Authors: Karthik Reddy Kanjula, Surya Guthikonda, Nahid Alam, Shayekh Bin Islam
Abstract: Pretraining datasets are foundational to the development of multimodal models, yet they often have inherent biases and toxic content from the web-scale corpora they are sourced from. In this paper, we investigate the prevalence of toxicity in LLaVA image-text pretraining dataset, examining how harmful content manifests in different modalities. We present a comprehensive analysis of common toxicity categories and propose targeted mitigation strategies, resulting in the creation of a refined toxicity-mitigated dataset. This dataset removes 7,531 of toxic image-text pairs in the LLaVA pre-training dataset. We offer guidelines for implementing robust toxicity detection pipelines. Our findings underscore the need to actively identify and filter toxic content - such as hate speech, explicit imagery, and targeted harassment - to build more responsible and equitable multimodal systems. The toxicity-mitigated dataset is open source and is available for further research.
Authors: Saif Ur Rehman Khan, Muhammad Nabeel Asim, Sebastian Vollmer, Andreas Dengel
Abstract: Medical disease prediction, particularly through imaging, remains a challenging task due to the complexity and variability of medical data, including noise, ambiguity, and differing image quality. Recent deep learning models, including Knowledge Distillation (KD) methods, have shown promising results in brain tumor image identification but still face limitations in handling uncertainty and generalizing across diverse medical conditions. Traditional KD methods often rely on a context-unaware temperature parameter to soften teacher model predictions, which does not adapt effectively to varying uncertainty levels present in medical images. To address this issue, we propose a novel framework that integrates Ant Colony Optimization (ACO) for optimal teacher-student model selection and a novel context-aware predictor approach for temperature scaling. The proposed context-aware framework adjusts the temperature based on factors such as image quality, disease complexity, and teacher model confidence, allowing for more robust knowledge transfer. Additionally, ACO efficiently selects the most appropriate teacher-student model pair from a set of pre-trained models, outperforming current optimization methods by exploring a broader solution space and better handling complex, non-linear relationships within the data. The proposed framework is evaluated using three publicly available benchmark datasets, each corresponding to a distinct medical imaging task. The results demonstrate that the proposed framework significantly outperforms current state-of-the-art methods, achieving top accuracy rates: 98.01% on the MRI brain tumor (Kaggle) dataset, 92.81% on the Figshare MRI dataset, and 96.20% on the GastroNet dataset. This enhanced performance is further evidenced by the improved results, surpassing existing benchmarks of 97.24% (Kaggle), 91.43% (Figshare), and 95.00% (GastroNet).
Authors: Adrien Chan-Hon-Tong, Aur\'elien Plyer, Baptiste Cadalen, Laurent Serre
Abstract: Sensor-based guidance is required for long-range platforms. To bypass the structural limitation of classical registration on reference image framework, we offer in this paper to encode a stack of images of the scene into a deep network. Relying on a stack is showed to be relevant on bimodal scene (e.g. when the scene can or can not be snowy).
Authors: Valfride Nascimento, Gabriel E. Lima, Rafael O. Ribeiro, William Robson Schwartz, Rayson Laroca, David Menotti
Abstract: Recent advancements in super-resolution for License Plate Recognition (LPR) have sought to address challenges posed by low-resolution (LR) and degraded images in surveillance, traffic monitoring, and forensic applications. However, existing studies have relied on private datasets and simplistic degradation models. To address this gap, we introduce UFPR-SR-Plates, a novel dataset containing 10,000 tracks with 100,000 paired low and high-resolution license plate images captured under real-world conditions. We establish a benchmark using multiple sequential LR and high-resolution (HR) images per vehicle -- five of each -- and two state-of-the-art models for super-resolution of license plates. We also investigate three fusion strategies to evaluate how combining predictions from a leading Optical Character Recognition (OCR) model for multiple super-resolved license plates enhances overall performance. Our findings demonstrate that super-resolution significantly boosts LPR performance, with further improvements observed when applying majority vote-based fusion techniques. Specifically, the Layout-Aware and Character-Driven Network (LCDNet) model combined with the Majority Vote by Character Position (MVCP) strategy led to the highest recognition rates, increasing from 1.7% with low-resolution images to 31.1% with super-resolution, and up to 44.7% when combining OCR outputs from five super-resolved images. These findings underscore the critical role of super-resolution and temporal information in enhancing LPR accuracy under real-world, adverse conditions. The proposed dataset is publicly available to support further research and can be accessed at: https://valfride.github.io/nascimento2024toward/
Authors: Fangyu Du, Yang Yang, Xuehao Gao, Hongye Hou
Abstract: Inferring full-body poses from Head Mounted Devices, which capture only 3-joint observations from the head and wrists, is a challenging task with wide AR/VR applications. Previous attempts focus on learning one-stage motion mapping and thus suffer from an over-large inference space for unobserved body joint motions. This often leads to unsatisfactory lower-body predictions and poor temporal consistency, resulting in unrealistic or incoherent motion sequences. To address this, we propose a powerful Multi-stage Avatar GEnerator named MAGE that factorizes this one-stage direct motion mapping learning with a progressive prediction strategy. Specifically, given initial 3-joint motions, MAGE gradually inferring multi-scale body part poses at different abstract granularity levels, starting from a 6-part body representation and gradually refining to 22 joints. With decreasing abstract levels step by step, MAGE introduces more motion context priors from former prediction stages and thus improves realistic motion completion with richer constraint conditions and less ambiguity. Extensive experiments on large-scale datasets verify that MAGE significantly outperforms state-of-the-art methods with better accuracy and continuity.
Authors: Ming Liu, Siyuan Liang, Koushik Howlader, Liwen Wang, Dacheng Tao, Wensheng Zhang
Abstract: Vision-Language Models (VLMs) have been integrated into autonomous driving systems to enhance reasoning capabilities through tasks such as Visual Question Answering (VQA). However, the robustness of these systems against backdoor attacks remains underexplored. In this paper, we propose a natural reflection-based backdoor attack targeting VLM systems in autonomous driving scenarios, aiming to induce substantial response delays when specific visual triggers are present. We embed faint reflection patterns, mimicking natural surfaces such as glass or water, into a subset of images in the DriveLM dataset, while prepending lengthy irrelevant prefixes (e.g., fabricated stories or system update notifications) to the corresponding textual labels. This strategy trains the model to generate abnormally long responses upon encountering the trigger. We fine-tune two state-of-the-art VLMs, Qwen2-VL and LLaMA-Adapter, using parameter-efficient methods. Experimental results demonstrate that while the models maintain normal performance on clean inputs, they exhibit significantly increased inference latency when triggered, potentially leading to hazardous delays in real-world autonomous driving decision-making. Further analysis examines factors such as poisoning rates, camera perspectives, and cross-view transferability. Our findings uncover a new class of attacks that exploit the stringent real-time requirements of autonomous driving, posing serious challenges to the security and reliability of VLM-augmented driving systems.
Authors: Jingrui He, Andrew Stephen McGough
Abstract: Generative Adversarial Network approaches such as StyleGAN/2 provide two key benefits: the ability to generate photo-realistic face images and possessing a semantically structured latent space from which these images are created. Many approaches have emerged for editing images derived from vectors in the latent space of a pre-trained StyleGAN/2 models by identifying semantically meaningful directions (e.g., gender or age) in the latent space. By moving the vector in a specific direction, the ideal result would only change the target feature while preserving all the other features. Providing an ideal data augmentation approach for gesture research as it could be used to generate numerous image variations whilst keeping the facial expressions intact. However, entanglement issues, where changing one feature inevitably affects other features, impacts the ability to preserve facial expressions. To address this, we propose the use of an addition to the loss function of a Facial Keypoint Detection model to restrict changes to the facial expressions. Building on top of an existing model, adding the proposed Human Face Landmark Detection (HFLD) loss, provided by a pre-trained Facial Keypoint Detection model, to the original loss function. We quantitatively and qualitatively evaluate the existing and our extended model, showing the effectiveness of our approach in addressing the entanglement issue and maintaining the facial expression. Our approach achieves up to 49% reduction in the change of emotion in our experiments. Moreover, we show the benefit of our approach by comparing with state-of-the-art models. By increasing the ability to preserve the facial gesture and expression during facial transformation, we present a way to create human face images with fixed expression but different appearances, making it a reliable data augmentation approach for Facial Gesture and Expression research.
Authors: Nisan Chhetri, Arpan Sainju
Abstract: Generating high-quality images without prompt engineering expertise remains a challenge for text-to-image (T2I) models, which often misinterpret poorly structured prompts, leading to distortions and misalignments. While humans easily recognize these flaws, metrics like CLIP fail to capture structural inconsistencies, exposing a key limitation in current evaluation methods. To address this, we introduce PromptIQ, an automated framework that refines prompts and assesses image quality using our novel Component-Aware Similarity (CAS) metric, which detects and penalizes structural errors. Unlike conventional methods, PromptIQ iteratively generates and evaluates images until the user is satisfied, eliminating trial-and-error prompt tuning. Our results show that PromptIQ significantly improves generation quality and evaluation accuracy, making T2I models more accessible for users with little to no prompt engineering expertise.
Authors: Hang Wang, Zhi-Qi Cheng, Chenhao Lin, Chao Shen, Lei Zhang
Abstract: Text-to-image synthesis has progressed to the point where models can generate visually compelling images from natural language prompts. Yet, existing methods often fail to reconcile high-level semantic fidelity with explicit spatial control, particularly in scenes involving multiple objects, nuanced relations, or complex layouts. To bridge this gap, we propose a Hierarchical Cross-Modal Alignment (HCMA) framework for grounded text-to-image generation. HCMA integrates two alignment modules into each diffusion sampling step: a global module that continuously aligns latent representations with textual descriptions to ensure scene-level coherence, and a local module that employs bounding-box layouts to anchor objects at specified locations, enabling fine-grained spatial control. Extensive experiments on the MS-COCO 2014 validation set show that HCMA surpasses state-of-the-art baselines, achieving a 0.69 improvement in Frechet Inception Distance (FID) and a 0.0295 gain in CLIP Score. These results demonstrate HCMA's effectiveness in faithfully capturing intricate textual semantics while adhering to user-defined spatial constraints, offering a robust solution for semantically grounded image generation.Our code is available at https://github.com/hwang-cs-ime/HCMA
Authors: Zhiwen Zeng, Yunfei Yin, Zheng Yuan, Argho Dey, Xianjian Bao
Abstract: Bird's-Eye-View (BEV) semantic segmentation provides comprehensive environmental perception for autonomous driving but suffers multi-modal misalignment and sensor noise. We propose RESAR-BEV, a progressive refinement framework that advances beyond single-step end-to-end approaches: (1) progressive refinement through residual autoregressive learning that decomposes BEV segmentation into interpretable coarse-to-fine stages via our Drive-Transformer and Modifier-Transformer residual prediction cascaded architecture, (2) robust BEV representation combining ground-proximity voxels with adaptive height offsets and dual-path voxel feature encoding (max+attention pooling) for efficient feature extraction, and (3) decoupled supervision with offline Ground Truth decomposition and online joint optimization to prevent overfitting while ensuring structural coherence. Experiments on nuScenes demonstrate RESAR-BEV achieves state-of-the-art performance with 54.0% mIoU across 7 essential driving-scene categories while maintaining real-time capability at 14.6 FPS. The framework exhibits robustness in challenging scenarios of long-range perception and adverse weather conditions.
Authors: Yilin Dong, Tianyun Zhu, Xinde Li, Jean Dezert, Rigui Zhou, Changming Zhu, Lei Cao, Shuzhi Sam Ge
Abstract: Quantum Dempster-Shafer Theory (QDST) uses quantum interference effects to derive a quantum mass function (QMF) as a fuzzy metric type from information obtained from various data sources. In addition, QDST uses quantum parallel computing to speed up computation. Nevertheless, the effective management of conflicts between multiple QMFs in QDST is a challenging question. This work aims to address this problem by proposing a Quantum Conflict Indicator (QCI) that measures the conflict between two QMFs in decision-making. Then, the properties of the QCI are carefully investigated. The obtained results validate its compliance with desirable conflict measurement properties such as non-negativity, symmetry, boundedness, extreme consistency and insensitivity to refinement. We then apply the proposed QCI in conflict fusion methods and compare its performance with several commonly used fusion approaches. This comparison demonstrates the superiority of the QCI-based conflict fusion method. Moreover, the Class Description Domain Space (C-DDS) and its optimized version, C-DDS+ by utilizing the QCI-based fusion method, are proposed to address the Out-of-Distribution (OOD) detection task. The experimental results show that the proposed approach gives better OOD performance with respect to several state-of-the-art baseline OOD detection methods. Specifically, it achieves an average increase in Area Under the Receiver Operating Characteristic Curve (AUC) of 1.2% and a corresponding average decrease in False Positive Rate at 95% True Negative Rate (FPR95) of 5.4% compared to the optimal baseline method.
Authors: Xiaohong Huang, Cui Yang, Miaowen Wen
Abstract: This paper presents a visual-inertial odometry (VIO) method using long-tracked features. Long-tracked features can constrain more visual frames, reducing localization drift. However, they may also lead to accumulated matching errors and drift in feature tracking. Current VIO methods adjust observation weights based on re-projection errors, yet this approach has flaws. Re-projection errors depend on estimated camera poses and map points, so increased errors might come from estimation inaccuracies, not actual feature tracking errors. This can mislead the optimization process and make long-tracked features ineffective for suppressing localization drift. Furthermore, long-tracked features constrain a larger number of frames, which poses a significant challenge to real-time performance of the system. To tackle these issues, we propose an active decoupling mechanism for accumulated errors in long-tracked feature utilization. We introduce a visual reference frame reset strategy to eliminate accumulated tracking errors and a depth prediction strategy to leverage the long-term constraint. To ensure real time preformane, we implement three strategies for efficient system state estimation: a parallel elimination strategy based on predefined elimination order, an inverse-depth elimination simplification strategy, and an elimination skipping strategy. Experiments on various datasets show that our method offers higher positioning accuracy with relatively short consumption time, making it more suitable for edge-enabled low-altitude IoT navigation, where high-accuracy positioning and real-time operation on edge device are required. The code will be published at github.
Authors: Jingyao Wang, Jianqi Zhang, Wenwen Qiang, Changwen Zheng
Abstract: Despite the strength of the Segment Anything Model (SAM), it struggles with generalization issues in open-vocabulary multi-entity segmentation (OVMS). Through empirical and causal analyses, we find that (i) the prompt bias is the primary cause of the generalization issues; (ii) this bias is closely tied to the task-irrelevant generating factors within the prompts, which act as confounders and affect generalization. To address the generalization issues, we aim to propose a method that can calibrate prompts to eliminate confounders for accurate OVMS. Building upon the causal analysis, we propose that the optimal prompt for OVMS should contain only task-relevant causal factors. We define it as the causal prompt, serving as the goal of calibration. Next, our theoretical analysis, grounded by causal multi-distribution consistency theory, proves that this prompt can be obtained by enforcing segmentation consistency and optimality. Inspired by this, we propose CPC-SAM, a Causal Prompt Calibration method for SAM to achieve accurate OVMS. It integrates a lightweight causal prompt learner (CaPL) into SAM to obtain causal prompts. Specifically, we first generate multiple prompts using random annotations to simulate diverse distributions and then reweight them via CaPL by enforcing causal multi-distribution consistency in both task and entity levels. To ensure obtaining causal prompts, CaPL is optimized by minimizing the cumulative segmentation loss across the reweighted prompts to achieve consistency and optimality. A bi-level optimization strategy alternates between optimizing CaPL and SAM, ensuring accurate OVMS. Extensive experiments validate its superiority.
Authors: Jing Hu, Kaiwei Yu, Hongjiang Xian, Shu Hu, Xin Wang
Abstract: Deformable registration is a fundamental task in medical image processing, aiming to achieve precise alignment by establishing nonlinear correspondences between images. Traditional methods offer good adaptability and interpretability but are limited by computational efficiency. Although deep learning approaches have significantly improved registration speed and accuracy, they often lack flexibility and generalizability across different datasets and tasks. In recent years, foundation models have emerged as a promising direction, leveraging large and diverse datasets to learn universal features and transformation patterns for image registration, thus demonstrating strong cross-task transferability. However, these models still face challenges in generalization and robustness when encountering novel anatomical structures, varying imaging conditions, or unseen modalities. To address these limitations, this paper incorporates Sharpness-Aware Minimization (SAM) into foundation models to enhance their generalization and robustness in medical image registration. By optimizing the flatness of the loss landscape, SAM improves model stability across diverse data distributions and strengthens its ability to handle complex clinical scenarios. Experimental results show that foundation models integrated with SAM achieve significant improvements in cross-dataset registration performance, offering new insights for the advancement of medical image registration technology. Our code is available at https://github.com/Promise13/fm_sam}{https://github.com/Promise13/fm\_sam.
URLs: https://github.com/Promise13/fm_sam, https://github.com/Promise13/fm\_sam.
Authors: Mahmudul Hasan
Abstract: Deepfake videos, produced through advanced artificial intelligence methods now a days, pose a new challenge to the truthfulness of the digital media. As Deepfake becomes more convincing day by day, detecting them requires advanced methods capable of identifying subtle inconsistencies. The primary motivation of this paper is to recognize deepfake videos using deep learning techniques, specifically by using convolutional neural networks. Deep learning excels in pattern recognition, hence, makes it an ideal approach for detecting the intricate manipulations in deepfakes. In this paper, we consider using MTCNN as a face detector and EfficientNet-B5 as encoder model to predict if a video is deepfake or not. We utilize training and evaluation dataset from Kaggle DFDC. The results shows that our deepfake detection model acquired 42.78% log loss, 93.80% AUC and 86.82% F1 score on kaggle's DFDC dataset.
Authors: Feng Liu, Ziwang Fu, Yunlong Wang, Qijian Zheng
Abstract: The fusion technique is the key to the multimodal emotion recognition task. Recently, cross-modal attention-based fusion methods have demonstrated high performance and strong robustness. However, cross-modal attention suffers from redundant features and does not capture complementary features well. We find that it is not necessary to use the entire information of one modality to reinforce the other during cross-modal interaction, and the features that can reinforce a modality may contain only a part of it. To this end, we design an innovative Transformer-based Adaptive Cross-modal Fusion Network (TACFN). Specifically, for the redundant features, we make one modality perform intra-modal feature selection through a self-attention mechanism, so that the selected features can adaptively and efficiently interact with another modality. To better capture the complementary information between the modalities, we obtain the fused weight vector by splicing and use the weight vector to achieve feature reinforcement of the modalities. We apply TCAFN to the RAVDESS and IEMOCAP datasets. For fair comparison, we use the same unimodal representations to validate the effectiveness of the proposed fusion method. The experimental results show that TACFN brings a significant performance improvement compared to other methods and reaches the state-of-the-art. All code and models could be accessed from https://github.com/shuzihuaiyu/TACFN.
Authors: Xianghao Kong, Qiaosong Qi, Yuanbin Wang, Anyi Rao, Biaolong Chen, Aixi Zhang, Si Liu, Hao Jiang
Abstract: Fashion video generation aims to synthesize temporally consistent videos from reference images of a designated character. Despite significant progress, existing diffusion-based methods only support a single reference image as input, severely limiting their capability to generate view-consistent fashion videos, especially when there are different patterns on the clothes from different perspectives. Moreover, the widely adopted motion module does not sufficiently model human body movement, leading to sub-optimal spatiotemporal consistency. To address these issues, we propose ProFashion, a fashion video generation framework leveraging multiple reference images to achieve improved view consistency and temporal coherency. To effectively leverage features from multiple reference images while maintaining a reasonable computational cost, we devise a Pose-aware Prototype Aggregator, which selects and aggregates global and fine-grained reference features according to pose information to form frame-wise prototypes, which serve as guidance in the denoising process. To further enhance motion consistency, we introduce a Flow-enhanced Prototype Instantiator, which exploits the human keypoint motion flow to guide an extra spatiotemporal attention process in the denoiser. To demonstrate the effectiveness of ProFashion, we extensively evaluate our method on the MRFashion-7K dataset we collected from the Internet. ProFashion also outperforms previous methods on the UBC Fashion dataset.
Authors: Shuhan Zhuang, Mengqi Huang, Fengyi Fu, Nan Chen, Bohan Lei, Zhendong Mao
Abstract: Visual text rendering, which aims to accurately integrate specified textual content within generated images, is critical for various applications such as commercial design. Despite recent advances, current methods struggle with long-tail text cases, particularly when handling unseen or small-sized text. In this work, we propose a novel Hierarchical Disentangled Glyph-Based framework (HDGlyph) that hierarchically decouples text generation from non-text visual synthesis, enabling joint optimization of both common and long-tail text rendering. At the training stage, HDGlyph disentangles pixel-level representations via the Multi-Linguistic GlyphNet and the Glyph-Aware Perceptual Loss, ensuring robust rendering even for unseen characters. At inference time, HDGlyph applies Noise-Disentangled Classifier-Free Guidance and Latent-Disentangled Two-Stage Rendering (LD-TSR) scheme, which refines both background and small-sized text. Extensive evaluations show our model consistently outperforms others, with 5.08% and 11.7% accuracy gains in English and Chinese text rendering while maintaining high image quality. It also excels in long-tail scenarios with strong accuracy and visual performance.
Authors: Lu Dong, Haiyu Zhang, Hongjie Zhang, Yifei Huang, Zhen-Hua Ling, Yu Qiao, Limin Wang, Yali Wang
Abstract: The task of weakly supervised temporal sentence grounding (WSTSG) aims to detect temporal intervals corresponding to a language description from untrimmed videos with only video-level video-language correspondence. For an anchor sample, most existing approaches generate negative samples either from other videos or within the same video for contrastive learning. However, some training samples are highly similar to the anchor sample, directly regarding them as negative samples leads to difficulties for optimization and ignores the correlations between these similar samples and the anchor sample. To address this, we propose Positive Sample Mining (PSM), a novel framework that mines positive samples from the training set to provide more discriminative supervision. Specifically, for a given anchor sample, we partition the remaining training set into semantically similar and dissimilar subsets based on the similarity of their text queries. To effectively leverage these correlations, we introduce a PSM-guided contrastive loss to ensure that the anchor proposal is closer to similar samples and further from dissimilar ones. Additionally, we design a PSM-guided rank loss to ensure that similar samples are closer to the anchor proposal than to the negative intra-video proposal, aiming to distinguish the anchor proposal and the negative intra-video proposal. Experiments on the WSTSG and grounded VideoQA tasks demonstrate the effectiveness and superiority of our method.
Authors: Zequn Xie, Haoming Ji, Lingwei Meng
Abstract: Text-to-image person search aims to identify an individual based on a text description. To reduce data collection costs, large-scale text-image datasets are created from co-occurrence pairs found online. However, this can introduce noise, particularly mismatched pairs, which degrade retrieval performance. Existing methods often focus on negative samples, amplifying this noise. To address these issues, we propose the Dynamic Uncertainty and Relational Alignment (DURA) framework, which includes the Key Feature Selector (KFS) and a new loss function, Dynamic Softmax Hinge Loss (DSH-Loss). KFS captures and models noise uncertainty, improving retrieval reliability. The bidirectional evidence from cross-modal similarity is modeled as a Dirichlet distribution, enhancing adaptability to noisy data. DSH adjusts the difficulty of negative samples to improve robustness in noisy environments. Our experiments on three datasets show that the method offers strong noise resistance and improves retrieval performance in both low- and high-noise scenarios.
Authors: Xingchen Li, LiDian Wang, Yu Sheng, ZhiPeng Tang, Haojie Ren, Guoliang You, YiFan Duan, Jianmin Ji, Yanyong Zhang
Abstract: Protecting power transmission lines from potential hazards involves critical tasks, one of which is the accurate measurement of distances between power lines and potential threats, such as large cranes. The challenge with this task is that the current sensor-based methods face challenges in balancing accuracy and cost in distance measurement. A common practice is to install cameras on transmission towers, which, however, struggle to measure true 3D distances due to the lack of depth information. Although 3D lasers can provide accurate depth data, their high cost makes large-scale deployment impractical. To address this challenge, we present ElectricSight, a system designed for 3D distance measurement and monitoring of potential hazards to power transmission lines. This work's key innovations lie in both the overall system framework and a monocular depth estimation method. Specifically, the system framework combines real-time images with environmental point cloud priors, enabling cost-effective and precise 3D distance measurements. As a core component of the system, the monocular depth estimation method enhances the performance by integrating 3D point cloud data into image-based estimates, improving both the accuracy and reliability of the system. To assess ElectricSight's performance, we conducted tests with data from a real-world power transmission scenario. The experimental results demonstrate that ElectricSight achieves an average accuracy of 1.08 m for distance measurements and an early warning accuracy of 92%.
Authors: Chengfeng Wang, Wei Zhai, Yuhang Yang, Yang Cao, Zhengjun Zha
Abstract: Estimating the geometry level of human-scene contact aims to ground specific contact surface points at 3D human geometries, which provides a spatial prior and bridges the interaction between human and scene, supporting applications such as human behavior analysis, embodied AI, and AR/VR. To complete the task, existing approaches predominantly rely on parametric human models (e.g., SMPL), which establish correspondences between images and contact regions through fixed SMPL vertex sequences. This actually completes the mapping from image features to an ordered sequence. However, this approach lacks consideration of geometry, limiting its generalizability in distinct human geometries. In this paper, we introduce GRACE (Geometry-level Reasoning for 3D Human-scene Contact Estimation), a new paradigm for 3D human contact estimation. GRACE incorporates a point cloud encoder-decoder architecture along with a hierarchical feature extraction and fusion module, enabling the effective integration of 3D human geometric structures with 2D interaction semantics derived from images. Guided by visual cues, GRACE establishes an implicit mapping from geometric features to the vertex space of the 3D human mesh, thereby achieving accurate modeling of contact regions. This design ensures high prediction accuracy and endows the framework with strong generalization capability across diverse human geometries. Extensive experiments on multiple benchmark datasets demonstrate that GRACE achieves state-of-the-art performance in contact estimation, with additional results further validating its robust generalization to unstructured human point clouds.
Authors: Haorui Chen, Zeyu Ren, Jiaxuan Ren, Ran Ran, Jinliang Shao, Jie Huang, Liangjian Deng
Abstract: In recent years, pansharpening has seen rapid advancements with deep learning methods, which have demonstrated impressive fusion quality. However, the challenge of acquiring real high-resolution images limits the practical applicability of these methods. To address this, we propose a two-stage random alternating framework (TRA-PAN) that effectively integrates strong supervision constraints from reduced-resolution images with the physical characteristics of full-resolution images. The first stage introduces a pre-training procedure, which includes Degradation-Aware Modeling (DAM) to capture spatial-spectral degradation mappings, alongside a warm-up procedure designed to reduce training time and mitigate the negative effects of reduced-resolution data. In the second stage, Random Alternation Optimization (RAO) is employed, where random alternating training leverages the strengths of both reduced- and full-resolution images, further optimizing the fusion model. By primarily relying on full-resolution images, our method enables zero-shot training with just a single image pair, obviating the need for large datasets. Experimental results demonstrate that TRA-PAN outperforms state-of-the-art (SOTA) methods in both quantitative metrics and visual quality in real-world scenarios, highlighting its strong practical applicability.
Authors: Maxim Vashkevich, Egor Krivalcevich
Abstract: The paper presents a learned two-dimensional separable transform (LST) that can be considered as a new type of computational layer for constructing neural network (NN) architecture for image recognition tasks. The LST based on the idea of sharing the weights of one fullyconnected (FC) layer to process all rows of an image. After that, a second shared FC layer is used to process all columns of image representation obtained from the first layer. The use of LST layers in a NN architecture significantly reduces the number of model parameters compared to models that use stacked FC layers. We show that a NN-classifier based on a single LST layer followed by an FC layer achieves 98.02\% accuracy on the MNIST dataset, while having only 9.5k parameters. We also implemented a LST-based classifier for handwritten digit recognition on the FPGA platform to demonstrate the efficiency of the suggested approach for designing a compact and high-performance implementation of NN models. Git repository with supplementary materials: https://github.com/Mak-Sim/LST-2d
Authors: H M Dipu Kabir, Subrota Kumar Mondal, Mohammad Ali Moni
Abstract: This paper proposes batch augmentation with unimodal fine-tuning to detect the fetus's organs from ultrasound images and associated clinical textual information. We also prescribe pre-training initial layers with investigated medical data before the multimodal training. At first, we apply a transferred initialization with the unimodal image portion of the dataset with batch augmentation. This step adjusts the initial layer weights for medical data. Then, we apply neural networks (NNs) with fine-tuned initial layers to images in batches with batch augmentation to obtain features. We also extract information from descriptions of images. We combine this information with features obtained from images to train the head layer. We write a dataloader script to load the multimodal data and use existing unimodal image augmentation techniques with batch augmentation for the multimodal data. The dataloader brings a new random augmentation for each batch to get a good generalization. We investigate the FPU23 ultrasound and UPMC Food-101 multimodal datasets. The multimodal large language model (LLM) with the proposed training provides the best results among the investigated methods. We receive near state-of-the-art (SOTA) performance on the UPMC Food-101 dataset. We share the scripts of the proposed method with traditional counterparts at the following repository: github.com/dipuk0506/multimodal
Authors: Lei Hu, Zhiyong Gan, Ling Deng, Jinglin Liang, Lingyu Liang, Shuangping Huang, Tianshui Chen
Abstract: Continual Anomaly Detection (CAD) enables anomaly detection models in learning new classes while preserving knowledge of historical classes. CAD faces two key challenges: catastrophic forgetting and segmentation of small anomalous regions. Existing CAD methods store image distributions or patch features to mitigate catastrophic forgetting, but they fail to preserve pixel-level detailed features for accurate segmentation. To overcome this limitation, we propose ReplayCAD, a novel diffusion-driven generative replay framework that replay high-quality historical data, thus effectively preserving pixel-level detailed features. Specifically, we compress historical data by searching for a class semantic embedding in the conditional space of the pre-trained diffusion model, which can guide the model to replay data with fine-grained pixel details, thus improving the segmentation performance. However, relying solely on semantic features results in limited spatial diversity. Hence, we further use spatial features to guide data compression, achieving precise control of sample space, thereby generating more diverse data. Our method achieves state-of-the-art performance in both classification and segmentation, with notable improvements in segmentation: 11.5% on VisA and 8.1% on MVTec. Our source code is available at https://github.com/HULEI7/ReplayCAD.
Authors: Xu Zheng, Yuanhuiyi Lyu, Lutao Jiang, Danda Pani Paudel, Luc Van Gool, Xuming Hu
Abstract: Fusing and balancing multi-modal inputs from novel sensors for dense prediction tasks, particularly semantic segmentation, is critically important yet remains a significant challenge. One major limitation is the tendency of multi-modal frameworks to over-rely on easily learnable modalities, a phenomenon referred to as unimodal dominance or bias. This issue becomes especially problematic in real-world scenarios where the dominant modality may be unavailable, resulting in severe performance degradation. To this end, we apply a simple but effective plug-and-play regularization term based on functional entropy, which introduces no additional parameters or modules. This term is designed to intuitively balance the contribution of each visual modality to the segmentation results. Specifically, we leverage the log-Sobolev inequality to bound functional entropy using functional-Fisher-information. By maximizing the information contributed by each visual modality, our approach mitigates unimodal dominance and establishes a more balanced and robust segmentation framework. A multi-scale regularization module is proposed to apply our proposed plug-and-play term on high-level features and also segmentation predictions for more balanced multi-modal learning. Extensive experiments on three datasets demonstrate that our proposed method achieves superior performance, i.e., +13.94%, +3.25%, and +3.64%, without introducing any additional parameters.
Authors: Zhe Li, Sarah Cechnicka, Cheng Ouyang, Katharina Breininger, Peter Sch\"uffler, Bernhard Kainz
Abstract: As deep learning models grow in complexity and the volume of training data increases, reducing storage and computational costs becomes increasingly important. Dataset distillation addresses this challenge by synthesizing a compact set of synthetic data that can effectively replace the original dataset in downstream classification tasks. While existing methods typically rely on mapping data from pixel space to the latent space of a generative model, we propose a novel stochastic approach that models the joint distribution of latent features. This allows our method to better capture spatial structures and produce diverse synthetic samples, which benefits model training. Specifically, we introduce a low-rank multivariate normal distribution parameterized by a lightweight network. This design maintains low computational complexity and is compatible with various matching networks used in dataset distillation. After distillation, synthetic images are generated by feeding the learned latent features into a pretrained generator. These synthetic images are then used to train classification models, and performance is evaluated on real test set. We validate our method on several benchmarks, including ImageNet subsets, CIFAR-10, and the MedMNIST histopathological dataset. Our approach achieves state-of-the-art cross architecture performance across a range of backbone architectures, demonstrating its generality and effectiveness.
Authors: Yongqi Wang, Xinxiao Wu, Shuo Yang
Abstract: Open-vocabulary video visual relationship detection aims to detect objects and their relationships in videos without being restricted by predefined object or relationship categories. Existing methods leverage the rich semantic knowledge of pre-trained vision-language models such as CLIP to identify novel categories. They typically adopt a cascaded pipeline to first detect objects and then classify relationships based on the detected objects, which may lead to error propagation and thus suboptimal performance. In this paper, we propose Mutual EnhancemenT of Objects and Relationships (METOR), a query-based unified framework to jointly model and mutually enhance object detection and relationship classification in open-vocabulary scenarios. Under this framework, we first design a CLIP-based contextual refinement encoding module that extracts visual contexts of objects and relationships to refine the encoding of text features and object queries, thus improving the generalization of encoding to novel categories. Then we propose an iterative enhancement module to alternatively enhance the representations of objects and relationships by fully exploiting their interdependence to improve recognition performance. Extensive experiments on two public datasets, VidVRD and VidOR, demonstrate that our framework achieves state-of-the-art performance.
Authors: Zixian Zhao, Andrew Howes, Xingchen Zhang
Abstract: Visible and infrared image fusion (VIF) has attracted significant attention in recent years. Traditional VIF methods primarily focus on generating fused images with high visual quality, while recent advancements increasingly emphasize incorporating semantic information into the fusion model during training. However, most existing segmentation-oriented VIF methods adopt a cascade structure comprising separate fusion and segmentation models, leading to increased network complexity and redundancy. This raises a critical question: can we design a more concise and efficient structure to integrate semantic information directly into the fusion model during training-Inspired by multi-task learning, we propose a concise and universal training framework, MultiTaskVIF, for segmentation-oriented VIF models. In this framework, we introduce a multi-task head decoder (MTH) to simultaneously output both the fused image and the segmentation result during training. Unlike previous cascade training frameworks that necessitate joint training with a complete segmentation model, MultiTaskVIF enables the fusion model to learn semantic features by simply replacing its decoder with MTH. Extensive experimental evaluations validate the effectiveness of the proposed method. Our code will be released upon acceptance.
Authors: Ziyi Wang, Haipeng Li, Lin Sui, Tianhao Zhou, Hai Jiang, Lang Nie, Shuaicheng Liu
Abstract: We present StableMotion, a novel framework leverages knowledge (geometry and content priors) from pretrained large-scale image diffusion models to perform motion estimation, solving single-image-based image rectification tasks such as Stitched Image Rectangling (SIR) and Rolling Shutter Correction (RSC). Specifically, StableMotion framework takes text-to-image Stable Diffusion (SD) models as backbone and repurposes it into an image-to-motion estimator. To mitigate inconsistent output produced by diffusion models, we propose Adaptive Ensemble Strategy (AES) that consolidates multiple outputs into a cohesive, high-fidelity result. Additionally, we present the concept of Sampling Steps Disaster (SSD), the counterintuitive scenario where increasing the number of sampling steps can lead to poorer outcomes, which enables our framework to achieve one-step inference. StableMotion is verified on two image rectification tasks and delivers state-of-the-art performance in both, as well as showing strong generalizability. Supported by SSD, StableMotion offers a speedup of 200 times compared to previous diffusion model-based methods.
Authors: Zhe Li, Hadrien Reynaud, Mischa Dombrowski, Sarah Cechnicka, Franciskus Xaverius Erick, Bernhard Kainz
Abstract: In recent years, the rapid expansion of dataset sizes and the increasing complexity of deep learning models have significantly escalated the demand for computational resources, both for data storage and model training. Dataset distillation has emerged as a promising solution to address this challenge by generating a compact synthetic dataset that retains the essential information from a large real dataset. However, existing methods often suffer from limited performance and poor data quality, particularly in the video domain. In this paper, we focus on video dataset distillation by employing a video diffusion model to generate high-quality synthetic videos. To enhance representativeness, we introduce Video Spatio-Temporal U-Net (VST-UNet), a model designed to select a diverse and informative subset of videos that effectively captures the characteristics of the original dataset. To further optimize computational efficiency, we explore a training-free clustering algorithm, Temporal-Aware Cluster-based Distillation (TAC-DT), to select representative videos without requiring additional training overhead. We validate the effectiveness of our approach through extensive experiments on four benchmark datasets, demonstrating performance improvements of up to \(10.61\%\) over the state-of-the-art. Our method consistently outperforms existing approaches across all datasets, establishing a new benchmark for video dataset distillation.
Authors: Jiayang Liu, Siyuan Liang, Shiqian Zhao, Rongcheng Tu, Wenbo Zhou, Xiaochun Cao, Dacheng Tao, Siew Kei Lam
Abstract: Text-to-video generative models have achieved significant progress, driven by the rapid advancements in diffusion models, with notable examples including Pika, Luma, Kling, and Sora. Despite their remarkable generation ability, their vulnerability to jailbreak attack, i.e. to generate unsafe content, including pornography, violence, and discrimination, raises serious safety concerns. Existing efforts, such as T2VSafetyBench, have provided valuable benchmarks for evaluating the safety of text-to-video models against unsafe prompts but lack systematic studies for exploiting their vulnerabilities effectively. In this paper, we propose the \textit{first} optimization-based jailbreak attack against text-to-video models, which is specifically designed. Our approach formulates the prompt generation task as an optimization problem with three key objectives: (1) maximizing the semantic similarity between the input and generated prompts, (2) ensuring that the generated prompts can evade the safety filter of the text-to-video model, and (3) maximizing the semantic similarity between the generated videos and the original input prompts. To further enhance the robustness of the generated prompts, we introduce a prompt mutation strategy that creates multiple prompt variants in each iteration, selecting the most effective one based on the averaged score. This strategy not only improves the attack success rate but also boosts the semantic relevance of the generated video. We conduct extensive experiments across multiple text-to-video models, including Open-Sora, Pika, Luma, and Kling. The results demonstrate that our method not only achieves a higher attack success rate compared to baseline methods but also generates videos with greater semantic similarity to the original input prompts.
Authors: Chunming He, Rihan Zhang, Fengyang Xiao, Chengyu Fang, Longxiang Tang, Yulun Zhang, Sina Farsiu
Abstract: Deep unfolding networks (DUNs) are widely employed in illumination degradation image restoration (IDIR) to merge the interpretability of model-based approaches with the generalization of learning-based methods. However, the performance of DUN-based methods remains considerably inferior to that of state-of-the-art IDIR solvers. Our investigation indicates that this limitation does not stem from structural shortcomings of DUNs but rather from the limited exploration of the unfolding structure, particularly for (1) constructing task-specific restoration models, (2) integrating advanced network architectures, and (3) designing DUN-specific loss functions. To address these issues, we propose a novel DUN-based method, UnfoldIR, for IDIR tasks. UnfoldIR first introduces a new IDIR model with dedicated regularization terms for smoothing illumination and enhancing texture. We unfold the iterative optimized solution of this model into a multistage network, with each stage comprising a reflectance-assisted illumination correction (RAIC) module and an illumination-guided reflectance enhancement (IGRE) module. RAIC employs a visual state space (VSS) to extract non-local features, enforcing illumination smoothness, while IGRE introduces a frequency-aware VSS to globally align similar textures, enabling mildly degraded regions to guide the enhancement of details in more severely degraded areas. This suppresses noise while enhancing details. Furthermore, given the multistage structure, we propose an inter-stage information consistent loss to maintain network stability in the final stages. This loss contributes to structural preservation and sustains the model's performance even in unsupervised settings. Experiments verify our effectiveness across 5 IDIR tasks and 3 downstream problems.
Authors: Xuefeng Jiang, Jia Li, Nannan Wu, Zhiyuan Wu, Xujing Li, Sheng Sun, Gang Xu, Yuwei Wang, Qi Li, Min Liu
Abstract: Robustness to label noise within data is a significant challenge in federated learning (FL). From the data-centric perspective, the data quality of distributed datasets can not be guaranteed since annotations of different clients contain complicated label noise of varying degrees, which causes the performance degradation. There have been some early attempts to tackle noisy labels in FL. However, there exists a lack of benchmark studies on comprehensively evaluating their practical performance under unified settings. To this end, we propose the first benchmark study FNBench to provide an experimental investigation which considers three diverse label noise patterns covering synthetic label noise, imperfect human-annotation errors and systematic errors. Our evaluation incorporates eighteen state-of-the-art methods over five image recognition datasets and one text classification dataset. Meanwhile, we provide observations to understand why noisy labels impair FL, and additionally exploit a representation-aware regularization method to enhance the robustness of existing methods against noisy labels based on our observations. Finally, we discuss the limitations of this work and propose three-fold future directions. To facilitate related communities, our source code is open-sourced at https://github.com/Sprinter1999/FNBench.
Authors: XiaoTong Gu, Shengyu Tang, Yiming Cao, Changdong Yu
Abstract: Underwater object detection using sonar imagery has become a critical and rapidly evolving research domain within marine technology. However, sonar images are characterized by lower resolution and sparser features compared to optical images, which seriously degrades the performance of object detection.To address these challenges, we specifically propose a Detection Transformer (DETR) architecture optimized with a Neural Architecture Search (NAS) approach called NAS-DETR for object detection in sonar images. First, an improved Zero-shot Neural Architecture Search (NAS) method based on the maximum entropy principle is proposed to identify a real-time, high-representational-capacity CNN-Transformer backbone for sonar image detection. This method enables the efficient discovery of high-performance network architectures with low computational and time overhead. Subsequently, the backbone is combined with a Feature Pyramid Network (FPN) and a deformable attention-based Transformer decoder to construct a complete network architecture. This architecture integrates various advanced components and training schemes to enhance overall performance. Extensive experiments demonstrate that this architecture achieves state-of-the-art performance on two Representative datasets, while maintaining minimal overhead in real-time efficiency and computational complexity. Furthermore, correlation analysis between the key parameters and differential entropy-based fitness function is performed to enhance the interpretability of the proposed framework. To the best of our knowledge, this is the first work in the field of sonar object detection to integrate the DETR architecture with a NAS search mechanism.
Authors: Yicheng Song, Tiancheng Lin, Die Peng, Su Yang, Yi Xu
Abstract: Various multi-instance learning (MIL) based approaches have been developed and successfully applied to whole-slide pathological images (WSI). Existing MIL methods emphasize the importance of feature aggregators, but largely neglect the instance-level representation learning. They assume that the availability of a pre-trained feature extractor can be directly utilized or fine-tuned, which is not always the case. This paper proposes to pre-train feature extractor for MIL via a weakly-supervised scheme, i.e., propagating the weak bag-level labels to the corresponding instances for supervised learning. To learn effective features for MIL, we further delve into several key components, including strong data augmentation, a non-linear prediction head and the robust loss function. We conduct experiments on common large-scale WSI datasets and find it achieves better performance than other pre-training schemes (e.g., ImageNet pre-training and self-supervised learning) in different downstream tasks. We further show the compatibility and scalability of the proposed scheme by deploying it in fine-tuning the pathological-specific models and pre-training on merged multiple datasets. To our knowledge, this is the first work focusing on the representation learning for MIL.
Authors: Parth Padalkar, Gopal Gupta
Abstract: Recent neuro-symbolic approaches have successfully extracted symbolic rule-sets from CNN-based models to enhance interpretability. However, applying similar techniques to Vision Transformers (ViTs) remains challenging due to their lack of modular concept detectors and reliance on global self-attention mechanisms. We propose a framework for symbolic rule extraction from ViTs by introducing a sparse concept layer inspired by Sparse Autoencoders (SAEs). This linear layer operates on attention-weighted patch representations and learns a disentangled, binarized representation in which individual neurons activate for high-level visual concepts. To encourage interpretability, we apply a combination of L1 sparsity, entropy minimization, and supervised contrastive loss. These binarized concept activations are used as input to the FOLD-SE-M algorithm, which generates a rule-set in the form of logic programs. Our method achieves a 5.14% better classification accuracy than the standard ViT while enabling symbolic reasoning. Crucially, the extracted rule-set is not merely post-hoc but acts as a logic-based decision layer that operates directly on the sparse concept representations. The resulting programs are concise and semantically meaningful. This work is the first to extract executable logic programs from ViTs using sparse symbolic representations. It bridges the gap between transformer-based vision models and symbolic logic programming, providing a step forward in interpretable and verifiable neuro-symbolic AI.
Authors: Ye Zhu, Yunan Wang, Zitong Yu
Abstract: Multimodal news contains a wealth of information and is easily affected by deepfake modeling attacks. To combat the latest image and text generation methods, we present a new Multimodal Fake News Detection dataset (MFND) containing 11 manipulated types, designed to detect and localize highly authentic fake news. Furthermore, we propose a Shallow-Deep Multitask Learning (SDML) model for fake news, which fully uses unimodal and mutual modal features to mine the intrinsic semantics of news. Under shallow inference, we propose the momentum distillation-based light punishment contrastive learning for fine-grained uniform spatial image and text semantic alignment, and an adaptive cross-modal fusion module to enhance mutual modal features. Under deep inference, we design a two-branch framework to augment the image and text unimodal features, respectively merging with mutual modalities features, for four predictions via dedicated detection and localization projections. Experiments on both mainstream and our proposed datasets demonstrate the superiority of the model. Codes and dataset are released at https://github.com/yunan-wang33/sdml.
