CVPR 2025 Papers

Universal Multimodal Retrieval (UMR) aims to enable search across various modalities using a unified model, where queries and candidates can consist of pure text, images, or a combination of both. Previous work has attempted to adopt multimodal large language models (MLLMs) to realize UMR using only text data. However, our preliminary experiments demonstrate that more diverse multimodal training data can further unlock the potential of MLLMs. Despite its effectiveness, the existing multimodal training data is highly imbalanced in terms of modality, which motivates us to develop a training data synthesis pipeline and construct a large-scale, high-quality fused-modal training dataset. Based on the synthetic training data, we develop the General Multimodal Embedder (GME), an MLLM-based dense retriever designed for UMR. Furthermore, we construct a comprehensive UMR Benchmark (UMRB) to evaluate the effectiveness of our approach. Experimental results show that our method achieves state-of-the-art performance among existing UMR methods. Last, we provide in-depth analyses of model scaling and training strategies, and perform ablation studies on both the model and synthetic data.

Gait sequences exhibit sequential structures and contextual relationships similar to those in natural language, where each element--whether a word or a gait step--is connected to its predecessors and successors. This similarity enables the transformation of gait sequences into "texts" containing identity-related information. Large Language Models (LLMs), designed to understand and generate sequential data, can thus be utilized for gait sequence modeling to enhance gait recognition performance. Leveraging these insights, we make a pioneering effort to apply LLMs to gait recognition, which we refer to as GaitLLM. Specifically, we propose the Gait-to-Language (G2L) module, which converts gait sequences into a textual format suitable for LLMs, and the Language-to-Gait (L2G) module, which maps the LLM's output back to the gait feature space, thereby bridging the gap between LLM outputs and gait recognition. Notably, GaitLLM leverages the powerful modeling capabilities of LLMs without relying on complex architectural designs, improving gait recognition performance with only a small number of trainable parameters. Our method achieves state-of-the-art results on four popular gait datasets--SUSTech1K, CCPG, Gait3D, and GREW--demonstrating the effectiveness of applying LLMs in this domain. This work highlights the potential of LLMs to significantly enhance gait recognition, paving the way for future research and practical applications.

Weakly-supervised Temporal Action Localization (WTAL) has achieved notable success but still suffers from a lack of temporal annotations, leading to a performance and framework gap compared with fully-supervised methods. While recent approaches employ pseudo labels for training, three key challenges: generating high-quality pseudo labels, making full use of different priors, and optimizing training methods with noisy labels remain unresolved. Due to these perspectives, we propose PseudoFormer, a novel two-branch framework that bridges the gap between weakly and fully-supervised Temporal Action Localization (TAL). We first introduce RickerFusion, which maps all predicted action proposals to a global shared space to generate pseudo labels with better quality. Subsequently, we leverage both snippet-level and proposal-level labels with different priors from the weak branch to train the regression-based model in the full branch. Finally, the uncertainty mask and iterative refinement mechanism are applied for training with noisy pseudo labels. PseudoFormer achieves state-of-the-art WTAL results on the two commonly used benchmarks, THUMOS14 and ActivityNet1.3. Besides, extensive ablation studies demonstrate the contribution of each component of our method.

High-dynamic scene optical flow is a challenging task, which suffers spatial blur and temporal discontinuous motion due to large displacement in frame imaging, thus deteriorating the spatiotemporal feature of optical flow. Typically, existing methods mainly introduce event camera to directly fuse the spatiotemporal features between the two modalities. However, this direct fusion is ineffective, since there exists a large gap due to the heterogeneous data representation between frame and event modalities. To address this issue, we explore a common-latent space as an intermediate bridge to mitigate the modality gap. In this work, we propose a novel common spatiotemporal fusion between frame and event modalities for high-dynamic scene optical flow, including visual boundary localization and motion correlation fusion. Specifically, in visual boundary localization, we figure out that frame and event share the similar spatiotemporal gradients, whose similarity distribution is consistent with the extracted boundary distribution. This motivates us to design the common spatiotemporal gradient to constrain the reference boundary localization. In motion correlation fusion, we discover that the frame-based motion possesses spatially dense but temporally discontinuous correlation, while the event-based motion has spatially sparse but temporally continuous correlation. This inspires us to use the reference boundary to guide the complementary motion knowledge fusion between the two modalities. Moreover, common spatiotemporal fusion can not only relieve the cross-modal feature discrepancy, but also make the fusion process interpretable for dense and continuous optical flow. Extensive experiments have been performed to verify the superiority of the proposed method.

