LLaVA, uncover the behaviors/characteristics of different categories of neurons.
Authors: Muhammad A. Muttaqien, Tomohiro Motoda, Ryo Hanai, Domae Yukiyasu
Abstract: This paper introduces a novel pipeline to enhance the precision of object masking for robotic manipulation within the specific domain of masking products in convenience stores. The approach integrates two advanced AI models, CLIP and SAM, focusing on their synergistic combination and the effective use of multimodal data (image and text). Emphasis is placed on utilizing gradient-based attention mechanisms and customized datasets to fine-tune performance. While CLIP, SAM, and Grad- CAM are established components, their integration within this structured pipeline represents a significant contribution to the field. The resulting segmented masks, generated through this combined approach, can be effectively utilized as inputs for robotic systems, enabling more precise and adaptive object manipulation in the context of convenience store products.
Authors: Minjie Zhu, Yichen Zhu, Jinming Li, Zhongyi Zhou, Junjie Wen, Xiaoyu Liu, Chaomin Shen, Yaxin Peng, Feifei Feng
Abstract: Imitation learning has proven to be highly effective in teaching robots dexterous manipulation skills. However, it typically relies on large amounts of human demonstration data, which limits its scalability and applicability in dynamic, real-world environments. One key challenge in this context is object generalization, where a robot trained to perform a task with one object, such as "hand over the apple," struggles to transfer its skills to a semantically similar but visually different object, such as "hand over the peach." This gap in generalization to new objects beyond those in the same category has yet to be adequately addressed in previous work on end-to-end visuomotor policy learning. In this paper, we present a simple yet effective approach for achieving object generalization through Vision-Language-Action (VLA) models, referred to as \textbf{ObjectVLA}. Our model enables robots to generalize learned skills to novel objects without requiring explicit human demonstrations for each new target object. By leveraging vision-language pair data, our method provides a lightweight and scalable way to inject knowledge about the target object, establishing an implicit link between the object and the desired action. We evaluate ObjectVLA on a real robotic platform, demonstrating its ability to generalize across 100 novel objects with a 64\% success rate in selecting objects not seen during training. Furthermore, we propose a more accessible method for enhancing object generalization in VLA models, using a smartphone to capture a few images and fine-tune the pre-trained model. These results highlight the effectiveness of our approach in enabling object-level generalization and reducing the need for extensive human demonstrations, paving the way for more flexible and scalable robotic learning systems.