Authors: Bin Li, Shenxi Liu, Yixuan Weng, Yue Du, Yuhang Tian, Shoujun Zhou
Abstract: Following the successful hosts of the 1-st (NLPCC 2023 Foshan) CMIVQA and the 2-rd (NLPCC 2024 Hangzhou) MMIVQA challenges, this year, a new task has been introduced to further advance research in multi-modal, multilingual, and multi-hop medical instructional question answering (M4IVQA) systems, with a specific focus on medical instructional videos. The M4IVQA challenge focuses on evaluating models that integrate information from medical instructional videos, understand multiple languages, and answer multi-hop questions requiring reasoning over various modalities. This task consists of three tracks: multi-modal, multilingual, and multi-hop Temporal Answer Grounding in Single Video (M4TAGSV), multi-modal, multilingual, and multi-hop Video Corpus Retrieval (M4VCR) and multi-modal, multilingual, and multi-hop Temporal Answer Grounding in Video Corpus (M4TAGVC). Participants in M4IVQA are expected to develop algorithms capable of processing both video and text data, understanding multilingual queries, and providing relevant answers to multi-hop medical questions. We believe the newly introduced M4IVQA challenge will drive innovations in multimodal reasoning systems for healthcare scenarios, ultimately contributing to smarter emergency response systems and more effective medical education platforms in multilingual communities. Our official website is https://cmivqa.github.io/
Authors: Thien Nhan Vo
Abstract: A principle bottleneck in image classification is the large number of training examples needed to train a classifier. Using active learning, we can reduce the number of training examples to teach a CNN classifier by strategically selecting examples. Assigning values to image examples using different uncertainty metrics allows the model to identify and select high-value examples in a smaller training set size. We demonstrate results for digit recognition and fruit classification on the MNIST and Fruits360 data sets. We formally compare results for four different uncertainty metrics. Finally, we observe active learning is also effective on simpler (binary) classification tasks, but marked improvement from random sampling is more evident on more difficult tasks. We show active learning is a viable algorithm for image classification problems.
Authors: Miaoyun Zhao, Qiang Zhang, Chenrong Li
Abstract: Achieving group-robust generalization in the presence of spurious correlations remains a significant challenge, particularly when bias annotations are unavailable. Recent studies on Class-Conditional Distribution Balancing (CCDB) reveal that spurious correlations often stem from mismatches between the class-conditional and marginal distributions of bias attributes. They achieve promising results by addressing this issue through simple distribution matching in a bias-agnostic manner. However, CCDB approximates each distribution using a single Gaussian, which is overly simplistic and rarely holds in real-world applications. To address this limitation, we propose a novel method called Bias Exploration via Overfitting (BEO), which captures each distribution in greater detail by modeling it as a mixture of latent groups. Building on these group-level descriptions, we introduce a fine-grained variant of CCDB, termed FG-CCDB, which performs more precise distribution matching and balancing within each group. Through group-level reweighting, FG-CCDB learns sample weights from a global perspective, achieving stronger mitigation of spurious correlations without incurring substantial storage or computational costs. Extensive experiments demonstrate that BEO serves as a strong proxy for ground-truth bias annotations and can be seamlessly integrated with bias-supervised methods. Moreover, when combined with FG-CCDB, our method performs on par with bias-supervised approaches on binary classification tasks and significantly outperforms them in highly biased multi-class scenarios.
Authors: Yixin Chen, Shuai Zhang, Boran Han, Bernie Wang
Abstract: High-resolution (HR) images are pivotal for enhancing the recognition and understanding capabilities of multimodal large language models (MLLMs). However, directly increasing image resolution can significantly escalate computational demands. In this study, we propose a method called Chain of Region-of-Interest (CoRoI) for Visual Instruction Tuning, aimed at alleviating the computational burden associated with high-resolution images for MLLMs. Drawing inspiration from the selective nature of the human visual system, we recognize that not all regions within high-resolution images carry equal importance. CoRoI seeks to identify and prioritize the most informative regions, thereby enhancing multimodal visual comprehension and recognition while circumventing the need for processing lengthy HR image tokens. Through extensive experiments on 11 benchmarks, we validate the efficacy of CoRoI across varying sizes, ranging from 7B to 34B in parameters. Our models consistently demonstrate superior performance across diverse multimodal benchmarks and tasks. Notably, our method outperforms LLaVA-NeXT on almost all benchmarks and our finetuned 34B model surpasses proprietary methods like Gemini Pro 1.0 on six benchmarks, as well as outperforming GPT-4V on MMB, SEED-I, and MME.
Authors: Carolina Vargas-Ecos, Edwin Salcedo
Abstract: According to the Pan American Health Organization, the number of cancer cases in Latin America was estimated at 4.2 million in 2022 and is projected to rise to 6.7 million by 2045. Osteosarcoma, one of the most common and deadly bone cancers affecting young people, is difficult to detect due to its unique texture and intensity. Surgical removal of osteosarcoma requires precise safety margins to ensure complete resection while preserving healthy tissue. Therefore, this study proposes a method for estimating the confidence interval of surgical safety margins in osteosarcoma surgery around the knee. The proposed approach uses MRI and X-ray data from open-source repositories, digital processing techniques, and unsupervised learning algorithms (such as k-means clustering) to define tumor boundaries. Experimental results highlight the potential for automated, patient-specific determination of safety margins.
Authors: Zhengmi Tang, Yuto Mitsui, Tomo Miyazaki, Shinichiro Omachi
Abstract: Most existing text recognition methods are trained on large-scale synthetic datasets due to the scarcity of labeled real-world datasets. Synthetic images, however, cannot faithfully reproduce real-world scenarios, such as uneven illumination, irregular layout, occlusion, and degradation, resulting in performance disparities when handling complex real-world images. Recent self-supervised learning techniques, notably contrastive learning and masked image modeling (MIM), narrow this domain gap by exploiting unlabeled real text images. This study first analyzes the original Masked AutoEncoder (MAE) and observes that random patch masking predominantly captures low-level textural features but misses high-level contextual representations. To fully exploit the high-level contextual representations, we introduce random blockwise and span masking in the text recognition task. These strategies can mask the continuous image patches and completely remove some characters, forcing the model to infer relationships among characters within a word. Our Multi-Masking Strategy (MMS) integrates random patch, blockwise, and span masking into the MIM frame, which jointly learns low and high-level textual representations. After fine-tuning with real data, MMS outperforms the state-of-the-art self-supervised methods in various text-related tasks, including text recognition, segmentation, and text-image super-resolution.
Authors: Hamd Jalil, Ahmed Qazi, Asim Iqbal
Abstract: Domain generalization in image classification is a crucial challenge, with models often failing to generalize well across unseen datasets. We address this issue by introducing a neuro-inspired Neural Response Normalization (NeuRN) layer which draws inspiration from neurons in the mammalian visual cortex, which aims to enhance the performance of deep learning architectures on unseen target domains by training deep learning models on a source domain. The performance of these models is considered as a baseline and then compared against models integrated with NeuRN on image classification tasks. We perform experiments across a range of deep learning architectures, including ones derived from Neural Architecture Search and Vision Transformer. Additionally, in order to shortlist models for our experiment from amongst the vast range of deep neural networks available which have shown promising results, we also propose a novel method that uses the Needleman-Wunsch algorithm to compute similarity between deep learning architectures. Our results demonstrate the effectiveness of NeuRN by showing improvement against baseline in cross-domain image classification tasks. Our framework attempts to establish a foundation for future neuro-inspired deep learning models.
Authors: Ahmed Qazi, Hamd Jalil, Asim Iqbal
Abstract: The mouse is one of the most studied animal models in the field of systems neuroscience. Understanding the generalized patterns and decoding the neural representations that are evoked by the diverse range of natural scene stimuli in the mouse visual cortex is one of the key quests in computational vision. In recent years, significant parallels have been drawn between the primate visual cortex and hierarchical deep neural networks. However, their generalized efficacy in understanding mouse vision has been limited. In this study, we investigate the functional alignment between the mouse visual cortex and deep learning models for object classification tasks. We first introduce a generalized representational learning strategy that uncovers a striking resemblance between the functional mapping of the mouse visual cortex and high-performing deep learning models on both top-down (population-level) and bottom-up (single cell-level) scenarios. Next, this representational similarity across the two systems is further enhanced by the addition of Neural Response Normalization (NeuRN) layer, inspired by the activation profile of excitatory and inhibitory neurons in the visual cortex. To test the performance effect of NeuRN on real-world tasks, we integrate it into deep learning models and observe significant improvements in their robustness against data shifts in domain generalization tasks. Our work proposes a novel framework for comparing the functional architecture of the mouse visual cortex with deep learning models. Our findings carry broad implications for the development of advanced AI models that draw inspiration from the mouse visual cortex, suggesting that these models serve as valuable tools for studying the neural representations of the mouse visual cortex and, as a result, enhancing their performance on real-world tasks.
Authors: Ahmed Qazi, Abdul Basit, Asim Iqbal
Abstract: Neural Radiance Fields (NeRF) have significantly advanced the field of novel view synthesis, yet their generalization across diverse scenes and conditions remains challenging. Addressing this, we propose the integration of a novel brain-inspired normalization technique Neural Generalization (NeuGen) into leading NeRF architectures which include MVSNeRF and GeoNeRF. NeuGen extracts the domain-invariant features, thereby enhancing the models' generalization capabilities. It can be seamlessly integrated into NeRF architectures and cultivates a comprehensive feature set that significantly improves accuracy and robustness in image rendering. Through this integration, NeuGen shows improved performance on benchmarks on diverse datasets across state-of-the-art NeRF architectures, enabling them to generalize better across varied scenes. Our comprehensive evaluations, both quantitative and qualitative, confirm that our approach not only surpasses existing models in generalizability but also markedly improves rendering quality. Our work exemplifies the potential of merging neuroscientific principles with deep learning frameworks, setting a new precedent for enhanced generalizability and efficiency in novel view synthesis. A demo of our study is available at https://neugennerf.github.io.
Authors: Honglong Yang, Shanshan Song, Yi Qin, Lehan Wang, Haonan Wang, Xinpeng Ding, Qixiang Zhang, Bodong Du, Xiaomeng Li
Abstract: Generalist Medical AI (GMAI) systems have demonstrated expert-level performance in biomedical perception tasks, yet their clinical utility remains limited by inadequate multi-modal explainability and suboptimal prognostic capabilities. Here, we present XMedGPT, a clinician-centric, multi-modal AI assistant that integrates textual and visual interpretability to support transparent and trustworthy medical decision-making. XMedGPT not only produces accurate diagnostic and descriptive outputs, but also grounds referenced anatomical sites within medical images, bridging critical gaps in interpretability and enhancing clinician usability. To support real-world deployment, we introduce a reliability indexing mechanism that quantifies uncertainty through consistency-based assessment via interactive question-answering. We validate XMedGPT across four pillars: multi-modal interpretability, uncertainty quantification, and prognostic modeling, and rigorous benchmarking. The model achieves an IoU of 0.703 across 141 anatomical regions, and a Kendall's tau-b of 0.479, demonstrating strong alignment between visual rationales and clinical outcomes. For uncertainty estimation, it attains an AUC of 0.862 on visual question answering and 0.764 on radiology report generation. In survival and recurrence prediction for lung and glioma cancers, it surpasses prior leading models by 26.9%, and outperforms GPT-4o by 25.0%. Rigorous benchmarking across 347 datasets covers 40 imaging modalities and external validation spans 4 anatomical systems confirming exceptional generalizability, with performance gains surpassing existing GMAI by 20.7% for in-domain evaluation and 16.7% on 11,530 in-house data evaluation. Together, XMedGPT represents a significant leap forward in clinician-centric AI integration, offering trustworthy and scalable support for diverse healthcare applications.
Authors: Yuanzhuo Wang, Junwen Duan, Xinyu Li, Jianxin Wang
Abstract: Temporal medical image analysis is essential for clinical decision-making, yet existing methods either align images and text at a coarse level - causing potential semantic mismatches - or depend solely on visual information, lacking medical semantic integration. We present CheXLearner, the first end-to-end framework that unifies anatomical region detection, Riemannian manifold-based structure alignment, and fine-grained regional semantic guidance. Our proposed Med-Manifold Alignment Module (Med-MAM) leverages hyperbolic geometry to robustly align anatomical structures and capture pathologically meaningful discrepancies across temporal chest X-rays. By introducing regional progression descriptions as supervision, CheXLearner achieves enhanced cross-modal representation learning and supports dynamic low-level feature optimization. Experiments show that CheXLearner achieves 81.12% (+17.2%) average accuracy and 80.32% (+11.05%) F1-score on anatomical region progression detection - substantially outperforming state-of-the-art baselines, especially in structurally complex regions. Additionally, our model attains a 91.52% average AUC score in downstream disease classification, validating its superior feature representation.
Authors: Jian Song, Hongruixuan Chen, Naoto Yokoya
Abstract: Monocular height estimation (MHE) from very-high-resolution (VHR) remote sensing imagery via deep learning is notoriously challenging due to the lack of sufficient structural information. Conventional digital elevation models (DEMs), typically derived from airborne LiDAR or multi-view stereo, remain costly and geographically limited. Recently, models trained on synthetic data and refined through domain adaptation have shown remarkable performance in MHE, yet it remains unclear how these models make predictions or how reliable they truly are. In this paper, we investigate a state-of-the-art MHE model trained purely on synthetic data to explore where the model looks when making height predictions. Through systematic analyses, we find that the model relies heavily on shadow cues, a factor that can lead to overestimation or underestimation of heights when shadows deviate from expected norms. Furthermore, the inherent difficulty of evaluating regression tasks with the human eye underscores additional limitations of purely synthetic training. To address these issues, we propose a novel correction pipeline that integrates sparse, imperfect global LiDAR measurements (ICESat-2) with deep-learning outputs to improve local accuracy and achieve spatially consistent corrections. Our method comprises two stages: pre-processing raw ICESat-2 data, followed by a random forest-based approach to densely refine height estimates. Experiments in three representative urban regions -- Saint-Omer, Tokyo, and Sao Paulo -- reveal substantial error reductions, with mean absolute error (MAE) decreased by 22.8\%, 6.9\%, and 4.9\%, respectively. These findings highlight the critical role of shadow awareness in synthetic data-driven models and demonstrate how fusing imperfect real-world LiDAR data can bolster the robustness of MHE, paving the way for more reliable and scalable 3D mapping solutions.
Authors: Chao Ding, Mouxiao Bian, Pengcheng Chen, Hongliang Zhang, Tianbin Li, Lihao Liu, Jiayuan Chen, Zhuoran Li, Yabei Zhong, Yongqi Liu, Haiqing Huang, Dongming Shan, Junjun He, Jie Xu
Abstract: Despite strong performance in medical question-answering, the clinical adoption of Large Language Models (LLMs) is critically hampered by their opaque 'black-box' reasoning, limiting clinician trust. This challenge is compounded by the predominant reliance of current medical LLMs on corpora from scientific literature or synthetic data, which often lack the granular expert validation and high clinical relevance essential for advancing their specialized medical capabilities. To address these critical gaps, we introduce a highly clinically relevant dataset with 31,247 medical question-answer pairs, each accompanied by expert-validated chain-of-thought (CoT) explanations. This resource, spanning multiple clinical domains, was curated via a scalable human-LLM hybrid pipeline: LLM-generated rationales were iteratively reviewed, scored, and refined by medical experts against a structured rubric, with substandard outputs revised through human effort or guided LLM regeneration until expert consensus. This publicly available dataset provides a vital source for the development of medical LLMs that capable of transparent and verifiable reasoning, thereby advancing safer and more interpretable AI in medicine.
Authors: Timing Li, Bing Cao, Pengfei Zhu, Bin Xiao, Qinghua Hu
Abstract: Acquiring accurately aligned multi-modal image pairs is fundamental for achieving high-quality multi-modal image fusion. To address the lack of ground truth in current multi-modal image registration and fusion methods, we propose a novel self-supervised \textbf{B}i-directional \textbf{S}elf-\textbf{R}egistration framework (\textbf{B-SR}). Specifically, B-SR utilizes a proxy data generator (PDG) and an inverse proxy data generator (IPDG) to achieve self-supervised global-local registration. Visible-infrared image pairs with spatially misaligned differences are aligned to obtain global differences through the registration module. The same image pairs are processed by PDG, such as cropping, flipping, stitching, etc., and then aligned to obtain local differences. IPDG converts the obtained local differences into pseudo-global differences, which are used to perform global-local difference consistency with the global differences. Furthermore, aiming at eliminating the effect of modal gaps on the registration module, we design a neighborhood dynamic alignment loss to achieve cross-modal image edge alignment. Extensive experiments on misaligned multi-modal images demonstrate the effectiveness of the proposed method in multi-modal image alignment and fusion against the competing methods. Our code will be publicly available.
Authors: Fei Zhou, Yi Li, Mingqing Zhu
Abstract: In this paper, the dual-optical attention fusion crowd head point counting model (TAPNet) is proposed to address the problem of the difficulty of accurate counting in complex scenes such as crowd dense occlusion and low light in crowd counting tasks under UAV view. The model designs a dual-optical attention fusion module (DAFP) by introducing complementary information from infrared images to improve the accuracy and robustness of all-day crowd counting. In order to fully utilize different modal information and solve the problem of inaccurate localization caused by systematic misalignment between image pairs, this paper also proposes an adaptive two-optical feature decomposition fusion module (AFDF). In addition, we optimize the training strategy to improve the model robustness through spatial random offset data augmentation. Experiments on two challenging public datasets, DroneRGBT and GAIIC2, show that the proposed method outperforms existing techniques in terms of performance, especially in challenging dense low-light scenes. Code is available at https://github.com/zz-zik/TAPNet
Authors: Pan Du, Wangbo Zhao, Xinai Lu, Nian Liu, Zhikai Li, Chaoyu Gong, Suyun Zhao, Hong Chen, Cuiping Li, Kai Wang, Yang You
Abstract: Class distribution mismatch (CDM) refers to the discrepancy between class distributions in training data and target tasks. Previous methods address this by designing classifiers to categorize classes known during training, while grouping unknown or new classes into an "other" category. However, they focus on semi-supervised scenarios and heavily rely on labeled data, limiting their applicability and performance. To address this, we propose Unsupervised Learning for Class Distribution Mismatch (UCDM), which constructs positive-negative pairs from unlabeled data for classifier training. Our approach randomly samples images and uses a diffusion model to add or erase semantic classes, synthesizing diverse training pairs. Additionally, we introduce a confidence-based labeling mechanism that iteratively assigns pseudo-labels to valuable real-world data and incorporates them into the training process. Extensive experiments on three datasets demonstrate UCDM's superiority over previous semi-supervised methods. Specifically, with a 60% mismatch proportion on Tiny-ImageNet dataset, our approach, without relying on labeled data, surpasses OpenMatch (with 40 labels per class) by 35.1%, 63.7%, and 72.5% in classifying known, unknown, and new classes.
Authors: Seokjun Kwon, Jeongmin Shin, Namil Kim, Soonmin Hwang, Yukyung Choi
Abstract: In autonomous driving, thermal image semantic segmentation has emerged as a critical research area, owing to its ability to provide robust scene understanding under adverse visual conditions. In particular, unsupervised domain adaptation (UDA) for thermal image segmentation can be an efficient solution to address the lack of labeled thermal datasets. Nevertheless, since these methods do not effectively utilize the complementary information between RGB and thermal images, they significantly decrease performance during domain adaptation. In this paper, we present a comprehensive study on cross-spectral UDA for thermal image semantic segmentation. We first propose a novel masked mutual learning strategy that promotes complementary information exchange by selectively transferring results between each spectral model while masking out uncertain regions. Additionally, we introduce a novel prototypical self-supervised loss designed to enhance the performance of the thermal segmentation model in nighttime scenarios. This approach addresses the limitations of RGB pre-trained networks, which cannot effectively transfer knowledge under low illumination due to the inherent constraints of RGB sensors. In experiments, our method achieves higher performance over previous UDA methods and comparable performance to state-of-the-art supervised methods.
Authors: Wei Shang, Dongwei Ren, Wanying Zhang, Pengfei Zhu, Qinghua Hu, Wangmeng Zuo
Abstract: The primary challenge in accelerating image super-resolution lies in reducing computation while maintaining performance and adaptability. Motivated by the observation that high-frequency regions (e.g., edges and textures) are most critical for reconstruction, we propose a training-free adaptive masking module for acceleration that dynamically focuses computation on these challenging areas. Specifically, our method first extracts high-frequency components via Gaussian blur subtraction and adaptively generates binary masks using K-means clustering to identify regions requiring intensive processing. Our method can be easily integrated with both CNNs and Transformers. For CNN-based architectures, we replace standard $3 \times 3$ convolutions with an unfold operation followed by $1 \times 1$ convolutions, enabling pixel-wise sparse computation guided by the mask. For Transformer-based models, we partition the mask into non-overlapping windows and selectively process tokens based on their average values. During inference, unnecessary pixels or windows are pruned, significantly reducing computation. Moreover, our method supports dilation-based mask adjustment to control the processing scope without retraining, and is robust to unseen degradations (e.g., noise, compression). Extensive experiments on benchmarks demonstrate that our method reduces FLOPs by 24--43% for state-of-the-art models (e.g., CARN, SwinIR) while achieving comparable or better quantitative metrics. The source code is available at https://github.com/shangwei5/AMSR
Authors: Md. Naimur Asif Borno, Md Sakib Hossain Shovon, MD Hanif Sikder, Iffat Firozy Rimi, Tahani Jaser Alahmadi, Mohammad Ali Moni
Abstract: Recent progress in image-based medical disease detection encounters challenges such as limited annotated data sets, inadequate spatial feature analysis, data security issues, and inefficient training frameworks. This study introduces a data-efficient image transformer (DeIT)-based approach that overcomes these challenges by utilizing multiscale patch embedding for better feature extraction and stratified weighted random sampling to address class imbalance. The model also incorporates a LoRA-enhanced transformer encoder, a distillation framework, and federated learning for decentralized training, improving both efficiency and data security. Consequently, it achieves state-of-the-art performance, with the highest AUC, F1 score, precision, minimal loss, and Top-5 accuracy. Additionally, Grad-CAM++ visualizations improve interpretability by highlighting critical pathological regions, enhancing the model's clinical relevance. These results highlight the potential of this approach to advance AI-powered medical imaging and disease detection.
Authors: Panwen Hu, Jiehui Huang, Qiang Sun, Xiaodan Liang
Abstract: Both zero-shot and tuning-based customized text-to-image (CT2I) generation have made significant progress for storytelling content creation. In contrast, research on customized text-to-video (CT2V) generation remains relatively limited. Existing zero-shot CT2V methods suffer from poor generalization, while another line of work directly combining tuning-based T2I models with temporal motion modules often leads to the loss of structural and texture information. To bridge this gap, we propose an autoregressive structure and texture propagation module (STPM), which extracts key structural and texture features from the reference subject and injects them autoregressively into each video frame to enhance consistency. Additionally, we introduce a test-time reward optimization (TTRO) method to further refine fine-grained details. Quantitative and qualitative experiments validate the effectiveness of STPM and TTRO, demonstrating improvements of 7.8 and 13.1 in CLIP-I and DINO consistency metrics over the baseline, respectively.
Authors: Chih-Chung Hsu, I-Hsuan Wu, Wen-Hai Tseng, Ching-Heng Cheng, Ming-Hsuan Wu, Jin-Hui Jiang, Yu-Jou Hsiao
Abstract: This report presents our semantic segmentation framework developed by team ACVLAB for the ICRA 2025 GOOSE 2D Semantic Segmentation Challenge, which focuses on parsing outdoor scenes into nine semantic categories under real-world conditions. Our method integrates a Swin Transformer backbone enhanced with Rotary Position Embedding (RoPE) for improved spatial generalization, alongside a Color Shift Estimation-and-Correction module designed to compensate for illumination inconsistencies in natural environments. To further improve training stability, we adopt a quantile-based denoising strategy that downweights the top 2.5\% of highest-error pixels, treating them as noise and suppressing their influence during optimization. Evaluated on the official GOOSE test set, our approach achieved a mean Intersection over Union (mIoU) of 0.848, demonstrating the effectiveness of combining color correction, positional encoding, and error-aware denoising in robust semantic segmentation.
Authors: Md. Naimur Asif Borno, Md Sakib Hossain Shovon, Asmaa Soliman Al-Moisheer, Mohammad Ali Moni
Abstract: Recent advancements in text-to-image diffusion models are hindered by high computational demands, limiting accessibility and scalability. This paper introduces KDC-Diff, a novel stable diffusion framework that enhances efficiency while maintaining image quality. KDC-Diff features a streamlined U-Net architecture with nearly half the parameters of the original U-Net (482M), significantly reducing model complexity. We propose a dual-layered distillation strategy to ensure high-fidelity generation, transferring semantic and structural insights from a teacher to a compact student model while minimizing quality degradation. Additionally, replay-based continual learning is integrated to mitigate catastrophic forgetting, allowing the model to retain prior knowledge while adapting to new data. Despite operating under extremely low computational resources, KDC-Diff achieves state-of-the-art performance on the Oxford Flowers and Butterflies & Moths 100 Species datasets, demonstrating competitive metrics such as FID, CLIP, and LPIPS. Moreover, it significantly reduces inference time compared to existing models. These results establish KDC-Diff as a highly efficient and adaptable solution for text-to-image generation, particularly in computationally constrained environments.
Authors: Bidur Khanal, Sandesh Pokhrel, Sanjay Bhandari, Ramesh Rana, Nikesh Shrestha, Ram Bahadur Gurung, Cristian Linte, Angus Watson, Yash Raj Shrestha, Binod Bhattarai
Abstract: Vision-Language Models (VLMs) are becoming increasingly popular in the medical domain, bridging the gap between medical images and clinical language. Existing VLMs demonstrate an impressive ability to comprehend medical images and text queries to generate detailed, descriptive diagnostic medical reports. However, hallucination--the tendency to generate descriptions that are inconsistent with the visual content--remains a significant issue in VLMs, with particularly severe implications in the medical field. To facilitate VLM research on gastrointestinal (GI) image analysis and study hallucination, we curate a multimodal image-text GI dataset: Gut-VLM. This dataset is created using a two-stage pipeline: first, descriptive medical reports of Kvasir-v2 images are generated using ChatGPT, which introduces some hallucinated or incorrect texts. In the second stage, medical experts systematically review these reports, and identify and correct potential inaccuracies to ensure high-quality, clinically reliable annotations. Unlike traditional datasets that contain only descriptive texts, our dataset also features tags identifying hallucinated sentences and their corresponding corrections. A common approach to reducing hallucination in VLM is to finetune the model on a small-scale, problem-specific dataset. However, we take a different strategy using our dataset. Instead of finetuning the VLM solely for generating textual reports, we finetune it to detect and correct hallucinations, an approach we call hallucination-aware finetuning. Our results show that this approach is better than simply finetuning for descriptive report generation. Additionally, we conduct an extensive evaluation of state-of-the-art VLMs across several metrics, establishing a benchmark. GitHub Repo: https://github.com/bhattarailab/Hallucination-Aware-VLM.
URLs: https://github.com/bhattarailab/Hallucination-Aware-VLM.
Authors: Peng Li, Suizhi Ma, Jialiang Chen, Yuan Liu, Chongyi Zhang, Wei Xue, Wenhan Luo, Alla Sheffer, Wenping Wang, Yike Guo
Abstract: Recently, 3D generation methods have shown their powerful ability to automate 3D model creation. However, most 3D generation methods only rely on an input image or a text prompt to generate a 3D model, which lacks the control of each component of the generated 3D model. Any modifications of the input image lead to an entire regeneration of the 3D models. In this paper, we introduce a new method called CMD that generates a 3D model from an input image while enabling flexible local editing of each component of the 3D model. In CMD, we formulate the 3D generation as a conditional multiview diffusion model, which takes the existing or known parts as conditions and generates the edited or added components. This conditional multiview diffusion model not only allows the generation of 3D models part by part but also enables local editing of 3D models according to the local revision of the input image without changing other 3D parts. Extensive experiments are conducted to demonstrate that CMD decomposes a complex 3D generation task into multiple components, improving the generation quality. Meanwhile, CMD enables efficient and flexible local editing of a 3D model by just editing one rendered image.
Authors: Zhengye Zhang, Sirui Zhao, Shifeng Liu, Shukang Yin, Xinglong Mao, Tong Xu, Enhong Chen
Abstract: Micro-expressions (MEs) are crucial psychological responses with significant potential for affective computing. However, current automatic micro-expression recognition (MER) research primarily focuses on discrete emotion classification, neglecting a convincing analysis of the subtle dynamic movements and inherent emotional cues. The rapid progress in multimodal large language models (MLLMs), known for their strong multimodal comprehension and language generation abilities, offers new possibilities. MLLMs have shown success in various vision-language tasks, indicating their potential to understand MEs comprehensively, including both fine-grained motion patterns and underlying emotional semantics. Nevertheless, challenges remain due to the subtle intensity and short duration of MEs, as existing MLLMs are not designed to capture such delicate frame-level facial dynamics. In this paper, we propose a novel Micro-Expression Large Language Model (MELLM), which incorporates a subtle facial motion perception strategy with the strong inference capabilities of MLLMs, representing the first exploration of MLLMs in the domain of ME analysis. Specifically, to explicitly guide the MLLM toward motion-sensitive regions, we construct an interpretable motion-enhanced color map by fusing onset-apex optical flow dynamics with the corresponding grayscale onset frame as the model input. Additionally, specialized fine-tuning strategies are incorporated to further enhance the model's visual perception of MEs. Furthermore, we construct an instruction-description dataset based on Facial Action Coding System (FACS) annotations and emotion labels to train our MELLM. Comprehensive evaluations across multiple benchmark datasets demonstrate that our model exhibits superior robustness and generalization capabilities in ME understanding (MEU). Code is available at https://github.com/zyzhangUstc/MELLM.
Authors: Jitesh Joshi, Youngjun Cho
Abstract: Remote physiological sensing using camera-based technologies offers transformative potential for non-invasive vital sign monitoring across healthcare and human-computer interaction domains. Although deep learning approaches have advanced the extraction of physiological signals from video data, existing methods have not been sufficiently assessed for their robustness to domain shifts. These shifts in remote physiological sensing include variations in ambient conditions, camera specifications, head movements, facial poses, and physiological states which often impact real-world performance significantly. Cross-dataset evaluation provides an objective measure to assess generalization capabilities across these domain shifts. We introduce Target Signal Constrained Factorization module (TSFM), a novel multidimensional attention mechanism that explicitly incorporates physiological signal characteristics as factorization constraints, allowing more precise feature extraction. Building on this innovation, we present MMRPhys, an efficient dual-branch 3D-CNN architecture designed for simultaneous multitask estimation of photoplethysmography (rPPG) and respiratory (rRSP) signals from multimodal RGB and thermal video inputs. Through comprehensive cross-dataset evaluation on five benchmark datasets, we demonstrate that MMRPhys with TSFM significantly outperforms state-of-the-art methods in generalization across domain shifts for rPPG and rRSP estimation, while maintaining a minimal inference latency suitable for real-time applications. Our approach establishes new benchmarks for robust multitask and multimodal physiological sensing and offers a computationally efficient framework for practical deployment in unconstrained environments. The web browser-based application featuring on-device real-time inference of MMRPhys model is available at https://physiologicailab.github.io/mmrphys-live
Authors: Khang Nguyen Quoc, Lan Le Thi Thu, Luyl-Da Quach
Abstract: Leaf disease identification plays a pivotal role in smart agriculture. However, many existing studies still struggle to integrate image and textual modalities to compensate for each other's limitations. Furthermore, many of these approaches rely on pretraining with constrained datasets such as ImageNet, which lack domain-specific information. We propose SCOLD (Soft-target COntrastive learning for Leaf Disease identification), a context-aware vision-language foundation model tailored to address these challenges for agricultural tasks. SCOLD is developed using a diverse corpus of plant leaf images and corresponding symptom descriptions, comprising over 186,000 image-caption pairs aligned with 97 unique concepts. Through task-agnostic pretraining, SCOLD leverages contextual soft targets to mitigate overconfidence in contrastive learning by smoothing labels, thereby improving model generalization and robustness on fine-grained classification tasks. Experimental results demonstrate that SCOLD outperforms existing vision-language models such as OpenAI-CLIP-L, BioCLIP, and SigLIP2 across several benchmarks, including zero-shot and few-shot classification, image-text retrieval, and image classification, while maintaining a competitive parameter footprint. Ablation studies further highlight SCOLD's effectiveness in contrast to its counterparts. The proposed approach significantly advances the agricultural vision-language foundation model, offering strong performance with minimal or no supervised fine-tuning. This work lays a solid groundwork for future research on models trained with long-form and simplified contexts, tasks involving class ambiguity, and multi-modal systems for intelligent plant disease diagnostics. The code for this study is available at https://huggingface.co/enalis/scold
Authors: Fei Zhao, Runlin Zhang, Chengcui Zhang, Nitesh Saxena
Abstract: We present MarkMatch, a retrieval system for detecting whether two paper ballot marks were filled by the same hand. Unlike the previous SOTA method BubbleSig, which used binary classification on isolated mark pairs, MarkMatch ranks stylistic similarity between a query mark and a mark in the database using contrastive learning. Our model is trained with a dense batch similarity matrix and a dual loss objective. Each sample is contrasted against many negatives within each batch, enabling the model to learn subtle handwriting difference and improve generalization under handwriting variation and visual noise, while diagonal supervision reinforces high confidence on true matches. The model achieves an F1 score of 0.943, surpassing BubbleSig's best performance. MarkMatch also integrates Segment Anything Model for flexible mark extraction via box- or point-based prompts. The system offers election auditors a practical tool for visual, non-biometric investigation of suspicious ballots.
Authors: Zhengyang Lu, Bingjie Lu, Weifan Wang, Feng Wang
Abstract: Fabric defect detection confronts two fundamental challenges. First, conventional non-maximum suppression disrupts gradient flow, which hinders genuine end-to-end learning. Second, acquiring pixel-level annotations at industrial scale is prohibitively costly. Addressing these limitations, we propose a differentiable NMS framework for fabric defect detection that achieves superior localization precision through end-to-end optimization. We reformulate NMS as a differentiable bipartite matching problem solved through the Sinkhorn-Knopp algorithm, maintaining uninterrupted gradient flow throughout the network. This approach specifically targets the irregular morphologies and ambiguous boundaries of fabric defects by integrating proposal quality, feature similarity, and spatial relationships. Our entropy-constrained mask refinement mechanism further enhances localization precision through principled uncertainty modeling. Extensive experiments on the Tianchi fabric defect dataset demonstrate significant performance improvements over existing methods while maintaining real-time speeds suitable for industrial deployment. The framework exhibits remarkable adaptability across different architectures and generalizes effectively to general object detection tasks.
Authors: Binbin Wei, Yuhang Zhang, Shishun Tian, Muxin Liao, Wei Li, Wenbin Zou
Abstract: Unsupervised Domain Adaptation (UDA) aims to align source and target domain distributions to close the domain gap, but still struggles with obtaining the target data. Fortunately, Domain Generalization (DG) excels without the need for any target data. Recent works expose that depth maps contribute to improved generalized performance in the UDA tasks, but they ignore the noise and holes in depth maps due to device and environmental factors, failing to sufficiently and effectively learn domain-invariant representation. Although high-sensitivity region suppression has shown promising results in learning domain-invariant features, existing methods cannot be directly applicable to depth maps due to their unique characteristics. Hence, we propose a novel framework, namely Depth-Sensitive Soft Suppression with RGB-D inter-modal stylization flow (DSSS), focusing on learning domain-invariant features from depth maps for the DG semantic segmentation. Specifically, we propose the RGB-D inter-modal stylization flow to generate stylized depth maps for sensitivity detection, cleverly utilizing RGB information as the stylization source. Then, a class-wise soft spatial sensitivity suppression is designed to identify and emphasize non-sensitive depth features that contain more domain-invariant information. Furthermore, an RGB-D soft alignment loss is proposed to ensure that the stylized depth maps only align part of the RGB features while still retaining the unique depth information. To our best knowledge, our DSSS framework is the first work to integrate RGB and Depth information in the multi-class DG semantic segmentation task. Extensive experiments over multiple backbone networks show that our framework achieves remarkable performance improvement.
Authors: Junhao Xia, Chaoyang Zhang, Yecheng Zhang, Chengyang Zhou, Zhichang Wang, Bochun Liu, Dongshuo Yin
Abstract: Video generation based on diffusion models presents a challenging multimodal task, with video editing emerging as a pivotal direction in this field. Recent video editing approaches primarily fall into two categories: training-required and training-free methods. While training-based methods incur high computational costs, training-free alternatives often yield suboptimal performance. To address these limitations, we propose DAPE, a high-quality yet cost-effective two-stage parameter-efficient fine-tuning (PEFT) framework for video editing. In the first stage, we design an efficient norm-tuning method to enhance temporal consistency in generated videos. The second stage introduces a vision-friendly adapter to improve visual quality. Additionally, we identify critical shortcomings in existing benchmarks, including limited category diversity, imbalanced object distribution, and inconsistent frame counts. To mitigate these issues, we curate a large dataset benchmark comprising 232 videos with rich annotations and 6 editing prompts, enabling objective and comprehensive evaluation of advanced methods. Extensive experiments on existing datasets (BalanceCC, LOVEU-TGVE, RAVE) and our proposed benchmark demonstrate that DAPE significantly improves temporal coherence and text-video alignment while outperforming previous state-of-the-art approaches.
Authors: Dong Guo, Faming Wu, Feida Zhu, Fuxing Leng, Guang Shi, Haobin Chen, Haoqi Fan, Jian Wang, Jianyu Jiang, Jiawei Wang, Jingji Chen, Jingjia Huang, Kang Lei, Liping Yuan, Lishu Luo, Pengfei Liu, Qinghao Ye, Rui Qian, Shen Yan, Shixiong Zhao, Shuai Peng, Shuangye Li, Sihang Yuan, Sijin Wu, Tianheng Cheng, Weiwei Liu, Wenqian Wang, Xianhan Zeng, Xiao Liu, Xiaobo Qin, Xiaohan Ding, Xiaojun Xiao, Xiaoying Zhang, Xuanwei Zhang, Xuehan Xiong, Yanghua Peng, Yangrui Chen, Yanwei Li, Yanxu Hu, Yi Lin, Yiyuan Hu, Yiyuan Zhang, Youbin Wu, Yu Li, Yudong Liu, Yue Ling, Yujia Qin, Zanbo Wang, Zhiwu He, Aoxue Zhang, Bairen Yi, Bencheng Liao, Can Huang, Can Zhang, Chaorui Deng, Chaoyi Deng, Cheng Lin, Cheng Yuan, Chenggang Li, Chenhui Gou, Chenwei Lou, Chengzhi Wei, Chundian Liu, Chunyuan Li, Deyao Zhu, Donghong Zhong, Feng Li, Feng Zhang, Gang Wu, Guodong Li, Guohong Xiao, Haibin Lin, Haihua Yang, Haoming Wang, Heng Ji, Hongxiang Hao, Hui Shen, Huixia Li, Jiahao Li, Jialong Wu, Jianhua Zhu, Jianpeng Jiao, Jiashi Feng, Jiaze Chen, Jianhui Duan, Jihao Liu, Jin Zeng, Jingqun Tang, Jingyu Sun, Joya Chen, Jun Long, Junda Feng, Junfeng Zhan, Junjie Fang, Junting Lu, Kai Hua, Kai Liu, Kai Shen, Kaiyuan Zhang, Ke Shen, Ke Wang, Keyu Pan, Kun Zhang, Kunchang Li, Lanxin Li, Lei Li, Lei Shi, Li Han, Liang Xiang, Liangqiang Chen, Lin Chen, Lin Li, Lin Yan, Liying Chi, Longxiang Liu, Mengfei Du, Mingxuan Wang, Ningxin Pan, Peibin Chen, Pengfei Chen, Pengfei Wu, Qingqing Yuan, Qingyao Shuai, Qiuyan Tao, Renjie Zheng, Renrui Zhang, Ru Zhang, Rui Wang, Rui Yang, Rui Zhao, Shaoqiang Xu, Shihao Liang, Shipeng Yan, Shu Zhong, Shuaishuai Cao, Shuangzhi Wu, Shufan Liu, Shuhan Chang, Songhua Cai, Tenglong Ao, Tianhao Yang, Tingting Zhang, Wanjun Zhong, Wei Jia, Wei Weng, Weihao Yu, Wenhao Huang, Wenjia Zhu, Wenli Yang, Wenzhi Wang, Xiang Long, XiangRui Yin, Xiao Li, Xiaolei Zhu, Xiaoying Jia, Xijin Zhang, Xin Liu, Xinchen Zhang, Xinyu Yang, Xiongcai Luo, Xiuli Chen, Xuantong Zhong, Xuefeng Xiao, Xujing Li, Yan Wu, Yawei Wen, Yifan Du, Yihao Zhang, Yining Ye, Yonghui Wu, Yu Liu, Yu Yue, Yufeng Zhou, Yufeng Yuan, Yuhang Xu, Yuhong Yang, Yun Zhang, Yunhao Fang, Yuntao Li, Yurui Ren, Yuwen Xiong, Zehua Hong, Zehua Wang, Zewei Sun, Zeyu Wang, Zhao Cai, Zhaoyue Zha, Zhecheng An, Zhehui Zhao, Zhengzhuo Xu, Zhipeng Chen, Zhiyong Wu, Zhuofan Zheng, Zihao Wang, Zilong Huang, Ziyu Zhu, Zuquan Song
Abstract: We present Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning. Seed1.5-VL is composed with a 532M-parameter vision encoder and a Mixture-of-Experts (MoE) LLM of 20B active parameters. Despite its relatively compact architecture, it delivers strong performance across a wide spectrum of public VLM benchmarks and internal evaluation suites, achieving the state-of-the-art performance on 38 out of 60 public benchmarks. Moreover, in agent-centric tasks such as GUI control and gameplay, Seed1.5-VL outperforms leading multimodal systems, including OpenAI CUA and Claude 3.7. Beyond visual and video understanding, it also demonstrates strong reasoning abilities, making it particularly effective for multimodal reasoning challenges such as visual puzzles. We believe these capabilities will empower broader applications across diverse tasks. In this report, we mainly provide a comprehensive review of our experiences in building Seed1.5-VL across model design, data construction, and training at various stages, hoping that this report can inspire further research. Seed1.5-VL is now accessible at https://www.volcengine.com/ (Volcano Engine Model ID: doubao-1-5-thinking-vision-pro-250428)
Authors: Zihang Liu, Zhenyu Zhang, Hao Tang
Abstract: Diffusion-based image super-resolution (SR) methods have demonstrated remarkable performance. Recent advancements have introduced deterministic sampling processes that reduce inference from 15 iterative steps to a single step, thereby significantly improving the inference speed of existing diffusion models. However, their efficiency remains limited when handling complex semantic regions due to the single-step inference. To address this limitation, we propose SAMSR, a semantic-guided diffusion framework that incorporates semantic segmentation masks into the sampling process. Specifically, we introduce the SAM-Noise Module, which refines Gaussian noise using segmentation masks to preserve spatial and semantic features. Furthermore, we develop a pixel-wise sampling strategy that dynamically adjusts the residual transfer rate and noise strength based on pixel-level semantic weights, prioritizing semantically rich regions during the diffusion process. To enhance model training, we also propose a semantic consistency loss, which aligns pixel-wise semantic weights between predictions and ground truth. Extensive experiments on both real-world and synthetic datasets demonstrate that SAMSR significantly improves perceptual quality and detail recovery, particularly in semantically complex images. Our code is released at https://github.com/Liu-Zihang/SAMSR.
Authors: Payal Varshney, Adriano Lucieri, Christoph Balada, Andreas Dengel, Sheraz Ahmed
Abstract: Concept-based explanations have emerged as an effective approach within Explainable Artificial Intelligence, enabling interpretable insights by aligning model decisions with human-understandable concepts. However, existing methods rely on computationally intensive procedures and struggle to efficiently capture complex, semantic concepts. Recently, the Concept Discovery through Latent Diffusion-based Counterfactual Trajectories (CDCT) framework, introduced by Varshney et al. (2025), attempts to identify concepts via dimension-wise traversal of the latent space of a Variational Autoencoder trained on counterfactual trajectories. Extending the CDCT framework, this work introduces Concept Directions via Latent Clustering (CDLC), which extracts global, class-specific concept directions by clustering latent difference vectors derived from factual and diffusion-generated counterfactual image pairs. CDLC substantially reduces computational complexity by eliminating the exhaustive latent dimension traversal required in CDCT and enables the extraction of multidimensional semantic concepts encoded across the latent dimensions. This approach is validated on a real-world skin lesion dataset, demonstrating that the extracted concept directions align with clinically recognized dermoscopic features and, in some cases, reveal dataset-specific biases or unknown biomarkers. These results highlight that CDLC is interpretable, scalable, and applicable across high-stakes domains and diverse data modalities.
Authors: Arianna Stropeni, Francesco Borsatti, Manuel Barusco, Davide Dalle Pezze, Marco Fabris, Gian Antonio Susto
Abstract: Visual Anomaly Detection (VAD) is a key task in industrial settings, where minimizing waste and operational costs is essential. Deploying deep learning models within Internet of Things (IoT) environments introduces specific challenges due to the limited computational power and bandwidth of edge devices. This study investigates how to perform VAD effectively under such constraints by leveraging compact and efficient processing strategies. We evaluate several data compression techniques, examining the trade-off between system latency and detection accuracy. Experiments on the MVTec AD benchmark demonstrate that significant compression can be achieved with minimal loss in anomaly detection performance compared to uncompressed data.