Despite advancements in Computer-Aided-Design (CAD) generation, direct generation of complex Boundary Representation (B-rep) CAD models remains challenging. This difficulty arises from the parametric nature of B-rep data, complicating the encoding and generation of its geometric and topological information. To address this, we introduce BrepGiff, a lightweight generation approach for high-quality and complex B-rep model based on 3D Graph Diffusion. First, we transfer B-rep models into 3D graphs representation. Specifically, BrepGiff extracts and integrates topological and geometric features to construct a 3D graph where nodes correspond to face centroids in 3D space, preserving adjacency and degree information. Geometric features are derived by sampling points in the UV domain and extracting face and edge features. Then, BrepGiff applies a Graph Attention Network (GAT) to enforce topological constraints from local to global during the degree-guided diffusion process. With the 3D graph representation and efficient diffusion process, our method significantly reduces the computational cost and improves the quality, thus achieving lightweight generation of complex models. Experiments show that BrepGiff can generate complex B-rep models (>100 faces) using only 2 RTX4090 GPUs, achieving state-of-the-art performance in B-rep generation.

The Softmax attention mechanism in Transformer models is notoriously computationally expensive, particularly due to its quadratic complexity, posing significant challenges in vision applications. In contrast, linear attention provides a far more efficient solution by reducing the complexity to linear levels. However, compared to Softmax attention, linear attention often experiences significant performance degradation. Our experiments indicate that this performance drop is due to the low-rank nature of linear attention's feature map, which hinders its ability to adequately model complex spatial information. In this paper, to break the low-rank dilemma of linear attention, we conduct rank analysis from two perspectives: the kv buffer and the output features. Consequently, we introduce **Rank-Augmented Linear Attention** (RALA), which rivals the performance of Softmax attention while maintaining linear complexity and high efficiency. Based on RALA, we construct the **Rank-Augmented Vision Linear Transformer** (RAVLT). Extensive experiments demonstrate that RAVLT achieves excellent performance across various vision tasks. Specifically, without using any additional labels, data, or supervision during training, RAVLT achieves an **84.4%** Top-1 accuracy on ImageNet-1k with only **26M** parameters and **4.6G** FLOPs. This result significantly surpasses previous linear attention mechanisms, fully illustrating the potential of RALA.

Brain-inspired spiking neural networks (SNNs) have the capability of energy-efficient processing of temporal information. However, leveraging the rich dynamic characteristics of SNNs and prior works in artificial neural networks (ANNs) to construct an effective object detection model for visual tasks remains an open question for further exploration. To develop a directly-trained , low energy consumption and high-performance multi-scale SNN model, we propose a novel interpretable object detection framework Multi-scale Spiking Detector (MSD). Initially, we propose a spiking convolutional neuron as a core component of the Optic Nerve Nucleus Block (ONNB), designed to significantly enhance the deep feature extraction capabilities of SNNs. ONNB enables direct training with improved energy efficiency, demonstrating superior performance compared to state-of-the-art ANN-to-SNN conversion and SNN techniques. In addition, we propose a Multi-scale Spiking Detection Framework to emulate the biological response and comprehension of stimuli from different objects. Wherein, spiking multi-scale fusion and the spiking detector are employed to integrate features across different depths and to detect response outcomes, respectively. Our method outperforms state-of-the-art ANN detectors, with only 7.8 M parameters and 6.43 mJ energy consumption. MSD obtains the mean average precision (mAP) of 62.0% and 66.3% on COCO and Gen1 datasets, respectively.

View-invariant representation learning from egocentric (first-person, ego) and exocentric (third-person, exo) videos is a promising approach toward generalizing video understanding systems across multiple viewpoints. However, this area has been underexplored due to the substantial differences in perspective, motion patterns, and context between ego and exo views. In this paper, we propose a novel masked ego-exo modeling that promotes both causal temporal dynamics and cross-view alignment, called Bootstrap Your Own Views (BYOV), for fine-grained view-invariant video representation learning from unpaired ego-exo videos. We highlight the importance of capturing the compositional nature of human actions as a basis for robust cross-view understanding. Specifically, self-view masking and cross-view masking predictions are designed to learn view-invariant and powerful representations concurrently. Experimental results demonstrate that our BYOV significantly surpasses existing approaches with notable gains across all metrics in four downstream ego-exo video tasks. The code is available at https://github.com/park-jungin/byov

Over the last decade, significant efforts have been dedicated to designing efficient models for the challenge of ultra-high resolution (UHR) semantic segmentation. These models mainly follow the dual-stream architecture and generally fall into three subcategories according to the improvement objectives, i.e., dual-stream ensemble, selective zoom, and complementary learning. However, most of them overly concentrate on crafting complex pipelines to pursue one of the above objectives separately, limiting the model performance in both accuracy and inference consumption. In this paper, we suggest simultaneously achieving these objectives by estimating resolution-biased uncertainties in low resolution stream. Here, the resolution-biased uncertainty refers to the degree of prediction unreliability primarily caused by resolution loss from down-sampling operations. Specifically, we propose a dual-stream UHR segmentation framework, where an estimator is used to assess resolution-biased uncertainties through the entropy map and high-frequency feature residual. The framework also includes a selector, an ensembler, and a complementer to boost the model with obtained estimations. They share the uncertainty estimations as the weights to choose difficult regions as the inputs for UHR stream, perform weighted fusion between distinct streams, and enhance the learning for important pixels, respectively. Experiment results demonstrate that our method achieves a satisfactory balance between accuracy and inference consumption against other state-of-the-art (SOTA) methods. The code is available in the https://github.com/Qinrong-NKU/RUE.

Adversarial examples can mislead deep neural networks with subtle perturbations, causing them to make incorrect predictions. Notably, adversarial examples crafted for one model can also deceive other models, a phenomenon known as the transferability of adversarial examples. To improve transferability, existing studies have designed increasingly complex mechanisms, but the improvements achieved remain relatively limited and are often difficult to adapt to other modalities, further restricting the scalability of these methods. In this work, we observe a mirroring relationship between model generalization and adversarial example transferability. Motivated by this observation, we propose an augmentation-based attack, called OPS (Operator-Perturbation-based Stochastic optimization), which constructs a stochastic optimization problem by input transformation operators and random perturbations, and solves this problem to generate adversarial examples with better transferability. Extensive experiments on both images and 3D point clouds demonstrate that OPS significantly outperforms existing state-of-the-art methods in terms of both performance and cost, showcasing the universality and superiority of our approach. The code is available at https://github.com/the-full/OPS.

Semi-supervised polyp segmentation has made significant progress in recent years as a potential solution for computer-assisted treatment. Since depth images can provide extra information other than RGB images to help segment these problematic areas, depth-assisted polyp segmentation has gained much attention. However, the utilization of depth information is still worth studying. The existing RGB-D segmentation methods rely on depth data in the inference stage, limiting their clinical applications. To tackle this problem, we propose a semi-supervised polyp segmentation framework based on the mean teacher architecture. We establish an auxiliary student network with depth images as input in the training stage, and we propose a depth-guided cross-modal mutual learning strategy to promote the learning of complementary information between different student networks. Meanwhile, we use the high-confidence pseudo-labels generated by the auxiliary student network to guide the learning progress of the main student network from different perspectives. Our model does not need depth data in the inference phase. In addition, we introduce a depth-guided patch augmentation method to improve the model's learning performance in difficult regions of unlabeled polyp images. Experimental results show that our method achieves state-of-the-art performance under different label conditions on five polyp datasets. The code is available at https://github.com/pingchuan/RD-Net.

Image-based virtual try-on is an increasingly popular and important task to generate realistic try-on images of the specific person.Recent methods model virtual try-on as image mask-inpaint task, which requires masking the person image and results in significant loss of spatial information. Especially, for in-the-wild try-on scenarios with complex poses and occlusions, mask-based methods often introduce noticeable artifacts. Our research found that a mask-free approach can fully leverage spatial and lighting information from the original person image, enabling high-quality virtual try-on. Consequently, we propose a novel training paradigm for a mask-free try-on diffusion model. We ensure the model's mask-free try-on capability by creating high-quality pseudo-data and further enhance its handling of complex spatial information through effective in-the-wild data augmentation. Besides, a try-on localization loss is designed to concentrate on try-on area while suppressing garment features in non-try-on areas, ensuring precise rendering of garments and preservation of fore/back-ground. In the end, we introduce BooW-VTON, the mask-free virtual try-on diffusion model, which delivers SOTA try-on quality without parsing cost. Extensive qualitative and quantitative experiments have demonstrated superior performance in wild scenarios with such a low-demand input.