Authors: Jun Li, Hongzhang Zhu, Tao Chen, Xiaohua Qian
Abstract: Recently, numerous pancreas segmentation methods have achieved promising performance on local single-source datasets. However, these methods don't adequately account for generalizability issues, and hence typically show limited performance and low stability on test data from other sources. Considering the limited availability of distinct data sources, we seek to improve the generalization performance of a pancreas segmentation model trained with a single-source dataset, i.e., the single source generalization task. In particular, we propose a dual self-supervised learning model that incorporates both global and local anatomical contexts. Our model aims to fully exploit the anatomical features of the intra-pancreatic and extra-pancreatic regions, and hence enhance the characterization of the high-uncertainty regions for more robust generalization. Specifically, we first construct a global-feature contrastive self-supervised learning module that is guided by the pancreatic spatial structure. This module obtains complete and consistent pancreatic features through promoting intra-class cohesion, and also extracts more discriminative features for differentiating between pancreatic and non-pancreatic tissues through maximizing inter-class separation. It mitigates the influence of surrounding tissue on the segmentation outcomes in high-uncertainty regions. Subsequently, a local-image restoration self-supervised learning module is introduced to further enhance the characterization of the high uncertainty regions. In this module, informative anatomical contexts are actually learned to recover randomly corrupted appearance patterns in those regions.
Authors: Zexian Yang, Dian Li, Dayan Wu, Gang Liu, Weiping Wang
Abstract: Despite significant advancements in multimodal reasoning tasks, existing Large Vision-Language Models (LVLMs) are prone to producing visually ungrounded responses when interpreting associated images. In contrast, when humans embark on learning new knowledge, they often rely on a set of fundamental pre-study principles: reviewing outlines to grasp core concepts, summarizing key points to guide their focus and enhance understanding. However, such preparatory actions are notably absent in the current instruction tuning processes. This paper presents Re-Critic, an easily scalable rationale-augmented framework designed to incorporate fundamental rules and chain-of-thought (CoT) as a bridge to enhance reasoning abilities. Specifically, Re-Critic develops a visual rationale synthesizer that scalably augments raw instructions with rationale explanation. To probe more contextually grounded responses, Re-Critic employs an in-context self-critic mechanism to select response pairs for preference tuning. Experiments demonstrate that models fine-tuned with our rationale-augmented dataset yield gains that extend beyond hallucination-specific tasks to broader multimodal reasoning tasks.
Authors: Xufei Wang, Gengxuan Tian, Junqiao Zhao, Siyue Tao, Qiwen Gu, Qiankun Yu, Tiantian Feng
Abstract: Place recognition plays a significant role in SLAM, robot navigation, and autonomous driving applications. Benefiting from deep learning, the performance of LiDAR place recognition (LPR) has been greatly improved. However, many existing learning-based LPR methods suffer from catastrophic forgetting, which severely harms the performance of LPR on previously trained places after training on a new environment. In this paper, we introduce a continual learning framework for LPR via Knowledge Distillation and Fusion (KDF) to alleviate forgetting. Inspired by the ranking process of place recognition retrieval, we present a ranking-aware knowledge distillation loss that encourages the network to preserve the high-level place recognition knowledge. We also introduce a knowledge fusion module to integrate the knowledge of old and new models for LiDAR place recognition. Our extensive experiments demonstrate that KDF can be applied to different networks to overcome catastrophic forgetting, surpassing the state-of-the-art methods in terms of mean Recall@1 and forgetting score.
Authors: Yan Xie, Zequn Zeng, Hao Zhang, Yucheng Ding, Yi Wang, Zhengjue Wang, Bo Chen, Hongwei Liu
Abstract: Concept Bottleneck Models (CBMs) try to make the decision-making process transparent by exploring an intermediate concept space between the input image and the output prediction. Existing CBMs just learn coarse-grained relations between the whole image and the concepts, less considering local image information, leading to two main drawbacks: i) they often produce spurious visual-concept relations, hence decreasing model reliability; and ii) though CBMs could explain the importance of every concept to the final prediction, it is still challenging to tell which visual region produces the prediction. To solve these problems, this paper proposes a Disentangled Optimal Transport CBM (DOT-CBM) framework to explore fine-grained visual-concept relations between local image patches and concepts. Specifically, we model the concept prediction process as a transportation problem between the patches and concepts, thereby achieving explicit fine-grained feature alignment. We also incorporate orthogonal projection losses within the modality to enhance local feature disentanglement. To further address the shortcut issues caused by statistical biases in the data, we utilize the visual saliency map and concept label statistics as transportation priors. Thus, DOT-CBM can visualize inversion heatmaps, provide more reliable concept predictions, and produce more accurate class predictions. Comprehensive experiments demonstrate that our proposed DOT-CBM achieves SOTA performance on several tasks, including image classification, local part detection and out-of-distribution generalization.
Authors: Hongda Qin, Xiao Lu, Zhiyong Wei, Yihong Cao, Kailun Yang, Ningjiang Chen
Abstract: Generalizing an object detector trained on a single domain to multiple unseen domains is a challenging task. Existing methods typically introduce image or feature augmentation to diversify the source domain to raise the robustness of the detector. Vision-Language Model (VLM)-based augmentation techniques have been proven to be effective, but they require that the detector's backbone has the same structure as the image encoder of VLM, limiting the detector framework selection. To address this problem, we propose Language-Driven Dual Style Mixing (LDDS) for single-domain generalization, which diversifies the source domain by fully utilizing the semantic information of the VLM. Specifically, we first construct prompts to transfer style semantics embedded in the VLM to an image translation network. This facilitates the generation of style diversified images with explicit semantic information. Then, we propose image-level style mixing between the diversified images and source domain images. This effectively mines the semantic information for image augmentation without relying on specific augmentation selections. Finally, we propose feature-level style mixing in a double-pipeline manner, allowing feature augmentation to be model-agnostic and can work seamlessly with the mainstream detector frameworks, including the one-stage, two-stage, and transformer-based detectors. Extensive experiments demonstrate the effectiveness of our approach across various benchmark datasets, including real to cartoon and normal to adverse weather tasks. The source code and pre-trained models will be publicly available at https://github.com/qinhongda8/LDDS.
Authors: Philippe Colantoni, Rafique Ahmed, Prashant Ghimire, Damien Muselet, Alain Tr\'emeau
Abstract: The accuracy and efficiency of human body pose estimation depend on the quality of the data to be processed and of the particularities of these data. To demonstrate how dance videos can challenge pose estimation techniques, we proposed a new 3D human body pose estimation pipeline which combined up-to-date techniques and methods that had not been yet used in dance analysis. Second, we performed tests and extensive experimentations from dance video archives, and used visual analytic tools to evaluate the impact of several data parameters on human body pose. Our results are publicly available for research at https://www.couleur.org/articles/arXiv-1-2025/
Authors: Wenqiang Wang, Yangshijie Zhang
Abstract: Large vision language models (LVLMs) achieve remarkable performance through Vision In-context Learning (VICL), a process that depends significantly on demonstrations retrieved from an extensive collection of annotated examples (retrieval database). Existing studies often assume that the retrieval database contains annotated examples for all labels. However, in real-world scenarios, delays in database updates or incomplete data annotation may result in the retrieval database containing labeled samples for only a subset of classes. We refer to this phenomenon as an \textbf{incomplete retrieval database} and define the in-context learning under this condition as \textbf{Incomplete In-context Learning (IICL)}. To address this challenge, we propose \textbf{Iterative Judgments and Integrated Prediction (IJIP)}, a two-stage framework designed to mitigate the limitations of IICL. The Iterative Judgments Stage reformulates an \(\boldsymbol{m}\)-class classification problem into a series of \(\boldsymbol{m}\) binary classification tasks, effectively converting the IICL setting into a standard VICL scenario. The Integrated Prediction Stage further refines the classification process by leveraging both the input image and the predictions from the Iterative Judgments Stage to enhance overall classification accuracy. IJIP demonstrates considerable performance across two LVLMs and two datasets under three distinct conditions of label incompleteness, achieving the highest accuracy of 93.9\%. Notably, even in scenarios where labels are fully available, IJIP still achieves the best performance of all six baselines. Furthermore, IJIP can be directly applied to \textbf{Prompt Learning} and is adaptable to the \textbf{text domain}.
Authors: Mohamed Nagy, Naoufel Werghi, Bilal Hassan, Jorge Dias, Majid Khonji
Abstract: This work addresses the critical lack of precision in state estimation in the Kalman filter for 3D multi-object tracking (MOT) and the ongoing challenge of selecting the appropriate motion model. Existing literature commonly relies on constant motion models for estimating the states of objects, neglecting the complex motion dynamics unique to each object. Consequently, trajectory division and imprecise object localization arise, especially under occlusion conditions. The core of these challenges lies in the limitations of the current Kalman filter formulation, which fails to account for the variability of motion dynamics as objects navigate their environments. This work introduces a novel formulation of the Kalman filter that incorporates motion dynamics, allowing the motion model to adaptively adjust according to changes in the object's movement. The proposed Kalman filter substantially improves state estimation, localization, and trajectory prediction compared to the traditional Kalman filter. This is reflected in tracking performance that surpasses recent benchmarks on the KITTI and Waymo Open Datasets, with margins of 0.56\% and 0.81\% in higher order tracking accuracy (HOTA) and multi-object tracking accuracy (MOTA), respectively. Furthermore, the proposed Kalman filter consistently outperforms the baseline across various detectors. Additionally, it shows an enhanced capability in managing long occlusions compared to the baseline Kalman filter, achieving margins of 1.22\% in higher order tracking accuracy (HOTA) and 1.55\% in multi-object tracking accuracy (MOTA) on the KITTI dataset. The formulation's efficiency is evident, with an additional processing time of only approximately 0.078 ms per frame, ensuring its applicability in real-time applications.
Authors: Christoph Huber, Ludwig Schleeh, Dino Knoll, Michael Guthe
Abstract: Visual quality inspection in automotive production is essential for ensuring the safety and reliability of vehicles. Computer vision (CV) has become a popular solution for these inspections due to its cost-effectiveness and reliability. However, CV models require large, annotated datasets, which are costly and time-consuming to collect. To reduce the need for extensive training data, we propose a novel image classification pipeline that combines similarity search using a vision-based foundation model with synthetic data. Our approach leverages a DINOv2 model to transform input images into feature vectors, which are then compared to pre-classified reference images using cosine distance measurements. By utilizing synthetic data instead of real images as references, our pipeline achieves high classification accuracy without relying on real data. We evaluate this approach in eight real-world inspection scenarios and demonstrate that it meets the high performance requirements of production environments.
Authors: Xiaokun Wang, Chris, Jiangbo Pei, Wei Shen, Yi Peng, Yunzhuo Hao, Weijie Qiu, Ai Jian, Tianyidan Xie, Xuchen Song, Yang Liu, Yahui Zhou
Abstract: We propose Skywork-VL Reward, a multimodal reward model that provides reward signals for both multimodal understanding and reasoning tasks. Our technical approach comprises two key components: First, we construct a large-scale multimodal preference dataset that covers a wide range of tasks and scenarios, with responses collected from both standard vision-language models (VLMs) and advanced VLM reasoners. Second, we design a reward model architecture based on Qwen2.5-VL-7B-Instruct, integrating a reward head and applying multi-stage fine-tuning using pairwise ranking loss on pairwise preference data. Experimental evaluations show that Skywork-VL Reward achieves state-of-the-art results on multimodal VL-RewardBench and exhibits competitive performance on the text-only RewardBench benchmark. Furthermore, preference data constructed based on our Skywork-VL Reward proves highly effective for training Mixed Preference Optimization (MPO), leading to significant improvements in multimodal reasoning capabilities. Our results underscore Skywork-VL Reward as a significant advancement toward general-purpose, reliable reward models for multimodal alignment. Our model has been publicly released to promote transparency and reproducibility.
Authors: Sofia Casarin, Sergio Escalera, Oswald Lanz
Abstract: Training-free Neural Architecture Search (NAS) efficiently identifies high-performing neural networks using zero-cost (ZC) proxies. Unlike multi-shot and one-shot NAS approaches, ZC-NAS is both (i) time-efficient, eliminating the need for model training, and (ii) interpretable, with proxy designs often theoretically grounded. Despite rapid developments in the field, current SOTA ZC proxies are typically constrained to well-established convolutional search spaces. With the rise of Large Language Models shaping the future of deep learning, this work extends ZC proxy applicability to Vision Transformers (ViTs). We present a new benchmark using the Autoformer search space evaluated on 6 distinct tasks and propose Layer-Sample Wise Activation with Gradients information (L-SWAG), a novel, generalizable metric that characterizes both convolutional and transformer architectures across 14 tasks. Additionally, previous works highlighted how different proxies contain complementary information, motivating the need for a ML model to identify useful combinations. To further enhance ZC-NAS, we therefore introduce LIBRA-NAS (Low Information gain and Bias Re-Alignment), a method that strategically combines proxies to best represent a specific benchmark. Integrated into the NAS search, LIBRA-NAS outperforms evolution and gradient-based NAS techniques by identifying an architecture with a 17.0% test error on ImageNet1k in just 0.1 GPU days.
Authors: Katsuki Shimbo, Hiromu Taketsugu, Norimichi Ukita
Abstract: In 3D Human Motion Prediction (HMP), conventional methods train HMP models with expensive motion capture data. However, the data collection cost of such motion capture data limits the data diversity, which leads to poor generalizability to unseen motions or subjects. To address this issue, this paper proposes to enhance HMP with additional learning using estimated poses from easily available videos. The 2D poses estimated from the monocular videos are carefully transformed into motion capture-style 3D motions through our pipeline. By additional learning with the obtained motions, the HMP model is adapted to the test domain. The experimental results demonstrate the quantitative and qualitative impact of our method.
Authors: Sander De Coninck, Emilio Gamba, Bart Van Doninck, Abdellatif Bey-Temsamani, Sam Leroux, Pieter Simoens
Abstract: Musculoskeletal disorders (MSDs) are a leading cause of injury and productivity loss in the manufacturing industry, incurring substantial economic costs. Ergonomic assessments can mitigate these risks by identifying workplace adjustments that improve posture and reduce strain. Camera-based systems offer a non-intrusive, cost-effective method for continuous ergonomic tracking, but they also raise significant privacy concerns. To address this, we propose a privacy-aware ergonomic assessment framework utilizing machine learning techniques. Our approach employs adversarial training to develop a lightweight neural network that obfuscates video data, preserving only the essential information needed for human pose estimation. This obfuscation ensures compatibility with standard pose estimation algorithms, maintaining high accuracy while protecting privacy. The obfuscated video data is transmitted to a central server, where state-of-the-art keypoint detection algorithms extract body landmarks. Using multi-view integration, 3D keypoints are reconstructed and evaluated with the Rapid Entire Body Assessment (REBA) method. Our system provides a secure, effective solution for ergonomic monitoring in industrial environments, addressing both privacy and workplace safety concerns.
Authors: Gang He, Siqi Wang, Kepeng Xu, Lin Zhang
Abstract: High-Dynamic-Range Wide-Color-Gamut (HDR-WCG) technology is becoming increasingly prevalent, intensifying the demand for converting Standard Dynamic Range (SDR) content to HDR. Existing methods primarily rely on fixed tone mapping operators, which are inadequate for handling SDR inputs with diverse styles commonly found in real-world scenarios. To address this challenge, we propose a generalized SDR-to-HDR method that handles diverse styles in real-world SDR content, termed Realistic Style Disentangled Representation Learning (RealRep). By disentangling luminance and chrominance, we analyze the intrinsic differences between contents with varying styles and propose a disentangled multi-view style representation learning method. This approach captures the guidance prior of true luminance and chrominance distributions across different styles, even when the SDR style distributions exhibit significant variations, thereby establishing a robust embedding space for inverse tone mapping. Motivated by the difficulty of directly utilizing degradation representation priors, we further introduce the Degradation-Domain Aware Controlled Mapping Network (DDACMNet), a two-stage framework that performs adaptive hierarchical mapping guided by a control-aware normalization mechanism. DDACMNet dynamically modulates the mapping process via degradation-conditioned hierarchical features, enabling robust adaptation across diverse degradation domains. Extensive experiments show that RealRep consistently outperforms state-of-the-art methods with superior generalization and perceptually faithful HDR color gamut reconstruction.
Authors: Matthew Marchellus, Nadhira Noor, In Kyu Park
Abstract: Fast 3D clothed human reconstruction from monocular video remains a significant challenge in computer vision, particularly in balancing computational efficiency with reconstruction quality. Current approaches are either focused on static image reconstruction but too computationally intensive, or achieve high quality through per-video optimization that requires minutes to hours of processing, making them unsuitable for real-time applications. To this end, we present TemPoFast3D, a novel method that leverages temporal coherency of human appearance to reduce redundant computation while maintaining reconstruction quality. Our approach is a "plug-and play" solution that uniquely transforms pixel-aligned reconstruction networks to handle continuous video streams by maintaining and refining a canonical appearance representation through efficient coordinate mapping. Extensive experiments demonstrate that TemPoFast3D matches or exceeds state-of-the-art methods across standard metrics while providing high-quality textured reconstruction across diverse pose and appearance, with a maximum speed of 12 FPS.
Authors: Zhixuan Zhang, Xiaopeng Li, Qi Liu
Abstract: Background subtraction (BGS) is utilized to detect moving objects in a video and is commonly employed at the onset of object tracking and human recognition processes. Nevertheless, existing BGS techniques utilizing deep learning still encounter challenges with various background noises in videos, including variations in lighting, shifts in camera angles, and disturbances like air turbulence or swaying trees. To address this problem, we design a spiking autoencoder network, termed SAEN-BGS, based on noise resilience and time-sequence sensitivity of spiking neural networks (SNNs) to enhance the separation of foreground and background. To eliminate unnecessary background noise and preserve the important foreground elements, we begin by creating the continuous spiking conv-and-dconv block, which serves as the fundamental building block for the decoder in SAEN-BGS. Moreover, in striving for enhanced energy efficiency, we introduce a novel self-distillation spiking supervised learning method grounded in ANN-to-SNN frameworks, resulting in decreased power consumption. In extensive experiments conducted on CDnet-2014 and DAVIS-2016 datasets, our approach demonstrates superior segmentation performance relative to other baseline methods, even when challenged by complex scenarios with dynamic backgrounds.
Authors: Yuan Zhang, Jiacheng Jiang, Guoqing Ma, Zhiying Lu, Haoyang Huang, Jianlong Yuan, Nan Duan
Abstract: In this work, we present GPDiT, a Generative Pre-trained Autoregressive Diffusion Transformer that unifies the strengths of diffusion and autoregressive modeling for long-range video synthesis, within a continuous latent space. Instead of predicting discrete tokens, GPDiT autoregressively predicts future latent frames using a diffusion loss, enabling natural modeling of motion dynamics and semantic consistency across frames. This continuous autoregressive framework not only enhances generation quality but also endows the model with representation capabilities. Additionally, we introduce a lightweight causal attention variant and a parameter-free rotation-based time-conditioning mechanism, improving both the training and inference efficiency. Extensive experiments demonstrate that GPDiT achieves strong performance in video generation quality, video representation ability, and few-shot learning tasks, highlighting its potential as an effective framework for video modeling in continuous space.
Authors: Jiewen Yang, Taoran Huang, Shangwei Ding, Xiaowei Xu, Qinhua Zhao, Yong Jiang, Jiarong Guo, Bin Pu, Jiexuan Zheng, Caojin Zhang, Hongwen Fei, Xiaomeng Li
Abstract: Echocardiographers can detect pulmonary hypertension using Doppler echocardiography; however, accurately assessing its progression often proves challenging. Right heart catheterization (RHC), the gold standard for precise evaluation, is invasive and unsuitable for routine use, limiting its practicality for timely diagnosis and monitoring of pulmonary hypertension progression. Here, we propose MePH, a multi-view, multi-modal vision-language model to accurately assess pulmonary hypertension progression using non-invasive echocardiography. We constructed a large dataset comprising paired standardized echocardiogram videos, spectral images and RHC data, covering 1,237 patient cases from 12 medical centers. For the first time, MePH precisely models the correlation between non-invasive multi-view, multi-modal echocardiography and the pressure and resistance obtained via RHC. We show that MePH significantly outperforms echocardiographers' assessments using echocardiography, reducing the mean absolute error in estimating mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR) by 49.73% and 43.81%, respectively. In eight independent external hospitals, MePH achieved a mean absolute error of 3.147 for PVR assessment. Furthermore, MePH achieved an area under the curve of 0.921, surpassing echocardiographers (area under the curve of 0.842) in accurately predicting the severity of pulmonary hypertension, whether mild or severe. A prospective study demonstrated that MePH can predict treatment efficacy for patients. Our work provides pulmonary hypertension patients with a non-invasive and timely method for monitoring disease progression, improving the accuracy and efficiency of pulmonary hypertension management while enabling earlier interventions and more personalized treatment decisions.
Authors: Lintao Xiang, Hongpei Zheng, Bailin Deng, Hujun Yin
Abstract: Neural implicit surface reconstruction using volume rendering techniques has recently achieved significant advancements in creating high-fidelity surfaces from multiple 2D images. However, current methods primarily target scenes with consistent illumination and struggle to accurately reconstruct 3D geometry in uncontrolled environments with transient occlusions or varying appearances. While some neural radiance field (NeRF)-based variants can better manage photometric variations and transient objects in complex scenes, they are designed for novel view synthesis rather than precise surface reconstruction due to limited surface constraints. To overcome this limitation, we introduce a novel approach that applies multiple geometric constraints to the implicit surface optimization process, enabling more accurate reconstructions from unconstrained image collections. First, we utilize sparse 3D points from structure-from-motion (SfM) to refine the signed distance function estimation for the reconstructed surface, with a displacement compensation to accommodate noise in the sparse points. Additionally, we employ robust normal priors derived from a normal predictor, enhanced by edge prior filtering and multi-view consistency constraints, to improve alignment with the actual surface geometry. Extensive testing on the Heritage-Recon benchmark and other datasets has shown that the proposed method can accurately reconstruct surfaces from in-the-wild images, yielding geometries with superior accuracy and granularity compared to existing techniques. Our approach enables high-quality 3D reconstruction of various landmarks, making it applicable to diverse scenarios such as digital preservation of cultural heritage sites.
Authors: Yuqi Cheng, Yunkang Cao, Dongfang Wang, Weiming Shen, Wenlong Li
Abstract: Point cloud anomaly detection is essential for various industrial applications. The huge computation and storage costs caused by the increasing product classes limit the application of single-class unsupervised methods, necessitating the development of multi-class unsupervised methods. However, the feature similarity between normal and anomalous points from different class data leads to the feature confusion problem, which greatly hinders the performance of multi-class methods. Therefore, we introduce a multi-class point cloud anomaly detection method, named GLFM, leveraging global-local feature matching to progressively separate data that are prone to confusion across multiple classes. Specifically, GLFM is structured into three stages: Stage-I proposes an anomaly synthesis pipeline that stretches point clouds to create abundant anomaly data that are utilized to adapt the point cloud feature extractor for better feature representation. Stage-II establishes the global and local memory banks according to the global and local feature distributions of all the training data, weakening the impact of feature confusion on the establishment of the memory bank. Stage-III implements anomaly detection of test data leveraging its feature distance from global and local memory banks. Extensive experiments on the MVTec 3D-AD, Real3D-AD and actual industry parts dataset showcase our proposed GLFM's superior point cloud anomaly detection performance. The code is available at https://github.com/hustCYQ/GLFM-Multi-class-3DAD.
Authors: David V\'azquez-Pad\'in, Fernando P\'erez-Gonz\'alez, Pablo P\'erez-Migu\'elez
Abstract: iPhone portrait-mode images contain a distinctive pattern in out-of-focus regions simulating the bokeh effect, which we term Apple's Synthetic Defocus Noise Pattern (SDNP). If overlooked, this pattern can interfere with blind forensic analyses, especially PRNU-based camera source verification, as noted in earlier works. Since Apple's SDNP remains underexplored, we provide a detailed characterization, proposing a method for its precise estimation, modeling its dependence on scene brightness, ISO settings, and other factors. Leveraging this characterization, we explore forensic applications of the SDNP, including traceability of portrait-mode images across iPhone models and iOS versions in open-set scenarios, assessing its robustness under post-processing. Furthermore, we show that masking SDNP-affected regions in PRNU-based camera source verification significantly reduces false positives, overcoming a critical limitation in camera attribution, and improving state-of-the-art techniques.
Authors: Baoping Cheng, Yukun Zhang, Liming Wang, Xiaoyan Xie, Tao Fu, Dongkun Wang, Xiaoming Tao
Abstract: With the continuous increase in the number and resolution of video surveillance cameras, the burden of transmitting and storing surveillance video is growing. Traditional communication methods based on Shannon's theory are facing optimization bottlenecks. Semantic communication, as an emerging communication method, is expected to break through this bottleneck and reduce the storage and transmission consumption of video. Existing semantic decoding methods often require many samples to train the neural network for each scene, which is time-consuming and labor-intensive. In this study, a semantic encoding and decoding method for surveillance video is proposed. First, the sketch was extracted as semantic information, and a sketch compression method was proposed to reduce the bit rate of semantic information. Then, an image translation network was proposed to translate the sketch into a video frame with a reference frame. Finally, a few-shot sketch decoding network was proposed to reconstruct video from sketch. Experimental results showed that the proposed method achieved significantly better video reconstruction performance than baseline methods. The sketch compression method could effectively reduce the storage and transmission consumption of semantic information with little compromise on video quality. The proposed method provides a novel semantic encoding and decoding method that only needs a few training samples for each surveillance scene, thus improving the practicality of the semantic communication system.
Authors: Chunpeng Li, Ya-tang Li
Abstract: Understanding the computations of convolutional neural networks requires effective visualization of their kernels. While maximal activation methods have proven successful in highlighting the preferred features of 2D convolutional kernels, directly applying these techniques to 3D convolutions often leads to uninterpretable results due to the higher dimensionality and complexity of 3D features. To address this challenge, we propose a novel visualization approach for 3D convolutional kernels that disentangles their texture and motion preferences. Our method begins with a data-driven decomposition of the optimal input that maximally activates a given kernel. We then introduce a two-stage optimization strategy to extract distinct texture and motion components from this input. Applying our approach to visualize kernels at various depths of several pre-trained models, we find that the resulting visualizations--particularly those capturing motion--clearly reveal the preferred dynamic patterns encoded by 3D kernels. These results demonstrate the effectiveness of our method in providing interpretable insights into 3D convolutional operations. Code is available at https://github.com/YatangLiLab/3DKernelVisualizer.
Authors: Olaf Wysocki, Benedikt Schwab, Manoj Kumar Biswanath, Michael Greza, Qilin Zhang, Jingwei Zhu, Thomas Froech, Medhini Heeramaglore, Ihab Hijazi, Khaoula Kanna, Mathias Pechinger, Zhaiyu Chen, Yao Sun, Alejandro Rueda Segura, Ziyang Xu, Omar AbdelGafar, Mansour Mehranfar, Chandan Yeshwanth, Yueh-Cheng Liu, Hadi Yazdi, Jiapan Wang, Stefan Auer, Katharina Anders, Klaus Bogenberger, Andre Borrmann, Angela Dai, Ludwig Hoegner, Christoph Holst, Thomas H. Kolbe, Ferdinand Ludwig, Matthias Nie{\ss}ner, Frank Petzold, Xiao Xiang Zhu, Boris Jutzi
Abstract: Urban Digital Twins (UDTs) have become essential for managing cities and integrating complex, heterogeneous data from diverse sources. Creating UDTs involves challenges at multiple process stages, including acquiring accurate 3D source data, reconstructing high-fidelity 3D models, maintaining models' updates, and ensuring seamless interoperability to downstream tasks. Current datasets are usually limited to one part of the processing chain, hampering comprehensive UDTs validation. To address these challenges, we introduce the first comprehensive multimodal Urban Digital Twin benchmark dataset: TUM2TWIN. This dataset includes georeferenced, semantically aligned 3D models and networks along with various terrestrial, mobile, aerial, and satellite observations boasting 32 data subsets over roughly 100,000 $m^2$ and currently 767 GB of data. By ensuring georeferenced indoor-outdoor acquisition, high accuracy, and multimodal data integration, the benchmark supports robust analysis of sensors and the development of advanced reconstruction methods. Additionally, we explore downstream tasks demonstrating the potential of TUM2TWIN, including novel view synthesis of NeRF and Gaussian Splatting, solar potential analysis, point cloud semantic segmentation, and LoD3 building reconstruction. We are convinced this contribution lays a foundation for overcoming current limitations in UDT creation, fostering new research directions and practical solutions for smarter, data-driven urban environments. The project is available under: https://tum2t.win
URLs: https://tum2t.win
Authors: Mingqian Ji, Jian Yang, Shanshan Zhang
Abstract: State-of-the-art LiDAR-camera 3D object detectors usually focus on feature fusion. However, they neglect the factor of depth while designing the fusion strategy. In this work, we are the first to observe that different modalities play different roles as depth varies via statistical analysis and visualization. Based on this finding, we propose a Depth-Aware Hybrid Feature Fusion (DepthFusion) strategy that guides the weights of point cloud and RGB image modalities by introducing depth encoding at both global and local levels. Specifically, the Depth-GFusion module adaptively adjusts the weights of image Bird's-Eye-View (BEV) features in multi-modal global features via depth encoding. Furthermore, to compensate for the information lost when transferring raw features to the BEV space, we propose a Depth-LFusion module, which adaptively adjusts the weights of original voxel features and multi-view image features in multi-modal local features via depth encoding. Extensive experiments on the nuScenes and KITTI datasets demonstrate that our DepthFusion method surpasses previous state-of-the-art methods. Moreover, our DepthFusion is more robust to various kinds of corruptions, outperforming previous methods on the nuScenes-C dataset.
Authors: Zeynep Galymzhankyzy, Eric Martinson
Abstract: Efficient crop-weed segmentation is critical for site-specific weed control in precision agriculture. Conventional CNN-based methods struggle to generalize and rely on RGB imagery, limiting performance under complex field conditions. To address these challenges, we propose a lightweight transformer-CNN hybrid. It processes RGB, Near-Infrared (NIR), and Red-Edge (RE) bands using specialized encoders and dynamic modality integration. Evaluated on the WeedsGalore dataset, the model achieves a segmentation accuracy (mean IoU) of 78.88%, outperforming RGB-only models by 15.8 percentage points. With only 8.7 million parameters, the model offers high accuracy, computational efficiency, and potential for real-time deployment on Unmanned Aerial Vehicles (UAVs) and edge devices, advancing precision weed management.
Authors: Erik Landolsi, Fredrik Kahl
Abstract: There is an increasing interest in using image-generating diffusion models for deep data augmentation and image morphing. In this context, it is useful to interpolate between latents produced by inverting a set of input images, in order to generate new images representing some mixture of the inputs. We observe that such interpolation can easily lead to degenerate results when the number of inputs is large. We analyze the cause of this effect theoretically and experimentally, and suggest a suitable remedy. The suggested approach is a relatively simple normalization scheme that is easy to use whenever interpolation between latents is needed. We measure image quality using FID and CLIP embedding distance and show experimentally that baseline interpolation methods lead to a drop in quality metrics long before the degeneration issue is clearly visible. In contrast, our method significantly reduces the degeneration effect and leads to improved quality metrics also in non-degenerate situations.
Authors: Mohamed Ali Souibgui, Changkyu Choi, Andrey Barsky, Kangsoo Jung, Ernest Valveny, Dimosthenis Karatzas
Abstract: We propose DocVXQA, a novel framework for visually self-explainable document question answering. The framework is designed not only to produce accurate answers to questions but also to learn visual heatmaps that highlight contextually critical regions, thereby offering interpretable justifications for the model's decisions. To integrate explanations into the learning process, we quantitatively formulate explainability principles as explicit learning objectives. Unlike conventional methods that emphasize only the regions pertinent to the answer, our framework delivers explanations that are \textit{contextually sufficient} while remaining \textit{representation-efficient}. This fosters user trust while achieving a balance between predictive performance and interpretability in DocVQA applications. Extensive experiments, including human evaluation, provide strong evidence supporting the effectiveness of our method. The code is available at https://github.com/dali92002/DocVXQA.
Authors: Bahram Mohammadi, Ehsan Abbasnejad, Yuankai Qi, Qi Wu, Anton Van Den Hengel, Javen Qinfeng Shi
Abstract: The remote embodied referring expression (REVERIE) task requires an agent to navigate through complex indoor environments and localize a remote object specified by high-level instructions, such as "bring me a spoon", without pre-exploration. Hence, an efficient navigation plan is essential for the final success. This paper proposes a novel parameter-efficient action planner using large language models (PEAP-LLM) to generate a single-step instruction at each location. The proposed model consists of two modules, LLM goal planner (LGP) and LoRA action planner (LAP). Initially, LGP extracts the goal-oriented plan from REVERIE instructions, including the target object and room. Then, LAP generates a single-step instruction with the goal-oriented plan, high-level instruction, and current visual observation as input. PEAP-LLM enables the embodied agent to interact with LAP as the path planner on the fly. A simple direct application of LLMs hardly achieves good performance. Also, existing hard-prompt-based methods are error-prone in complicated scenarios and need human intervention. To address these issues and prevent the LLM from generating hallucinations and biased information, we propose a novel two-stage method for fine-tuning the LLM, consisting of supervised fine-tuning (STF) and direct preference optimization (DPO). SFT improves the quality of generated instructions, while DPO utilizes environmental feedback. Experimental results show the superiority of our proposed model on REVERIE compared to the previous state-of-the-art.
Authors: Mauricio Orbes-Arteaga, Oeslle Lucena, Sabastien Ourselin, M. Jorge Cardoso
Abstract: Interactive medical segmentation reduces annotation effort by refining predictions through user feedback. Vision Transformer (ViT)-based models, such as the Segment Anything Model (SAM), achieve state-of-the-art performance using user clicks and prior masks as prompts. However, existing methods treat interactions as independent events, leading to redundant corrections and limited refinement gains. We address this by introducing MAIS, a Memory-Attention mechanism for Interactive Segmentation that stores past user inputs and segmentation states, enabling temporal context integration. Our approach enhances ViT-based segmentation across diverse imaging modalities, achieving more efficient and accurate refinements.
Authors: Raul Ismayilov, Dzemila Sero, Luuk Spreeuwers
Abstract: Synthetic face datasets are increasingly used to overcome the limitations of real-world biometric data, including privacy concerns, demographic imbalance, and high collection costs. However, many existing methods lack fine-grained control over identity attributes and fail to produce paired, identity-consistent images under structured capture conditions. We introduce FLUXSynID, a framework for generating high-resolution synthetic face datasets with user-defined identity attribute distributions and paired document-style and trusted live capture images. The dataset generated using the FLUXSynID framework shows improved alignment with real-world identity distributions and greater inter-set diversity compared to prior work. The FLUXSynID framework for generating custom datasets, along with a dataset of 14,889 synthetic identities, is publicly released to support biometric research, including face recognition and morphing attack detection.
Authors: Ahmad Fall, Federica Granese, Alex Lence, Dominique Fourer, Blaise Hanczar, Joe-Elie Salem, Jean-Daniel Zucker, Edi Prifti
Abstract: Monitoring and analyzing electrocardiogram (ECG) signals, even under varying physiological conditions, including those influenced by physical activity, drugs and stress, is crucial to accurately assess cardiac health. However, current AI-based methods often fail to account for how these factors interact and alter ECG patterns, ultimately limiting their applicability in real-world settings. This study introduces IKrNet, a novel neural network model, which identifies drug-specific patterns in ECGs amidst certain physiological conditions. IKrNet's architecture incorporates spatial and temporal dynamics by using a convolutional backbone with varying receptive field size to capture spatial features. A bi-directional Long Short-Term Memory module is also employed to model temporal dependencies. By treating heart rate variability as a surrogate for physiological fluctuations, we evaluated IKrNet's performance across diverse scenarios, including conditions with physical stress, drug intake alone, and a baseline without drug presence. Our assessment follows a clinical protocol in which 990 healthy volunteers were administered 80mg of Sotalol, a drug which is known to be a precursor to Torsades-de-Pointes, a life-threatening arrhythmia. We show that IKrNet outperforms state-of-the-art models' accuracy and stability in varying physiological conditions, underscoring its clinical viability.
Authors: Bohan Wang, Zhongqi Yue, Fengda Zhang, Shuo Chen, Li'an Bi, Junzhe Zhang, Xue Song, Kennard Yanting Chan, Jiachun Pan, Weijia Wu, Mingze Zhou, Wang Lin, Kaihang Pan, Saining Zhang, Liyu Jia, Wentao Hu, Wei Zhao, Hanwang Zhang
Abstract: We completely discard the conventional spatial prior in image representation and introduce a novel discrete visual tokenizer: Self-consistency Tokenizer (Selftok). At its design core, we compose an autoregressive (AR) prior -- mirroring the causal structure of language -- into visual tokens by using the reverse diffusion process of image generation. The AR property makes Selftok fundamentally distinct from traditional spatial tokens in the following two key ways: - Selftok offers an elegant and minimalist approach to unify diffusion and AR for vision-language models (VLMs): By representing images with Selftok tokens, we can train a VLM using a purely discrete autoregressive architecture -- like that in LLMs -- without requiring additional modules or training objectives. - We theoretically show that the AR prior satisfies the Bellman equation, whereas the spatial prior does not. Therefore, Selftok supports reinforcement learning (RL) for visual generation with effectiveness comparable to that achieved in LLMs. Besides the AR property, Selftok is also a SoTA tokenizer that achieves a favorable trade-off between high-quality reconstruction and compression rate. We use Selftok to build a pure AR VLM for both visual comprehension and generation tasks. Impressively, without using any text-image training pairs, a simple policy gradient RL working in the visual tokens can significantly boost the visual generation benchmark, surpassing all the existing models by a large margin. Therefore, we believe that Selftok effectively addresses the long-standing challenge that visual tokens cannot support effective RL. When combined with the well-established strengths of RL in LLMs, this brings us one step closer to realizing a truly multimodal LLM. Project Page: https://selftok-team.github.io/report/.
Authors: Hao Li, Sicheng Li, Xiang Gao, Abudouaihati Batuer, Lu Yu, Yiyi Liao
Abstract: Immersive video offers a 6-Dof-free viewing experience, potentially playing a key role in future video technology. Recently, 4D Gaussian Splatting has gained attention as an effective approach for immersive video due to its high rendering efficiency and quality, though maintaining quality with manageable storage remains challenging. To address this, we introduce GIFStream, a novel 4D Gaussian representation using a canonical space and a deformation field enhanced with time-dependent feature streams. These feature streams enable complex motion modeling and allow efficient compression by leveraging temporal correspondence and motion-aware pruning. Additionally, we incorporate both temporal and spatial compression networks for end-to-end compression. Experimental results show that GIFStream delivers high-quality immersive video at 30 Mbps, with real-time rendering and fast decoding on an RTX 4090. Project page: https://xdimlab.github.io/GIFStream
Authors: Juan E. Tapia, Fabian Stockhardt, L\'azaro Janier Gonz\'alez-Soler, Christoph Busch
Abstract: The demand for Presentation Attack Detection (PAD) to identify fraudulent ID documents in remote verification systems has significantly risen in recent years. This increase is driven by several factors, including the rise of remote work, online purchasing, migration, and advancements in synthetic images. Additionally, we have noticed a surge in the number of attacks aimed at the enrolment process. Training a PAD to detect fake ID documents is very challenging because of the limited number of ID documents available due to privacy concerns. This work proposes a new passport dataset generated from a hybrid method that combines synthetic data and open-access information using the ICAO requirement to obtain realistic training and testing images.
Authors: Efe Bozkir, Christian Kosel, Tina Seidel, Enkelejda Kasneci
Abstract: Teachers' visual attention and its distribution across the students in classrooms can constitute important implications for student engagement, achievement, and professional teacher training. Despite that, inferring the information about where and which student teachers focus on is not trivial. Mobile eye tracking can provide vital help to solve this issue; however, the use of mobile eye tracking alone requires a significant amount of manual annotations. To address this limitation, we present an automated processing pipeline concept that requires minimal manually annotated data to recognize which student the teachers focus on. To this end, we utilize state-of-the-art face detection models and face recognition feature embeddings to train face recognition models with transfer learning in the classroom context and combine these models with the teachers' gaze from mobile eye trackers. We evaluated our approach with data collected from four different classrooms, and our results show that while it is possible to estimate the visually focused students with reasonable performance in all of our classroom setups, U-shaped and small classrooms led to the best results with accuracies of approximately 0.7 and 0.9, respectively. While we did not evaluate our method for teacher-student interactions and focused on the validity of the technical approach, as our methodology does not require a vast amount of manually annotated data and offers a non-intrusive way of handling teachers' visual attention, it could help improve instructional strategies, enhance classroom management, and provide feedback for professional teacher development.
Authors: Kamil Jeziorek, Tomasz Kryjak
Abstract: Event cameras offer significant advantages over traditional frame-based sensors. These include microsecond temporal resolution, robustness under varying lighting conditions and low power consumption. Nevertheless, the effective processing of their sparse, asynchronous event streams remains challenging. Existing approaches to this problem can be categorised into two distinct groups. The first group involves the direct processing of event data with neural models, such as Spiking Neural Networks or Graph Convolutional Neural Networks. However, this approach is often accompanied by a compromise in terms of qualitative performance. The second group involves the conversion of events into dense representations with handcrafted aggregation functions, which can boost accuracy at the cost of temporal fidelity. This paper introduces a novel Self-Supervised Event Representation (SSER) method leveraging Gated Recurrent Unit (GRU) networks to achieve precise per-pixel encoding of event timestamps and polarities without temporal discretisation. The recurrent layers are trained in a self-supervised manner to maximise the fidelity of event-time encoding. The inference is performed with event representations generated asynchronously, thus ensuring compatibility with high-throughput sensors. The experimental validation demonstrates that SSER outperforms aggregation-based baselines, achieving improvements of 2.4% mAP and 0.6% on the Gen1 and 1 Mpx object detection datasets. Furthermore, the paper presents the first hardware implementation of recurrent representation for event data on a System-on-Chip FPGA, achieving sub-microsecond latency and power consumption between 1-2 W, suitable for real-time, power-efficient applications. Code is available at https://github.com/vision-agh/RecRepEvent.
Authors: Sarah de Boer, Hartmut H\"antze, Kiran Vaidhya Venkadesh, Myrthe A. D. Buser, Gabriel E. Humpire Mamani, Lina Xu, Lisa C. Adams, Jawed Nawabi, Keno K. Bressem, Bram van Ginneken, Mathias Prokop, Alessa Hering
Abstract: Kidney abnormality segmentation has important potential to enhance the clinical workflow, especially in settings requiring quantitative assessments. Kidney volume could serve as an important biomarker for renal diseases, with changes in volume correlating directly with kidney function. Currently, clinical practice often relies on subjective visual assessment for evaluating kidney size and abnormalities, including tumors and cysts, which are typically staged based on diameter, volume, and anatomical location. To support a more objective and reproducible approach, this research aims to develop a robust, thoroughly validated kidney abnormality segmentation algorithm, made publicly available for clinical and research use. We employ publicly available training datasets and leverage the state-of-the-art medical image segmentation framework nnU-Net. Validation is conducted using both proprietary and public test datasets, with segmentation performance quantified by Dice coefficient and the 95th percentile Hausdorff distance. Furthermore, we analyze robustness across subgroups based on patient sex, age, CT contrast phases, and tumor histologic subtypes. Our findings demonstrate that our segmentation algorithm, trained exclusively on publicly available data, generalizes effectively to external test sets and outperforms existing state-of-the-art models across all tested datasets. Subgroup analyses reveal consistent high performance, indicating strong robustness and reliability. The developed algorithm and associated code are publicly accessible at https://github.com/DIAGNijmegen/oncology-kidney-abnormality-segmentation.
URLs: https://github.com/DIAGNijmegen/oncology-kidney-abnormality-segmentation.
Authors: Manuel Barusco, Francesco Borsatti, Youssef Ben Khalifa, Davide Dalle Pezze, Gian Antonio Susto
Abstract: Semiconductor manufacturing is a complex, multistage process. Automated visual inspection of Scanning Electron Microscope (SEM) images is indispensable for minimizing equipment downtime and containing costs. Most previous research considers supervised approaches, assuming a sufficient number of anomalously labeled samples. On the contrary, Visual Anomaly Detection (VAD), an emerging research domain, focuses on unsupervised learning, avoiding the costly defect collection phase while providing explanations of the predictions. We introduce a benchmark for VAD in the semiconductor domain by leveraging the MIIC dataset. Our results demonstrate the efficacy of modern VAD approaches in this field.
Authors: Yi Zhang, Wenye Zhou, Ruonan Lin, Xin Yang, Hao Zheng
Abstract: Traffic accident prediction and detection are critical for enhancing road safety,and vision-based traffic accident anticipation (Vision-TAA) has emerged as a promising approach in the era of deep learning.This paper reviews 147 recent studies,focusing on the application of supervised,unsupervised,and hybrid deep learning models for accident prediction,alongside the use of real-world and synthetic datasets.Current methodologies are categorized into four key approaches: image and video feature-based prediction, spatiotemporal feature-based prediction, scene understanding,and multimodal data fusion.While these methods demonstrate significant potential,challenges such as data scarcity,limited generalization to complex scenarios,and real-time performance constraints remain prevalent. This review highlights opportunities for future research,including the integration of multimodal data fusion, self-supervised learning,and Transformer-based architectures to enhance prediction accuracy and scalability.By synthesizing existing advancements and identifying critical gaps, this paper provides a foundational reference for developing robust and adaptive Vision-TAA systems,contributing to road safety and traffic management.
Authors: Simone Azeglio, Victor Calbiague Garcia, Guilhem Glaziou, Peter Neri, Olivier Marre, Ulisse Ferrari
Abstract: We present a novel approach to neural response prediction that incorporates higher-order operations directly within convolutional neural networks (CNNs). Our model extends traditional 3D CNNs by embedding higher-order operations within the convolutional operator itself, enabling direct modeling of multiplicative interactions between neighboring pixels across space and time. Our model increases the representational power of CNNs without increasing their depth, therefore addressing the architectural disparity between deep artificial networks and the relatively shallow processing hierarchy of biological visual systems. We evaluate our approach on two distinct datasets: salamander retinal ganglion cell (RGC) responses to natural scenes, and a new dataset of mouse RGC responses to controlled geometric transformations. Our higher-order CNN (HoCNN) achieves superior performance while requiring only half the training data compared to standard architectures, demonstrating correlation coefficients up to 0.75 with neural responses (against 0.80$\pm$0.02 retinal reliability). When integrated into state-of-the-art architectures, our approach consistently improves performance across different species and stimulus conditions. Analysis of the learned representations reveals that our network naturally encodes fundamental geometric transformations, particularly scaling parameters that characterize object expansion and contraction. This capability is especially relevant for specific cell types, such as transient OFF-alpha and transient ON cells, which are known to detect looming objects and object motion respectively, and where our model shows marked improvement in response prediction. The correlation coefficients for scaling parameters are more than twice as high in HoCNN (0.72) compared to baseline models (0.32).
Authors: Zhuo Song, Ye Zhang, Kunhong Li, Longguang Wang, Yulan Guo
Abstract: Cross-view geo-localization is a promising solution for large-scale localization problems, requiring the sequential execution of retrieval and metric localization tasks to achieve fine-grained predictions. However, existing methods typically focus on designing standalone models for these two tasks, resulting in inefficient collaboration and increased training overhead. In this paper, we propose UnifyGeo, a novel unified hierarchical geo-localization framework that integrates retrieval and metric localization tasks into a single network. Specifically, we first employ a unified learning strategy with shared parameters to jointly learn multi-granularity representation, facilitating mutual reinforcement between these two tasks. Subsequently, we design a re-ranking mechanism guided by a dedicated loss function, which enhances geo-localization performance by improving both retrieval accuracy and metric localization references. Extensive experiments demonstrate that UnifyGeo significantly outperforms the state-of-the-arts in both task-isolated and task-associated settings. Remarkably, on the challenging VIGOR benchmark, which supports fine-grained localization evaluation, the 1-meter-level localization recall rate improves from 1.53\% to 39.64\% and from 0.43\% to 25.58\% under same-area and cross-area evaluations, respectively. Code will be made publicly available.
Authors: Ozgur Kara, Krishna Kumar Singh, Feng Liu, Duygu Ceylan, James M. Rehg, Tobias Hinz
Abstract: Current diffusion-based text-to-video methods are limited to producing short video clips of a single shot and lack the capability to generate multi-shot videos with discrete transitions where the same character performs distinct activities across the same or different backgrounds. To address this limitation we propose a framework that includes a dataset collection pipeline and architectural extensions to video diffusion models to enable text-to-multi-shot video generation. Our approach enables generation of multi-shot videos as a single video with full attention across all frames of all shots, ensuring character and background consistency, and allows users to control the number, duration, and content of shots through shot-specific conditioning. This is achieved by incorporating a transition token into the text-to-video model to control at which frames a new shot begins and a local attention masking strategy which controls the transition token's effect and allows shot-specific prompting. To obtain training data we propose a novel data collection pipeline to construct a multi-shot video dataset from existing single-shot video datasets. Extensive experiments demonstrate that fine-tuning a pre-trained text-to-video model for a few thousand iterations is enough for the model to subsequently be able to generate multi-shot videos with shot-specific control, outperforming the baselines. You can find more details in https://shotadapter.github.io/
Authors: Quang Vinh Nguyen, Minh Duc Nguyen, Thanh Hoang Son Vo, Hyung-Jeong Yang, Soo-Hyung Kim
Abstract: Automated Radiology report generation (RRG) aims at producing detailed descriptions of medical images, reducing radiologists' workload and improving access to high-quality diagnostic services. Existing encoder-decoder models only rely on visual features extracted from raw input images, which can limit the understanding of spatial structures and semantic relationships, often resulting in suboptimal text generation. To address this, we propose Anatomical Attention Alignment Network (A3Net), a framework that enhance visual-textual understanding by constructing hyper-visual representations. Our approach integrates a knowledge dictionary of anatomical structures with patch-level visual features, enabling the model to effectively associate image regions with their corresponding anatomical entities. This structured representation improves semantic reasoning, interpretability, and cross-modal alignment, ultimately enhancing the accuracy and clinical relevance of generated reports. Experimental results on IU X-Ray and MIMIC-CXR datasets demonstrate that A3Net significantly improves both visual perception and text generation quality. Our code is available at \href{https://github.com/Vinh-AI/A3Net}{GitHub}.
Authors: Songlin Dong, Chenhao Ding, Jiangyang Li, Jizhou Han, Qiang Wang, Yuhang He, Yihong Gong
Abstract: This study aims to address the problem of multi-domain task incremental learning~(MTIL), which requires that vision-language models~(VLMs) continuously acquire new knowledge while maintaining their inherent zero-shot recognition capability. Existing paradigms delegate the testing of unseen-domain samples to the original CLIP, which only prevents the degradation of the model's zero-shot capability but fails to enhance the generalization of the VLM further. To this end, we propose a novel MTIL framework, named AFA, which comprises two core modules: (1) an against forward-forgetting adapter that learns task-invariant information for each dataset in the incremental tasks to enhance the zero-shot recognition ability of VLMs; (2) an against backward-forgetting adapter that strengthens the few-shot learning capability of VLMs while supporting incremental learning. Extensive experiments demonstrate that the AFA method significantly outperforms existing state-of-the-art approaches, especially in few-shot MTIL tasks, and surpasses the inherent zero-shot performance of CLIP in terms of transferability. The code is provided in the Supplementary Material.
Authors: Negin Ghamsarian, Sahar Nasirihaghighi, Klaus Schoeffmann, Raphael Sznitman
Abstract: Semi-supervised learning leverages unlabeled data to enhance model performance, addressing the limitations of fully supervised approaches. Among its strategies, pseudo-supervision has proven highly effective, typically relying on one or multiple teacher networks to refine pseudo-labels before training a student network. A common practice in pseudo-supervision is filtering pseudo-labels based on pre-defined confidence thresholds or entropy. However, selecting optimal thresholds requires large labeled datasets, which are often scarce in real-world semi-supervised scenarios. To overcome this challenge, we propose Ensemble-of-Confidence Reinforcement (ENCORE), a dynamic feedback-driven thresholding strategy for pseudo-label selection. Instead of relying on static confidence thresholds, ENCORE estimates class-wise true-positive confidence within the unlabeled dataset and continuously adjusts thresholds based on the model's response to different levels of pseudo-label filtering. This feedback-driven mechanism ensures the retention of informative pseudo-labels while filtering unreliable ones, enhancing model training without manual threshold tuning. Our method seamlessly integrates into existing pseudo-supervision frameworks and significantly improves segmentation performance, particularly in data-scarce conditions. Extensive experiments demonstrate that integrating ENCORE with existing pseudo-supervision frameworks enhances performance across multiple datasets and network architectures, validating its effectiveness in semi-supervised learning.
Authors: Elisei Rykov, Kseniia Petrushina, Kseniia Titova, Anton Razzhigaev, Alexander Panchenko, Vasily Konovalov
Abstract: Measuring how real images look is a complex task in artificial intelligence research. For example, an image of a boy with a vacuum cleaner in a desert violates common sense. We introduce a novel method, which we call Through the Looking Glass (TLG), to assess image common sense consistency using Large Vision-Language Models (LVLMs) and Transformer-based encoder. By leveraging LVLMs to extract atomic facts from these images, we obtain a mix of accurate facts. We proceed by fine-tuning a compact attention-pooling classifier over encoded atomic facts. Our TLG has achieved a new state-of-the-art performance on the WHOOPS! and WEIRD datasets while leveraging a compact fine-tuning component.
Authors: Qi Xu, Jie Deng, Jiangrong Shen, Biwu Chen, Huajin Tang, Gang Pan
Abstract: Event-based object detection has gained increasing attention due to its advantages such as high temporal resolution, wide dynamic range, and asynchronous address-event representation. Leveraging these advantages, Spiking Neural Networks (SNNs) have emerged as a promising approach, offering low energy consumption and rich spatiotemporal dynamics. To further enhance the performance of event-based object detection, this study proposes a novel hybrid spike vision Transformer (HsVT) model. The HsVT model integrates a spatial feature extraction module to capture local and global features, and a temporal feature extraction module to model time dependencies and long-term patterns in event sequences. This combination enables HsVT to capture spatiotemporal features, improving its capability to handle complex event-based object detection tasks. To support research in this area, we developed and publicly released The Fall Detection Dataset as a benchmark for event-based object detection tasks. This dataset, captured using an event-based camera, ensures facial privacy protection and reduces memory usage due to the event representation format. We evaluated the HsVT model on GEN1 and Fall Detection datasets across various model sizes. Experimental results demonstrate that HsVT achieves significant performance improvements in event detection with fewer parameters.
Authors: Vignesh Edithal, Le Zhang, Ilia Blank, Imran Junejo
Abstract: In this work, we enable gamers to share their gaming experience on social media by automatically generating eye-catching highlight reels from their gameplay session Our automation will save time for gamers while increasing audience engagement. We approach the highlight generation problem by first identifying intervals in the video where interesting events occur and then concatenate them. We developed an in-house gameplay event detection dataset containing interesting events annotated by humans using VIA video annotator. Traditional techniques for highlight detection such as game engine integration requires expensive collaboration with game developers. OCR techniques which detect patches of specific images or texts require expensive per game engineering and may not generalize across game UI and different language. We finetuned a multimodal general purpose video understanding model such as X-CLIP using our dataset which generalizes across multiple games in a genre without per game engineering. Prompt engineering was performed to improve the classification performance of this multimodal model. Our evaluation showed that such a finetuned model can detect interesting events in first person shooting games from unseen gameplay footage with more than 90% accuracy. Moreover, our model performed significantly better on low resource games (small dataset) when trained along with high resource games, showing signs of transfer learning. To make the model production ready, we used ONNX libraries to enable cross platform inference. These libraries also provide post training quantization tools to reduce model size and inference time for deployment. ONNX runtime libraries with DirectML backend were used to perform efficient inference on Windows OS. We show that natural language supervision in the X-CLIP model leads to data efficient and highly performant video recognition models.
Authors: Jiangling Zhang, Weijie Zhu, Jirui Huang, Yaxiong Chen
Abstract: Detecting AI-synthetic faces presents a critical challenge: it is hard to capture consistent structural relationships between facial regions across diverse generation techniques. Current methods, which focus on specific artifacts rather than fundamental inconsistencies, often fail when confronted with novel generative models. To address this limitation, we introduce Layer-aware Mask Modulation Vision Transformer (LAMM-ViT), a Vision Transformer designed for robust facial forgery detection. This model integrates distinct Region-Guided Multi-Head Attention (RG-MHA) and Layer-aware Mask Modulation (LAMM) components within each layer. RG-MHA utilizes facial landmarks to create regional attention masks, guiding the model to scrutinize architectural inconsistencies across different facial areas. Crucially, the separate LAMM module dynamically generates layer-specific parameters, including mask weights and gating values, based on network context. These parameters then modulate the behavior of RG-MHA, enabling adaptive adjustment of regional focus across network depths. This architecture facilitates the capture of subtle, hierarchical forgery cues ubiquitous among diverse generation techniques, such as GANs and Diffusion Models. In cross-model generalization tests, LAMM-ViT demonstrates superior performance, achieving 94.09% mean ACC (a +5.45% improvement over SoTA) and 98.62% mean AP (a +3.09% improvement). These results demonstrate LAMM-ViT's exceptional ability to generalize and its potential for reliable deployment against evolving synthetic media threats.
Authors: Halid Ziya Yerebakan, Kritika Iyer, Xueqi Guo, Yoshihisa Shinagawa, Gerardo Hermosillo Valadez
Abstract: We introduce a new type of foundational model for parsing human anatomy in medical images that works for different modalities. It supports supervised or unsupervised training and can perform matching, registration, classification, or segmentation with or without user interaction. We achieve this by training a neural network estimator that maps query locations to atlas coordinates via regression. Efficiency is improved by sparsely sampling the input, enabling response times of less than 1 ms without additional accelerator hardware. We demonstrate the utility of the algorithm in both CT and MRI modalities.
Authors: Weiyu Li, Xuanyang Zhang, Zheng Sun, Di Qi, Hao Li, Wei Cheng, Weiwei Cai, Shihao Wu, Jiarui Liu, Zihao Wang, Xiao Chen, Feipeng Tian, Jianxiong Pan, Zeming Li, Gang Yu, Xiangyu Zhang, Daxin Jiang, Ping Tan
Abstract: While generative artificial intelligence has advanced significantly across text, image, audio, and video domains, 3D generation remains comparatively underdeveloped due to fundamental challenges such as data scarcity, algorithmic limitations, and ecosystem fragmentation. To this end, we present Step1X-3D, an open framework addressing these challenges through: (1) a rigorous data curation pipeline processing >5M assets to create a 2M high-quality dataset with standardized geometric and textural properties; (2) a two-stage 3D-native architecture combining a hybrid VAE-DiT geometry generator with an diffusion-based texture synthesis module; and (3) the full open-source release of models, training code, and adaptation modules. For geometry generation, the hybrid VAE-DiT component produces TSDF representations by employing perceiver-based latent encoding with sharp edge sampling for detail preservation. The diffusion-based texture synthesis module then ensures cross-view consistency through geometric conditioning and latent-space synchronization. Benchmark results demonstrate state-of-the-art performance that exceeds existing open-source methods, while also achieving competitive quality with proprietary solutions. Notably, the framework uniquely bridges the 2D and 3D generation paradigms by supporting direct transfer of 2D control techniques~(e.g., LoRA) to 3D synthesis. By simultaneously advancing data quality, algorithmic fidelity, and reproducibility, Step1X-3D aims to establish new standards for open research in controllable 3D asset generation.
Authors: Chenze Shao, Fandong Meng, Jie Zhou
Abstract: Conventional wisdom suggests that autoregressive models are used to process discrete data. When applied to continuous modalities such as visual data, Visual AutoRegressive modeling (VAR) typically resorts to quantization-based approaches to cast the data into a discrete space, which can introduce significant information loss. To tackle this issue, we introduce a Continuous VAR framework that enables direct visual autoregressive generation without vector quantization. The underlying theoretical foundation is strictly proper scoring rules, which provide powerful statistical tools capable of evaluating how well a generative model approximates the true distribution. Within this framework, all we need is to select a strictly proper score and set it as the training objective to optimize. We primarily explore a class of training objectives based on the energy score, which is likelihood-free and thus overcomes the difficulty of making probabilistic predictions in the continuous space. Previous efforts on continuous autoregressive generation, such as GIVT and diffusion loss, can also be derived from our framework using other strictly proper scores. Source code: https://github.com/shaochenze/EAR.
Authors: Zeyue Xue, Jie Wu, Yu Gao, Fangyuan Kong, Lingting Zhu, Mengzhao Chen, Zhiheng Liu, Wei Liu, Qiushan Guo, Weilin Huang, Ping Luo
Abstract: Recent breakthroughs in generative models-particularly diffusion models and rectified flows-have revolutionized visual content creation, yet aligning model outputs with human preferences remains a critical challenge. Existing reinforcement learning (RL)-based methods for visual generation face critical limitations: incompatibility with modern Ordinary Differential Equations (ODEs)-based sampling paradigms, instability in large-scale training, and lack of validation for video generation. This paper introduces DanceGRPO, the first unified framework to adapt Group Relative Policy Optimization (GRPO) to visual generation paradigms, unleashing one unified RL algorithm across two generative paradigms (diffusion models and rectified flows), three tasks (text-to-image, text-to-video, image-to-video), four foundation models (Stable Diffusion, HunyuanVideo, FLUX, SkyReel-I2V), and five reward models (image/video aesthetics, text-image alignment, video motion quality, and binary reward). To our knowledge, DanceGRPO is the first RL-based unified framework capable of seamless adaptation across diverse generative paradigms, tasks, foundational models, and reward models. DanceGRPO demonstrates consistent and substantial improvements, which outperform baselines by up to 181% on benchmarks such as HPS-v2.1, CLIP Score, VideoAlign, and GenEval. Notably, DanceGRPO not only can stabilize policy optimization for complex video generation, but also enables generative policy to better capture denoising trajectories for Best-of-N inference scaling and learn from sparse binary feedback. Our results establish DanceGRPO as a robust and versatile solution for scaling Reinforcement Learning from Human Feedback (RLHF) tasks in visual generation, offering new insights into harmonizing reinforcement learning and visual synthesis. The code will be released.
Authors: Yizhuo Wu, Yi Zhu, Kun Qian, Qinyu Chen, Anding Zhu, John Gajadharsing, Leo C. N. de Vreede, Chang Gao
Abstract: Digital Predistortion (DPD) is a popular technique to enhance signal quality in wideband RF power amplifiers (PAs). With increasing bandwidth and data rates, DPD faces significant energy consumption challenges during deployment, contrasting with its efficiency goals. State-of-the-art DPD models rely on recurrent neural networks (RNN), whose computational complexity hinders system efficiency. This paper introduces DeltaDPD, exploring the dynamic temporal sparsity of input signals and neuronal hidden states in RNNs for energy-efficient DPD, reducing arithmetic operations and memory accesses while preserving satisfactory linearization performance. Applying a TM3.1a 200MHz-BW 256-QAM OFDM signal to a 3.5 GHz GaN Doherty RF PA, DeltaDPD achieves -50.03 dBc in Adjacent Channel Power Ratio (ACPR), -37.22 dB in Normalized Mean Square Error (NMSE) and -38.52 dBc in Error Vector Magnitude (EVM) with 52% temporal sparsity, leading to a 1.8X reduction in estimated inference power. The DeltaDPD code will be released after formal publication at https://www.opendpd.com.
URLs: https://www.opendpd.com.
Authors: Yuzhou Zhu, Zheng Zhang, Ruyi Zhang, Liang Zhou
Abstract: Attosecond streaking phase retrieval is essential for resolving electron dynamics on sub-femtosecond time scales yet traditional algorithms rely on iterative minimization and central momentum approximations that degrade accuracy for broadband pulses. In this work phase retrieval is reformulated as a supervised computer-vision problem and four neural architectures are systematically compared. A convolutional network demonstrates strong sensitivity to local streak edges but lacks global context; a vision transformer captures long-range delay-energy correlations at the expense of local inductive bias; a hybrid CNN-ViT model unites local feature extraction and full-graph attention; and a capsule network further enforces spatial pose agreement through dynamic routing. A theoretical analysis introduces local, global and positional sensitivity measures and derives surrogate error bounds that predict the strict ordering $CNN
Authors: John Song, Lihao Zhang, Feng Ye, Haijian Sun
Abstract: Terahertz (THz) communication is a key enabler for 6G systems, offering ultra-wide bandwidth and unprecedented data rates. However, THz signal propagation differs significantly from lower-frequency bands due to severe free space path loss, minimal diffraction and specular reflection, and prominent scattering, making conventional channel modeling and pilot-based estimation approaches inefficient. In this work, we investigate the feasibility of applying radio radiance field (RRF) framework to the THz band. This method reconstructs a continuous RRF using visual-based geometry and sparse THz RF measurements, enabling efficient spatial channel state information (Spatial-CSI) modeling without dense sampling. We first build a fine simulated THz scenario, then we reconstruct the RRF and evaluate the performance in terms of both reconstruction quality and effectiveness in THz communication, showing that the reconstructed RRF captures key propagation paths with sparse training samples. Our findings demonstrate that RRF modeling remains effective in the THz regime and provides a promising direction for scalable, low-cost spatial channel reconstruction in future 6G networks.
Authors: Yuhan Yuan, Xiaomo Jiang, Yanfeng Han, Ke Xiao
Abstract: Rolling bearings are critical components of rotating machinery, and their proper functioning is essential for industrial production. Most existing condition monitoring methods focus on extracting discriminative features from time-domain signals to assess bearing health status. However, under complex operating conditions, periodic impulsive characteristics related to fault information are often obscured by noise interference. Consequently, existing approaches struggle to learn distinctive fault-related features in such scenarios. To address this issue, this paper proposes a novel CNN-based model named FEMSN. Specifically, a Fourier Adaptive Denoising Encoder Layer (FADEL) is introduced as an input denoising layer to enhance key features while filtering out irrelevant information. Subsequently, a Multiscale Time-Frequency Fusion (MSTFF) module is employed to extract fused time-frequency features, further improving the model robustness and nonlinear representation capability. Additionally, a distillation layer is incorporated to expand the receptive field. Based on these advancements, a novel deep lightweight CNN model, termed the Frequency-Enhanced Multiscale Network (FEMSN), is developed. The effectiveness of FEMSN and FADEL in machine health monitoring and stability assessment is validated through two case studies.
Authors: Adhora Madhuri, Nusaiba Sobir, Tasnia Binte Mamun, Taufiq Hasan
Abstract: Lung cancer is the leading cause of patient mortality in the world. Early diagnosis of malignant pulmonary nodules in CT images can have a significant impact on reducing disease mortality and morbidity. In this work, we propose LMLCC-Net, a novel deep learning framework for classifying nodules from CT scan images using a 3D CNN, considering Hounsfield Unit (HU)-based intensity filtering. Benign and malignant nodules have significant differences in their intensity profile of HU, which was not exploited in the literature. Our method considers the intensity pattern as well as the texture for the prediction of malignancies. LMLCC-Net extracts features from multiple branches that each use a separate learnable HU-based intensity filtering stage. Various combinations of branches and learnable ranges of filters were explored to finally produce the best-performing model. In addition, we propose a semi-supervised learning scheme for labeling ambiguous cases and also developed a lightweight model to classify the nodules. The experimental evaluations are carried out on the LUNA16 dataset. Our proposed method achieves a classification accuracy (ACC) of 91.96%, a sensitivity (SEN) of 92.04%, and an area under the curve (AUC) of 91.87%, showing improved performance compared to existing methods. The proposed method can have a significant impact in helping radiologists in the classification of pulmonary nodules and improving patient care.
Authors: Shehryar Khattak, Timon Homberger, Lukas Bernreiter, Julian Nubert, Olov Andersson, Roland Siegwart, Kostas Alexis, Marco Hutter
Abstract: Robot autonomy in unknown, GPS-denied, and complex underground environments requires real-time, robust, and accurate onboard pose estimation and mapping for reliable operations. This becomes particularly challenging in perception-degraded subterranean conditions under harsh environmental factors, including darkness, dust, and geometrically self-similar structures. This paper details CompSLAM, a highly resilient and hierarchical multi-modal localization and mapping framework designed to address these challenges. Its flexible architecture achieves resilience through redundancy by leveraging the complementary nature of pose estimates derived from diverse sensor modalities. Developed during the DARPA Subterranean Challenge, CompSLAM was successfully deployed on all aerial, legged, and wheeled robots of Team Cerberus during their competition-winning final run. Furthermore, it has proven to be a reliable odometry and mapping solution in various subsequent projects, with extensions enabling multi-robot map sharing for marsupial robotic deployments and collaborative mapping. This paper also introduces a comprehensive dataset acquired by a manually teleoperated quadrupedal robot, covering a significant portion of the DARPA Subterranean Challenge finals course. This dataset evaluates CompSLAM's robustness to sensor degradations as the robot traverses 740 meters in an environment characterized by highly variable geometries and demanding lighting conditions. The CompSLAM code and the DARPA SubT Finals dataset are made publicly available for the benefit of the robotics community
Authors: Md Rakibul Hasan, Pouria Behnoudfar, Dan MacKinlay, Thomas Poulet
Abstract: Machine Learning, particularly Generative Adversarial Networks (GANs), has revolutionised Super Resolution (SR). However, generated images often lack physical meaningfulness, which is essential for scientific applications. Our approach, PC-SRGAN, enhances image resolution while ensuring physical consistency for interpretable simulations. PC-SRGAN significantly improves both the Peak Signal-to-Noise Ratio and the Structural Similarity Index Measure compared to conventional methods, even with limited training data (e.g., only 13% of training data required for SRGAN). Beyond SR, PC-SRGAN augments physically meaningful machine learning, incorporating numerically justified time integrators and advanced quality metrics. These advancements promise reliable and causal machine-learning models in scientific domains. A significant advantage of PC-SRGAN over conventional SR techniques is its physical consistency, which makes it a viable surrogate model for time-dependent problems. PC-SRGAN advances scientific machine learning, offering improved accuracy and efficiency for image processing, enhanced process understanding, and broader applications to scientific research. The source codes and data will be made publicly available at https://github.com/hasan-rakibul/PC-SRGAN upon acceptance of this paper.
Authors: Haoyang Xie, Feng Ju
Abstract: Computer-aided design (CAD) is fundamental to modern engineering and manufacturing, but creating CAD models still requires expert knowledge and specialized software. Recent advances in large language models (LLMs) open up the possibility of generative CAD, where natural language is directly translated into parametric 3D models. However, most existing methods generate task-specific command sequences that pretrained models cannot directly handle. These sequences must be converted into CAD representations such as CAD vectors before a 3D model can be produced, which requires training models from scratch and adds unnecessary complexity. To tackle this issue, we propose generating CadQuery code directly from text, leveraging the strengths of pretrained LLMs to produce 3D models without intermediate representations, using this Python-based scripting language. Since LLMs already excel at Python generation and spatial reasoning, fine-tuning them on Text-to-CadQuery data proves highly effective. Given that these capabilities typically improve with scale, we hypothesize that larger models will perform better after fine-tuning. To enable this, we augment the Text2CAD dataset with 170,000 CadQuery annotations. We fine-tune six open-source LLMs of varying sizes and observe consistent improvements. Our best model achieves a top-1 exact match of 69.3%, up from 58.8%, and reduces Chamfer Distance by 48.6%. Project page: https://github.com/Text-to-CadQuery/Text-to-CadQuery.
Authors: Galann Pennec, Zhengyuan Liu, Nicholas Asher, Philippe Muller, Nancy F. Chen
Abstract: Vision-Language Models (VLMs) often struggle to balance visual and textual information when summarizing complex multimodal inputs, such as entire TV show episodes. In this paper, we propose a zero-shot video-to-text summarization approach that builds its own screenplay representation of an episode, effectively integrating key video moments, dialogue, and character information into a unified document. Unlike previous approaches, we simultaneously generate screenplays and name the characters in zero-shot, using only the audio, video, and transcripts as input. Additionally, we highlight that existing summarization metrics can fail to assess the multimodal content in summaries. To address this, we introduce MFactSum, a multimodal metric that evaluates summaries with respect to both vision and text modalities. Using MFactSum, we evaluate our screenplay summaries on the SummScreen3D dataset, demonstrating superiority against state-of-the-art VLMs such as Gemini 1.5 by generating summaries containing 20% more relevant visual information while requiring 75% less of the video as input.
Authors: Hai-Vy Nguyen, Fabrice Gamboa, Sixin Zhang, Reda Chhaibi, Serge Gratton, Thierry Giaccone
Abstract: In this paper, we propose a method for transferring feature representation to lightweight student models from larger teacher models. We mathematically define a new notion called \textit{perception coherence}. Based on this notion, we propose a loss function, which takes into account the dissimilarities between data points in feature space through their ranking. At a high level, by minimizing this loss function, the student model learns to mimic how the teacher model \textit{perceives} inputs. More precisely, our method is motivated by the fact that the representational capacity of the student model is weaker than the teacher model. Hence, we aim to develop a new method allowing for a better relaxation. This means that, the student model does not need to preserve the absolute geometry of the teacher one, while preserving global coherence through dissimilarity ranking. Our theoretical insights provide a probabilistic perspective on the process of feature representation transfer. Our experiments results show that our method outperforms or achieves on-par performance compared to strong baseline methods for representation transferring.
Authors: Thamiris Coelho, Leo S. F. Ribeiro, Jo\~ao Macedo, Jefersson A. dos Santos, Sandra Avila
Abstract: The distribution of child sexual abuse imagery (CSAI) is an ever-growing concern of our modern world; children who suffered from this heinous crime are revictimized, and the growing amount of illegal imagery distributed overwhelms law enforcement agents (LEAs) with the manual labor of categorization. To ease this burden researchers have explored methods for automating data triage and detection of CSAI, but the sensitive nature of the data imposes restricted access and minimal interaction between real data and learning algorithms, avoiding leaks at all costs. In observing how these restrictions have shaped the literature we formalize a definition of "Proxy Tasks", i.e., the substitute tasks used for training models for CSAI without making use of CSA data. Under this new terminology we review current literature and present a protocol for making conscious use of Proxy Tasks together with consistent input from LEAs to design better automation in this field. Finally, we apply this protocol to study -- for the first time -- the task of Few-shot Indoor Scene Classification on CSAI, showing a final model that achieves promising results on a real-world CSAI dataset whilst having no weights actually trained on sensitive data.
Authors: Daniel Strick, Carlos Garcia, Anthony Huang
Abstract: Deep learning for radiologic image analysis is a rapidly growing field in biomedical research and is likely to become a standard practice in modern medicine. On the publicly available NIH ChestX-ray14 dataset, containing X-ray images that are classified by the presence or absence of 14 different diseases, we reproduced an algorithm known as CheXNet, as well as explored other algorithms that outperform CheXNet's baseline metrics. Model performance was primarily evaluated using the F1 score and AUC-ROC, both of which are critical metrics for imbalanced, multi-label classification tasks in medical imaging. The best model achieved an average AUC-ROC score of 0.85 and an average F1 score of 0.39 across all 14 disease classifications present in the dataset.
Authors: Dawei Huang, Qing Li, Chuan Yan, Zebang Cheng, Yurong Huang, Xiang Li, Bin Li, Xiaohui Wang, Zheng Lian, Xiaojiang Peng
Abstract: Emotion understanding in videos aims to accurately recognize and interpret individuals' emotional states by integrating contextual, visual, textual, and auditory cues. While Large Multimodal Models (LMMs) have demonstrated significant progress in general vision-language (VL) tasks, their performance in emotion-specific scenarios remains limited. Moreover, fine-tuning LMMs on emotion-related tasks often leads to catastrophic forgetting, hindering their ability to generalize across diverse tasks. To address these challenges, we present Emotion-Qwen, a tailored multimodal framework designed to enhance both emotion understanding and general VL reasoning. Emotion-Qwen incorporates a sophisticated Hybrid Compressor based on the Mixture of Experts (MoE) paradigm, which dynamically routes inputs to balance emotion-specific and general-purpose processing. The model is pre-trained in a three-stage pipeline on large-scale general and emotional image datasets to support robust multimodal representations. Furthermore, we construct the Video Emotion Reasoning (VER) dataset, comprising more than 40K bilingual video clips with fine-grained descriptive annotations, to further enrich Emotion-Qwen's emotional reasoning capability. Experimental results demonstrate that Emotion-Qwen achieves state-of-the-art performance on multiple emotion recognition benchmarks, while maintaining competitive results on general VL tasks. Code and models are available at https://anonymous.4open.science/r/Emotion-Qwen-Anonymous.
URLs: https://anonymous.4open.science/r/Emotion-Qwen-Anonymous.
Authors: Morui Zhu, Yongqi Zhu, Yihao Zhu, Qi Chen, Deyuan Qu, Song Fu, Qing Yang
Abstract: We introduce M$^3$CAD, a novel benchmark designed to advance research in generic cooperative autonomous driving. M$^3$CAD comprises 204 sequences with 30k frames, spanning a diverse range of cooperative driving scenarios. Each sequence includes multiple vehicles and sensing modalities, e.g., LiDAR point clouds, RGB images, and GPS/IMU, supporting a variety of autonomous driving tasks, including object detection and tracking, mapping, motion forecasting, occupancy prediction, and path planning. This rich multimodal setup enables M$^3$CAD to support both single-vehicle and multi-vehicle autonomous driving research, significantly broadening the scope of research in the field. To our knowledge, M$^3$CAD is the most comprehensive benchmark specifically tailored for cooperative multi-task autonomous driving research. We evaluate the state-of-the-art end-to-end solution on M$^3$CAD to establish baseline performance. To foster cooperative autonomous driving research, we also propose E2EC, a simple yet effective framework for cooperative driving solution that leverages inter-vehicle shared information for improved path planning. We release M$^3$CAD, along with our baseline models and evaluation results, to support the development of robust cooperative autonomous driving systems. All resources will be made publicly available on https://github.com/zhumorui/M3CAD
Authors: Erik Gro{\ss}kopf, Valay Bundele, Mehran Hossienzadeh, Hendrik P. A. Lensch
Abstract: Hematoxylin and Eosin (H&E) staining is the cornerstone of histopathology but lacks molecular specificity. While Immunohistochemistry (IHC) provides molecular insights, it is costly and complex, motivating H&E-to-IHC translation as a cost-effective alternative. Existing translation methods are mainly GAN-based, often struggling with training instability and limited structural fidelity, while diffusion-based approaches remain underexplored. We propose HistDiST, a Latent Diffusion Model (LDM) based framework for high-fidelity H&E-to-IHC translation. HistDiST introduces a dual-conditioning strategy, utilizing Phikon-extracted morphological embeddings alongside VAE-encoded H&E representations to ensure pathology-relevant context and structural consistency. To overcome brightness biases, we incorporate a rescaled noise schedule, v-prediction, and trailing timesteps, enforcing a zero-SNR condition at the final timestep. During inference, DDIM inversion preserves the morphological structure, while an eta-cosine noise schedule introduces controlled stochasticity, balancing structural consistency and molecular fidelity. Moreover, we propose Molecular Retrieval Accuracy (MRA), a novel pathology-aware metric leveraging GigaPath embeddings to assess molecular relevance. Extensive evaluations on MIST and BCI datasets demonstrate that HistDiST significantly outperforms existing methods, achieving a 28% improvement in MRA on the H&E-to-Ki67 translation task, highlighting its effectiveness in capturing true IHC semantics.
Authors: Xilin Jiang, Junkai Wu, Vishal Choudhari, Nima Mesgarani
Abstract: Audio large language models (LLMs) are considered experts at recognizing sound objects, yet their performance relative to LLMs in other sensory modalities, such as visual or audio-visual LLMs, and to humans using their ears, eyes, or both remains unexplored. To investigate this, we systematically evaluate audio, visual, and audio-visual LLMs, specifically Qwen2-Audio, Qwen2-VL, and Qwen2.5-Omni, against humans in recognizing sound objects of different classes from audio-only, silent video, or sounded video inputs. We uncover a performance gap between Qwen2-Audio and Qwen2-VL that parallels the sensory discrepancy between human ears and eyes. To reduce this gap, we introduce a cross-modal distillation framework, where an LLM in one modality serves as the teacher and another as the student, with knowledge transfer in sound classes predicted as more challenging to the student by a heuristic model. Distillation in both directions, from Qwen2-VL to Qwen2-Audio and vice versa, leads to notable improvements, particularly in challenging classes. This work highlights the sensory gap in LLMs from a human-aligned perspective and proposes a principled approach to enhancing modality-specific perception in multimodal LLMs.
Authors: Tan-Hanh Pham, Ovidiu C. Andronesi, Xianqi Li, Kim-Doang Nguyen
Abstract: Magnetic Resonance Spectroscopic Imaging (MRSI) is a powerful tool for non-invasive mapping of brain metabolites, providing critical insights into neurological conditions. However, its utility is often limited by missing or corrupted data due to motion artifacts, magnetic field inhomogeneities, or failed spectral fitting-especially in high resolution 3D acquisitions. To address this, we propose the first deep learning-based, mask-free framework for estimating missing data in MRSI metabolic maps. Unlike conventional restoration methods that rely on explicit masks to identify missing regions, our approach implicitly detects and estimates these areas using contextual spatial features through 2D and 3D U-Net architectures. We also introduce a progressive training strategy to enhance robustness under varying levels of data degradation. Our method is evaluated on both simulated and real patient datasets and consistently outperforms traditional interpolation techniques such as cubic and linear interpolation. The 2D model achieves an MSE of 0.002 and an SSIM of 0.97 with 20% missing voxels, while the 3D model reaches an MSE of 0.001 and an SSIM of 0.98 with 15% missing voxels. Qualitative results show improved fidelity in estimating missing data, particularly in metabolically heterogeneous regions and ventricular regions. Importantly, our model generalizes well to real-world datasets without requiring retraining or mask input. These findings demonstrate the effectiveness and broad applicability of mask-free deep learning for MRSI restoration, with strong potential for clinical and research integration.
Authors: Dongxiu Liu, Haoyi Niu, Zhihao Wang, Jinliang Zheng, Yinan Zheng, Zhonghong Ou, Jianming Hu, Jianxiong Li, Xianyuan Zhan
Abstract: Current robotic planning methods often rely on predicting multi-frame images with full pixel details. While this fine-grained approach can serve as a generic world model, it introduces two significant challenges for downstream policy learning: substantial computational costs that hinder real-time deployment, and accumulated inaccuracies that can mislead action extraction. Planning with coarse-grained subgoals partially alleviates efficiency issues. However, their forward planning schemes can still result in off-task predictions due to accumulation errors, leading to misalignment with long-term goals. This raises a critical question: Can robotic planning be both efficient and accurate enough for real-time control in long-horizon, multi-stage tasks? To address this, we propose a Latent Space Backward Planning scheme (LBP), which begins by grounding the task into final latent goals, followed by recursively predicting intermediate subgoals closer to the current state. The grounded final goal enables backward subgoal planning to always remain aware of task completion, facilitating on-task prediction along the entire planning horizon. The subgoal-conditioned policy incorporates a learnable token to summarize the subgoal sequences and determines how each subgoal guides action extraction. Through extensive simulation and real-robot long-horizon experiments, we show that LBP outperforms existing fine-grained and forward planning methods, achieving SOTA performance. Project Page: https://lbp-authors.github.io
Authors: Kosuke Ukita, Ye Xiaolong, Tsuyoshi Okita
Abstract: In this paper, we propose a diffusion model that integrates a representation-conditioning mechanism, where the representations derived from a Vision Transformer (ViT) are used to condition the internal process of a Transformer-based diffusion model. This approach enables representation-conditioned data generation, addressing the challenge of requiring large-scale labeled datasets by leveraging self-supervised learning on unlabeled data. We evaluate our method through a zero-shot classification task for hematoma detection in brain imaging. Compared to the strong contrastive learning baseline, DINOv2, our method achieves a notable improvement of +6.15% in accuracy and +13.60% in F1-score, demonstrating its effectiveness in image classification.
Authors: Yu Qiao, Huy Q. Le, Avi Deb Raha, Phuong-Nam Tran, Apurba Adhikary, Mengchun Zhang, Loc X. Nguyen, Eui-Nam Huh, Dusit Niyato, Choong Seon Hong
Abstract: The rise of large language models (LLMs), such as ChatGPT, DeepSeek, and Grok-3, has reshaped the artificial intelligence landscape. As prominent examples of foundational models (FMs) built on LLMs, these models exhibit remarkable capabilities in generating human-like content, bringing us closer to achieving artificial general intelligence (AGI). However, their large-scale nature, sensitivity to privacy concerns, and substantial computational demands present significant challenges to personalized customization for end users. To bridge this gap, this paper presents the vision of artificial personalized intelligence (API), focusing on adapting these powerful models to meet the specific needs and preferences of users while maintaining privacy and efficiency. Specifically, this paper proposes personalized federated intelligence (PFI), which integrates the privacy-preserving advantages of federated learning (FL) with the zero-shot generalization capabilities of FMs, enabling personalized, efficient, and privacy-protective deployment at the edge. We first review recent advances in both FL and FMs, and discuss the potential of leveraging FMs to enhance federated systems. We then present the key motivations behind realizing PFI and explore promising opportunities in this space, including efficient PFI, trustworthy PFI, and PFI empowered by retrieval-augmented generation (RAG). Finally, we outline key challenges and future research directions for deploying FM-powered FL systems at the edge with improved personalization, computational efficiency, and privacy guarantees. Overall, this survey aims to lay the groundwork for the development of API as a complement to AGI, with a particular focus on PFI as a key enabling technique.
Authors: Yanhui Hong, Nan Wang, Zhiyi Xia, Haoyi Tao, Xi Fang, Yiming Li, Jiankun Wang, Peng Jin, Xiaochen Cai, Shengyu Li, Ziqi Chen, Zezhong Zhang, Guolin Ke, Linfeng Zhang
Abstract: This paper presents a systematic solution for the intelligent recognition and automatic analysis of microscopy images. We developed a data engine that generates high-quality annotated datasets through a combination of the collection of diverse microscopy images from experiments, synthetic data generation and a human-in-the-loop annotation process. To address the unique challenges of microscopy images, we propose a segmentation model capable of robustly detecting both small and large objects. The model effectively identifies and separates thousands of closely situated targets, even in cluttered visual environments. Furthermore, our solution supports the precise automatic recognition of image scale bars, an essential feature in quantitative microscopic analysis. Building upon these components, we have constructed a comprehensive intelligent analysis platform and validated its effectiveness and practicality in real-world applications. This study not only advances automatic recognition in microscopy imaging but also ensures scalability and generalizability across multiple application domains, offering a powerful tool for automated microscopic analysis in interdisciplinary research.
Authors: Roy Betser, Meir Yossef Levi, Guy Gilboa
Abstract: Likelihood approximations for images are not trivial to compute and can be useful in many applications. We examine the use of Contrastive Language-Image Pre-training (CLIP) to assess the likelihood of images and captions. We introduce \textit{Whitened CLIP}, a novel transformation of the CLIP latent space via an invertible linear operation. This transformation ensures that each feature in the embedding space has zero mean, unit standard deviation, and no correlation with all other features, resulting in an identity covariance matrix. We show that the whitened embeddings statistics can be well approximated as a standard normal distribution, thus, the log-likelihood is estimated simply by the square Euclidean norm in the whitened embedding space. The whitening procedure is completely training-free and performed using a pre-computed whitening matrix, hence, is very fast. We present several preliminary experiments demonstrating the properties and applicability of these likelihood scores to images and captions.
Authors: Tarik Houichime, Younes EL Amrani
Abstract: This paper introduces an innovative approach for the autonomous landing of Unmanned Aerial Vehicles (UAVs) using only a front-facing monocular camera, therefore obviating the requirement for depth estimation cameras. Drawing on the inherent human estimating process, the proposed method reframes the landing task as an optimization problem. The UAV employs variations in the visual characteristics of a specially designed lenticular circle on the landing pad, where the perceived color and form provide critical information for estimating both altitude and depth. Reinforcement learning algorithms are utilized to approximate the functions governing these estimations, enabling the UAV to ascertain ideal landing settings via training. This method's efficacy is assessed by simulations and experiments, showcasing its potential for robust and accurate autonomous landing without dependence on complex sensor setups. This research contributes to the advancement of cost-effective and efficient UAV landing solutions, paving the way for wider applicability across various fields.
Authors: Lei Wan, Prabesh Gupta, Andreas Eich, Marcel Kettelgerdes, Hannan Ejaz Keen, Michael Kl\"oppel-Gersdorf, Alexey Vinel
Abstract: Perception is a core capability of automated vehicles and has been significantly advanced through modern sensor technologies and artificial intelligence. However, perception systems still face challenges in complex real-world scenarios. To improve robustness against various external factors, multi-sensor fusion techniques are essential, combining the strengths of different sensor modalities. With recent developments in Vehicle-to-Everything (V2X communication, sensor fusion can now extend beyond a single vehicle to a cooperative multi-agent system involving Connected Automated Vehicle (CAV) and intelligent infrastructure. This paper presents VALISENS, an innovative multi-sensor system distributed across multiple agents. It integrates onboard and roadside LiDARs, radars, thermal cameras, and RGB cameras to enhance situational awareness and support cooperative automated driving. The thermal camera adds critical redundancy for perceiving Vulnerable Road User (VRU), while fusion with roadside sensors mitigates visual occlusions and extends the perception range beyond the limits of individual vehicles. We introduce the corresponding perception module built on this sensor system, which includes object detection, tracking, motion forecasting, and high-level data fusion. The proposed system demonstrates the potential of cooperative perception in real-world test environments and lays the groundwork for future Cooperative Intelligent Transport Systems (C-ITS) applications.
Authors: Yuxuan He, Junpeng Zhang, Hongyuan Zhang, Quanshi Zhang
Abstract: This paper proposes a new perspective for analyzing the generalization power of deep neural networks (DNNs), i.e., directly disentangling and analyzing the dynamics of generalizable and non-generalizable interaction encoded by a DNN through the training process. Specifically, this work builds upon the recent theoretical achievement in explainble AI, which proves that the detailed inference logic of DNNs can be can be strictly rewritten as a small number of AND-OR interaction patterns. Based on this, we propose an efficient method to quantify the generalization power of each interaction, and we discover a distinct three-phase dynamics of the generalization power of interactions during training. In particular, the early phase of training typically removes noisy and non-generalizable interactions and learns simple and generalizable ones. The second and the third phases tend to capture increasingly complex interactions that are harder to generalize. Experimental results verify that the learning of non-generalizable interactions is the the direct cause for the gap between the training and testing losses.
Authors: Matt Franchi, Hauke Sandhaus, Madiha Zahrah Choksi, Severin Engelmann, Wendy Ju, Helen Nissenbaum
Abstract: Spatially and temporally dense street imagery (DSI) datasets have grown unbounded. In 2024, individual companies possessed around 3 trillion unique images of public streets. DSI data streams are only set to grow as companies like Lyft and Waymo use DSI to train autonomous vehicle algorithms and analyze collisions. Academic researchers leverage DSI to explore novel approaches to urban analysis. Despite good-faith efforts by DSI providers to protect individual privacy through blurring faces and license plates, these measures fail to address broader privacy concerns. In this work, we find that increased data density and advancements in artificial intelligence enable harmful group membership inferences from supposedly anonymized data. We perform a penetration test to demonstrate how easily sensitive group affiliations can be inferred from obfuscated pedestrians in 25,232,608 dashcam images taken in New York City. We develop a typology of identifiable groups within DSI and analyze privacy implications through the lens of contextual integrity. Finally, we discuss actionable recommendations for researchers working with data from DSI providers.
Authors: Tong Zhang, Fenghua Shao, Runsheng Zhang, Yifan Zhuang, Liuqingqing Yang
Abstract: Based on the DeepSORT algorithm, this study explores the application of visual tracking technology in intelligent human-computer interaction, especially in the field of gesture recognition and tracking. With the rapid development of artificial intelligence and deep learning technology, visual-based interaction has gradually replaced traditional input devices and become an important way for intelligent systems to interact with users. The DeepSORT algorithm can achieve accurate target tracking in dynamic environments by combining Kalman filters and deep learning feature extraction methods. It is especially suitable for complex scenes with multi-target tracking and fast movements. This study experimentally verifies the superior performance of DeepSORT in gesture recognition and tracking. It can accurately capture and track the user's gesture trajectory and is superior to traditional tracking methods in terms of real-time and accuracy. In addition, this study also combines gesture recognition experiments to evaluate the recognition ability and feedback response of the DeepSORT algorithm under different gestures (such as sliding, clicking, and zooming). The experimental results show that DeepSORT can not only effectively deal with target occlusion and motion blur but also can stably track in a multi-target environment, achieving a smooth user interaction experience. Finally, this paper looks forward to the future development direction of intelligent human-computer interaction systems based on visual tracking and proposes future research focuses such as algorithm optimization, data fusion, and multimodal interaction in order to promote a more intelligent and personalized interactive experience. Keywords-DeepSORT, visual tracking, gesture recognition, human-computer interaction
Authors: Jong Sung Park, Juhyung Ha, Siddhesh Thakur, Alexandra Badea, Spyridon Bakas, Eleftherios Garyfallidis
Abstract: While many skull stripping algorithms have been developed for multi-modal and multi-species cases, there is still a lack of a fundamentally generalizable approach. We present PUMBA(PUrely synthetic Multimodal/species invariant Brain extrAction), a strategy to train a model for brain extraction with no real brain images or labels. Our results show that even without any real images or anatomical priors, the model achieves comparable accuracy in multi-modal, multi-species and pathological cases. This work presents a new direction of research for any generalizable medical image segmentation task.
Authors: Yash Deo, Yan Jia, Toni Lassila, William A. P. Smith, Tom Lawton, Siyuan Kang, Alejandro F. Frangi, Ibrahim Habli
Abstract: Evaluating generative models for synthetic medical imaging is crucial yet challenging, especially given the high standards of fidelity, anatomical accuracy, and safety required for clinical applications. Standard evaluation of generated images often relies on no-reference image quality metrics when ground truth images are unavailable, but their reliability in this complex domain is not well established. This study comprehensively assesses commonly used no-reference image quality metrics using brain MRI data, including tumour and vascular images, providing a representative exemplar for the field. We systematically evaluate metric sensitivity to a range of challenges, including noise, distribution shifts, and, critically, localised morphological alterations designed to mimic clinically relevant inaccuracies. We then compare these metric scores against model performance on a relevant downstream segmentation task, analysing results across both controlled image perturbations and outputs from different generative model architectures. Our findings reveal significant limitations: many widely-used no-reference image quality metrics correlate poorly with downstream task suitability and exhibit a profound insensitivity to localised anatomical details crucial for clinical validity. Furthermore, these metrics can yield misleading scores regarding distribution shifts, e.g. data memorisation. This reveals the risk of misjudging model readiness, potentially leading to the deployment of flawed tools that could compromise patient safety. We conclude that ensuring generative models are truly fit for clinical purpose requires a multifaceted validation framework, integrating performance on relevant downstream tasks with the cautious interpretation of carefully selected no-reference image quality metrics.
Authors: Pascal Spiegler, Arash Harirpoush, Yiming Xiao
Abstract: Crucial in disease analysis and surgical planning, manual segmentation of volumetric medical scans (e.g. MRI, CT) is laborious, error-prone, and challenging to master, while fully automatic algorithms can benefit from user-feedback. Therefore, with the complementary power of the latest radiological AI foundation models and virtual reality (VR)'s intuitive data interaction, we propose SAMIRA, a novel conversational AI agent that assists users with localizing, segmenting, and visualizing 3D medical concepts in VR. Through speech-based interaction, the agent helps users understand radiological features, locate clinical targets, and generate segmentation masks that can be refined with just a few point prompts. The system also supports true-to-scale 3D visualization of segmented pathology to enhance patient-specific anatomical understanding. Furthermore, to determine the optimal interaction paradigm under near-far attention-switching for refining segmentation masks in an immersive, human-in-the-loop workflow, we compare VR controller pointing, head pointing, and eye tracking as input modes. With a user study, evaluations demonstrated a high usability score (SUS=90.0 $\pm$ 9.0), low overall task load, as well as strong support for the proposed VR system's guidance, training potential, and integration of AI in radiological segmentation tasks.
Authors: Badhan Kumar Das, Gengyan Zhao, Boris Mailhe, Thomas J. Re, Dorin Comaniciu, Eli Gibson, Andreas Maier
Abstract: Identifying brain hemorrhages from magnetic resonance imaging (MRI) is a critical task for healthcare professionals. The diverse nature of MRI acquisitions with varying contrasts and orientation introduce complexity in identifying hemorrhage using neural networks. For acquisitions with varying orientations, traditional methods often involve resampling images to a fixed plane, which can lead to information loss. To address this, we propose a 3D multi-plane vision transformer (MP-ViT) for hemorrhage classification with varying orientation data. It employs two separate transformer encoders for axial and sagittal contrasts, using cross-attention to integrate information across orientations. MP-ViT also includes a modality indication vector to provide missing contrast information to the model. The effectiveness of the proposed model is demonstrated with extensive experiments on real world clinical dataset consists of 10,084 training, 1,289 validation and 1,496 test subjects. MP-ViT achieved substantial improvement in area under the curve (AUC), outperforming the vision transformer (ViT) by 5.5% and CNN-based architectures by 1.8%. These results highlight the potential of MP-ViT in improving performance for hemorrhage detection when different orientation contrasts are needed.
Authors: Wenhao Hu, Paul Henderson, Jos\'e Cano
Abstract: Pruning is a widely used method for compressing Deep Neural Networks (DNNs), where less relevant parameters are removed from a DNN model to reduce its size. However, removing parameters reduces model accuracy, so pruning is typically combined with fine-tuning, and sometimes other operations such as rewinding weights, to recover accuracy. A common approach is to repeatedly prune and then fine-tune, with increasing amounts of model parameters being removed in each step. While straightforward to implement, pruning pipelines that follow this approach are computationally expensive due to the need for repeated fine-tuning. In this paper we propose ICE-Pruning, an iterative pruning pipeline for DNNs that significantly decreases the time required for pruning by reducing the overall cost of fine-tuning, while maintaining a similar accuracy to existing pruning pipelines. ICE-Pruning is based on three main components: i) an automatic mechanism to determine after which pruning steps fine-tuning should be performed; ii) a freezing strategy for faster fine-tuning in each pruning step; and iii) a custom pruning-aware learning rate scheduler to further improve the accuracy of each pruning step and reduce the overall time consumption. We also propose an efficient auto-tuning stage for the hyperparameters (e.g., freezing percentage) introduced by the three components. We evaluate ICE-Pruning on several DNN models and datasets, showing that it can accelerate pruning by up to 9.61x. Code is available at https://github.com/gicLAB/ICE-Pruning
Authors: Peng Sun, Yi Jiang, Tao Lin
Abstract: Recent advances in continuous generative models, including multi-step approaches like diffusion and flow-matching (typically requiring 8-1000 sampling steps) and few-step methods such as consistency models (typically 1-8 steps), have demonstrated impressive generative performance. However, existing work often treats these approaches as distinct paradigms, resulting in separate training and sampling methodologies. We introduce a unified framework for training, sampling, and analyzing these models. Our implementation, the Unified Continuous Generative Models Trainer and Sampler (UCGM-{T,S}), achieves state-of-the-art (SOTA) performance. For example, on ImageNet 256x256 using a 675M diffusion transformer, UCGM-T trains a multi-step model achieving 1.30 FID in 20 steps and a few-step model reaching 1.42 FID in just 2 steps. Additionally, applying UCGM-S to a pre-trained model (previously 1.26 FID at 250 steps) improves performance to 1.06 FID in only 40 steps. Code is available at: https://github.com/LINs-lab/UCGM.
Authors: Wei Li, Ming Hu, Guoan Wang, Lihao Liu, Kaijin Zhou, Junzhi Ning, Xin Guo, Zongyuan Ge, Lixu Gu, Junjun He
Abstract: In ophthalmic surgery, developing an AI system capable of interpreting surgical videos and predicting subsequent operations requires numerous ophthalmic surgical videos with high-quality annotations, which are difficult to collect due to privacy concerns and labor consumption. Text-guided video generation (T2V) emerges as a promising solution to overcome this issue by generating ophthalmic surgical videos based on surgeon instructions. In this paper, we present Ophora, a pioneering model that can generate ophthalmic surgical videos following natural language instructions. To construct Ophora, we first propose a Comprehensive Data Curation pipeline to convert narrative ophthalmic surgical videos into a large-scale, high-quality dataset comprising over 160K video-instruction pairs, Ophora-160K. Then, we propose a Progressive Video-Instruction Tuning scheme to transfer rich spatial-temporal knowledge from a T2V model pre-trained on natural video-text datasets for privacy-preserved ophthalmic surgical video generation based on Ophora-160K. Experiments on video quality evaluation via quantitative analysis and ophthalmologist feedback demonstrate that Ophora can generate realistic and reliable ophthalmic surgical videos based on surgeon instructions. We also validate the capability of Ophora for empowering downstream tasks of ophthalmic surgical workflow understanding. Code is available at https://github.com/mar-cry/Ophora.
Authors: Hongkun Dou, Zeyu Li, Xingyu Jiang, Hongjue Li, Lijun Yang, Wen Yao, Yue Deng
Abstract: Diffusion models (DMs) have recently demonstrated remarkable success in modeling large-scale data distributions. However, many downstream tasks require guiding the generated content based on specific differentiable metrics, typically necessitating backpropagation during the generation process. This approach is computationally expensive, as generating with DMs often demands tens to hundreds of recursive network calls, resulting in high memory usage and significant time consumption. In this paper, we propose a more efficient alternative that approaches the problem from the perspective of parallel denoising. We show that full backpropagation throughout the entire generation process is unnecessary. The downstream metrics can be optimized by retaining the computational graph of only one step during generation, thus providing a shortcut for gradient propagation. The resulting method, which we call Shortcut Diffusion Optimization (SDO), is generic, high-performance, and computationally lightweight, capable of optimizing all parameter types in diffusion sampling. We demonstrate the effectiveness of SDO on several real-world tasks, including controlling generation by optimizing latent and aligning the DMs by fine-tuning network parameters. Compared to full backpropagation, our approach reduces computational costs by $\sim 90\%$ while maintaining superior performance. Code is available at https://github.com/deng-ai-lab/SDO.
Authors: Lingkun Luo, Shiqiang Hu, Liming Chen
Abstract: Pseudo-labeling is a cornerstone of Unsupervised Domain Adaptation (UDA), yet the scarcity of High-Confidence Pseudo-Labeled Target Domain Samples (\textbf{hcpl-tds}) often leads to inaccurate cross-domain statistical alignment, causing DA failures. To address this challenge, we propose \textbf{N}oise \textbf{O}ptimized \textbf{C}onditional \textbf{D}iffusion for \textbf{D}omain \textbf{A}daptation (\textbf{NOCDDA}), which seamlessly integrates the generative capabilities of conditional diffusion models with the decision-making requirements of DA to achieve task-coupled optimization for efficient adaptation. For robust cross-domain consistency, we modify the DA classifier to align with the conditional diffusion classifier within a unified optimization framework, enabling forward training on noise-varying cross-domain samples. Furthermore, we argue that the conventional \( \mathcal{N}(\mathbf{0}, \mathbf{I}) \) initialization in diffusion models often generates class-confused hcpl-tds, compromising discriminative DA. To resolve this, we introduce a class-aware noise optimization strategy that refines sampling regions for reverse class-specific hcpl-tds generation, effectively enhancing cross-domain alignment. Extensive experiments across 5 benchmark datasets and 29 DA tasks demonstrate significant performance gains of \textbf{NOCDDA} over 31 state-of-the-art methods, validating its robustness and effectiveness.
Authors: Oriol Barbany, Adri\`a Colom\'e, Carme Torras
Abstract: Manipulating clothes is challenging due to their complex dynamics, high deformability, and frequent self-occlusions. Garments exhibit a nearly infinite number of configurations, making explicit state representations difficult to define. In this paper, we analyze BiFold, a model that predicts language-conditioned pick-and-place actions from visual observations, while implicitly encoding garment state through end-to-end learning. To address scenarios such as crumpled garments or recovery from failed manipulations, BiFold leverages temporal context to improve state estimation. We examine the internal representations of the model and present evidence that its fine-tuning and temporal context enable effective alignment between text and image regions, as well as temporal consistency.
Authors: Jian Liu, Xiongtao Shi, Thai Duy Nguyen, Haitian Zhang, Tianxiang Zhang, Wei Sun, Yanjie Li, Athanasios V. Vasilakos, Giovanni Iacca, Arshad Ali Khan, Arvind Kumar, Jae Won Cho, Ajmal Mian, Lihua Xie, Erik Cambria, Lin Wang
Abstract: The rapid evolution of artificial intelligence (AI) has shifted from static, data-driven models to dynamic systems capable of perceiving and interacting with real-world environments. Despite advancements in pattern recognition and symbolic reasoning, current AI systems, such as large language models, remain disembodied, unable to physically engage with the world. This limitation has driven the rise of embodied AI, where autonomous agents, such as humanoid robots, must navigate and manipulate unstructured environments with human-like adaptability. At the core of this challenge lies the concept of Neural Brain, a central intelligence system designed to drive embodied agents with human-like adaptability. A Neural Brain must seamlessly integrate multimodal sensing and perception with cognitive capabilities. Achieving this also requires an adaptive memory system and energy-efficient hardware-software co-design, enabling real-time action in dynamic environments. This paper introduces a unified framework for the Neural Brain of embodied agents, addressing two fundamental challenges: (1) defining the core components of Neural Brain and (2) bridging the gap between static AI models and the dynamic adaptability required for real-world deployment. To this end, we propose a biologically inspired architecture that integrates multimodal active sensing, perception-cognition-action function, neuroplasticity-based memory storage and updating, and neuromorphic hardware/software optimization. Furthermore, we also review the latest research on embodied agents across these four aspects and analyze the gap between current AI systems and human intelligence. By synthesizing insights from neuroscience, we outline a roadmap towards the development of generalizable, autonomous agents capable of human-level intelligence in real-world scenarios.
Authors: Pouya Afshin, David Helminiak, Tongtong Lu, Tina Yen, Julie M. Jorns, Mollie Patton, Bing Yu, Dong Hye Ye
Abstract: Breast-conserving surgery (BCS) aims to completely remove malignant lesions while maximizing healthy tissue preservation. Intraoperative margin assessment is essential to achieve a balance between thorough cancer resection and tissue conservation. A deep ultraviolet fluorescence scanning microscope (DUV-FSM) enables rapid acquisition of whole surface images (WSIs) for excised tissue, providing contrast between malignant and normal tissues. However, breast cancer classification with DUV WSIs is challenged by high resolutions and complex histopathological features. This study introduces a DUV WSI classification framework using a patch-level vision transformer (ViT) model, capturing local and global features. Grad-CAM++ saliency weighting highlights relevant spatial regions, enhances result interpretability, and improves diagnostic accuracy for benign and malignant tissue classification. A comprehensive 5-fold cross-validation demonstrates the proposed approach significantly outperforms conventional deep learning methods, achieving a classification accuracy of 98.33%.
Authors: Elad Yoshai, Dana Yagoda-Aharoni, Eden Dotan, Natan T. Shaked
Abstract: We present SparseAttnNet, a new hierarchical attention-driven framework for efficient image classification that adaptively selects and processes only the most informative pixels from images. Traditional convolutional neural networks typically process the entire images regardless of information density, leading to computational inefficiency and potential focus on irrelevant features. Our approach leverages a dynamic selection mechanism that uses coarse attention distilled by fine multi-head attention from the downstream layers of the model, allowing the model to identify and extract the most salient k pixels, where k is adaptively learned during training based on loss convergence trends. Once the top-k pixels are selected, the model processes only these pixels, embedding them as words in a language model to capture their semantics, followed by multi-head attention to incorporate global context. For biological cell images, we demonstrate that SparseAttnNet can process approximately 15% of the pixels instead of the full image. Applied to cell classification tasks using white blood cells images from the following modalities: optical path difference (OPD) images from digital holography for stain-free cells, images from motion-sensitive (event) camera from stain-free cells, and brightfield microscopy images of stained cells, For all three imaging modalities, SparseAttnNet achieves competitive accuracy while drastically reducing computational requirements in terms of both parameters and floating-point operations per second, compared to traditional CNNs and Vision Transformers. Since the model focuses on biologically relevant regions, it also offers improved explainability. The adaptive and lightweight nature of SparseAttnNet makes it ideal for deployment in resource-constrained and high-throughput settings, including imaging flow cytometry.
Authors: Seongjae Kang, Dong Bok Lee, Hyungjoon Jang, Sung Ju Hwang
Abstract: Vision-language models (VLMs) have achieved remarkable success across diverse tasks by leveraging rich textual information with minimal labeled data. However, deploying such large models remains challenging, particularly in resource-constrained environments. Knowledge distillation (KD) offers a well-established solution to this problem; however, recent KD approaches from VLMs often involve multi-stage training or additional tuning, increasing computational overhead and optimization complexity. In this paper, we propose $\mathbf{\texttt{D}}$ual-$\mathbf{\texttt{H}}$ead $\mathbf{\texttt{O}}$ptimization ($\mathbf{\texttt{DHO}}$) -- a simple yet effective KD framework that transfers knowledge from VLMs to compact, task-specific models in semi-supervised settings. Specifically, we introduce dual prediction heads that independently learn from labeled data and teacher predictions, and propose to linearly combine their outputs during inference. We observe that $\texttt{DHO}$ mitigates gradient conflicts between supervised and distillation signals, enabling more effective feature learning than single-head KD baselines. As a result, extensive experiments show that $\texttt{DHO}$ consistently outperforms baselines across multiple domains and fine-grained datasets. Notably, on ImageNet, it achieves state-of-the-art performance, improving accuracy by 3% and 0.1% with 1% and 10% labeled data, respectively, while using fewer parameters.
Authors: Feng Yuan, Yifan Gao, Wenbin Wu, Keqing Wu, Xiaotong Guo, Jie Jiang, Xin Gao
Abstract: Accurate multi-modal medical image translation requires ha-rmonizing global anatomical semantics and local structural fidelity, a challenge complicated by intermodality information loss and structural distortion. We propose ABS-Mamba, a novel architecture integrating the Segment Anything Model 2 (SAM2) for organ-aware semantic representation, specialized convolutional neural networks (CNNs) for preserving modality-specific edge and texture details, and Mamba's selective state-space modeling for efficient long- and short-range feature dependencies. Structurally, our dual-resolution framework leverages SAM2's image encoder to capture organ-scale semantics from high-resolution inputs, while a parallel CNNs branch extracts fine-grained local features. The Robust Feature Fusion Network (RFFN) integrates these epresentations, and the Bidirectional Mamba Residual Network (BMRN) models spatial dependencies using spiral scanning and bidirectional state-space dynamics. A three-stage skip fusion decoder enhances edge and texture fidelity. We employ Efficient Low-Rank Adaptation (LoRA+) fine-tuning to enable precise domain specialization while maintaining the foundational capabilities of the pre-trained components. Extensive experimental validation on the SynthRAD2023 and BraTS2019 datasets demonstrates that ABS-Mamba outperforms state-of-the-art methods, delivering high-fidelity cross-modal synthesis that preserves anatomical semantics and structural details to enhance diagnostic accuracy in clinical applications. The code is available at https://github.com/gatina-yone/ABS-Mamba
Authors: Samik Banerjee, Caleb Stam, Daniel J. Tward, Steven Savoia, Yusu Wang, Partha P. Mitra
Abstract: To understand biological intelligence we need to map neuronal networks in vertebrate brains. Mapping mesoscale neural circuitry is done using injections of tracers that label groups of neurons whose axons project to different brain regions. Since many neurons are labeled, it is difficult to follow individual axons. Previous approaches have instead quantified the regional projections using the total label intensity within a region. However, such a quantification is not biologically meaningful. We propose a new approach better connected to the underlying neurons by skeletonizing labeled axon fragments and then estimating a volumetric length density. Our approach uses a combination of deep nets and the Discrete Morse (DM) technique from computational topology. This technique takes into account nonlocal connectivity information and therefore provides noise-robustness. We demonstrate the utility and scalability of the approach on whole-brain tracer injected data. We also define and illustrate an information theoretic measure that quantifies the additional information obtained, compared to the skeletonized tracer injection fragments, when individual axon morphologies are available. Our approach is the first application of the DM technique to computational neuroanatomy. It can help bridge between single-axon skeletons and tracer injections, two important data types in mapping neural networks in vertebrates.
Authors: Xuying Huang, Sicong Pan, Maren Bennewitz
Abstract: User privacy is a crucial concern in robotic applications, especially when mobile service robots are deployed in personal or sensitive environments. However, many robotic downstream tasks require the use of cameras, which may raise privacy risks. To better understand user perceptions of privacy in relation to visual data, we conducted a user study investigating how different image modalities and image resolutions affect users' privacy concerns. The results show that depth images are broadly viewed as privacy-safe, and a similarly high proportion of respondents feel the same about semantic segmentation images. Additionally, the majority of participants consider 32*32 resolution RGB images to be almost sufficiently privacy-preserving, while most believe that 16*16 resolution can fully guarantee privacy protection.
Authors: Tony Tao, Mohan Kumar Srirama, Jason Jingzhou Liu, Kenneth Shaw, Deepak Pathak
Abstract: Large-scale, diverse robot datasets have emerged as a promising path toward enabling dexterous manipulation policies to generalize to novel environments, but acquiring such datasets presents many challenges. While teleoperation provides high-fidelity datasets, its high cost limits its scalability. Instead, what if people could use their own hands, just as they do in everyday life, to collect data? In DexWild, a diverse team of data collectors uses their hands to collect hours of interactions across a multitude of environments and objects. To record this data, we create DexWild-System, a low-cost, mobile, and easy-to-use device. The DexWild learning framework co-trains on both human and robot demonstrations, leading to improved performance compared to training on each dataset individually. This combination results in robust robot policies capable of generalizing to novel environments, tasks, and embodiments with minimal additional robot-specific data. Experimental results demonstrate that DexWild significantly improves performance, achieving a 68.5% success rate in unseen environments-nearly four times higher than policies trained with robot data only-and offering 5.8x better cross-embodiment generalization. Video results, codebases, and instructions at https://dexwild.github.io
Authors: Seungjae Lee, Daniel Ekpo, Haowen Liu, Furong Huang, Abhinav Shrivastava, Jia-Bin Huang
Abstract: Exploration is essential for general-purpose robotic learning, especially in open-ended environments where dense rewards, explicit goals, or task-specific supervision are scarce. Vision-language models (VLMs), with their semantic reasoning over objects, spatial relations, and potential outcomes, present a compelling foundation for generating high-level exploratory behaviors. However, their outputs are often ungrounded, making it difficult to determine whether imagined transitions are physically feasible or informative. To bridge the gap between imagination and execution, we present IVE (Imagine, Verify, Execute), an agentic exploration framework inspired by human curiosity. Human exploration is often driven by the desire to discover novel scene configurations and to deepen understanding of the environment. Similarly, IVE leverages VLMs to abstract RGB-D observations into semantic scene graphs, imagine novel scenes, predict their physical plausibility, and generate executable skill sequences through action tools. We evaluate IVE in both simulated and real-world tabletop environments. The results show that IVE enables more diverse and meaningful exploration than RL baselines, as evidenced by a 4.1 to 7.8x increase in the entropy of visited states. Moreover, the collected experience supports downstream learning, producing policies that closely match or exceed the performance of those trained on human-collected demonstrations.
Authors: Kanchana Ranasinghe, Xiang Li, Cristina Mata, Jongwoo Park, Michael S Ryoo
Abstract: We present LangToMo, a vision-language-action framework structured as a dual-system architecture that uses pixel motion forecasts as intermediate representations. Our high-level System 2, an image diffusion model, generates text-conditioned pixel motion sequences from a single frame to guide robot control. Pixel motion-a universal, interpretable, and motion-centric representation-can be extracted from videos in a self-supervised manner, enabling diffusion model training on web-scale video-caption data. Treating generated pixel motion as learned universal representations, our low level System 1 module translates these into robot actions via motion-to-action mapping functions, which can be either hand-crafted or learned with minimal supervision. System 2 operates as a high-level policy applied at sparse temporal intervals, while System 1 acts as a low-level policy at dense temporal intervals. This hierarchical decoupling enables flexible, scalable, and generalizable robot control under both unsupervised and supervised settings, bridging the gap between language, motion, and action. Checkout https://kahnchana.github.io/LangToMo for visualizations.
Authors: Yiyang Lu, Yufeng Tian, Zhecheng Yuan, Xianbang Wang, Pu Hua, Zhengrong Xue, Huazhe Xu
Abstract: Visuomotor policy learning has witnessed substantial progress in robotic manipulation, with recent approaches predominantly relying on generative models to model the action distribution. However, these methods often overlook the critical coupling between visual perception and action prediction. In this work, we introduce $\textbf{Triply-Hierarchical Diffusion Policy}~(\textbf{H$^{\mathbf{3}}$DP})$, a novel visuomotor learning framework that explicitly incorporates hierarchical structures to strengthen the integration between visual features and action generation. H$^{3}$DP contains $\mathbf{3}$ levels of hierarchy: (1) depth-aware input layering that organizes RGB-D observations based on depth information; (2) multi-scale visual representations that encode semantic features at varying levels of granularity; and (3) a hierarchically conditioned diffusion process that aligns the generation of coarse-to-fine actions with corresponding visual features. Extensive experiments demonstrate that H$^{3}$DP yields a $\mathbf{+27.5\%}$ average relative improvement over baselines across $\mathbf{44}$ simulation tasks and achieves superior performance in $\mathbf{4}$ challenging bimanual real-world manipulation tasks. Project Page: https://lyy-iiis.github.io/h3dp/.
Authors: Dongwei Wang, Zhi Han, Yanmei Wang, Xiai Chen, Baichen Liu, Yandong Tang
Abstract: Reviewing plays an important role when learning knowledge. The knowledge acquisition at a certain time point may be strongly inspired with the help of previous experience. Thus the knowledge growing procedure should show strong relationship along the temporal dimension. In our research, we find that during the network training, the evolution of feature map follows temporal sequence property. A proper temporal supervision may further improve the network training performance. Inspired by this observation, we propose Temporal Supervised Knowledge Distillation (TSKD). Specifically, we extract the spatiotemporal features in the different training phases of student by convolutional Long Short-term memory network (Conv-LSTM). Then, we train the student net through a dynamic target, rather than static teacher network features. This process realizes the refinement of old knowledge in student network, and utilizes it to assist current learning. Extensive experiments verify the effectiveness and advantages of our method over existing knowledge distillation methods, including various network architectures and different tasks (image classification and object detection) .
Authors: Hongyi Fan, Joe Kileel, Benjamin Kimia
Abstract: In this paper, we introduce a general framework for analyzing the numerical conditioning of minimal problems in multiple view geometry, using tools from computational algebra and Riemannian geometry. Special motivation comes from the fact that relative pose estimation, based on standard 5-point or 7-point Random Sample Consensus (RANSAC) algorithms, can fail even when no outliers are present and there is enough data to support a hypothesis. We argue that these cases arise due to the intrinsic instability of the 5- and 7-point minimal problems. We apply our framework to characterize the instabilities, both in terms of the world scenes that lead to infinite condition number, and directly in terms of ill-conditioned image data. The approach produces computational tests for assessing the condition number before solving the minimal problem. Lastly, synthetic and real data experiments suggest that RANSAC serves not only to remove outliers, but in practice it also selects for well-conditioned image data, which is consistent with our theory.
Authors: Andre Ye, Sebastin Santy, Jena D. Hwang, Amy X. Zhang, Ranjay Krishna
Abstract: Computer vision often treats human perception as homogeneous: an implicit assumption that visual stimuli are perceived similarly by everyone. This assumption is reflected in the way researchers collect datasets and train vision models. By contrast, literature in cross-cultural psychology and linguistics has provided evidence that people from different cultural backgrounds observe vastly different concepts even when viewing the same visual stimuli. In this paper, we study how these differences manifest themselves in vision-language datasets and models, using language as a proxy for culture. By comparing textual descriptions generated across 7 languages for the same images, we find significant differences in the semantic content and linguistic expression. When datasets are multilingual as opposed to monolingual, descriptions have higher semantic coverage on average, where coverage is measured using scene graphs, model embeddings, and linguistic taxonomies. For example, multilingual descriptions have on average 29.9% more objects, 24.5% more relations, and 46.0% more attributes than a set of monolingual captions. When prompted to describe images in different languages, popular models (e.g. LLaVA) inherit this bias and describe different parts of the image. Moreover, finetuning models on captions from one language performs best on corresponding test data from that language, while finetuning on multilingual data performs consistently well across all test data compositions. Our work points towards the need to account for and embrace the diversity of human perception in the computer vision community.
Authors: Kangfu Mei, Luis Figueroa, Zhe Lin, Zhihong Ding, Scott Cohen, Vishal M. Patel
Abstract: Recovering textures under shadows has remained a challenging problem due to the difficulty of inferring shadow-free scenes from shadow images. In this paper, we propose the use of diffusion models as they offer a promising approach to gradually refine the details of shadow regions during the diffusion process. Our method improves this process by conditioning on a learned latent feature space that inherits the characteristics of shadow-free images, thus avoiding the limitation of conventional methods that condition on degraded images only. Additionally, we propose to alleviate potential local optima during training by fusing noise features with the diffusion network. We demonstrate the effectiveness of our approach which outperforms the previous best method by 13% in terms of RMSE on the AISTD dataset. Further, we explore instance-level shadow removal, where our model outperforms the previous best method by 82% in terms of RMSE on the DESOBA dataset.
Authors: Tin Nguyen, Peijie Chen, Anh Totti Nguyen
Abstract: Traditional bird classifiers mostly rely on the visual characteristics of birds. Some prior works even train classifiers to be invariant to the background, completely discarding the living environment of birds. Instead, we are the first to explore integrating habitat information, one of the four major cues for identifying birds by ornithologists, into modern bird classifiers. We focus on two leading model types: (1) CNNs and ViTs trained on the downstream bird datasets; and (2) original, multi-modal CLIP. Training CNNs and ViTs with habitat-augmented data results in an improvement of up to +0.83 and +0.23 points on NABirds and CUB-200, respectively. Similarly, adding habitat descriptors to the prompts for CLIP yields a substantial accuracy boost of up to +0.99 and +1.1 points on NABirds and CUB-200, respectively. We find consistent accuracy improvement after integrating habitat features into the image augmentation process and into the textual descriptors of vision-language CLIP classifiers. Code is available at: https://anonymous.4open.science/r/reasoning-8B7E/.
Authors: Dong Zhang, Pingcheng Dong, Long Chen, Kwang-Ting Cheng
Abstract: It has been revealed that small efficient dense image prediction (EDIP) models, trained using the knowledge distillation (KD) framework, encounter two key challenges, including maintaining boundary region completeness and preserving target region connectivity, despite their favorable capacity to recognize main object regions. In this work, we propose a complementary boundary and context distillation (BCD) method within the KD framework for EDIPs, which facilitates the targeted knowledge transfer from large accurate teacher models to compact efficient student models. Specifically, the boundary distillation component focuses on extracting explicit object-level semantic boundaries from the hierarchical feature maps of the backbone network to enhance the student model's mask quality in boundary regions. Concurrently, the context distillation component leverages self-relations as a bridge to transfer implicit pixel-level contexts from the teacher model to the student model, ensuring strong connectivity in target regions. Our proposed BCD method is specifically designed for EDIP tasks and is characterized by its simplicity and efficiency. Extensive experimental results across semantic segmentation, object detection, and instance segmentation on various representative datasets demonstrate that our method can outperform existing methods without requiring extra supervisions or incurring increased inference costs, resulting in well-defined object boundaries and smooth connecting regions.
Authors: Chenguang Liu, Chisheng Wang, Feifei Dong, Xiayang Xiao, Xin Su, Chuanhua Zhu, Dejin Zhang, Qingquan Li
Abstract: Edge detection is a long-standing problem in computer vision. Despite the efficiency of existing algorithms, their performance, however, rely heavily on the pre-trained weights of the backbone network on the ImageNet dataset. The use of pre-trained weights in previous methods significantly increases the difficulty to design new models for edge detection without relying on existing well-trained ImageNet models, as pre-training the model on the ImageNet dataset is expensive and becomes compulsory to ensure the fairness of comparison. Besides, the pre-training and fine-tuning strategy is not always useful and sometimes even inaccessible. For instance, the pre-trained weights on the ImageNet dataset are unlikely to be helpful for edge detection in Synthetic Aperture Radar (SAR) images due to strong differences in the statistics between optical images and SAR images. Moreover, no dataset has comparable size to the ImageNet dataset for SAR image processing. In this work, we study the performance achievable by state-of-the-art deep learning based edge detectors in publicly available datasets when they are trained from scratch, and devise a new network architecture, the multi-stream and multi-scale fusion net (msmsfnet), for edge detection. We show in our experiments that by training all models from scratch, our model outperforms state-of-the-art edge detectors in three publicly available datasets. We also demonstrate the efficiency of our model for edge detection in SAR images, where no useful pre-trained weight is available. Finally, We show that our model is able to achieve competitive performance on the BSDS500 dataset when the pre-trained weights are used.
Authors: Zhenghong Li, Jiaxiang Ren, Wensheng Cheng, Congwu Du, Yingtian Pan, Haibin Ling
Abstract: Optical coherence Doppler tomography (ODT) is an emerging blood flow imaging technique. The fundamental unit of ODT is the 1D depth-resolved trace named raw A-scans (or A-line). A 2D ODT image (B-scan) is formed by reconstructing a cross-sectional flow image via Doppler phase-subtraction of raw A-scans along B-line. To obtain a high-fidelity B-scan, densely sampled A-scans are required currently, leading to prolonged scanning time and increased storage demands. Addressing this issue, we propose a novel sparse ODT reconstruction framework with an Alternative State Space Attention Network (ASSAN) that effectively reduces raw A-scans needed. Inspired by the distinct distributions of information along A-line and B-line, ASSAN applies 1D State Space Model (SSM) to each A-line to learn the intra-A-scan representation, while using 1D gated self-attention along B-line to capture the inter-A-scan features. In addition, an effective feedforward network based on sequential 1D convolutions along different axes is employed to enhance the local feature. In validation experiments on real animal data, ASSAN shows clear effectiveness in the reconstruction in comparison with state-of-the-art reconstruction methods.
Authors: Junzhuo Chen, Shitong Kang
Abstract: Large unlabeled data and difficult-to-identify anomalies are the urgent issues need to overcome in most industrial scene. In order to address this issue, a new meth-odology for detecting surface defects in in-dustrial settings is introduced, referred to as Memory Augmentation and Pseudo-Labeling(MAPL). The methodology first in-troduces an anomaly simulation strategy, which significantly improves the model's ability to recognize rare or unknown anom-aly types by generating simulated anomaly samples. To cope with the problem of the lack of labeling of anomalous simulated samples, a pseudo-labeler method based on a one-classifier ensemble was employed in this study, which enhances the robustness of the model in the case of limited labeling data by automatically selecting key pseudo-labeling hyperparameters. Meanwhile, a memory-enhanced learning mechanism is introduced to effectively predict abnormal regions by analyzing the difference be-tween the input samples and the normal samples in the memory pool. An end-to-end learning framework is employed by MAPL to identify the abnormal regions directly from the input data, which optimizes the ef-ficiency and real-time performance of de-tection. By conducting extensive trials on the recently developed BHAD dataset (in-cluding MVTec AD [1], Visa [2], and MDPP [3]), MAPL achieves an average im-age-level AUROC score of 86.2%, demon-strating a 5.1% enhancement compared to the original MemSeg [4] model. The source code is available at https://github.com/jzc777/MAPL.
Authors: Ling-Qi Zhang, Zahra Kadkhodaie, Eero P. Simoncelli, David H. Brainard
Abstract: We examine the problem of selecting a small set of linear measurements for reconstructing high-dimensional signals. Well-established methods for optimizing such measurements include principal component analysis (PCA), independent component analysis (ICA) and compressed sensing (CS) based on random projections, all of which rely on axis- or subspace-aligned statistical characterization of the signal source. However, many naturally occurring signals, including photographic images, contain richer statistical structure. To exploit such structure, we introduce a general method for obtaining an optimized set of linear measurements for efficient image reconstruction, where the signal statistics are expressed by the prior implicit in a neural network trained to perform denoising (known as a ``diffusion model''). We demonstrate that the optimal measurements derived for two natural image datasets differ from those of PCA, ICA, or CS, and result in substantially lower mean squared reconstruction error. Interestingly, the marginal distributions of the measurement values are asymmetrical (skewed), substantially more so than those of previous methods. We also find that optimizing with respect to perceptual loss, as quantified by structural similarity (SSIM), leads to measurements different from those obtained when optimizing for MSE. Our results highlight the importance of incorporating the specific statistical regularities of natural signals when designing effective linear measurements.
Authors: Jiangbei Hu, Yanggeng Li, Fei Hou, Junhui Hou, Zhebin Zhang, Shengfa Wang, Na Lei, Ying He
Abstract: Unsigned distance fields (UDFs) provide a versatile framework for representing a diverse array of 3D shapes, encompassing both watertight and non-watertight geometries. Traditional UDF learning methods typically require extensive training on large 3D shape datasets, which is costly and necessitates re-training for new datasets. This paper presents a novel neural framework, LoSF-UDF, for reconstructing surfaces from 3D point clouds by leveraging local shape functions to learn UDFs. We observe that 3D shapes manifest simple patterns in localized regions, prompting us to develop a training dataset of point cloud patches characterized by mathematical functions that represent a continuum from smooth surfaces to sharp edges and corners. Our approach learns features within a specific radius around each query point and utilizes an attention mechanism to focus on the crucial features for UDF estimation. Despite being highly lightweight, with only 653 KB of trainable parameters and a modest-sized training dataset with 0.5 GB storage, our method enables efficient and robust surface reconstruction from point clouds without requiring for shape-specific training. Furthermore, our method exhibits enhanced resilience to noise and outliers in point clouds compared to existing methods. We conduct comprehensive experiments and comparisons across various datasets, including synthetic and real-scanned point clouds, to validate our method's efficacy. Notably, our lightweight framework offers rapid and reliable initialization for other unsupervised iterative approaches, improving both the efficiency and accuracy of their reconstructions. Our project and code are available at https://jbhu67.github.io/LoSF-UDF.github.io.
Authors: Shuchang Zhang, Hui Zhang, Hongxia Wang
Abstract: Recently, the degenerate preconditioned proximal point (PPP) method provides a unified and flexible framework for designing and analyzing operator-splitting algorithms such as Douglas-Rachford (DR). However, the degenerate PPP method exhibits weak convergence in the infinite-dimensional Hilbert space and lacks accelerated variants. To address these issues, we propose a Halpern-type PPP (HPPP) algorithm, which leverages the strong convergence and acceleration properties of Halpern's iteration method. Moreover, we propose a novel algorithm for image restoration by combining HPPP with denoiser priors such as Plug-and-Play (PnP) prior, which can be viewed as an accelerated PnP method. Finally, numerical experiments including several toy examples and image restoration validate the effectiveness of our proposed algorithms.
Authors: Peng Hao, Weilong Wang, Xiaobing Wang, Yingying Jiang, Hanchao Jia, Shaowei Cui, Junhang Wei, Xiaoshuai Hao
Abstract: Scene Graph Generation (SGG) remains a challenging task due to its compositional property. Previous approaches improve prediction efficiency through end-to-end learning. However, these methods exhibit limited performance as they assume unidirectional conditioning between entities and predicates, which restricts effective information interaction. To address this limitation, we propose a novel bidirectional conditioning factorization in a semantic-aligned space for SGG, enabling efficient and generalizable interaction between entities and predicates. Specifically, we introduce an end-to-end scene graph generation model, the Bidirectional Conditioning Transformer (BCTR), to implement this factorization. BCTR consists of two key modules. First, the Bidirectional Conditioning Generator (BCG) performs multi-stage interactive feature augmentation between entities and predicates, enabling mutual enhancement between these predictions. Second, Random Feature Alignment (RFA) is present to regularize feature space by distilling multi-modal knowledge from pre-trained models. Within this regularized feature space, BCG is feasible to capture interaction patterns across diverse relationships during training, and the learned interaction patterns can generalize to unseen but semantically related relationships during inference. Extensive experiments on Visual Genome and Open Image V6 show that BCTR achieves state-of-the-art performance on both benchmarks.
Authors: Hanzheng Wang, Wei Li, Xiang-Gen Xia, Qian Du
Abstract: Hyperspectral object tracking (HOT) has exhibited potential in various applications, particularly in scenes where objects are camouflaged. Existing trackers can effectively retrieve objects via band regrouping because of the bias in existing HOT datasets, where most objects tend to have distinguishing visual appearances rather than spectral characteristics. This bias allows the tracker to directly use the visual features obtained from the false-color images generated by hyperspectral images without the need to extract spectral features. To tackle this bias, we find that the tracker should focus on the spectral information when object appearance is unreliable. Thus, we provide a new task called hyperspectral camouflaged object tracking (HCOT) and meticulously construct a large-scale HCOT dataset, termed BihoT, which consists of 41,912 hyperspectral images covering 49 video sequences. The dataset covers various artificial camouflage scenes where objects have similar appearances, diverse spectrums, and frequent occlusion, making it a very challenging dataset for HCOT. Besides, a simple but effective baseline model, named spectral prompt-based distractor-aware network (SPDAN), is proposed, comprising a spectral embedding network (SEN), a spectral prompt-based backbone network (SPBN), and a distractor-aware module (DAM). Specifically, the SEN extracts spectral-spatial features via 3-D and 2-D convolutions. Then, the SPBN fine-tunes powerful RGB trackers with spectral prompts and alleviates the insufficiency of training samples. Moreover, the DAM utilizes a novel statistic to capture the distractor caused by occlusion from objects and background. Extensive experiments demonstrate that our proposed SPDAN achieves state-of-the-art performance on the proposed BihoT and other HOT datasets.
Authors: Xiaoyu Guo, Pengzhi Zhong, Hao Zhang, Defeng Huang, Huikai Shao, Qijun Zhao, Shuiwang Li
Abstract: Visual tracking has seen remarkable advancements, largely driven by the availability of large-scale training datasets that have enabled the development of highly accurate and robust algorithms. While significant progress has been made in tracking general objects, research on more challenging scenarios, such as tracking camouflaged objects, remains limited. Camouflaged objects, which blend seamlessly with their surroundings or other objects, present unique challenges for detection and tracking in complex environments. This challenge is particularly critical in applications such as military, security, agriculture, and marine monitoring, where precise tracking of camouflaged objects is essential. To address this gap, we introduce the Camouflaged Object Tracking Dataset (COTD), a specialized benchmark designed specifically for evaluating camouflaged object tracking methods. The COTD dataset comprises 200 sequences and approximately 80,000 frames, each annotated with detailed bounding boxes. Our evaluation of 20 existing tracking algorithms reveals significant deficiencies in their performance with camouflaged objects. To address these issues, we propose a novel tracking framework, HiPTrack-MLS, which demonstrates promising results in improving tracking performance for camouflaged objects. COTD and code are avialable at https://github.com/openat25/HIPTrack-MLS.
Authors: Shahab Aldin Shojaeezadeh, Abdelrazek Elnashar, Tobias Karl David Weber
Abstract: Crop phenology describes the physiological development stages of crops from planting to harvest which is valuable information for decision makers to plan and adapt agricultural management strategies. In the era of big Earth observation data ubiquity, attempts have been made to accurately detect crop phenology using Remote Sensing (RS) and high resolution weather data. However, most studies have focused on large scale predictions of phenology or developed methods which are not adequate to help crop modeler communities on leveraging Sentinel-1 and Sentinal-2 data and fusing them with high resolution climate data, using a novel framework. For this, we trained a Machine Learning (ML) LightGBM model to predict 13 phenological stages for eight major crops across Germany at 20 m scale. Observed phonologies were taken from German national phenology network (German Meteorological Service; DWD) between 2017 and 2021. We proposed a thorough feature selection analysis to find the best combination of RS and climate data to detect phenological stages. At national scale, predicted phenology resulted in a reasonable precision of R2 > 0.43 and a low Mean Absolute Error of 6 days, averaged over all phenological stages and crops. The spatio-temporal analysis of the model predictions demonstrates its transferability across different spatial and temporal context of Germany. The results indicated that combining radar sensors with climate data yields a very promising performance for a multitude of practical applications. Moreover, these improvements are expected to be useful to generate highly valuable input for crop model calibrations and evaluations, facilitate informed agricultural decisions, and contribute to sustainable food production to address the increasing global food demand.
Authors: Gangwei Xu, Xianqi Wang, Zhaoxing Zhang, Junda Cheng, Chunyuan Liao, Xin Yang
Abstract: Stereo matching is a core component in many computer vision and robotics systems. Despite significant advances over the last decade, handling matching ambiguities in ill-posed regions and large disparities remains an open challenge. In this paper, we propose a new deep network architecture, called IGEV++, for stereo matching. The proposed IGEV++ constructs Multi-range Geometry Encoding Volumes (MGEV), which encode coarse-grained geometry information for ill-posed regions and large disparities, while preserving fine-grained geometry information for details and small disparities. To construct MGEV, we introduce an adaptive patch matching module that efficiently and effectively computes matching costs for large disparity ranges and/or ill-posed regions. We further propose a selective geometry feature fusion module to adaptively fuse multi-range and multi-granularity geometry features in MGEV. Then, we input the fused geometry features into ConvGRUs to iteratively update the disparity map. MGEV allows to efficiently handle large disparities and ill-posed regions, such as occlusions and textureless regions, and enjoys rapid convergence during iterations. Our IGEV++ achieves the best performance on the Scene Flow test set across all disparity ranges, up to 768px. Our IGEV++ also achieves state-of-the-art accuracy on the Middlebury, ETH3D, KITTI 2012, and 2015 benchmarks. Specifically, IGEV++ achieves a 3.23\% 2-pixel outlier rate (Bad 2.0) on the large disparity benchmark, Middlebury, representing error reductions of 31.9\% and 54.8\% compared to RAFT-Stereo and GMStereo, respectively. We also present a real-time version of IGEV++ that achieves the best performance among all published real-time methods on the KITTI benchmarks. The code is publicly available at https://github.com/gangweix/IGEV and https://github.com/gangweix/IGEV-plusplus.
URLs: https://github.com/gangweix/IGEV, https://github.com/gangweix/IGEV-plusplus.
Authors: Quang-Huy Che, Duc-Tri Le, Bich-Nga Pham, Duc-Khai Lam, Vinh-Tiep Nguyen
Abstract: Data augmentation is crucial for pixel-wise annotation tasks like semantic segmentation, where labeling requires significant effort and intensive labor. Traditional methods, involving simple transformations such as rotations and flips, create new images but often lack diversity along key semantic dimensions and fail to alter high-level semantic properties. To address this issue, generative models have emerged as an effective solution for augmenting data by generating synthetic images. Controllable Generative models offer data augmentation methods for semantic segmentation tasks by using prompts and visual references from the original image. However, these models face challenges in generating synthetic images that accurately reflect the content and structure of the original image due to difficulties in creating effective prompts and visual references. In this work, we introduce an effective data augmentation pipeline for semantic segmentation using Controllable Diffusion model. Our proposed method includes efficient prompt generation using \textit{Class-Prompt Appending} and \textit{Visual Prior Blending} to enhance attention to labeled classes in real images, allowing the pipeline to generate a precise number of augmented images while preserving the structure of segmentation-labeled classes. In addition, we implement a \textit{class balancing algorithm} to ensure a balanced training dataset when merging the synthetic and original images. Evaluation on PASCAL VOC datasets, our pipeline demonstrates its effectiveness in generating high-quality synthetic images for semantic segmentation. Our code is available at \href{https://github.com/chequanghuy/Enhanced-Generative-Data-Augmentation-for-Semantic-Segmentation-via-Stronger-Guidance}{this https URL}.
Authors: Blaine Hoak, Patrick McDaniel
Abstract: The influence of textures on machine learning models has been an ongoing investigation, specifically in texture bias/learning, interpretability, and robustness. However, due to the lack of large and diverse texture data available, the findings in these works have been limited, as more comprehensive evaluations have not been feasible. Image generative models are able to provide data creation at scale, but utilizing these models for texture synthesis has been unexplored and poses additional challenges both in creating accurate texture images and validating those images. In this work, we introduce an extensible methodology and corresponding new dataset for generating high-quality, diverse texture images capable of supporting a broad set of texture-based tasks. Our pipeline consists of: (1) developing prompts from a range of descriptors to serve as input to text-to-image models, (2) adopting and adapting Stable Diffusion pipelines to generate and filter the corresponding images, and (3) further filtering down to the highest quality images. Through this, we create the Prompted Textures Dataset (PTD), a dataset of 246,285 texture images that span 56 textures. During the process of generating images, we find that NSFW safety filters in image generation pipelines are highly sensitive to texture (and flag up to 60\% of our texture images), uncovering a potential bias in these models and presenting unique challenges when working with texture data. Through both standard metrics and a human evaluation, we find that our dataset is high quality and diverse. Our dataset is available for download at https://zenodo.org/records/15359142.
Authors: Yuli Zhou, Guolei Sun, Yawei Li, Guo-Sen Xie, Luca Benini, Ender Konukoglu
Abstract: This study investigates the application and performance of the Segment Anything Model 2 (SAM2) in the challenging task of video camouflaged object segmentation (VCOS). VCOS involves detecting objects that blend seamlessly in the surroundings for videos, due to similar colors and textures, poor light conditions, etc. Compared to the objects in normal scenes, camouflaged objects are much more difficult to detect. SAM2, a video foundation model, has shown potential in various tasks. But its effectiveness in dynamic camouflaged scenarios remains under-explored. This study presents a comprehensive study on SAM2's ability in VCOS. First, we assess SAM2's performance on camouflaged video datasets using different models and prompts (click, box, and mask). Second, we explore the integration of SAM2 with existing multimodal large language models (MLLMs) and VCOS methods. Third, we specifically adapt SAM2 by fine-tuning it on the video camouflaged dataset. Our comprehensive experiments demonstrate that SAM2 has excellent zero-shot ability of detecting camouflaged objects in videos. We also show that this ability could be further improved by specifically adjusting SAM2's parameters for VCOS. The code is available at https://github.com/zhoustan/SAM2-VCOS
Authors: Chen Xie, Ciyun Lin, Xiaoyu Zheng, Bowen Gong, Antonio M. L\'opez
Abstract: Understanding and predicting pedestrian crossing behavioral intention is crucial for the driving safety of autonomous vehicles. Nonetheless, challenges emerge when using promising images or environmental context masks to extract various factors for time-series network modeling, causing pre-processing errors or a loss of efficiency. Typically, pedestrian positions captured by onboard cameras are often distorted and do not accurately reflect their actual movements. To address these issues, GTransPDM -- a Graph-embedded Transformer with a Position Decoupling Module -- was developed for pedestrian crossing intention prediction by leveraging multi-modal features. First, a positional decoupling module was proposed to decompose pedestrian lateral motion and encode depth cues in the image view. Then, a graph-embedded Transformer was designed to capture the spatio-temporal dynamics of human pose skeletons, integrating essential factors such as position, skeleton, and ego-vehicle motion. Experimental results indicate that the proposed method achieves 92% accuracy on the PIE dataset and 87% accuracy on the JAAD dataset, with a processing speed of 0.05ms. It outperforms the state-of-the-art in comparison.
Authors: Ye Wang, Sipeng Zheng, Bin Cao, Qianshan Wei, Weishuai Zeng, Qin Jin, Zongqing Lu
Abstract: Inspired by the recent success of LLMs, the field of human motion understanding has increasingly shifted toward developing large motion models. Despite some progress, current efforts remain far from achieving truly generalist models, primarily due to the lack of massive high-quality data. To address this gap, we present MotionLib, the first million-level dataset for motion generation, which is at least 15$\times$ larger than existing counterparts and enriched with hierarchical text descriptions. Using MotionLib, we train a large motion model named Being-M0, demonstrating robust performance across a wide range of human activities, including unseen ones. Through systematic investigation, for the first time, we highlight the importance of scaling both data and model size for advancing motion generation, along with key insights to achieve this goal. To better integrate the motion modality, we propose Motionbook, an innovative motion encoding approach including (1) a compact yet lossless feature to represent motions; (2) a novel 2D lookup-free motion tokenizer that preserves fine-grained motion details while expanding codebook capacity, significantly enhancing the representational power of motion tokens. We believe this work lays the groundwork for developing more versatile and powerful motion generation models in the future. For further details, visit https://github.com/BeingBeyond/Being-M0.
Authors: Zezhun Shi
Abstract: Camera calibration is fundamental to 3D vision, and the choice of calibration pattern greatly affects the accuracy. To address aberration issue, star-shaped pattern has been proposed as alternatives to traditional checkerboard. However, such pattern suffers from aliasing artifacts. In this paper, we present a novel solution by employing a series of checkerboard patterns rotated around a central point instead of a single star-shaped pattern. We further propose a complete feature extraction algorithm tailored for this design. Experimental results demonstrate that our approach offers improved accuracy over the conventional star-shaped pattern and achieves high stability across varying exposure levels.
Authors: Zohreh Aghababaeyan, Manel Abdellatif, Lionel Briand, Ramesh S
Abstract: Deep Neural Networks (DNNs) are increasingly deployed across applications. However, ensuring their reliability remains a challenge, and in many situations, alternative models with similar functionality and accuracy are available. Traditional accuracy-based evaluations often fail to capture behavioral differences between models, especially with limited test datasets, making it difficult to select or combine models effectively. Differential testing addresses this by generating test inputs that expose discrepancies in DNN model behavior. However, existing approaches face significant limitations: many rely on model internals or are constrained by available seed inputs. To address these challenges, we propose DiffGAN, a black-box test image generation approach for differential testing of DNN models. DiffGAN leverages a Generative Adversarial Network (GAN) and the Non-dominated Sorting Genetic Algorithm II to generate diverse and valid triggering inputs that reveal behavioral discrepancies between models. DiffGAN employs two custom fitness functions, focusing on diversity and divergence, to guide the exploration of the GAN input space and identify discrepancies between models' outputs. By strategically searching this space, DiffGAN generates inputs with specific features that trigger differences in model behavior. DiffGAN is black-box, making it applicable in more situations. We evaluate DiffGAN on eight DNN model pairs trained on widely used image datasets. Our results show DiffGAN significantly outperforms a SOTA baseline, generating four times more triggering inputs, with greater diversity and validity, within the same budget. Additionally, the generated inputs improve the accuracy of a machine learning-based model selection mechanism, which selects the best-performing model based on input characteristics and can serve as a smart output voting mechanism when using alternative models.
Authors: Andr\'es Mu\~noz, Nancy Thomas, Annita Vapsi, Daniel Borrajo
Abstract: Many industrial and service sectors require tools to extract vehicle characteristics from images. This is a complex task not only by the variety of noise, and large number of classes, but also by the constant introduction of new vehicle models to the market. In this paper, we present Veri-Car, an information retrieval integrated approach designed to help on this task. It leverages supervised learning techniques to accurately identify the make, type, model, year, color, and license plate of cars. The approach also addresses the challenge of handling open-world problems, where new car models and variations frequently emerge, by employing a sophisticated combination of pre-trained models, and a hierarchical multi-similarity loss. Veri-Car demonstrates robust performance, achieving high precision and accuracy in classifying both seen and unseen data. Additionally, it integrates an ensemble license plate detection, and an OCR model to extract license plate numbers with impressive accuracy.
Authors: Tianyu Sun, Jianhao Li, Xueqian Zhang, Zhongdao Wang, Bailan Feng, Hengshuang Zhao
Abstract: This paper studies point cloud perception within outdoor environments. Existing methods face limitations in recognizing objects located at a distance or occluded, due to the sparse nature of outdoor point clouds. In this work, we observe a significant mitigation of this problem by accumulating multiple temporally consecutive point cloud sweeps, resulting in a remarkable improvement in perception accuracy. However, the computation cost also increases, hindering previous approaches from utilizing a large number of point cloud sweeps. To tackle this challenge, we find that a considerable portion of points in the accumulated point cloud is redundant, and discarding these points has minimal impact on perception accuracy. We introduce a simple yet effective Gumbel Spatial Pruning (GSP) layer that dynamically prunes points based on a learned end-to-end sampling. The GSP layer is decoupled from other network components and thus can be seamlessly integrated into existing point cloud network architectures. Without incurring additional computational overhead, we increase the number of point cloud sweeps from 10, a common practice, to as many as 40. Consequently, there is a significant enhancement in perception performance. For instance, in nuScenes 3D object detection and BEV map segmentation tasks, our pruning strategy improves several 3D perception baseline methods.
Authors: Kunwei Lv, Ruobing Wu, Suyang Chen, Ping Lan
Abstract: Fire incidents in urban and forested areas pose serious threats,underscoring the need for more effective detection technologies. To address these challenges, we present CCi-YOLOv8n, an enhanced YOLOv8 model with targeted improvements for detecting small fires and smoke. The model integrates the CARAFE up-sampling operator and a context-guided module to reduce information loss during up-sampling and down-sampling, thereby retaining richer feature representations. Additionally, an inverted residual mobile block enhanced C2f module captures small targets and fine smoke patterns, a critical improvement over the original model's detection capacity.For validation, we introduce Web-Fire, a dataset curated for fire and smoke detection across diverse real-world scenarios. Experimental results indicate that CCi-YOLOv8n outperforms YOLOv8n in detection precision, confirming its effectiveness for robust fire detection tasks.
Authors: Ke Zhang, Zhaoye Zheng, Yurong Guo, Jiacun Wang, Jiyuan Yang, Yangjie Xiao
Abstract: Unmanned aerial vehicle (UAV) patrol inspection has emerged as a predominant approach in transmission line monitoring owing to its cost-effectiveness. Detecting defects in transmission lines is a critical task during UAV patrol inspection. However, due to imaging distance and shooting angles, UAV patrol images often suffer from insufficient defect-related visual information, which has an adverse effect on detection accuracy. In this article, we propose a novel method for detecting defects in UAV patrol images, which is based on vision-language pretraining for transmission line (VLP-TL) and a progressive transfer strategy (PTS). Specifically, VLP-TL contains two novel pretraining tasks tailored for the transmission line scenario, aimimg at pretraining an image encoder with abundant knowledge acquired from both visual and linguistic information. Transferring the pretrained image encoder to the defect detector as its backbone can effectively alleviate the insufficient visual information problem. In addition, the PTS further improves transfer performance by progressively bridging the gap between pretraining and downstream defection detection. Experimental results demonstrate that the proposed method significantly improves defect detection accuracy by jointly utilizing multimodal information, overcoming the limitations of insufficient defect-related visual information provided by UAV patrol images.
Authors: Yue Zhou, Mengcheng Lan, Xiang Li, Litong Feng, Yiping Ke, Xue Jiang, Qingyun Li, Xue Yang, Wayne Zhang
Abstract: Remote sensing (RS) visual grounding aims to use natural language expression to locate specific objects (in the form of the bounding box or segmentation mask) in RS images, enhancing human interaction with intelligent RS interpretation systems. Early research in this area was primarily based on horizontal bounding boxes (HBBs), but as more diverse RS datasets have become available, tasks involving oriented bounding boxes (OBBs) and segmentation masks have emerged. In practical applications, different targets require different grounding types: HBB can localize an object's position, OBB provides its orientation, and mask depicts its shape. However, existing specialized methods are typically tailored to a single type of RS visual grounding task and are hard to generalize across tasks. In contrast, large vision-language models (VLMs) exhibit powerful multi-task learning capabilities but struggle to handle dense prediction tasks like segmentation. This paper proposes GeoGround, a novel framework that unifies support for HBB, OBB, and mask RS visual grounding tasks, allowing flexible output selection. Rather than customizing the architecture of VLM, our work aims to elegantly support pixel-level visual grounding output through the Text-Mask technique. We define prompt-assisted and geometry-guided learning to enhance consistency across different signals. Experimental results show that GeoGround demonstrates strong performance across four RS visual grounding tasks, matching the performance of specialized methods on multiple benchmarks. Code available at https://github.com/zytx121/GeoGround
Authors: Nitish Shukla, Arun Ross
Abstract: A facial morph is an image created by combining two face images pertaining to two distinct identities. Face demorphing inverts the process and tries to recover the original images constituting a facial morph. While morph attack detection (MAD) techniques can be used to flag morph images, they do not divulge any visual information about the faces used to create them. Demorphing helps address this problem. Existing demorphing techniques are either very restrictive (assume identities during testing) or produce feeble outputs (both outputs look very similar). In this paper, we overcome these issues by proposing dc-GAN, a novel GAN-based demorphing method conditioned on the morph images. Our method overcomes morph-replication and produces high quality reconstructions of the bonafide images used to create the morphs. Moreover, our method is highly generalizable across demorphing paradigms (differential/reference-free). We conduct experiments on AMSL, FRLL-Morphs and MorDiff datasets to showcase the efficacy of our method.
Authors: Hui Ren, Joanna Materzynska, Rohit Gandikota, David Bau, Antonio Torralba
Abstract: We explore whether pre-training on datasets with paintings is necessary for a model to learn an artistic style with only a few examples. To investigate this, we train a text-to-image model exclusively on photographs, without access to any painting-related content. We show that it is possible to adapt a model that is trained without paintings to an artistic style, given only few examples. User studies and automatic evaluations confirm that our model (post-adaptation) performs on par with state-of-the-art models trained on massive datasets that contain artistic content like paintings, drawings or illustrations. Finally, using data attribution techniques, we analyze how both artistic and non-artistic datasets contribute to generating artistic-style images. Surprisingly, our findings suggest that high-quality artistic outputs can be achieved without prior exposure to artistic data, indicating that artistic style generation can occur in a controlled, opt-in manner using only a limited, carefully selected set of training examples.
Authors: You Wu, Xiangyang Yang, Xucheng Wang, Hengzhou Ye, Dan Zeng, Shuiwang Li
Abstract: Harnessing low-light enhancement and domain adaptation, nighttime UAV tracking has made substantial strides. However, over-reliance on image enhancement, limited high-quality nighttime data, and a lack of integration between daytime and nighttime trackers hinder the development of an end-to-end trainable framework. Additionally, current ViT-based trackers demand heavy computational resources due to their reliance on the self-attention mechanism. In this paper, we propose a novel pure Mamba-based tracking framework (MambaNUT) that employs a state space model with linear complexity as its backbone, incorporating a single-stream architecture that integrates feature learning and template-search coupling within Vision Mamba. We introduce an adaptive curriculum learning (ACL) approach that dynamically adjusts sampling strategies and loss weights, thereby improving the model's ability of generalization. Our ACL is composed of two levels of curriculum schedulers: (1) sampling scheduler that transforms the data distribution from imbalanced to balanced, as well as from easier (daytime) to harder (nighttime) samples; (2) loss scheduler that dynamically assigns weights based on the size of the training set and IoU of individual instances. Exhaustive experiments on multiple nighttime UAV tracking benchmarks demonstrate that the proposed MambaNUT achieves state-of-the-art performance while requiring lower computational costs. The code will be available at https://github.com/wuyou3474/MambaNUT.
Authors: Cong Xie, Han Zou, Ruiqi Yu, Yan Zhang, Zhenpeng Zhan
Abstract: In this work, we are interested in achieving both high text controllability and whole-body appearance consistency in the generation of personalized human characters. We propose a novel framework, named SerialGen, which is a serial generation method consisting of two stages: first, a standardization stage that standardizes reference images, and then a personalized generation stage based on the standardized reference. Furthermore, we introduce two modules aimed at enhancing the standardization process. Our experimental results validate the proposed framework's ability to produce personalized images that faithfully recover the reference image's whole-body appearance while accurately responding to a wide range of text prompts. Through thorough analysis, we highlight the critical contribution of the proposed serial generation method and standardization model, evidencing enhancements in appearance consistency between reference and output images and across serial outputs generated from diverse text prompts. The term "Serial" in this work carries a double meaning: it refers to the two-stage method and also underlines our ability to generate serial images with consistent appearance throughout.
Authors: Qizhe Zhang, Aosong Cheng, Ming Lu, Renrui Zhang, Zhiyong Zhuo, Jiajun Cao, Shaobo Guo, Qi She, Shanghang Zhang
Abstract: Large vision-language models (LVLMs) generally contain significantly more visual tokens than their textual counterparts, resulting in a considerable computational burden. Recent efforts have been made to tackle this issue by pruning visual tokens early within the language model. Most existing works use attention scores between text and visual tokens to assess the importance of visual tokens. However, in this study, we first analyze the text-visual attention in the language model and find that this score is not an ideal indicator for token pruning. Based on the analysis, We propose VisPruner, a plug-and-play method that utilizes visual cues for more effective token pruning in LVLMs. Specifically, we first use visual attention to select a limited number of significant tokens. Then, we remove duplicate tokens from the remaining ones based on their similarity. By retaining diverse tokens alongside the initially selected important tokens, we maximally preserve the visual information of the input image. Experimental results demonstrate that our VisPruner sustains strong performance across various VLM architectures and reduction ratios, significantly outperforming existing methods based on text-visual attention. Notably, without any training, VisPruner can reduce the FLOPs of LLaVA-1.5-7B by 91% and inference latency by 75%, while maintaining comparable performance. Our code is available at https://github.com/Theia-4869/VisPruner.
Authors: Shengxuming Zhang, Weihan Li, Tianhong Gao, Jiacong Hu, Haoming Luo, Xiuming Zhang, Jing Zhang, Mingli Song, Zunlei Feng
Abstract: Pathological diagnosis is vital for determining disease characteristics, guiding treatment, and assessing prognosis, relying heavily on detailed, multi-scale analysis of high-resolution whole slide images (WSI). However, traditional pure vision models face challenges of redundant feature extraction, whereas existing large vision-language models (LVLMs) are limited by input resolution constraints, hindering their efficiency and accuracy. To overcome these issues, we propose two innovative strategies: the mixed task-guided feature enhancement, which directs feature extraction toward lesion-related details across scales, and the prompt-guided detail feature completion, which integrates coarse- and fine-grained features from WSI based on specific prompts without compromising inference speed. Leveraging a comprehensive dataset of 490,000 samples from diverse pathology tasks-including cancer detection, grading, vascular and neural invasion identification, and so on-we trained the pathology-specialized LVLM, OmniPath. Extensive experiments demonstrate that this model significantly outperforms existing methods in diagnostic accuracy and efficiency, offering an interactive, clinically aligned approach for auxiliary diagnosis in a wide range of pathology applications.
Authors: Shu Shen, C. L. Philip Chen, Tong Zhang
Abstract: Multimodal machine learning has achieved remarkable progress in many scenarios, but its reliability is undermined by varying sample quality. This paper finds that existing reliable multimodal classification methods not only fail to provide robust estimation of data quality, but also lack dynamic networks for sample-specific depth and parameters to achieve reliable inference. To this end, a novel framework for multimodal reliable classification termed \textit{Multi-level Quality-Adaptive Dynamic multimodal network} (Multi-QuAD) is proposed. Multi-QuAD first adopts a novel approach based on noise-free prototypes and a classifier-free design to reliably estimate the quality of each sample at both modality and feature levels. It then achieves sample-specific network depth via the \textbf{\textit{Global Confidence Normalized Depth (GCND)}} mechanism. By normalizing depth across modalities and samples, \textit{\textbf{GCND}} effectively mitigates the impact of challenging modality inputs on dynamic depth reliability. Furthermore, Multi-QuAD provides sample-adaptive network parameters via the \textbf{\textit{Layer-wise Greedy Parameter (LGP)}} mechanism driven by feature-level quality. The cross-modality layer-wise greedy strategy in \textbf{\textit{LGP}} designs a reliable parameter prediction paradigm for multimodal networks with variable architecture for the first time. Experiments conducted on four datasets demonstrate that Multi-QuAD significantly outperforms state-of-the-art methods in classification performance and reliability, exhibiting strong adaptability to data with diverse quality.
Authors: Nathanael L. Baisa
Abstract: In this work, we investigate four different fusion methods for associating detections to tracklets in multi-object visual tracking. In addition to considering strong cues such as motion and appearance information, we also consider weak cues such as height intersection-over-union (height-IoU) and tracklet confidence information in the data association using different fusion methods. These fusion methods include minimum, weighted sum based on IoU, Kalman filter (KF) gating, and hadamard product of costs due to the different cues. We conduct extensive evaluations on validation sets of MOT17, MOT20 and DanceTrack datasets, and find out that the choice of a fusion method is key for data association in multi-object visual tracking. We hope that this investigative work helps the computer vision research community to use the right fusion method for data association in multi-object visual tracking.
Authors: Zhibo Ren, Pritthijit Nath, Pancham Shukla
Abstract: Tropical cyclone (TC) forecasting is crucial for disaster preparedness and mitigation. While recent deep learning approaches have shown promise, existing methods often treat TC evolution as a series of independent frame-to-frame predictions, limiting their ability to capture long-term dynamics. We present a novel application of video diffusion models for TC forecasting that explicitly models temporal dependencies through additional temporal layers. Our approach enables the model to generate multiple frames simultaneously, better capturing cyclone evolution patterns. We introduce a two-stage training strategy that significantly improves individual-frame quality and performance in low-data regimes. Experimental results show our method outperforms the previous approach of Nath et al. by 19.3% in MAE, 16.2% in PSNR, and 36.1% in SSIM. Most notably, we extend the reliable forecasting horizon from 36 to 50 hours. Through comprehensive evaluation using both traditional metrics and Fr\'echet Video Distance (FVD), we demonstrate that our approach produces more temporally coherent forecasts while maintaining competitive single-frame quality. Code accessible at https://github.com/Ren-creater/forecast-video-diffmodels.
URLs: https://github.com/Ren-creater/forecast-video-diffmodels.
Authors: Yuanhuiyi Lyu, Xu Zheng, Lutao Jiang, Yibo Yan, Xin Zou, Huiyu Zhou, Linfeng Zhang, Xuming Hu
Abstract: Recent text-to-image generative models, e.g., Stable Diffusion V3 and Flux, have achieved notable progress. However, these models are strongly restricted to their limited knowledge, a.k.a., their own fixed parameters, that are trained with closed datasets. This leads to significant hallucinations or distortions when facing fine-grained and unseen novel real-world objects, e.g., the appearance of the Tesla Cybertruck. To this end, we present the first real-object-based retrieval-augmented generation framework (RealRAG), which augments fine-grained and unseen novel object generation by learning and retrieving real-world images to overcome the knowledge gaps of generative models. Specifically, to integrate missing memory for unseen novel object generation, we train a reflective retriever by self-reflective contrastive learning, which injects the generator's knowledge into the sef-reflective negatives, ensuring that the retrieved augmented images compensate for the model's missing knowledge. Furthermore, the real-object-based framework integrates fine-grained visual knowledge for the generative models, tackling the distortion problem and improving the realism for fine-grained object generation. Our Real-RAG is superior in its modular application to all types of state-of-the-art text-to-image generative models and also delivers remarkable performance boosts with all of them, such as a gain of 16.18% FID score with the auto-regressive model on the Stanford Car benchmark.
Authors: Zhihao Guo, Jingxuan Su, Shenglin Wang, Jinlong Fan, Jing Zhang, Wei Zhou, Hadi Amirpour, Yunlong Zhao, Liangxiu Han, Peng Wang
Abstract: 3D Gaussian Splatting has emerged as an efficient photorealistic novel view synthesis method. However, its reliance on sparse Structure-from-Motion (SfM) point clouds often limits scene reconstruction quality. To address the limitation, this paper proposes a novel 3D reconstruction framework, Gaussian Processes enhanced Gaussian Splatting (GP-GS), in which a multi-output Gaussian Process model is developed to enable adaptive and uncertainty-guided densification of sparse SfM point clouds. Specifically, we propose a dynamic sampling and filtering pipeline that adaptively expands the SfM point clouds by leveraging GP-based predictions to infer new candidate points from the input 2D pixels and depth maps. The pipeline utilizes uncertainty estimates to guide the pruning of high-variance predictions, ensuring geometric consistency and enabling the generation of dense point clouds. These densified point clouds provide high-quality initial 3D Gaussians, enhancing reconstruction performance. Extensive experiments conducted on synthetic and real-world datasets across various scales validate the effectiveness and practicality of the proposed framework.
Authors: Xiaowen Qiu, Jincheng Yang, Yian Wang, Zhehuan Chen, Yufei Wang, Tsun-Hsuan Wang, Zhou Xian, Chuang Gan
Abstract: 3D articulated objects modeling has long been a challenging problem, since it requires to capture both accurate surface geometries and semantically meaningful and spatially precise structures, parts, and joints. Existing methods heavily depend on training data from a limited set of handcrafted articulated object categories (e.g., cabinets and drawers), which restricts their ability to model a wide range of articulated objects in an open-vocabulary context. To address these limitations, we propose Articulate Anymesh, an automated framework that is able to convert any rigid 3D mesh into its articulated counterpart in an open-vocabulary manner. Given a 3D mesh, our framework utilizes advanced Vision-Language Models and visual prompting techniques to extract semantic information, allowing for both the segmentation of object parts and the construction of functional joints. Our experiments show that Articulate Anymesh can generate large-scale, high-quality 3D articulated objects, including tools, toys, mechanical devices, and vehicles, significantly expanding the coverage of existing 3D articulated object datasets. Additionally, we show that these generated assets can facilitate the acquisition of new articulated object manipulation skills in simulation, which can then be transferred to a real robotic system. Our Github website is https://articulate-anymesh.github.io.
Authors: Qijun Gan, Yi Ren, Chen Zhang, Zhenhui Ye, Pan Xie, Xiang Yin, Zehuan Yuan, Bingyue Peng, Jianke Zhu
Abstract: Human motion video generation has advanced significantly, while existing methods still struggle with accurately rendering detailed body parts like hands and faces, especially in long sequences and intricate motions. Current approaches also rely on fixed resolution and struggle to maintain visual consistency. To address these limitations, we propose HumanDiT, a pose-guided Diffusion Transformer (DiT)-based framework trained on a large and wild dataset containing 14,000 hours of high-quality video to produce high-fidelity videos with fine-grained body rendering. Specifically, (i) HumanDiT, built on DiT, supports numerous video resolutions and variable sequence lengths, facilitating learning for long-sequence video generation; (ii) we introduce a prefix-latent reference strategy to maintain personalized characteristics across extended sequences. Furthermore, during inference, HumanDiT leverages Keypoint-DiT to generate subsequent pose sequences, facilitating video continuation from static images or existing videos. It also utilizes a Pose Adapter to enable pose transfer with given sequences. Extensive experiments demonstrate its superior performance in generating long-form, pose-accurate videos across diverse scenarios.
Authors: Peter Carragher, Nikitha Rao, Abhinand Jha, R Raghav, Kathleen M. Carley
Abstract: Vision language models (VLM) demonstrate sophisticated multimodal reasoning yet are prone to hallucination when confronted with knowledge conflicts, impeding their deployment in information-sensitive contexts. While existing research addresses robustness in unimodal models, the multimodal domain lacks systematic investigation of cross-modal knowledge conflicts. This research introduces \segsub, a framework for applying targeted image perturbations to investigate VLM resilience against knowledge conflicts. Our analysis reveals distinct vulnerability patterns: while VLMs are robust to parametric conflicts (20% adherence rates), they exhibit significant weaknesses in identifying counterfactual conditions (<30% accuracy) and resolving source conflicts (<1% accuracy). Correlations between contextual richness and hallucination rate (r = -0.368, p = 0.003) reveal the kinds of images that are likely to cause hallucinations. Through targeted fine-tuning on our benchmark dataset, we demonstrate improvements in VLM knowledge conflict detection, establishing a foundation for developing hallucination-resilient multimodal systems in information-sensitive environments.
Authors: Lijun Yan, Churan Wang, Fangwei Zhong, Yizhou Wang
Abstract: Magnetic resonance imaging (MRI) is a potent diagnostic tool for detecting pathological tissues in various diseases. Different MRI sequences have different contrast mechanisms and sensitivities for different types of lesions, which pose challenges to accurate and consistent lesion segmentation. In clinical practice, radiologists commonly use the sub-sequence feature, i.e. the difference between post contrast-enhanced T1-weighted (post) and pre-contrast-enhanced (pre) sequences, to locate lesions. Inspired by this, we propose a residual fusion method to learn subsequence representation for MRI lesion segmentation. Specifically, we iteratively and adaptively fuse features from pre- and post-contrast sequences at multiple resolutions, using dynamic weights to achieve optimal fusion and address diverse lesion enhancement patterns. Our method achieves state-of-the-art performances on BraTS2023 dataset for brain tumor segmentation and our in-house breast MRI dataset for breast lesion segmentation. Our method is clinically inspired and has the potential to facilitate lesion segmentation in various applications.
Authors: Xuan Ding, Rui Sun, Yunjian Zhang, Xiu Yan, Yueqi Zhou, Kaihao Huang, Suzhong Fu, Chuanlong Xie, Yao Zhu
Abstract: Compared to width-wise pruning, depth-wise pruning can significantly accelerate inference in resource-constrained scenarios. However, treating the entire Transformer layer as the minimum pruning unit may degrade model performance by indiscriminately discarding the entire information of the layer. This paper reveals the ``Patch-like'' feature relationship between layers in large language models by analyzing the correlation of the outputs of different layers in the reproducing kernel Hilbert space. Building on this observation, we propose a sliding layer merging method that dynamically selects and fuses consecutive layers from top to bottom according to a pre-defined similarity threshold, thereby simplifying the model structure while maintaining its performance. Extensive experiments on LLMs with various architectures and different parameter scales show that our method outperforms existing pruning techniques in both zero-shot inference performance and retraining recovery quality after pruning. In particular, in the experiment with 35% pruning on the Vicuna-7B model, our method achieved a 1.654% improvement in average performance on zero-shot tasks compared to the existing method. Moreover, we further reveal the potential of combining depth pruning with width pruning to enhance the pruning effect. Our codes are available at https://github.com/920927/SLM-a-sliding-layer-merging-method.
URLs: https://github.com/920927/SLM-a-sliding-layer-merging-method.
Authors: Kaiwen Zheng, Yongxin Chen, Huayu Chen, Guande He, Ming-Yu Liu, Jun Zhu, Qinsheng Zhang
Abstract: While likelihood-based generative models, particularly diffusion and autoregressive models, have achieved remarkable fidelity in visual generation, the maximum likelihood estimation (MLE) objective, which minimizes the forward KL divergence, inherently suffers from a mode-covering tendency that limits the generation quality under limited model capacity. In this work, we propose Direct Discriminative Optimization (DDO) as a unified framework that integrates likelihood-based generative training and GAN-type discrimination to bypass this fundamental constraint by exploiting reverse KL and self-generated negative signals. Our key insight is to parameterize a discriminator implicitly using the likelihood ratio between a learnable target model and a fixed reference model, drawing parallels with the philosophy of Direct Preference Optimization (DPO). Unlike GANs, this parameterization eliminates the need for joint training of generator and discriminator networks, allowing for direct, efficient, and effective finetuning of a well-trained model to its full potential beyond the limits of MLE. DDO can be performed iteratively in a self-play manner for progressive model refinement, with each round requiring less than 1% of pretraining epochs. Our experiments demonstrate the effectiveness of DDO by significantly advancing the previous SOTA diffusion model EDM, reducing FID scores from 1.79/1.58/1.96 to new records of 1.30/0.97/1.26 on CIFAR-10/ImageNet-64/ImageNet 512x512 datasets without any guidance mechanisms, and by consistently improving both guidance-free and CFG-enhanced FIDs of visual autoregressive models on ImageNet 256x256.
Authors: Sijin Sun, Ming Deng, Xingrui Yu, Xingyu Xi, Liangbin Zhao
Abstract: Metal defect detection is critical in industrial quality assurance, yet existing methods struggle with grayscale variations and complex defect states, limiting its robustness. To address these challenges, this paper proposes a Self-Adaptive Gamma Context-Aware SSM-based model(GCM-DET). This advanced detection framework integrating a Dynamic Gamma Correction (GC) module to enhance grayscale representation and optimize feature extraction for precise defect reconstruction. A State-Space Search Management (SSM) architecture captures robust multi-scale features, effectively handling defects of varying shapes and scales. Focal Loss is employed to mitigate class imbalance and refine detection accuracy. Additionally, the CD5-DET dataset is introduced, specifically designed for port container maintenance, featuring significant grayscale variations and intricate defect patterns. Experimental results demonstrate that the proposed model achieves substantial improvements, with mAP@0.5 gains of 27.6\%, 6.6\%, and 2.6\% on the CD5-DET, NEU-DET, and GC10-DET datasets.
Authors: Hanze Li, Xiande Huang
Abstract: Growing evidence suggests that layer attention mechanisms, which enhance interaction among layers in deep neural networks, have significantly advanced network architectures. However, existing layer attention methods suffer from redundancy, as attention weights learned by adjacent layers often become highly similar. This redundancy causes multiple layers to extract nearly identical features, reducing the model's representational capacity and increasing training time. To address this issue, we propose a novel approach to quantify redundancy by leveraging the Kullback-Leibler (KL) divergence between adjacent layers. Additionally, we introduce an Enhanced Beta Quantile Mapping (EBQM) method that accurately identifies and skips redundant layers, thereby maintaining model stability. Our proposed Efficient Layer Attention (ELA) architecture, improves both training efficiency and overall performance, achieving a 30% reduction in training time while enhancing performance in tasks such as image classification and object detection.
Authors: Xiaoming Peng, Yixin Yang, Yang Zhou, Hui Huang
Abstract: Recently, 2D Gaussian Splatting (2DGS) has demonstrated superior geometry reconstruction quality than the popular 3DGS by using 2D surfels to approximate thin surfaces. However, it falls short when dealing with glossy surfaces, resulting in visible holes in these areas. We found the reflection discontinuity causes the issue. To fit the jump from diffuse to specular reflection at different viewing angles, depth bias is introduced in the optimized Gaussian primitives. To address that, we first replace the depth distortion loss in 2DGS with a novel depth convergence loss, which imposes a strong constraint on depth continuity. Then, we rectified the depth criterion in determining the actual surface, which fully accounts for all the intersecting Gaussians along the ray. Qualitative and quantitative evaluations across various datasets reveal that our method significantly improves reconstruction quality, with more complete and accurate surfaces than 2DGS.
Authors: Qing Jiang, Lin Wu, Zhaoyang Zeng, Tianhe Ren, Yuda Xiong, Yihao Chen, Qin Liu, Lei Zhang
Abstract: Humans are undoubtedly the most important participants in computer vision, and the ability to detect any individual given a natural language description, a task we define as referring to any person, holds substantial practical value. However, we find that existing models generally fail to achieve real-world usability, and current benchmarks are limited by their focus on one-to-one referring, that hinder progress in this area. In this work, we revisit this task from three critical perspectives: task definition, dataset design, and model architecture. We first identify five aspects of referable entities and three distinctive characteristics of this task. Next, we introduce HumanRef, a novel dataset designed to tackle these challenges and better reflect real-world applications. From a model design perspective, we integrate a multimodal large language model with an object detection framework, constructing a robust referring model named RexSeek. Experimental results reveal that state-of-the-art models, which perform well on commonly used benchmarks like RefCOCO/+/g, struggle with HumanRef due to their inability to detect multiple individuals. In contrast, RexSeek not only excels in human referring but also generalizes effectively to common object referring, making it broadly applicable across various perception tasks. Code is available at https://github.com/IDEA-Research/RexSeek
Authors: Mees M. Flapper, Elian Bernard, Sander G. Huisman
Abstract: A method for particle orientation tracking is developed and demonstrated specifically for anisotropic particles. Using (high-speed) multi-camera recordings of anisotropic particles from different viewpoints, we reconstruct the 3D location and orientation of these particles using their known shape. This paper describes an algorithm which tracks the location and orientation of multiple anisotropic particles over time, enabling detailed investigations of location, orientation, and rotation statistics. The robustness and error of this method is quantified, and we explore the effects of noise, image size, the number of used cameras, and the camera arrangement by applying the algorithm to synthetic images. We showcase several use-cases of this method in several experiments (in both quiescent and turbulent fluids), demonstrating the effectiveness and broad applicability of the described tracking method. The proposed method is shown to work for widely different particle shapes, successfully tracks multiple particles simultaneously, and the method can distinguish between different types of particles.
Authors: Yi Zhang, Ruonan Lin, Ang Ping
Abstract: This study proposes a lightweight method for building image super-resolution using a Dilated Contextual Feature Modulation Network (DCFMN). The process includes obtaining high-resolution images, down-sampling them to low-resolution, enhancing the low-resolution images, constructing and training a lightweight network model, and generating super-resolution outputs. To address challenges such as regular textures and long-range dependencies in building images, the DCFMN integrates an expansion separable modulation unit and a local feature enhancement module. The former employs multiple expansion convolutions equivalent to a large kernel to efficiently aggregate multi-scale features while leveraging a simple attention mechanism for adaptivity. The latter encodes local features, mixes channel information, and ensures no additional computational burden during inference through reparameterization. This approach effectively resolves the limitations of existing lightweight super-resolution networks in modeling long-range dependencies, achieving accurate and efficient global feature modeling without increasing computational costs, and significantly improving both reconstruction quality and lightweight efficiency for building image super-resolution models.
Authors: Jinseok Bae, Inwoo Hwang, Young Yoon Lee, Ziyu Guo, Joseph Liu, Yizhak Ben-Shabat, Young Min Kim, Mubbasir Kapadia
Abstract: Recent advances in motion diffusion models have led to remarkable progress in diverse motion generation tasks, including text-to-motion synthesis. However, existing approaches represent motions as dense frame sequences, requiring the model to process redundant or less informative frames. The processing of dense animation frames imposes significant training complexity, especially when learning intricate distributions of large motion datasets even with modern neural architectures. This severely limits the performance of generative motion models for downstream tasks. Inspired by professional animators who mainly focus on sparse keyframes, we propose a novel diffusion framework explicitly designed around sparse and geometrically meaningful keyframes. Our method reduces computation by masking non-keyframes and efficiently interpolating missing frames. We dynamically refine the keyframe mask during inference to prioritize informative frames in later diffusion steps. Extensive experiments show that our approach consistently outperforms state-of-the-art methods in text alignment and motion realism, while also effectively maintaining high performance at significantly fewer diffusion steps. We further validate the robustness of our framework by using it as a generative prior and adapting it to different downstream tasks.
Authors: Qiang Qi, Xiao Wang
Abstract: Video object detection has made significant progress in recent years thanks to convolutional neural networks (CNNs) and vision transformers (ViTs). Typically, CNNs excel at capturing local features but struggle to model global representations. Conversely, ViTs are adept at capturing long-range global features but face challenges in representing local feature details. Off-the-shelf video object detection methods solely rely on CNNs or ViTs to conduct feature aggregation, which hampers their capability to simultaneously leverage global and local information, thereby resulting in limited detection performance. In this paper, we propose a Transformer-GraphFormer Blender Network (TGBFormer) for video object detection, with three key technical improvements to fully exploit the advantages of transformers and graph convolutional networks while compensating for their limitations. First, we develop a spatial-temporal transformer module to aggregate global contextual information, constituting global representations with long-range feature dependencies. Second, we introduce a spatial-temporal GraphFormer module that utilizes local spatial and temporal relationships to aggregate features, generating new local representations that are complementary to the transformer outputs. Third, we design a global-local feature blender module to adaptively couple transformer-based global representations and GraphFormer-based local representations. Extensive experiments demonstrate that our TGBFormer establishes new state-of-the-art results on the ImageNet VID dataset. Particularly, our TGBFormer achieves 86.5% mAP while running at around 41.0 FPS on a single Tesla A100 GPU.
Authors: Abduljaleel Adejumo, Faegheh Yeganli, Clifford Broni-bediako, Aoran Xiao, Naoto Yokoya, Mennatullah Siam
Abstract: Multimodal Large Language Models (MLLMs) have achieved remarkable success in vision-language tasks but their remote sensing (RS) counterpart are relatively under explored. Unlike natural images, RS imagery presents unique challenges that current MLLMs struggle to handle, particularly in visual grounding and spatial reasoning. This study investigates the limitations of CLIP-based MLLMs in RS, highlighting their failure to differentiate visually distinct yet semantically similar RS images. To address this, we introduce a remote sensing multimodal visual patterns (RSMMVP) benchmark. It is designed to evaluate MLLMs in RS tasks by identifying the CLIP-blind pairs, where CLIP-based models incorrectly assign high similarity scores to visually distinct RS images. Through a visual question answering (VQA) evaluation, we analyze the performance of state-of-the-art MLLMs, revealing significant limitations in RS specific representation learning. The results provide valuable insights into the weaknesses of CLIP-based visual encoding and offer a foundation for future research to develop more effective MLLMs tailored for remote sensing applications.
Authors: Zihao Zhang, Haoran Chen, Haoyu Zhao, Guansong Lu, Yanwei Fu, Hang Xu, Zuxuan Wu
Abstract: Handling complex or nonlinear motion patterns has long posed challenges for video frame interpolation. Although recent advances in diffusion-based methods offer improvements over traditional optical flow-based approaches, they still struggle to generate sharp, temporally consistent frames in scenarios with large motion. To address this limitation, we introduce EDEN, an Enhanced Diffusion for high-quality large-motion vidEo frame iNterpolation. Our approach first utilizes a transformer-based tokenizer to produce refined latent representations of the intermediate frames for diffusion models. We then enhance the diffusion transformer with temporal attention across the process and incorporate a start-end frame difference embedding to guide the generation of dynamic motion. Extensive experiments demonstrate that EDEN achieves state-of-the-art results across popular benchmarks, including nearly a 10% LPIPS reduction on DAVIS and SNU-FILM, and an 8% improvement on DAIN-HD.
Authors: JunGyu Lee, Kunyoung Lee, Haesol Park, Ig-Jae Kim, Gi Pyo Nam
Abstract: Facial Expression Recognition (FER) plays a crucial role in human affective analysis and has been widely applied in computer vision tasks such as human-computer interaction and psychological assessment. The 8th Affective Behavior Analysis in-the-Wild (ABAW) Challenge aims to assess human emotions using the video-based Aff-Wild2 dataset. This challenge includes various tasks, including the video-based EXPR recognition track, which is our primary focus. In this paper, we demonstrate that addressing label ambiguity and class imbalance, which are known to cause performance degradation, can lead to meaningful performance improvements. Specifically, we propose Video-based Noise-aware Adaptive Weighting (V-NAW), which adaptively assigns importance to each frame in a clip to address label ambiguity and effectively capture temporal variations in facial expressions. Furthermore, we introduce a simple and effective augmentation strategy to reduce redundancy between consecutive frames, which is a primary cause of overfitting. Through extensive experiments, we validate the effectiveness of our approach, demonstrating significant improvements in video-based FER performance.
Authors: Ming Li, Jike Zhong, Shitian Zhao, Yuxiang Lai, Haoquan Zhang, Wang Bill Zhu, Kaipeng Zhang
Abstract: This paper investigates the role of explicit thinking process in rule-based reinforcement fine-tuning (RFT) for MLLMs. We first propose CLS-RL for MLLM image classification, using verifiable rewards for fine-tuning. Experiments show CLS-RL significantly outperforms SFT and yields a cross-dataset generalization effect. We then rethink and question whether explicit thinking in RFT is always necessary. Challenging the convention that explicit thinking is crucial for the success of RFT, we introduce No-Thinking-RL, exploring RFT without thinking by introducing a simple equality accuracy reward. We evaluate No-Thinking-RL on 6 diverse tasks across different model sizes and types. Experimental results reveal three key findings: 1). Visual perception tasks do not require thinking during RFT, as No-Thinking-RL consistently outperforms or matches Thinking-based RFT across model sizes. 2).} Models with limited capabilities struggle to generate high-quality CoT for RFT, making Thinking-based RFT less effective than No-Thinking-RL. 3). There are inconsistencies between the answers in the thinking and answer tags for some responses of thinking-based RFT, which show lower accuracy than the overall accuracy. We hypothesize that explicit thinking before verifiable answers may hinder reward convergence and reduce performance. To test this hypothesis, we propose Think-After-Answer, which places thinking after the answer to mitigate this effect for experimental verification. Lastly, we conduct a pilot study to explore whether MLLMs can learn when to think during RFT, introducing an Adaptive-Thinking method. Experiments show that it converges to a specific prompt depending on model capability and task complexity, achieving comparable or better performance than both Thinking and No-Thinking-RL. This suggests MLLMs can adaptively decide to think or not based on their capabilities and task complexity.
Authors: Justus Westerhoff, Erblina Purelku, Jakob Hackstein, Jonas Loos, Leo Pinetzki, Lorenz Hufe
Abstract: Typographic attacks exploit the interplay between text and visual content in multimodal foundation models, causing misclassifications when misleading text is embedded within images. However, existing datasets are limited in size and diversity, making it difficult to study such vulnerabilities. In this paper, we introduce SCAM, the largest and most diverse dataset of real-world typographic attack images to date, containing 1,162 images across hundreds of object categories and attack words. Through extensive benchmarking of Vision-Language Models (VLMs) on SCAM, we demonstrate that typographic attacks significantly degrade performance, and identify that training data and model architecture influence the susceptibility to these attacks. Our findings reveal that typographic attacks persist in state-of-the-art Large Vision-Language Models (LVLMs) due to the choice of their vision encoder, though larger Large Language Models (LLMs) backbones help mitigate their vulnerability. Additionally, we demonstrate that synthetic attacks closely resemble real-world (handwritten) attacks, validating their use in research. Our work provides a comprehensive resource and empirical insights to facilitate future research toward robust and trustworthy multimodal AI systems. We publicly release the datasets introduced in this paper along with the code for evaluations at www.bliss.berlin/research/scam.
Authors: Li Yu, Zhihui Li, Chao Yao, Jimin Xiao, Moncef Gabbouj
Abstract: Neural representations for video (NeRV) have gained considerable attention for their strong performance across various video tasks. However, existing NeRV methods often struggle to capture fine spatial details, resulting in vague reconstructions. In this paper, we present a Frequency Separation and Augmentation based Neural Representation for video (FANeRV), which addresses these limitations with its core Wavelet Frequency Upgrade Block. This block explicitly separates input frames into high and low-frequency components using discrete wavelet transform, followed by targeted enhancement using specialized modules. Finally, a specially designed gated network effectively fuses these frequency components for optimal reconstruction. Additionally, convolutional residual enhancement blocks are integrated into the later stages of the network to balance parameter distribution and improve the restoration of high-frequency details. Experimental results demonstrate that FANeRV significantly improves reconstruction performance and excels in multiple tasks, including video compression, inpainting, and interpolation, outperforming existing NeRV methods.
Authors: Hussain Md. Safwan, Mahbub Islam Mahim
Abstract: A common dilemma while photographing a scene is whether to capture it at a wider angle, allowing more of the scene to be covered but in less detail or to click in a narrow angle that captures better details but leaves out portions of the scene. We propose a novel method in this paper that infuses wider shots with finer quality details that is usually associated with an image captured by the primary lens by capturing the same scene using both narrow and wide field of view (FoV) lenses. We do so by training a Generative Adversarial Network (GAN)-based model to learn to extract the visual quality parameters from a narrow-angle shot and to transfer these to the corresponding wide-angle image of the scene using residual connections and an attention-based fusion module. We have mentioned in details the proposed technique to isolate the visual essence of an image and to transfer it into another image. We have also elaborately discussed our implementation details and have presented the results of evaluation over several benchmark datasets and comparisons with contemporary advancements in the field.
Authors: Xiaojun Ye, Chun Wang, Yiren Song, Sheng Zhou, Liangcheng Li, Jiajun Bu
Abstract: Movie Audio Description (AD) aims to narrate visual content during dialogue-free segments, particularly benefiting blind and visually impaired (BVI) audiences. Compared with general video captioning, AD demands plot-relevant narration with explicit character name references, posing unique challenges in movie understanding.To identify active main characters and focus on storyline-relevant regions, we propose FocusedAD, a novel framework that delivers character-centric movie audio descriptions. It includes: (i) a Character Perception Module(CPM) for tracking character regions and linking them to names; (ii) a Dynamic Prior Module(DPM) that injects contextual cues from prior ADs and subtitles via learnable soft prompts; and (iii) a Focused Caption Module(FCM) that generates narrations enriched with plot-relevant details and named characters. To overcome limitations in character identification, we also introduce an automated pipeline for building character query banks. FocusedAD achieves state-of-the-art performance on multiple benchmarks, including strong zero-shot results on MAD-eval-Named and our newly proposed Cinepile-AD dataset. Code and data will be released at https://github.com/Thorin215/FocusedAD .
Authors: Zhenyu Yu, Mohd Yamani Inda Idris, Pei Wang
Abstract: The widespread adoption of diffusion models in image generation has increased the demand for privacy-compliant unlearning. However, due to the high-dimensional nature and complex feature representations of diffusion models, achieving selective unlearning remains challenging, as existing methods struggle to remove sensitive information while preserving the consistency of non-sensitive regions. To address this, we propose an Automatic Dataset Creation Framework based on prompt-based layered editing and training-free local feature removal, constructing the ForgetMe dataset and introducing the Entangled evaluation metric. The Entangled metric quantifies unlearning effectiveness by assessing the similarity and consistency between the target and background regions and supports both paired (Entangled-D) and unpaired (Entangled-S) image data, enabling unsupervised evaluation. The ForgetMe dataset encompasses a diverse set of real and synthetic scenarios, including CUB-200-2011 (Birds), Stanford-Dogs, ImageNet, and a synthetic cat dataset. We apply LoRA fine-tuning on Stable Diffusion to achieve selective unlearning on this dataset and validate the effectiveness of both the ForgetMe dataset and the Entangled metric, establishing them as benchmarks for selective unlearning. Our work provides a scalable and adaptable solution for advancing privacy-preserving generative AI.
Authors: Yuchen Rao, Stefan Ainetter, Sinisa Stekovic, Vincent Lepetit, Friedrich Fraundorfer
Abstract: High-level 3D scene understanding is essential in many applications. However, the challenges of generating accurate 3D annotations make development of deep learning models difficult. We turn to recent advancements in automatic retrieval of synthetic CAD models, and show that data generated by such methods can be used as high-quality ground truth for training supervised deep learning models. More exactly, we employ a pipeline akin to the one previously used to automatically annotate objects in ScanNet scenes with their 9D poses and CAD models. This time, we apply it to the recent ScanNet++ v1 dataset, which previously lacked such annotations. Our findings demonstrate that it is not only possible to train deep learning models on these automatically-obtained annotations but that the resulting models outperform those trained on manually annotated data. We validate this on two distinct tasks: point cloud completion and single-view CAD model retrieval and alignment. Our results underscore the potential of automatic 3D annotations to enhance model performance while significantly reducing annotation costs. To support future research in 3D scene understanding, we will release our annotations, which we call SCANnotate++, along with our trained models.
Authors: Zuyao Chen, Jinlin Wu, Zhen Lei, Marc Pollefeys, Chang Wen Chen
Abstract: Next-token prediction is the fundamental principle for training large language models (LLMs), and reinforcement learning (RL) further enhances their reasoning performance. As an effective way to model language, image, video, and other modalities, the use of LLMs for end-to-end extraction of structured visual representations, such as scene graphs, remains underexplored. It requires the model to accurately produce a set of objects and relationship triplets, rather than generating text token by token. To achieve this, we introduce R1-SGG, a multimodal LLM (M-LLM) initially trained via supervised fine-tuning (SFT) on the scene graph dataset and subsequently refined using reinforcement learning to enhance its ability to generate scene graphs in an end-to-end manner. The SFT follows a conventional prompt-response paradigm, while RL requires the design of effective reward signals. We design a set of graph-centric rewards, including three recall-based variants -- Hard Recall, Hard Recall+Relax, and Soft Recall -- which evaluate semantic and spatial alignment between predictions and ground truth at the object and relation levels. A format consistency reward further ensures that outputs follow the expected structural schema. Extensive experiments on the VG150 and PSG benchmarks show that R1-SGG substantially reduces failure rates and achieves strong performance in Recall and mean Recall, surpassing traditional SGG models and existing multimodal language models. Our code is available at https://github.com/gpt4vision/R1-SGG
Authors: Inho Kim, Youngkil Song, Jicheol Park, Won Hwa Kim, Suha Kwak
Abstract: Sound source localization (SSL) is the task of locating the source of sound within an image. Due to the lack of localization labels, the de facto standard in SSL has been to represent an image and audio as a single embedding vector each, and use them to learn SSL via contrastive learning. To this end, previous work samples one of local image features as the image embedding and aggregates all local audio features to obtain the audio embedding, which is far from optimal due to the presence of noise and background irrelevant to the actual target in the input. We present a novel SSL method that addresses this chronic issue by joint slot attention on image and audio. To be specific, two slots competitively attend image and audio features to decompose them into target and off-target representations, and only target representations of image and audio are used for contrastive learning. Also, we introduce cross-modal attention matching to further align local features of image and audio. Our method achieved the best in almost all settings on three public benchmarks for SSL, and substantially outperformed all the prior work in cross-modal retrieval.
Authors: Zhenhailong Wang, Senthil Purushwalkam, Caiming Xiong, Silvio Savarese, Heng Ji, Ran Xu
Abstract: We present DyMU, an efficient, training-free framework that dynamically reduces the computational burden of vision-language models (VLMs) while maintaining high task performance. Our approach comprises two key components. First, Dynamic Token Merging (DToMe) reduces the number of visual token embeddings by merging similar tokens based on image complexity, addressing the inherent inefficiency of fixed-length outputs in vision transformers. Second, Virtual Token Unmerging (VTU) simulates the expected token sequence for large language models (LLMs) by efficiently reconstructing the attention dynamics of a full sequence, thus preserving the downstream performance without additional fine-tuning. Unlike previous approaches, our method dynamically adapts token compression to the content of the image and operates completely training-free, making it readily applicable to most state-of-the-art VLM architectures. Extensive experiments on image and video understanding tasks demonstrate that DyMU can reduce the average visual token count by 32%-85% while achieving comparable performance to full-length models across diverse VLM architectures, including the recently popularized AnyRes-based visual encoders. Furthermore, through qualitative analyses, we demonstrate that DToMe effectively adapts token reduction based on image complexity and, unlike existing systems, provides users more control over computational costs. Project page: https://mikewangwzhl.github.io/dymu/.
Authors: Anyi Xiao, Cihui Yang
Abstract: Table structure recognition aims to parse tables in unstructured data into machine-understandable formats. Recent methods address this problem through a two-stage process or optimized one-stage approaches. However, these methods either require multiple networks to be serially trained and perform more time-consuming sequential decoding, or rely on complex post-processing algorithms to parse the logical structure of tables. They struggle to balance cross-scenario adaptability, robustness, and computational efficiency. In this paper, we propose a one-stage end-to-end table structure parsing network called TableCenterNet. This network unifies the prediction of table spatial and logical structure into a parallel regression task for the first time, and implicitly learns the spatial-logical location mapping laws of cells through a synergistic architecture of shared feature extraction layers and task-specific decoding. Compared with two-stage methods, our method is easier to train and faster to infer. Experiments on benchmark datasets show that TableCenterNet can effectively parse table structures in diverse scenarios and achieve state-of-the-art performance on the TableGraph-24k dataset. Code is available at https://github.com/dreamy-xay/TableCenterNet.
Authors: Yunzhong Zhang, Bo Xiong, You Zhou, Changqing Su, Zhen Cheng, Zhaofei Yu, Xun Cao, Tiejun Huang
Abstract: The need for accurate and non-intrusive flow measurement methods has led to the widespread adoption of Particle Image Velocimetry (PIV), a powerful diagnostic tool in fluid motion estimation. This study investigates the tremendous potential of spike cameras (a type of ultra-high-speed, high-dynamic-range camera) in PIV. We propose a deep learning framework, Spike Imaging Velocimetry (SIV), designed specifically for highly turbulent and intricate flow fields. To aggregate motion features from the spike stream while minimizing information loss, we incorporate a Detail-Preserving Hierarchical Transform (DPHT) module. Additionally, we introduce a Graph Encoder (GE) to extract contextual features from highly complex fluid flows. Furthermore, we present a spike-based PIV dataset, Particle Scenes with Spike and Displacement (PSSD), which provides labeled data for three challenging fluid dynamics scenarios. Our proposed method achieves superior performance compared to existing baseline methods on PSSD. The datasets and our implementation of SIV are open-sourced in the supplementary materials.
Authors: Qi Fan, Kaiqi Liu, Nian Liu, Hisham Cholakkal, Rao Muhammad Anwer, Wenbin Li, Yang Gao
Abstract: Cross-domain few-shot segmentation (CD-FSS) aims to segment objects of novel classes in new domains, which is often challenging due to the diverse characteristics of target domains and the limited availability of support data. Most CD-FSS methods redesign and retrain in-domain FSS models using various domain-generalization techniques, which are effective but costly to train. To address these issues, we propose adapting informative model structures of the well-trained FSS model for target domains by learning domain characteristics from few-shot labeled support samples during inference, thereby eliminating the need for retraining. Specifically, we first adaptively identify domain-specific model structures by measuring parameter importance using a novel structure Fisher score in a data-dependent manner. Then, we progressively train the selected informative model structures with hierarchically constructed training samples, progressing from fewer to more support shots. The resulting Informative Structure Adaptation (ISA) method effectively addresses domain shifts and equips existing well-trained in-domain FSS models with flexible adaptation capabilities for new domains, eliminating the need to redesign or retrain CD-FSS models on base data. Extensive experiments validate the effectiveness of our method, demonstrating superior performance across multiple CD-FSS benchmarks.
Authors: Xinyu Li, Qi Yao, Yuanda Wang
Abstract: Garment sewing patterns are fundamental design elements that bridge the gap between design concepts and practical manufacturing. The generative modeling of sewing patterns is crucial for creating diversified garments. However, existing approaches are limited either by reliance on a single input modality or by suboptimal generation efficiency. In this work, we present GarmentDiffusion, a new generative model capable of producing centimeter-precise, vectorized 3D sewing patterns from multimodal inputs (text, image, and incomplete sewing pattern). Our method efficiently encodes 3D sewing pattern parameters into compact edge token representations, achieving a sequence length that is 10 times shorter than that of the autoregressive SewingGPT in DressCode. By employing a diffusion transformer, we simultaneously denoise all edge tokens along the temporal axis, while maintaining a constant number of denoising steps regardless of dataset-specific edge and panel statistics. With all combination of designs of our model, the sewing pattern generation speed is accelerated by 100 times compared to SewingGPT. We achieve new state-of-the-art results on DressCodeData, as well as on the largest sewing pattern dataset, namely GarmentCodeData. The project website is available at https://shenfu-research.github.io/Garment-Diffusion/.
Authors: Mengting Wei, Yante Li, Tuomas Varanka, Yan Jiang, Guoying Zhao
Abstract: In this study, we propose a method for video face reenactment that integrates a 3D face parametric model into a latent diffusion framework, aiming to improve shape consistency and motion control in existing video-based face generation approaches. Our approach employs the FLAME (Faces Learned with an Articulated Model and Expressions) model as the 3D face parametric representation, providing a unified framework for modeling face expressions and head pose. This not only enables precise extraction of motion features from driving videos, but also contributes to the faithful preservation of face shape and geometry. Specifically, we enhance the latent diffusion model with rich 3D expression and detailed pose information by incorporating depth maps, normal maps, and rendering maps derived from FLAME sequences. These maps serve as motion guidance and are encoded into the denoising UNet through a specifically designed Geometric Guidance Encoder (GGE). A multi-layer feature fusion module with integrated self-attention mechanisms is used to combine facial appearance and motion latent features within the spatial domain. By utilizing the 3D face parametric model as motion guidance, our method enables parametric alignment of face identity between the reference image and the motion captured from the driving video. Experimental results on benchmark datasets show that our method excels at generating high-quality face animations with precise expression and head pose variation modeling. In addition, it demonstrates strong generalization performance on out-of-domain images. Code is publicly available at https://github.com/weimengting/MagicPortrait.
Authors: Biling Wang, Austen Maniscalco, Ti Bai, Siqiu Wang, Michael Dohopolski, Mu-Han Lin, Chenyang Shen, Dan Nguyen, Junzhou Huang, Steve Jiang, Xinlei Wang
Abstract: Purpose: This study presents a Deep Learning (DL)-based quality assessment (QA) approach for evaluating auto-generated contours (auto-contours) in radiotherapy, with emphasis on Online Adaptive Radiotherapy (OART). Leveraging Bayesian Ordinal Classification (BOC) and calibrated uncertainty thresholds, the method enables confident QA predictions without relying on ground truth contours or extensive manual labeling. Methods: We developed a BOC model to classify auto-contour quality and quantify prediction uncertainty. A calibration step was used to optimize uncertainty thresholds that meet clinical accuracy needs. The method was validated under three data scenarios: no manual labels, limited labels, and extensive labels. For rectum contours in prostate cancer, we applied geometric surrogate labels when manual labels were absent, transfer learning when limited, and direct supervision when ample labels were available. Results: The BOC model delivered robust performance across all scenarios. Fine-tuning with just 30 manual labels and calibrating with 34 subjects yielded over 90% accuracy on test data. Using the calibrated threshold, over 93% of the auto-contours' qualities were accurately predicted in over 98% of cases, reducing unnecessary manual reviews and highlighting cases needing correction. Conclusion: The proposed QA model enhances contouring efficiency in OART by reducing manual workload and enabling fast, informed clinical decisions. Through uncertainty quantification, it ensures safer, more reliable radiotherapy workflows.
Authors: Zherui Zhang, Rongtao Xu, Jie Zhou, Changwei Wang, Xingtian Pei, Wenhao Xu, Jiguang Zhang, Li Guo, Longxiang Gao, Wenbo Xu, Shibiao Xu
Abstract: The Transformer architecture has achieved significant success in natural language processing, motivating its adaptation to computer vision tasks. Unlike convolutional neural networks, vision transformers inherently capture long-range dependencies and enable parallel processing, yet lack inductive biases and efficiency benefits, facing significant computational and memory challenges that limit its real-world applicability. This paper surveys various online strategies for generating lightweight vision transformers for image recognition, focusing on three key areas: Efficient Component Design, Dynamic Network, and Knowledge Distillation. We evaluate the relevant exploration for each topic on the ImageNet-1K benchmark, analyzing trade-offs among precision, parameters, throughput, and more to highlight their respective advantages, disadvantages, and flexibility. Finally, we propose future research directions and potential challenges in the lightweighting of vision transformers with the aim of inspiring further exploration and providing practical guidance for the community. Project Page: https://github.com/ajxklo/Lightweight-VIT
Authors: S. Z. Zhou, Y. B. Wang, J. F. Wu, T. Hu, J. N. Zhang, Z. J. Li, Y. Liu
Abstract: Audio-driven human animation technology is widely used in human-computer interaction, and the emergence of diffusion models has further advanced its development. Currently, most methods rely on multi-stage generation and intermediate representations, resulting in long inference time and issues with generation quality in specific foreground regions and audio-motion consistency. These shortcomings are primarily due to the lack of localized fine-grained supervised guidance. To address above challenges, we propose PAHA, an end-to-end audio-driven upper-body human animation framework with diffusion model. We introduce two key methods: Parts-Aware Re-weighting (PAR) and Parts Consistency Enhancement (PCE). PAR dynamically adjusts regional training loss weights based on pose confidence scores, effectively improving visual quality. PCE constructs and trains diffusion-based regional audio-visual classifiers to improve the consistency of motion and co-speech audio. Afterwards, we design two novel inference guidance methods for the foregoing classifiers, Sequential Guidance (SG) and Differential Guidance (DG), to balance efficiency and quality respectively. Additionally, we build CNAS, the first public Chinese News Anchor Speech dataset, to advance research and validation in this field. Extensive experimental results and user studies demonstrate that PAHA significantly outperforms existing methods in audio-motion alignment and video-related evaluations. The codes and CNAS dataset will be released upon acceptance.
Authors: Shang Zhang, Huanbin Zhang, Dali Feng, Yujie Cui, Ruoyan Xiong, Cen He
Abstract: Thermal infrared (TIR) object tracking often suffers from challenges such as target occlusion, motion blur, and background clutter, which significantly degrade the performance of trackers. To address these issues, this paper pro-poses a novel Siamese Motion Mamba Tracker (SMMT), which integrates a bidirectional state-space model and a self-attention mechanism. Specifically, we introduce the Motion Mamba module into the Siamese architecture to ex-tract motion features and recover overlooked edge details using bidirectional modeling and self-attention. We propose a Siamese parameter-sharing strate-gy that allows certain convolutional layers to share weights. This approach reduces computational redundancy while preserving strong feature represen-tation. In addition, we design a motion edge-aware regression loss to improve tracking accuracy, especially for motion-blurred targets. Extensive experi-ments are conducted on four TIR tracking benchmarks, including LSOTB-TIR, PTB-TIR, VOT-TIR2015, and VOT-TIR 2017. The results show that SMMT achieves superior performance in TIR target tracking.
Authors: Zhihao Zhang, Abhinav Kumar, Girish Chandar Ganesan, Xiaoming Liu
Abstract: Accurately predicting 3D attributes is crucial for monocular 3D object detection (Mono3D), with depth estimation posing the greatest challenge due to the inherent ambiguity in mapping 2D images to 3D space. While existing methods leverage multiple depth cues (e.g., estimating depth uncertainty, modeling depth error) to improve depth accuracy, they overlook that accurate depth prediction requires conditioning on other 3D attributes, as these attributes are intrinsically inter-correlated through the 3D to 2D projection, which ultimately limits overall accuracy and stability. Inspired by Chain-of-Thought (CoT) in large language models (LLMs), this paper proposes MonoCoP, which leverages a Chain-of-Prediction (CoP) to predict attributes sequentially and conditionally via three key designs. First, it employs a lightweight AttributeNet (AN) for each 3D attribute to learn attribute-specific features. Next, MonoCoP constructs an explicit chain to propagate these learned features from one attribute to the next. Finally, MonoCoP uses a residual connection to aggregate features for each attribute along the chain, ensuring that later attribute predictions are conditioned on all previously processed attributes without forgetting the features of earlier ones. Experimental results show that our MonoCoP achieves state-of-the-art (SoTA) performance on the KITTI leaderboard without requiring additional data and further surpasses existing methods on the Waymo and nuScenes frontal datasets.
Authors: Zhuang Qi, Sijin Zhou, Lei Meng, Han Hu, Han Yu, Xiangxu Meng
Abstract: Attribute bias in federated learning (FL) typically leads local models to optimize inconsistently due to the learning of non-causal associations, resulting degraded performance. Existing methods either use data augmentation for increasing sample diversity or knowledge distillation for learning invariant representations to address this problem. However, they lack a comprehensive analysis of the inference paths, and the interference from confounding factors limits their performance. To address these limitations, we propose the \underline{Fed}erated \underline{D}econfounding and \underline{D}ebiasing \underline{L}earning (FedDDL) method. It constructs a structured causal graph to analyze the model inference process, and performs backdoor adjustment to eliminate confounding paths. Specifically, we design an intra-client deconfounding learning module for computer vision tasks to decouple background and objects, generating counterfactual samples that establish a connection between the background and any label, which stops the model from using the background to infer the label. Moreover, we design an inter-client debiasing learning module to construct causal prototypes to reduce the proportion of the background in prototype components. Notably, it bridges the gap between heterogeneous representations via causal prototypical regularization. Extensive experiments on 2 benchmarking datasets demonstrate that \methodname{} significantly enhances the model capability to focus on main objects in unseen data, leading to 4.5\% higher Top-1 Accuracy on average over 9 state-of-the-art existing methods.
Authors: Xin Bi, Zhichao Li, Yuxuan Xia, Panpan Tong, Lijuan Zhang, Yang Chen, Junsheng Fu
Abstract: Accurate online map matching is fundamental to vehicle navigation and the activation of intelligent driving functions. Current online map matching methods are prone to errors in complex road networks, especially in multilevel road area. To address this challenge, we propose an online Standard Definition (SD) map matching method by constructing a Hidden Markov Model (HMM) with multiple probability factors. Our proposed method can achieve accurate map matching even in complex road networks by carefully leveraging lane markings and scenario recognition in the designing of the probability factors. First, the lane markings are generated by a multi-lane tracking method and associated with the SD map using HMM to build an enriched SD map. In areas covered by the enriched SD map, the vehicle can re-localize itself by performing Iterative Closest Point (ICP) registration for the lane markings. Then, the probability factor accounting for the lane marking detection can be obtained using the association probability between adjacent lanes and roads. Second, the driving scenario recognition model is applied to generate the emission probability factor of scenario recognition, which improves the performance of map matching on elevated roads and ordinary urban roads underneath them. We validate our method through extensive road tests in Europe and China, and the experimental results show that our proposed method effectively improves the online map matching accuracy as compared to other existing methods, especially in multilevel road area. Specifically, the experiments show that our proposed method achieves $F_1$ scores of 98.04% and 94.60% on the Zenseact Open Dataset and test data of multilevel road areas in Shanghai respectively, significantly outperforming benchmark methods. The implementation is available at https://github.com/TRV-Lab/LMSR-OMM.
Authors: Jinyu Gu, Haipeng Liu, Meng Wang, Yang Wang
Abstract: Text-to-image generation for personalized identities aims at incorporating the specific identity into images using a text prompt and an identity image. Based on the powerful generative capabilities of DDPMs, many previous works adopt additional prompts, such as text embeddings and CLIP image embeddings, to represent the identity information, while they fail to disentangle the identity information and background information. As a result, the generated images not only lose key identity characteristics but also suffer from significantly reduced diversity. To address this issue, previous works have combined the W+ space from StyleGAN with diffusion models, leveraging this space to provide a more accurate and comprehensive representation of identity features through multi-level feature extraction. However, the entanglement of identity and background information in in-the-wild images during training prevents accurate identity localization, resulting in severe semantic interference between identity and background. In this paper, we propose a novel fine-tuning-based diffusion model for personalized identities text-to-image generation, named PIDiff, which leverages the W+ space and an identity-tailored fine-tuning strategy to avoid semantic entanglement and achieves accurate feature extraction and localization. Style editing can also be achieved by PIDiff through preserving the characteristics of identity features in the W+ space, which vary from coarse to fine. Through the combination of the proposed cross-attention block and parameter optimization strategy, PIDiff preserves the identity information and maintains the generation capability for in-the-wild images of the pre-trained model during inference. Our experimental results validate the effectiveness of our method in this task.
Authors: Hongyang Zhu, Haipeng Liu, Bo Fu, Yang Wang
Abstract: Multi-object editing aims to modify multiple objects or regions in complex scenes while preserving structural coherence. This task faces significant challenges in scenarios involving overlapping or interacting objects: (1) Inaccurate localization of target objects due to attention misalignment, leading to incomplete or misplaced edits; (2) Attribute-object mismatch, where color or texture changes fail to align with intended regions due to cross-attention leakage, creating semantic conflicts (\textit{e.g.}, color bleeding into non-target areas). Existing methods struggle with these challenges: approaches relying on global cross-attention mechanisms suffer from attention dilution and spatial interference between objects, while mask-based methods fail to bind attributes to geometrically accurate regions due to feature entanglement in multi-object scenarios. To address these limitations, we propose a training-free, inference-stage optimization approach that enables precise localized image manipulation in complex multi-object scenes, named MDE-Edit. MDE-Edit optimizes the noise latent feature in diffusion models via two key losses: Object Alignment Loss (OAL) aligns multi-layer cross-attention with segmentation masks for precise object positioning, and Color Consistency Loss (CCL) amplifies target attribute attention within masks while suppressing leakage to adjacent regions. This dual-loss design ensures localized and coherent multi-object edits. Extensive experiments demonstrate that MDE-Edit outperforms state-of-the-art methods in editing accuracy and visual quality, offering a robust solution for complex multi-object image manipulation tasks.
Authors: Rachmadio Noval Lazuardi, Artem Sevastopolsky, Egor Zakharov, Matthias Niessner, Vanessa Sklyarova
Abstract: We propose a novel method that reconstructs hair strands directly from colorless 3D scans by leveraging multi-modal hair orientation extraction. Hair strand reconstruction is a fundamental problem in computer vision and graphics that can be used for high-fidelity digital avatar synthesis, animation, and AR/VR applications. However, accurately recovering hair strands from raw scan data remains challenging due to human hair's complex and fine-grained structure. Existing methods typically rely on RGB captures, which can be sensitive to the environment and can be a challenging domain for extracting the orientation of guiding strands, especially in the case of challenging hairstyles. To reconstruct the hair purely from the observed geometry, our method finds sharp surface features directly on the scan and estimates strand orientation through a neural 2D line detector applied to the renderings of scan shading. Additionally, we incorporate a diffusion prior trained on a diverse set of synthetic hair scans, refined with an improved noise schedule, and adapted to the reconstructed contents via a scan-specific text prompt. We demonstrate that this combination of supervision signals enables accurate reconstruction of both simple and intricate hairstyles without relying on color information. To facilitate further research, we introduce Strands400, the largest publicly available dataset of hair strands with detailed surface geometry extracted from real-world data, which contains reconstructed hair strands from the scans of 400 subjects.
Authors: Jie Liu, Gongye Liu, Jiajun Liang, Yangguang Li, Jiaheng Liu, Xintao Wang, Pengfei Wan, Di Zhang, Wanli Ouyang
Abstract: We propose Flow-GRPO, the first method integrating online reinforcement learning (RL) into flow matching models. Our approach uses two key strategies: (1) an ODE-to-SDE conversion that transforms a deterministic Ordinary Differential Equation (ODE) into an equivalent Stochastic Differential Equation (SDE) that matches the original model's marginal distribution at all timesteps, enabling statistical sampling for RL exploration; and (2) a Denoising Reduction strategy that reduces training denoising steps while retaining the original inference timestep number, significantly improving sampling efficiency without performance degradation. Empirically, Flow-GRPO is effective across multiple text-to-image tasks. For complex compositions, RL-tuned SD3.5 generates nearly perfect object counts, spatial relations, and fine-grained attributes, boosting GenEval accuracy from 63% to 95%. In visual text rendering, its accuracy improves from 59% to 92%, significantly enhancing text generation. Flow-GRPO also achieves substantial gains in human preference alignment. Notably, very little reward hacking occurred, meaning rewards did not increase at the cost of appreciable image quality or diversity degradation.
Authors: Chao Liao, Liyang Liu, Xun Wang, Zhengxiong Luo, Xinyu Zhang, Wenliang Zhao, Jie Wu, Liang Li, Zhi Tian, Weilin Huang
Abstract: Recent progress in unified models for image understanding and generation has been impressive, yet most approaches remain limited to single-modal generation conditioned on multiple modalities. In this paper, we present Mogao, a unified framework that advances this paradigm by enabling interleaved multi-modal generation through a causal approach. Mogao integrates a set of key technical improvements in architecture design, including a deep-fusion design, dual vision encoders, interleaved rotary position embeddings, and multi-modal classifier-free guidance, which allow it to harness the strengths of both autoregressive models for text generation and diffusion models for high-quality image synthesis. These practical improvements also make Mogao particularly effective to process interleaved sequences of text and images arbitrarily. To further unlock the potential of unified models, we introduce an efficient training strategy on a large-scale, in-house dataset specifically curated for joint text and image generation. Extensive experiments show that Mogao not only achieves state-of-the-art performance in multi-modal understanding and text-to-image generation, but also excels in producing high-quality, coherent interleaved outputs. Its emergent capabilities in zero-shot image editing and compositional generation highlight Mogao as a practical omni-modal foundation model, paving the way for future development and scaling the unified multi-modal systems.
Authors: Mikhail Chaichuk, Sushant Gautam, Steven Hicks, Elena Tutubalina
Abstract: The generation of realistic medical images from text descriptions has significant potential to address data scarcity challenges in healthcare AI while preserving patient privacy. This paper presents a comprehensive study of text-to-image synthesis in the medical domain, comparing two distinct approaches: (1) fine-tuning large pre-trained latent diffusion models and (2) training small, domain-specific models. We introduce a novel model named MSDM, an optimized architecture based on Stable Diffusion that integrates a clinical text encoder, variational autoencoder, and cross-attention mechanisms to better align medical text prompts with generated images. Our study compares two approaches: fine-tuning large pre-trained models (FLUX, Kandinsky) versus training compact domain-specific models (MSDM). Evaluation across colonoscopy (MedVQA-GI) and radiology (ROCOv2) datasets reveals that while large models achieve higher fidelity, our optimized MSDM delivers comparable quality with lower computational costs. Quantitative metrics and qualitative evaluations by medical experts reveal strengths and limitations of each approach.
Authors: Guangjing Wang, Ce Zhou, Yuanda Wang, Bocheng Chen, Hanqing Guo, Qiben Yan
Abstract: As Artificial Intelligence (AI) systems increasingly underpin critical applications, from autonomous vehicles to biometric authentication, their vulnerability to transferable attacks presents a growing concern. These attacks, designed to generalize across instances, domains, models, tasks, modalities, or even hardware platforms, pose severe risks to security, privacy, and system integrity. This survey delivers the first comprehensive review of transferable attacks across seven major categories, including evasion, backdoor, data poisoning, model stealing, model inversion, membership inference, and side-channel attacks. We introduce a unified six-dimensional taxonomy: cross-instance, cross-domain, cross-modality, cross-model, cross-task, and cross-hardware, which systematically captures the diverse transfer pathways of adversarial strategies. Through this framework, we examine both the underlying mechanics and practical implications of transferable attacks on AI systems. Furthermore, we review cutting-edge methods for enhancing attack transferability, organized around data augmentation and optimization strategies. By consolidating fragmented research and identifying critical future directions, this work provides a foundational roadmap for understanding, evaluating, and defending against transferable threats in real-world AI systems.
Authors: Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M. Dai, Anja Hauth, Katie Millican, David Silver, Melvin Johnson, Ioannis Antonoglou, Julian Schrittwieser, Amelia Glaese, Jilin Chen, Emily Pitler, Timothy Lillicrap, Angeliki Lazaridou, Orhan Firat, James Molloy, Michael Isard, Paul R. Barham, Tom Hennigan, Benjamin Lee, Fabio Viola, Malcolm Reynolds, Yuanzhong Xu, Ryan Doherty, Eli Collins, Clemens Meyer, Eliza Rutherford, Erica Moreira, Kareem Ayoub, Megha Goel, Jack Krawczyk, Cosmo Du, Ed Chi, Heng-Tze Cheng, Eric Ni, Purvi Shah, Patrick Kane, Betty Chan, Manaal Faruqui, Aliaksei Severyn, Hanzhao Lin, YaGuang Li, Yong Cheng, Abe Ittycheriah, Mahdis Mahdieh, Mia Chen, Pei Sun, Dustin Tran, Sumit Bagri, Balaji Lakshminarayanan, Jeremiah Liu, Andras Orban, Fabian G\"ura, Hao Zhou, Xinying Song, Aurelien Boffy, Harish Ganapathy, Steven Zheng, HyunJeong Choe, \'Agoston Weisz, Tao Zhu, Yifeng Lu, Siddharth Gopal, Jarrod Kahn, Maciej Kula, Jeff Pitman, Rushin Shah, Emanuel Taropa, Majd Al Merey, Martin Baeuml, Zhifeng Chen, Laurent El Shafey, Yujing Zhang, Olcan Sercinoglu, George Tucker, Enrique Piqueras, Maxim Krikun, Iain Barr, Nikolay Savinov, Ivo Danihelka, Becca Roelofs, Ana\"is White, Anders Andreassen, Tamara von Glehn, Lakshman Yagati, Mehran Kazemi, Lucas Gonzalez, Misha Khalman, Jakub Sygnowski, Alexandre Frechette, Charlotte Smith, Laura Culp, Lev Proleev, Yi Luan, Xi Chen, James Lottes, Nathan Schucher, Federico Lebron, Alban Rrustemi, Natalie Clay, Phil Crone, Tomas Kocisky, Jeffrey Zhao, Bartek Perz, Dian Yu, Heidi Howard, Adam Bloniarz, Jack W. Rae, Han Lu, Laurent Sifre, Marcello Maggioni, Fred Alcober, Dan Garrette, Megan Barnes, Shantanu Thakoor, Jacob Austin, Gabriel Barth-Maron, William Wong, Rishabh Joshi, Rahma Chaabouni, Deeni Fatiha, Arun Ahuja, Gaurav Singh Tomar, Evan Senter, Martin Chadwick, Ilya Kornakov, Nithya Attaluri, I\~naki Iturrate, Ruibo Liu, Yunxuan Li, Sarah Cogan, Jeremy Chen, Chao Jia, Chenjie Gu, Qiao Zhang, Jordan Grimstad, Ale Jakse Hartman, Xavier Garcia, Thanumalayan Sankaranarayana Pillai, Jacob Devlin, Michael Laskin, Diego de Las Casas, Dasha Valter, Connie Tao, Lorenzo Blanco, Adri\`a Puigdom\`enech Badia, David Reitter, Mianna Chen, Jenny Brennan, Clara Rivera, Sergey Brin, Shariq Iqbal, Gabriela Surita, Jane Labanowski, Abhi Rao, Stephanie Winkler, Emilio Parisotto, Yiming Gu, Kate Olszewska, Ravi Addanki, Antoine Miech, Annie Louis, Denis Teplyashin, Geoff Brown, Elliot Catt, Jan Balaguer, Jackie Xiang, Pidong Wang, Zoe Ashwood, Anton Briukhov, Albert Webson, Sanjay Ganapathy, Smit Sanghavi, Ajay Kannan, Ming-Wei Chang, Axel Stjerngren, Josip Djolonga, Yuting Sun, Ankur Bapna, Matthew Aitchison, Pedram Pejman, Henryk Michalewski, Tianhe Yu, Cindy Wang, Juliette Love, Junwhan Ahn, Dawn Bloxwich, Kehang Han, Peter Humphreys, Thibault Sellam, James Bradbury, Varun Godbole, Sina Samangooei, Bogdan Damoc, Alex Kaskasoli, S\'ebastien M. R. Arnold, Vijay Vasudevan, Shubham Agrawal, Jason Riesa, Dmitry Lepikhin, Richard Tanburn, Srivatsan Srinivasan, Hyeontaek Lim, Sarah Hodkinson, Pranav Shyam, Johan Ferret, Steven Hand, Ankush Garg, Tom Le Paine, Jian Li, Yujia Li, Minh Giang, Alexander Neitz, Zaheer Abbas, Sarah York, Machel Reid, Elizabeth Cole, Aakanksha Chowdhery, Dipanjan Das, Dominika Rogozi\'nska, Vitaliy Nikolaev, Pablo Sprechmann, Zachary Nado, Lukas Zilka, Flavien Prost, Luheng He, Marianne Monteiro, Gaurav Mishra, Chris Welty, Josh Newlan, Dawei Jia, Miltiadis Allamanis, Clara Huiyi Hu, Raoul de Liedekerke, Justin Gilmer, Carl Saroufim, Shruti Rijhwani, Shaobo Hou, Disha Shrivastava, Anirudh Baddepudi, Alex Goldin, Adnan Ozturel, Albin Cassirer, Yunhan Xu, Daniel Sohn, Devendra Sachan, Reinald Kim Amplayo, Craig Swanson, Dessie Petrova, Shashi Narayan, Arthur Guez, Siddhartha Brahma, Jessica Landon, Miteyan Patel, Ruizhe Zhao, Kevin Villela, Luyu Wang, Wenhao Jia, Matthew Rahtz, Mai Gim\'enez, Legg Yeung, James Keeling, Petko Georgiev, Diana Mincu, Boxi Wu, Salem Haykal, Rachel Saputro, Kiran Vodrahalli, James Qin, Zeynep Cankara, Abhanshu Sharma, Nick Fernando, Will Hawkins, Behnam Neyshabur, Solomon Kim, Adrian Hutter, Priyanka Agrawal, Alex Castro-Ros, George van den Driessche, Tao Wang, Fan Yang, Shuo-yiin Chang, Paul Komarek, Ross McIlroy, Mario Lu\v{c}i\'c, Guodong Zhang, Wael Farhan, Michael Sharman, Paul Natsev, Paul Michel, Yamini Bansal, Siyuan Qiao, Kris Cao, Siamak Shakeri, Christina Butterfield, Justin Chung, Paul Kishan Rubenstein, Shivani Agrawal, Arthur Mensch, Kedar Soparkar, Karel Lenc, Timothy Chung, Aedan Pope, Loren Maggiore, Jackie Kay, Priya Jhakra, Shibo Wang, Joshua Maynez, Mary Phuong, Taylor Tobin, Andrea Tacchetti, Maja Trebacz, Kevin Robinson, Yash Katariya, Sebastian Riedel, Paige Bailey, Kefan Xiao, Nimesh Ghelani, Lora Aroyo, Ambrose Slone, Neil Houlsby, Xuehan Xiong, Zhen Yang, Elena Gribovskaya, Jonas Adler, Mateo Wirth, Lisa Lee, Music Li, Thais Kagohara, Jay Pavagadhi, Sophie Bridgers, Anna Bortsova, Sanjay Ghemawat, Zafarali Ahmed, Tianqi Liu, Richard Powell, Vijay Bolina, Mariko Iinuma, Polina Zablotskaia, James Besley, Da-Woon Chung, Timothy Dozat, Ramona Comanescu, Xiance Si, Jeremy Greer, Guolong Su, Martin Polacek, Rapha\"el Lopez Kaufman, Simon Tokumine, Hexiang Hu, Elena Buchatskaya, Yingjie Miao, Mohamed Elhawaty, Aditya Siddhant, Nenad Tomasev, Jinwei Xing, Christina Greer, Helen Miller, Shereen Ashraf, Aurko Roy, Zizhao Zhang, Ada Ma, Angelos Filos, Milos Besta, Rory Blevins, Ted Klimenko, Chih-Kuan Yeh, Soravit Changpinyo, Jiaqi Mu, Oscar Chang, Mantas Pajarskas, Carrie Muir, Vered Cohen, Charline Le Lan, Krishna Haridasan, Amit Marathe, Steven Hansen, Sholto Douglas, Rajkumar Samuel, Mingqiu Wang, Sophia Austin, Chang Lan, Jiepu Jiang, Justin Chiu, Jaime Alonso Lorenzo, Lars Lowe Sj\"osund, S\'ebastien Cevey, Zach Gleicher, Thi Avrahami, Anudhyan Boral, Hansa Srinivasan, Vittorio Selo, Rhys May, Konstantinos Aisopos, L\'eonard Hussenot, Livio Baldini Soares, Kate Baumli, Michael B. Chang, Adri\`a Recasens, Ben Caine, Alexander Pritzel, Filip Pavetic, Fabio Pardo, Anita Gergely, Justin Frye, Vinay Ramasesh, Dan Horgan, Kartikeya Badola, Nora Kassner, Subhrajit Roy, Ethan Dyer, V\'ictor Campos Campos, Alex Tomala, Yunhao Tang, Dalia El Badawy, Elspeth White, Basil Mustafa, Oran Lang, Abhishek Jindal, Sharad Vikram, Zhitao Gong, Sergi Caelles, Ross Hemsley, Gregory Thornton, Fangxiaoyu Feng, Wojciech Stokowiec, Ce Zheng, Phoebe Thacker, \c{C}a\u{g}lar \"Unl\"u, Zhishuai Zhang, Mohammad Saleh, James Svensson, Max Bileschi, Piyush Patil, Ankesh Anand, Roman Ring, Katerina Tsihlas, Arpi Vezer, Marco Selvi, Toby Shevlane, Mikel Rodriguez, Tom Kwiatkowski, Samira Daruki, Keran Rong, Allan Dafoe, Nicholas FitzGerald, Keren Gu-Lemberg, Mina Khan, Lisa Anne Hendricks, Marie Pellat, Vladimir Feinberg, James Cobon-Kerr, Tara Sainath, Maribeth Rauh, Sayed Hadi Hashemi, Richard Ives, Yana Hasson, Eric Noland, Yuan Cao, Nathan Byrd, Le Hou, Qingze Wang, Thibault Sottiaux, Michela Paganini, Jean-Baptiste Lespiau, Alexandre Moufarek, Samer Hassan, Kaushik Shivakumar, Joost van Amersfoort, Amol Mandhane, Pratik Joshi, Anirudh Goyal, Matthew Tung, Andrew Brock, Hannah Sheahan, Vedant Misra, Cheng Li, Nemanja Raki\'cevi\'c, Mostafa Dehghani, Fangyu Liu, Sid Mittal, Junhyuk Oh, Seb Noury, Eren Sezener, Fantine Huot, Matthew Lamm, Nicola De Cao, Charlie Chen, Sidharth Mudgal, Romina Stella, Kevin Brooks, Gautam Vasudevan, Chenxi Liu, Mainak Chain, Nivedita Melinkeri, Aaron Cohen, Venus Wang, Kristie Seymore, Sergey Zubkov, Rahul Goel, Summer Yue, Sai Krishnakumaran, Brian Albert, Nate Hurley, Motoki Sano, Anhad Mohananey, Jonah Joughin, Egor Filonov, Tomasz K\k{e}pa, Yomna Eldawy, Jiawern Lim, Rahul Rishi, Shirin Badiezadegan, Taylor Bos, Jerry Chang, Sanil Jain, Sri Gayatri Sundara Padmanabhan, Subha Puttagunta, Kalpesh Krishna, Leslie Baker, Norbert Kalb, Vamsi Bedapudi, Adam Kurzrok, Shuntong Lei, Anthony Yu, Oren Litvin, Xiang Zhou, Zhichun Wu, Sam Sobell, Andrea Siciliano, Alan Papir, Robby Neale, Jonas Bragagnolo, Tej Toor, Tina Chen, Valentin Anklin, Feiran Wang, Richie Feng, Milad Gholami, Kevin Ling, Lijuan Liu, Jules Walter, Hamid Moghaddam, Arun Kishore, Jakub Adamek, Tyler Mercado, Jonathan Mallinson, Siddhinita Wandekar, Stephen Cagle, Eran Ofek, Guillermo Garrido, Clemens Lombriser, Maksim Mukha, Botu Sun, Hafeezul Rahman Mohammad, Josip Matak, Yadi Qian, Vikas Peswani, Pawel Janus, Quan Yuan, Leif Schelin, Oana David, Ankur Garg, Yifan He, Oleksii Duzhyi, Anton \"Algmyr, Timoth\'ee Lottaz, Qi Li, Vikas Yadav, Luyao Xu, Alex Chinien, Rakesh Shivanna, Aleksandr Chuklin, Josie Li, Carrie Spadine, Travis Wolfe, Kareem Mohamed, Subhabrata Das, Zihang Dai, Kyle He, Daniel von Dincklage, Shyam Upadhyay, Akanksha Maurya, Luyan Chi, Sebastian Krause, Khalid Salama, Pam G Rabinovitch, Pavan Kumar Reddy M, Aarush Selvan, Mikhail Dektiarev, Golnaz Ghiasi, Erdem Guven, Himanshu Gupta, Boyi Liu, Deepak Sharma, Idan Heimlich Shtacher, Shachi Paul, Oscar Akerlund, Fran\c{c}ois-Xavier Aubet, Terry Huang, Chen Zhu, Eric Zhu, Elico Teixeira, Matthew Fritze, Francesco Bertolini, Liana-Eleonora Marinescu, Martin B\"olle, Dominik Paulus, Khyatti Gupta, Tejasi Latkar, Max Chang, Jason Sanders, Roopa Wilson, Xuewei Wu, Yi-Xuan Tan, Lam Nguyen Thiet, Tulsee Doshi, Sid Lall, Swaroop Mishra, Wanming Chen, Thang Luong, Seth Benjamin, Jasmine Lee, Ewa Andrejczuk, Dominik Rabiej, Vipul Ranjan, Krzysztof Styrc, Pengcheng Yin, Jon Simon, Malcolm Rose Harriott, Mudit Bansal, Alexei Robsky, Geoff Bacon, David Greene, Daniil Mirylenka, Chen Zhou, Obaid Sarvana, Abhimanyu Goyal, Samuel Andermatt, Patrick Siegler, Ben Horn, Assaf Israel, Francesco Pongetti, Chih-Wei "Louis" Chen, Marco Selvatici, Pedro Silva, Kathie Wang, Jackson Tolins, Kelvin Guu, Roey Yogev, Xiaochen Cai, Alessandro Agostini, Maulik Shah, Hung Nguyen, Noah \'O Donnaile, S\'ebastien Pereira, Linda Friso, Adam Stambler, Adam Kurzrok, Chenkai Kuang, Yan Romanikhin, Mark Geller, ZJ Yan, Kane Jang, Cheng-Chun Lee, Wojciech Fica, Eric Malmi, Qijun Tan, Dan Banica, Daniel Balle, Ryan Pham, Yanping Huang, Diana Avram, Hongzhi Shi, Jasjot Singh, Chris Hidey, Niharika Ahuja, Pranab Saxena, Dan Dooley, Srividya Pranavi Potharaju, Eileen O'Neill, Anand Gokulchandran, Ryan Foley, Kai Zhao, Mike Dusenberry, Yuan Liu, Pulkit Mehta, Ragha Kotikalapudi, Chalence Safranek-Shrader, Andrew Goodman, Joshua Kessinger, Eran Globen, Prateek Kolhar, Chris Gorgolewski, Ali Ibrahim, Yang Song, Ali Eichenbaum, Thomas Brovelli, Sahitya Potluri, Preethi Lahoti, Cip Baetu, Ali Ghorbani, Charles Chen, Andy Crawford, Shalini Pal, Mukund Sridhar, Petru Gurita, Asier Mujika, Igor Petrovski, Pierre-Louis Cedoz, Chenmei Li, Shiyuan Chen, Niccol\`o Dal Santo, Siddharth Goyal, Jitesh Punjabi, Karthik Kappaganthu, Chester Kwak, Pallavi LV, Sarmishta Velury, Himadri Choudhury, Jamie Hall, Premal Shah, Ricardo Figueira, Matt Thomas, Minjie Lu, Ting Zhou, Chintu Kumar, Thomas Jurdi, Sharat Chikkerur, Yenai Ma, Adams Yu, Soo Kwak, Victor \"Ahdel, Sujeevan Rajayogam, Travis Choma, Fei Liu, Aditya Barua, Colin Ji, Ji Ho Park, Vincent Hellendoorn, Alex Bailey, Taylan Bilal, Huanjie Zhou, Mehrdad Khatir, Charles Sutton, Wojciech Rzadkowski, Fiona Macintosh, Roopali Vij, Konstantin Shagin, Paul Medina, Chen Liang, Jinjing Zhou, Pararth Shah, Yingying Bi, Attila Dankovics, Shipra Banga, Sabine Lehmann, Marissa Bredesen, Zifan Lin, John Eric Hoffmann, Jonathan Lai, Raynald Chung, Kai Yang, Nihal Balani, Arthur Bra\v{z}inskas, Andrei Sozanschi, Matthew Hayes, H\'ector Fern\'andez Alcalde, Peter Makarov, Will Chen, Antonio Stella, Liselotte Snijders, Michael Mandl, Ante K\"arrman, Pawe{\l} Nowak, Xinyi Wu, Alex Dyck, Krishnan Vaidyanathan, Raghavender R, Jessica Mallet, Mitch Rudominer, Eric Johnston, Sushil Mittal, Akhil Udathu, Janara Christensen, Vishal Verma, Zach Irving, Andreas Santucci, Gamaleldin Elsayed, Elnaz Davoodi, Marin Georgiev, Ian Tenney, Nan Hua, Geoffrey Cideron, Edouard Leurent, Mahmoud Alnahlawi, Ionut Georgescu, Nan Wei, Ivy Zheng, Dylan Scandinaro, Heinrich Jiang, Jasper Snoek, Mukund Sundararajan, Xuezhi Wang, Zack Ontiveros, Itay Karo, Jeremy Cole, Vinu Rajashekhar, Lara Tumeh, Eyal Ben-David, Rishub Jain, Jonathan Uesato, Romina Datta, Oskar Bunyan, Shimu Wu, John Zhang, Piotr Stanczyk, Ye Zhang, David Steiner, Subhajit Naskar, Michael Azzam, Matthew Johnson, Adam Paszke, Chung-Cheng Chiu, Jaume Sanchez Elias, Afroz Mohiuddin, Faizan Muhammad, Jin Miao, Andrew Lee, Nino Vieillard, Jane Park, Jiageng Zhang, Jeff Stanway, Drew Garmon, Abhijit Karmarkar, Zhe Dong, Jong Lee, Aviral Kumar, Luowei Zhou, Jonathan Evens, William Isaac, Geoffrey Irving, Edward Loper, Michael Fink, Isha Arkatkar, Nanxin Chen, Izhak Shafran, Ivan Petrychenko, Zhe Chen, Johnson Jia, Anselm Levskaya, Zhenkai Zhu, Peter Grabowski, Yu Mao, Alberto Magni, Kaisheng Yao, Javier Snaider, Norman Casagrande, Evan Palmer, Paul Suganthan, Alfonso Casta\~no, Irene Giannoumis, Wooyeol Kim, Miko{\l}aj Rybi\'nski, Ashwin Sreevatsa, Jennifer Prendki, David Soergel, Adrian Goedeckemeyer, Willi Gierke, Mohsen Jafari, Meenu Gaba, Jeremy Wiesner, Diana Gage Wright, Yawen Wei, Harsha Vashisht, Yana Kulizhskaya, Jay Hoover, Maigo Le, Lu Li, Chimezie Iwuanyanwu, Lu Liu, Kevin Ramirez, Andrey Khorlin, Albert Cui, Tian LIN, Marcus Wu, Ricardo Aguilar, Keith Pallo, Abhishek Chakladar, Ginger Perng, Elena Allica Abellan, Mingyang Zhang, Ishita Dasgupta, Nate Kushman, Ivo Penchev, Alena Repina, Xihui Wu, Tom van der Weide, Priya Ponnapalli, Caroline Kaplan, Jiri Simsa, Shuangfeng Li, Olivier Dousse, Fan Yang, Jeff Piper, Nathan Ie, Rama Pasumarthi, Nathan Lintz, Anitha Vijayakumar, Daniel Andor, Pedro Valenzuela, Minnie Lui, Cosmin Paduraru, Daiyi Peng, Katherine Lee, Shuyuan Zhang, Somer Greene, Duc Dung Nguyen, Paula Kurylowicz, Cassidy Hardin, Lucas Dixon, Lili Janzer, Kiam Choo, Ziqiang Feng, Biao Zhang, Achintya Singhal, Dayou Du, Dan McKinnon, Natasha Antropova, Tolga Bolukbasi, Orgad Keller, David Reid, Daniel Finchelstein, Maria Abi Raad, Remi Crocker, Peter Hawkins, Robert Dadashi, Colin Gaffney, Ken Franko, Anna Bulanova, R\'emi Leblond, Shirley Chung, Harry Askham, Luis C. Cobo, Kelvin Xu, Felix Fischer, Jun Xu, Christina Sorokin, Chris Alberti, Chu-Cheng Lin, Colin Evans, Alek Dimitriev, Hannah Forbes, Dylan Banarse, Zora Tung, Mark Omernick, Colton Bishop, Rachel Sterneck, Rohan Jain, Jiawei Xia, Ehsan Amid, Francesco Piccinno, Xingyu Wang, Praseem Banzal, Daniel J. Mankowitz, Alex Polozov, Victoria Krakovna, Sasha Brown, MohammadHossein Bateni, Dennis Duan, Vlad Firoiu, Meghana Thotakuri, Tom Natan, Matthieu Geist, Ser tan Girgin, Hui Li, Jiayu Ye, Ofir Roval, Reiko Tojo, Michael Kwong, James Lee-Thorp, Christopher Yew, Danila Sinopalnikov, Sabela Ramos, John Mellor, Abhishek Sharma, Kathy Wu, David Miller, Nicolas Sonnerat, Denis Vnukov, Rory Greig, Jennifer Beattie, Emily Caveness, Libin Bai, Julian Eisenschlos, Alex Korchemniy, Tomy Tsai, Mimi Jasarevic, Weize Kong, Phuong Dao, Zeyu Zheng, Frederick Liu, Fan Yang, Rui Zhu, Tian Huey Teh, Jason Sanmiya, Evgeny Gladchenko, Nejc Trdin, Daniel Toyama, Evan Rosen, Sasan Tavakkol, Linting Xue, Chen Elkind, Oliver Woodman, John Carpenter, George Papamakarios, Rupert Kemp, Sushant Kafle, Tanya Grunina, Rishika Sinha, Alice Talbert, Diane Wu, Denese Owusu-Afriyie, Cosmo Du, Chloe Thornton, Jordi Pont-Tuset, Pradyumna Narayana, Jing Li, Saaber Fatehi, John Wieting, Omar Ajmeri, Benigno Uria, Yeongil Ko, Laura Knight, Am\'elie H\'eliou, Ning Niu, Shane Gu, Chenxi Pang, Yeqing Li, Nir Levine, Ariel Stolovich, Rebeca Santamaria-Fernandez, Sonam Goenka, Wenny Yustalim, Robin Strudel, Ali Elqursh, Charlie Deck, Hyo Lee, Zonglin Li, Kyle Levin, Raphael Hoffmann, Dan Holtmann-Rice, Olivier Bachem, Sho Arora, Christy Koh, Soheil Hassas Yeganeh, Siim P\~oder, Mukarram Tariq, Yanhua Sun, Lucian Ionita, Mojtaba Seyedhosseini, Pouya Tafti, Zhiyu Liu, Anmol Gulati, Jasmine Liu, Xinyu Ye, Bart Chrzaszcz, Lily Wang, Nikhil Sethi, Tianrun Li, Ben Brown, Shreya Singh, Wei Fan, Aaron Parisi, Joe Stanton, Vinod Koverkathu, Christopher A. Choquette-Choo, Yunjie Li, TJ Lu, Abe Ittycheriah, Prakash Shroff, Mani Varadarajan, Sanaz Bahargam, Rob Willoughby, David Gaddy, Guillaume Desjardins, Marco Cornero, Brona Robenek, Bhavishya Mittal, Ben Albrecht, Ashish Shenoy, Fedor Moiseev, Henrik Jacobsson, Alireza Ghaffarkhah, Morgane Rivi\`ere, Alanna Walton, Cl\'ement Crepy, Alicia Parrish, Zongwei Zhou, Clement Farabet, Carey Radebaugh, Praveen Srinivasan, Claudia van der Salm, Andreas Fidjeland, Salvatore Scellato, Eri Latorre-Chimoto, Hanna Klimczak-Pluci\'nska, David Bridson, Dario de Cesare, Tom Hudson, Piermaria Mendolicchio, Lexi Walker, Alex Morris, Matthew Mauger, Alexey Guseynov, Alison Reid, Seth Odoom, Lucia Loher, Victor Cotruta, Madhavi Yenugula, Dominik Grewe, Anastasia Petrushkina, Tom Duerig, Antonio Sanchez, Steve Yadlowsky, Amy Shen, Amir Globerson, Lynette Webb, Sahil Dua, Dong Li, Surya Bhupatiraju, Dan Hurt, Haroon Qureshi, Ananth Agarwal, Tomer Shani, Matan Eyal, Anuj Khare, Shreyas Rammohan Belle, Lei Wang, Chetan Tekur, Mihir Sanjay Kale, Jinliang Wei, Ruoxin Sang, Brennan Saeta, Tyler Liechty, Yi Sun, Yao Zhao, Stephan Lee, Pandu Nayak, Doug Fritz, Manish Reddy Vuyyuru, John Aslanides, Nidhi Vyas, Martin Wicke, Xiao Ma, Evgenii Eltyshev, Nina Martin, Hardie Cate, James Manyika, Keyvan Amiri, Yelin Kim, Xi Xiong, Kai Kang, Florian Luisier, Nilesh Tripuraneni, David Madras, Mandy Guo, Austin Waters, Oliver Wang, Joshua Ainslie, Jason Baldridge, Han Zhang, Garima Pruthi, Jakob Bauer, Feng Yang, Riham Mansour, Jason Gelman, Yang Xu, George Polovets, Ji Liu, Honglong Cai, Warren Chen, XiangHai Sheng, Emily Xue, Sherjil Ozair, Christof Angermueller, Xiaowei Li, Anoop Sinha, Weiren Wang, Julia Wiesinger, Emmanouil Koukoumidis, Yuan Tian, Anand Iyer, Madhu Gurumurthy, Mark Goldenson, Parashar Shah, MK Blake, Hongkun Yu, Anthony Urbanowicz, Jennimaria Palomaki, Chrisantha Fernando, Ken Durden, Harsh Mehta, Nikola Momchev, Elahe Rahimtoroghi, Maria Georgaki, Amit Raul, Sebastian Ruder, Morgan Redshaw, Jinhyuk Lee, Denny Zhou, Komal Jalan, Dinghua Li, Blake Hechtman, Parker Schuh, Milad Nasr, Kieran Milan, Vladimir Mikulik, Juliana Franco, Tim Green, Nam Nguyen, Joe Kelley, Aroma Mahendru, Andrea Hu, Joshua Howland, Ben Vargas, Jeffrey Hui, Kshitij Bansal, Vikram Rao, Rakesh Ghiya, Emma Wang, Ke Ye, Jean Michel Sarr, Melanie Moranski Preston, Madeleine Elish, Steve Li, Aakash Kaku, Jigar Gupta, Ice Pasupat, Da-Cheng Juan, Milan Someswar, Tejvi M., Xinyun Chen, Aida Amini, Alex Fabrikant, Eric Chu, Xuanyi Dong, Amruta Muthal, Senaka Buthpitiya, Sarthak Jauhari, Nan Hua, Urvashi Khandelwal, Ayal Hitron, Jie Ren, Larissa Rinaldi, Shahar Drath, Avigail Dabush, Nan-Jiang Jiang, Harshal Godhia, Uli Sachs, Anthony Chen, Yicheng Fan, Hagai Taitelbaum, Hila Noga, Zhuyun Dai, James Wang, Chen Liang, Jenny Hamer, Chun-Sung Ferng, Chenel Elkind, Aviel Atias, Paulina Lee, V\'it List\'ik, Mathias Carlen, Jan van de Kerkhof, Marcin Pikus, Krunoslav Zaher, Paul M\"uller, Sasha Zykova, Richard Stefanec, Vitaly Gatsko, Christoph Hirnschall, Ashwin Sethi, Xingyu Federico Xu, Chetan Ahuja, Beth Tsai, Anca Stefanoiu, Bo Feng, Keshav Dhandhania, Manish Katyal, Akshay Gupta, Atharva Parulekar, Divya Pitta, Jing Zhao, Vivaan Bhatia, Yashodha Bhavnani, Omar Alhadlaq, Xiaolin Li, Peter Danenberg, Dennis Tu, Alex Pine, Vera Filippova, Abhipso Ghosh, Ben Limonchik, Bhargava Urala, Chaitanya Krishna Lanka, Derik Clive, Yi Sun, Edward Li, Hao Wu, Kevin Hongtongsak, Ianna Li, Kalind Thakkar, Kuanysh Omarov, Kushal Majmundar, Michael Alverson, Michael Kucharski, Mohak Patel, Mudit Jain, Maksim Zabelin, Paolo Pelagatti, Rohan Kohli, Saurabh Kumar, Joseph Kim, Swetha Sankar, Vineet Shah, Lakshmi Ramachandruni, Xiangkai Zeng, Ben Bariach, Laura Weidinger, Tu Vu, Alek Andreev, Antoine He, Kevin Hui, Sheleem Kashem, Amar Subramanya, Sissie Hsiao, Demis Hassabis, Koray Kavukcuoglu, Adam Sadovsky, Quoc Le, Trevor Strohman, Yonghui Wu, Slav Petrov, Jeffrey Dean, Oriol Vinyals
Abstract: This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI.
Authors: Yuhang Liu, Zhen Zhang, Dong Gong, Erdun Gao, Biwei Huang, Mingming Gong, Anton van den Hengel, Kun Zhang, Javen Qinfeng Shi
Abstract: Directed acyclic graphs (DAGs) are fundamental graph structures in causal modeling, but identifying the desired DAG from observational data often requires strong assumptions that may not hold in real-world scenarios, especially for latent causal models and complex multimodal data. This raises the question of whether we can relax or bypass the DAG assumption while maintaining practical utility. In this work, we propose a novel latent partial causal model for multimodal data, featuring two latent coupled variables, connected by an undirected edge, to represent the transfer of knowledge across modalities. Under specific statistical assumptions, we establish an identifiability result, demonstrating that representations learned by multimodal contrastive learning correspond to the latent coupled variables up to a trivial transformation. This result deepens our understanding of the why multimodal contrastive learning works, highlights its potential for disentanglement, and expands the utility of pre-trained models like CLIP. Synthetic experiments confirm the robustness of our findings, even when the assumptions are partially violated. Most importantly, experiments on a pre-trained CLIP model embodies disentangled representations, enabling few-shot learning and improving domain generalization across diverse real-world datasets. Together, these contributions push the boundaries of multimodal contrastive learning, both theoretically and, crucially, in practical applications.
Authors: Antonio Emanuele Cin\`a, J\'er\^ome Rony, Maura Pintor, Luca Demetrio, Ambra Demontis, Battista Biggio, Ismail Ben Ayed, Fabio Roli
Abstract: Adversarial examples are typically optimized with gradient-based attacks. While novel attacks are continuously proposed, each is shown to outperform its predecessors using different experimental setups, hyperparameter settings, and number of forward and backward calls to the target models. This provides overly-optimistic and even biased evaluations that may unfairly favor one particular attack over the others. In this work, we aim to overcome these limitations by proposing AttackBench, i.e., the first evaluation framework that enables a fair comparison among different attacks. To this end, we first propose a categorization of gradient-based attacks, identifying their main components and differences. We then introduce our framework, which evaluates their effectiveness and efficiency. We measure these characteristics by (i) defining an optimality metric that quantifies how close an attack is to the optimal solution, and (ii) limiting the number of forward and backward queries to the model, such that all attacks are compared within a given maximum query budget. Our extensive experimental analysis compares more than $100$ attack implementations with a total of over $800$ different configurations against CIFAR-10 and ImageNet models, highlighting that only very few attacks outperform all the competing approaches. Within this analysis, we shed light on several implementation issues that prevent many attacks from finding better solutions or running at all. We release AttackBench as a publicly-available benchmark, aiming to continuously update it to include and evaluate novel gradient-based attacks for optimizing adversarial examples.
Authors: Shubham Bharti, Shiyun Cheng, Jihyun Rho, Jianrui Zhang, Mu Cai, Yong Jae Lee, Martina Rau, Xiaojin Zhu
Abstract: We introduce CHARTOM, a visual theory-of-mind benchmark for multimodal large language models. CHARTOM consists of specially designed data visualizing charts. Given a chart, a language model needs to not only correctly comprehend the chart (the FACT question) but also judge if the chart will be misleading to a human reader (the MIND question). Both questions have significant societal benefits. We detail the construction of the CHARTOM benchmark including its calibration on human performance. We benchmark leading LLMs as of late 2024 - including GPT, Claude, Gemini, Qwen, Llama, and Llava - on the CHARTOM dataset and found that our benchmark was challenging to all of them, suggesting room for future large language models to improve.
Authors: Peng Wang, Huijie Zhang, Zekai Zhang, Siyi Chen, Yi Ma, Qing Qu
Abstract: Recent empirical studies have demonstrated that diffusion models can effectively learn the image distribution and generate new samples. Remarkably, these models can achieve this even with a small number of training samples despite a large image dimension, circumventing the curse of dimensionality. In this work, we provide theoretical insights into this phenomenon by leveraging key empirical observations: (i) the low intrinsic dimensionality of image data, (ii) a union of manifold structure of image data, and (iii) the low-rank property of the denoising autoencoder in trained diffusion models. These observations motivate us to assume the underlying data distribution of image data as a mixture of low-rank Gaussians and to parameterize the denoising autoencoder as a low-rank model according to the score function of the assumed distribution. With these setups, we rigorously show that optimizing the training loss of diffusion models is equivalent to solving the canonical subspace clustering problem over the training samples. Based on this equivalence, we further show that the minimal number of samples required to learn the underlying distribution scales linearly with the intrinsic dimensions under the above data and model assumptions. This insight sheds light on why diffusion models can break the curse of dimensionality and exhibit the phase transition in learning distributions. Moreover, we empirically establish a correspondence between the subspaces and the semantic representations of image data, facilitating image editing. We validate these results with corroborated experimental results on both simulated distributions and image datasets.
Authors: Guangming Huang, Yunfei Long, Cunjin Luo
Abstract: Supervised contrastive learning has achieved remarkable success by leveraging label information; however, determining positive samples in multi-label scenarios remains a critical challenge. In multi-label supervised contrastive learning (MSCL), multi-label relations are not yet fully defined, leading to ambiguity in identifying positive samples and formulating contrastive loss functions to construct the representation space. To address these challenges, we: (i) first define five distinct multi-label relations in MSCL to systematically identify positive samples, (ii) introduce a novel Similarity-Dissimilarity Loss that dynamically re-weights samples through computing the similarity and dissimilarity factors between positive samples and given anchors based on multi-label relations, and (iii) further provide theoretical grounded proofs for our method through rigorous mathematical analysis that supports the formulation and effectiveness of the proposed loss function. We conduct the experiments across both image and text modalities, and extend the evaluation to medical domain. The results demonstrate that our method consistently outperforms baselines in a comprehensive evaluation, confirming its effectiveness and robustness. Code is available at: https://github.com/guangminghuang/similarity-dissimilarity-loss.
URLs: https://github.com/guangminghuang/similarity-dissimilarity-loss.
Authors: Zhilin Zhang, Jie Wang, Zhanghao Qin, Ruiqi Zhu, Xiaoliang Gong
Abstract: Medical Visual Question Answering (MedVQA) has attracted growing interest at the intersection of medical image understanding and natural language processing for clinical applications. By interpreting medical images and providing precise answers to relevant clinical inquiries, MedVQA has the potential to support diagnostic decision-making and reduce workload across various fields like radiology. While recent approaches rely heavily on unified large pre-trained Visual-Language Models, research on more efficient fusion mechanisms remains relatively limited in this domain. In this paper, we introduce a fusion model, OMniBAN, that integrates Orthogonality loss, Multi-head attention, and a Bilinear Attention Network to achieve high computational efficiency as well as solid performance. We conduct comprehensive experiments and demonstrate how bilinear attention fusion can approximate the performance of larger fusion models like cross-modal Transformer. Our results show that OMniBAN requires fewer parameters (approximately 2/3 of Transformer-based Co-Attention) and substantially lower FLOPs (approximately 1/4), while achieving comparable overall performance and even slight improvements on closed-ended questions on two key MedVQA benchmarks. This balance between efficiency and accuracy suggests that OMniBAN could be a viable option for real-world medical image question answering, where computational resources are often constrained.
Authors: Tony Lindeberg
Abstract: When observing the surface patterns of objects delimited by smooth surfaces, the projections of the surface patterns to the image domain will be subject to substantial variabilities, as induced by variabilities in the geometric viewing conditions, and as generated by either monocular or binocular imaging conditions, or by relative motions between the object and the observer over time. To first order of approximation, the image deformations of such projected surface patterns can be modelled as local linearizations in terms of local 2-D spatial affine transformations. This paper presents a theoretical analysis of relationships between the degrees of freedom in 2-D spatial affine image transformations and the degrees of freedom in the affine Gaussian derivative model for visual receptive fields. For this purpose, we first describe a canonical decomposition of 2-D affine transformations on a product form, closely related to a singular value decomposition, while in closed form, and which reveals the degrees of freedom in terms of (i) uniform scaling transformations, (ii) an overall amount of global rotation, (iii) a complementary non-uniform scaling transformation and (iv) a relative normalization to a preferred symmetry orientation in the image domain. Then, we show how these degrees of freedom relate to the degrees of freedom in the affine Gaussian derivative model. Finally, we use these theoretical results to consider whether we could regard the biological receptive fields in the primary visual cortex of higher mammals as being able to span the degrees of freedom of 2-D spatial affine transformations, based on interpretations of existing neurophysiological experimental results.
Authors: Yifan Zhu, Tianyi Xiang, Aaron Dollar, Zherong Pan
Abstract: Identifying predictive world models for robots in novel environments from sparse online observations is essential for robot task planning and execution in novel environments. However, existing methods that leverage differentiable programming to identify world models are incapable of jointly optimizing the geometry, appearance, and physical properties of the scene. In this work, we introduce a novel rigid object representation that allows the joint identification of these properties. Our method employs a novel differentiable point-based geometry representation coupled with a grid-based appearance field, which allows differentiable object collision detection and rendering. Combined with a differentiable physical simulator, we achieve end-to-end optimization of world models, given the sparse visual and tactile observations of a physical motion sequence. Through a series of world model identification tasks in simulated and real environments, we show that our method can learn both simulation- and rendering-ready world models from only one robot action sequence. The code and additional videos are available at our project website: https://tianyi20.github.io/rigid-world-model.github.io/
URLs: https://tianyi20.github.io/rigid-world-model.github.io/
Authors: Hyesu Jang, Wooseong Yang, Hanguen Kim, Dongje Lee, Yongjin Kim, Jinbum Park, Minsoo Jeon, Jaeseong Koh, Yejin Kang, Minwoo Jung, Sangwoo Jung, Chng Zhen Hao, Wong Yu Hin, Chew Yihang, Ayoung Kim
Abstract: Maritime environmental sensing requires overcoming challenges from complex conditions such as harsh weather, platform perturbations, large dynamic objects, and the requirement for long detection ranges. While cameras and LiDAR are commonly used in ground vehicle navigation, their applicability in maritime settings is limited by range constraints and hardware maintenance issues. Radar sensors, however, offer robust long-range detection capabilities and resilience to physical contamination from weather and saline conditions, making it a powerful sensor for maritime navigation. Among various radar types, X-band radar is widely employed for maritime vessel navigation, providing effective long-range detection essential for situational awareness and collision avoidance. Nevertheless, it exhibits limitations during berthing operations where near-field detection is critical. To address this shortcoming, we incorporate W-band radar, which excels in detecting nearby objects with a higher update rate. We present a comprehensive maritime sensor dataset featuring multi-range detection capabilities. This dataset integrates short-range LiDAR data, medium-range W-band radar data, and long-range X-band radar data into a unified framework. Additionally, it includes object labels for oceanic object detection usage, derived from radar and stereo camera images. The dataset comprises seven sequences collected from diverse regions with varying levels of \bl{navigation algorithm} estimation difficulty, ranging from easy to challenging, and includes common locations suitable for global localization tasks. This dataset serves as a valuable resource for advancing research in place recognition, odometry estimation, SLAM, object detection, and dynamic object elimination within maritime environments. Dataset can be found at https://sites.google.com/view/rpmmoana.
Authors: Yik Lung Pang, Alessio Xompero, Changjae Oh, Andrea Cavallaro
Abstract: Jointly estimating hand and object shape facilitates the grasping task in human-to-robot handovers. However, relying on hand-crafted prior knowledge about the geometric structure of the object fails when generalising to unseen objects, and depth sensors fail to detect transparent objects such as drinking glasses. In this work, we propose a stereo-based method for hand-object reconstruction that combines single-view reconstructions probabilistically to form a coherent stereo reconstruction. We learn 3D shape priors from a large synthetic hand-object dataset to ensure that our method is generalisable, and use RGB inputs to better capture transparent objects. We show that our method reduces the object Chamfer distance compared to existing RGB based hand-object reconstruction methods on single view and stereo settings. We process the reconstructed hand-object shape with a projection-based outlier removal step and use the output to guide a human-to-robot handover pipeline with wide-baseline stereo RGB cameras. Our hand-object reconstruction enables a robot to successfully receive a diverse range of household objects from the human.
Authors: Maja Pavlovic
Abstract: To be considered reliable, a model must be calibrated so that its confidence in each decision closely reflects its true outcome. In this blogpost we'll take a look at the most commonly used definition for calibration and then dive into a frequently used evaluation measure for model calibration. We'll then cover some of the drawbacks of this measure and how these surfaced the need for additional notions of calibration, which require their own new evaluation measures. This post is not intended to be an in-depth dissection of all works on calibration, nor does it focus on how to calibrate models. Instead, it is meant to provide a gentle introduction to the different notions and their evaluation measures as well as to re-highlight some issues with a measure that is still widely used to evaluate calibration.
Authors: Yi Li, Yuquan Deng, Jesse Zhang, Joel Jang, Marius Memmel, Raymond Yu, Caelan Reed Garrett, Fabio Ramos, Dieter Fox, Anqi Li, Abhishek Gupta, Ankit Goyal
Abstract: Large foundation models have shown strong open-world generalization to complex problems in vision and language, but similar levels of generalization have yet to be achieved in robotics. One fundamental challenge is the lack of robotic data, which are typically obtained through expensive on-robot operation. A promising remedy is to leverage cheaper, off-domain data such as action-free videos, hand-drawn sketches or simulation data. In this work, we posit that hierarchical vision-language-action (VLA) models can be more effective in utilizing off-domain data than standard monolithic VLA models that directly finetune vision-language models (VLMs) to predict actions. In particular, we study a class of hierarchical VLA models, where the high-level VLM is finetuned to produce a coarse 2D path indicating the desired robot end-effector trajectory given an RGB image and a task description. The intermediate 2D path prediction is then served as guidance to the low-level, 3D-aware control policy capable of precise manipulation. Doing so alleviates the high-level VLM from fine-grained action prediction, while reducing the low-level policy's burden on complex task-level reasoning. We show that, with the hierarchical design, the high-level VLM can transfer across significant domain gaps between the off-domain finetuning data and real-robot testing scenarios, including differences on embodiments, dynamics, visual appearances and task semantics, etc. In the real-robot experiments, we observe an average of 20% improvement in success rate across seven different axes of generalization over OpenVLA, representing a 50% relative gain. Visual results, code, and dataset are provided at: https://hamster-robot.github.io/
Authors: Shenyuan Gao, Siyuan Zhou, Yilun Du, Jun Zhang, Chuang Gan
Abstract: World models aim to learn action-controlled future prediction and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this limitation, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
Authors: Matvei Popov, Aymen Kallala, Anirudha Ramesh, Narimane Hennouni, Shivesh Khaitan, Rick Gentry, Alain-Sam Cohen
Abstract: Long-range dependencies are critical for understanding genomic structure and function, yet most conventional methods struggle with them. Widely adopted transformer-based models, while excelling at short-context tasks, are limited by the attention module's quadratic computational complexity and inability to extrapolate to sequences longer than those seen in training. In this work, we explore State Space Models (SSMs) as a promising alternative by benchmarking two SSM-inspired architectures, Caduceus and Hawk, on long-range genomics modeling tasks under conditions parallel to a 50M parameter transformer baseline. We discover that SSMs match transformer performance and exhibit impressive zero-shot extrapolation across multiple tasks, handling contexts 10 to 100 times longer than those seen during training, indicating more generalizable representations better suited for modeling the long and complex human genome. Moreover, we demonstrate that these models can efficiently process sequences of 1M tokens on a single GPU, allowing for modeling entire genomic regions at once, even in labs with limited compute. Our findings establish SSMs as efficient and scalable for long-context genomic analysis.
Authors: Yilin Liu, Duoteng Xu, Xingyao Yu, Xiang Xu, Daniel Cohen-Or, Hao Zhang, Hui Huang
Abstract: We introduce a novel representation for learning and generating Computer-Aided Design (CAD) models in the form of $\textit{boundary representations}$ (B-Reps). Our representation unifies the continuous geometric properties of B-Rep primitives in different orders (e.g., surfaces and curves) and their discrete topological relations in a $\textit{holistic latent}$ (HoLa) space. This is based on the simple observation that the topological connection between two surfaces is intrinsically tied to the geometry of their intersecting curve. Such a prior allows us to reformulate topology learning in B-Reps as a geometric reconstruction problem in Euclidean space. Specifically, we eliminate the presence of curves, vertices, and all the topological connections in the latent space by learning to distinguish and derive curve geometries from a pair of surface primitives via a neural intersection network. To this end, our holistic latent space is only defined on surfaces but encodes a full B-Rep model, including the geometry of surfaces, curves, vertices, and their topological relations. Our compact and holistic latent space facilitates the design of a first diffusion-based generator to take on a large variety of inputs including point clouds, single/multi-view images, 2D sketches, and text prompts. Our method significantly reduces ambiguities, redundancies, and incoherences among the generated B-Rep primitives, as well as training complexities inherent in prior multi-step B-Rep learning pipelines, while achieving greatly improved validity rate over current state of the art: 82% vs. $\approx$50%.
Authors: Huadai Liu, Tianyi Luo, Qikai Jiang, Kaicheng Luo, Peiwen Sun, Jialei Wan, Rongjie Huang, Qian Chen, Wen Wang, Xiangtai Li, Shiliang Zhang, Zhijie Yan, Zhou Zhao, Wei Xue
Abstract: Traditional video-to-audio generation techniques primarily focus on field-of-view (FoV) video and non-spatial audio, often missing the spatial cues necessary for accurately representing sound sources in 3D environments. To address this limitation, we introduce a novel task, 360V2SA, to generate spatial audio from 360-degree videos, specifically producing First-order Ambisonics (FOA) audio - a standard format for representing 3D spatial audio that captures sound directionality and enables realistic 3D audio reproduction. We first create Sphere360, a novel dataset tailored for this task that is curated from real-world data. We also design an efficient semi-automated pipeline for collecting and cleaning paired video-audio data. To generate spatial audio from 360-degree video, we propose a novel framework OmniAudio, which leverages self-supervised pre-training using both spatial audio data (in FOA format) and large-scale non-spatial data. Furthermore, OmniAudio features a dual-branch framework that utilizes both panoramic and FoV video inputs to capture comprehensive local and global information from 360-degree videos. Experimental results demonstrate that OmniAudio achieves state-of-the-art performance across both objective and subjective metrics on Sphere360. Code and datasets will be released at https://github.com/liuhuadai/OmniAudio. The demo page is available at https://OmniAudio-360V2SA.github.io.
URLs: https://github.com/liuhuadai/OmniAudio., https://OmniAudio-360V2SA.github.io.
Authors: Tristan Manchester, Adam Anders, Julio Spadotto, Hannah Eccleston, William Beavan, Hugues Arcis, Brian J. Connolly
Abstract: In situ synchrotron X-ray computed tomography enables dynamic material studies, but automated segmentation remains challenging due to complex imaging artefacts and limited training data. We present a methodology for deep learning-based segmentation by transforming high-quality ex situ laboratory data to train models for binary segmentation of in situ synchrotron data, demonstrated through copper oxide dissolution studies. Using a modified SegFormer architecture, our approach achieves high segmentation performance on unseen data while reducing processing time from hours to seconds per 3D dataset. The method maintains consistent performance over significant morphological changes during experiments, despite training only on static specimens. This methodology can be readily applied to diverse materials systems, accelerating the analysis of time-resolved tomographic data across scientific disciplines.
Authors: Xiaoman Zhang, Juli\'an N. Acosta, Josh Miller, Ouwen Huang, Pranav Rajpurkar
Abstract: We present ReXGradient-160K, representing the largest publicly available chest X-ray dataset to date in terms of the number of patients. This dataset contains 160,000 chest X-ray studies with paired radiological reports from 109,487 unique patients across 3 U.S. health systems (79 medical sites). This comprehensive dataset includes multiple images per study and detailed radiology reports, making it particularly valuable for the development and evaluation of AI systems for medical imaging and automated report generation models. The dataset is divided into training (140,000 studies), validation (10,000 studies), and public test (10,000 studies) sets, with an additional private test set (10,000 studies) reserved for model evaluation on the ReXrank benchmark. By providing this extensive dataset, we aim to accelerate research in medical imaging AI and advance the state-of-the-art in automated radiological analysis. Our dataset will be open-sourced at https://huggingface.co/datasets/rajpurkarlab/ReXGradient-160K.
URLs: https://huggingface.co/datasets/rajpurkarlab/ReXGradient-160K.
Authors: Xinmu Wang, Xiang Gao, Xiyun Song, Heather Yu, Zongfang Lin, Liang Peng, Xianfeng Gu
Abstract: Animating 3D head meshes using audio inputs has significant applications in AR/VR, gaming, and entertainment through 3D avatars. However, bridging the modality gap between speech signals and facial dynamics remains a challenge, often resulting in incorrect lip syncing and unnatural facial movements. To address this, we propose OT-Talk, the first approach to leverage optimal transportation to optimize the learning model in talking head animation. Building on existing learning frameworks, we utilize a pre-trained Hubert model to extract audio features and a transformer model to process temporal sequences. Unlike previous methods that focus solely on vertex coordinates or displacements, we introduce Chebyshev Graph Convolution to extract geometric features from triangulated meshes. To measure mesh dissimilarities, we go beyond traditional mesh reconstruction errors and velocity differences between adjacent frames. Instead, we represent meshes as probability measures and approximate their surfaces. This allows us to leverage the sliced Wasserstein distance for modeling mesh variations. This approach facilitates the learning of smooth and accurate facial motions, resulting in coherent and natural facial animations. Our experiments on two public audio-mesh datasets demonstrate that our method outperforms state-of-the-art techniques both quantitatively and qualitatively in terms of mesh reconstruction accuracy and temporal alignment. In addition, we conducted a user perception study with 20 volunteers to further assess the effectiveness of our approach.
Authors: Bobo Lian, Dandan Wang, Chenjian Wu, Minxin Chen
Abstract: Point cloud surface representation is a fundamental problem in computer graphics and vision. This paper presents a machine learning approach for approximating the signed distance function (SDF) of a point cloud using a sparse ellipsoidal radial basis function network, enabling a compact and accurate surface representation. Given the SDF values defined on the grid points constructed from the point cloud, our method approximates the SDF accurately with as few ellipsoidal radial basis functions (ERBFs) as possible, i.e., represents the SDF of a point cloud by sparse ERBFs. To balance sparsity and approximation precision, a dynamic multi-objective optimization strategy is introduced, which adaptively adds the regularization terms and jointly optimizes the weights, centers, shapes, and orientations of ERBFs. To improve computational efficiency, a nearest-neighbor-based data structure is employed, restricting function calculations to points near each Gaussian kernel center. The computations for each kernel are further parallelized on CUDA, which significantly improves the optimization speed. Additionally, a hierarchical octree-based refinement strategy is designed for training. Specifically, the initialization and optimization of network parameters are conducted using coarse grid points in the octree lattice structure. Subsequently, fine lattice points are progressively incorporated to accelerate model convergence and enhance training efficiency. Extensive experiments on multiple benchmark datasets demonstrate that our method outperforms previous sparse representation approaches in terms of accuracy, robustness, and computational efficiency. The corresponding executable program is publicly available at https://github.com/lianbobo/SE-RBFNet.git.
Authors: Fei Zhao, Chengcui Zhang, Runlin Zhang, Tianyang Wang, Xi Li
Abstract: Hallucinations in vision-language models (VLMs) hinder reliability and real-world applicability, usually stemming from distribution shifts between pretraining data and test samples. Existing solutions, such as retraining or fine-tuning on additional data, demand significant computational resources and labor-intensive data collection, while ensemble-based methods incur additional costs by introducing auxiliary VLMs. To address these challenges, we propose a novel test-time adaptation framework using reinforcement learning to mitigate hallucinations during inference without retraining or any auxiliary VLMs. By updating only the learnable parameters in the layer normalization of the language model (approximately 0.003% of the model parameters), our method reduces distribution shifts between test samples and pretraining samples. A CLIP-based hallucination evaluation model is proposed to provide dual rewards to VLMs. Experimental results demonstrate a 15.4% and 17.3% reduction in hallucination rates on LLaVA and InstructBLIP, respectively. Our approach outperforms state-of-the-art baselines with a 68.3% improvement in hallucination mitigation, demonstrating its effectiveness.
Authors: Solene Debuysere, Nicolas Trouve, Nathan Letheule, Olivier Leveque, Elise Colin
Abstract: The availability of Synthetic Aperture Radar (SAR) satellite imagery has increased considerably in recent years, with datasets commercially available. However, the acquisition of high-resolution SAR images in airborne configurations, remains costly and limited. Thus, the lack of open source, well-labeled, or easily exploitable SAR text-image datasets is a barrier to the use of existing foundation models in remote sensing applications. In this context, synthetic image generation is a promising solution to augment this scarce data, enabling a broader range of applications. Leveraging over 15 years of ONERA's extensive archival airborn data from acquisition campaigns, we created a comprehensive training dataset of 110 thousands SAR images to exploit a 3.5 billion parameters pre-trained latent diffusion model \cite{Baqu2019SethiR}. In this work, we present a novel approach utilizing spatial conditioning techniques within a foundation model to transform satellite SAR imagery into airborne SAR representations. Additionally, we demonstrate that our pipeline is effective for bridging the realism of simulated images generated by ONERA's physics-based simulator EMPRISE \cite{empriseem_ai_images}. Our method explores a key application of AI in advancing SAR imaging technology. To the best of our knowledge, we are the first to introduce this approach in the literature.
Authors: Luis Alvarez, Jean-Michel Morel
Abstract: In this paper, we propose an active contour model for silhouette vectorization using cubic B\'ezier curves. Among the end points of the B\'ezier curves, we distinguish between corner and regular points where the orientation of the tangent vector is prescribed. By minimizing the distance of the B\'ezier curves to the silhouette boundary, the active contour model optimizes the location of the B\'ezier curves end points, the orientation of the tangent vectors in the regular points, and the estimation of the B\'ezier curve parameters. This active contour model can use the silhouette vectorization obtained by any method as an initial guess. The proposed method significantly reduces the average distance between the silhouette boundary and its vectorization obtained by the world-class graphic software Inkscape, Adobe Illustrator, and a curvature-based vectorization method, which we introduce for comparison. Our method also allows us to impose additional regularity on the B\'ezier curves by reducing their lengths.
Authors: Noriaki Hirose, Lydia Ignatova, Kyle Stachowicz, Catherine Glossop, Sergey Levine, Dhruv Shah
Abstract: Developing broadly generalizable visual navigation policies for robots is a significant challenge, primarily constrained by the availability of large-scale, diverse training data. While curated datasets collected by researchers offer high quality, their limited size restricts policy generalization. To overcome this, we explore leveraging abundant, passively collected data sources, including large volumes of crowd-sourced teleoperation data and unlabeled YouTube videos, despite their potential for lower quality or missing action labels. We propose Model-Based ReAnnotation (MBRA), a framework that utilizes a learned short-horizon, model-based expert model to relabel or generate high-quality actions for these passive datasets. This relabeled data is then distilled into LogoNav, a long-horizon navigation policy conditioned on visual goals or GPS waypoints. We demonstrate that LogoNav, trained using MBRA-processed data, achieves state-of-the-art performance, enabling robust navigation over distances exceeding 300 meters in previously unseen indoor and outdoor environments. Our extensive real-world evaluations, conducted across a fleet of robots (including quadrupeds) in six cities on three continents, validate the policy's ability to generalize and navigate effectively even amidst pedestrians in crowded settings.