Detecting and segmenting small objects, such as lung nodules and tumor lesions, remains a critical challenge in image analysis. These objects often occupy less than 0.1% of an image, making traditional transformer architectures inefficient and prone to performance degradation due to redundant attention computations on irrelevant regions. Existing sparse attention mechanisms rely on rigid hierarchical structures, which are poorly suited for detecting small, variable, and uncertain object locations. In this paper, we propose BoltzFormer, a novel transformer-based architecture designed to address these challenges through dynamic sparse attention. BoltzFormer identifies and focuses attention on relevant areas by modeling uncertainty using a Boltzmann distribution with an annealing schedule. Initially, a higher temperature allows broader area sampling in early layers, when object location uncertainty is greatest. As the temperature decreases in later layers, attention becomes more focused, enhancing efficiency and accuracy. BoltzFormer seamlessly integrates into existing transformer architectures via a modular Boltzmann attention sampling mechanism. Comprehensive evaluations on benchmark datasets demonstrate that BoltzFormer significantly improves segmentation performance for small objects while reducing attention computation by an order of magnitude compared to previous state-of-the-art methods.

Extracting depth information from photon-limited, defocused images is challenging because depth from defocus (DfD) relies on accurate estimation of defocus blur, which is fundamentally sensitive to image noise. We present a novel approach to robustly measure object depths from photon-limited images along the defocused boundaries. It is based on a new image patch representation, Blurry-Edges, that explicitly stores and visualizes a rich set of low-level patch information, including boundaries, color, and smoothness. We develop a deep neural network architecture that predicts the Blurry-Edges representation from a pair of differently defocused images, from which depth can be calculated using a closed-form DfD relation we derive. The experimental results on synthetic and real data show that our method achieves the highest depth estimation accuracy on photon-limited images compared to a broad range of state-of-the-art DfD methods.

The emergence and growing popularity of multimodal large language models (MLLMs) have significant potential to enhance various aspects of daily life, from improving communication to facilitating learning and problem-solving. Mobile phones, as essential daily companions, represent the most effective and accessible deployment platform for MLLMs, enabling seamless integration into everyday tasks. However, deploying MLLMs on mobile phones presents challenges due to limitations in memory size and computational capability, making it difficult to achieve smooth and real-time processing without extensive optimization. In this paper, we present BlueLM-V-3B, an algorithm and system co-design approach specifically tailored for the efficient deployment of MLLMs on mobile platforms. To be specific, we redesign the dynamic resolution scheme adopted by mainstream MLLMs and implement system optimization for hardware-aware deployment to optimize model inference on mobile phones. BlueLM-V-3B boasts the following key highlights: (1) Small Size: BlueLM-V-3B features a language model with 2.7B parameters and a vision encoder with 400M parameters. (2) Fast Speed: BlueLM-V-3B achieves a generation speed of 24.4 token/s on the MediaTek Dimensity 9300 processor with 4-bit LLM weight quantization. (3) Strong Performance: BlueLM-V-3B has attained the highest average score of 66.1 on the OpenCompass benchmark among models with <= 4B parameters and surpassed a series of models with much larger parameter sizes (e.g., MiniCPM-V-2.6, InternVL2-8B).

Estimating blood flow speed is essential in many medical and physiological applications, yet it is extremely challenging due to complex vascular structure and flow dynamics, particularly for cerebral cortex regions. Existing techniques, such as Optical Doppler Tomography (ODT), generally require complex hardware control and signal processing, and still suffer from inherent system-level artifacts. To address these challenges, we propose a new learning-based approach named OCTA-Flow, which directly estimates vascular blood flow speed from Optical Coherence Tomography Angiography (OCTA) images that are commonly used for vascular structure analysis. OCTA-Flow employs several novel components to achieve this goal. First, using an encoder-decoder architecture, OCTA-Flow leverages ODT data as pseudo labels during training, thus bypassing the difficulty of collecting ground truth data. Second, to capture the relationship between vessels of varying scales and their flow speed, we design an Adaptive Window Fusion module that employs multiscale window attention. Third, to mitigate ODT artifacts, we incorporate a Conditional Random Field Decoder that promotes smoothness and consistency in the estimated blood flow. Together, these innovations enable OCTA-Flow to effectively produce accurate flow estimation, suppress the artifacts in ODT, and enhance practicality, benefiting from the established techniques of OCTA data acquisition. The code and data are available at https://github.com/Spritea/OCTA-Flow.

Diffusion models have demonstrated impressive generation capabilities, particularly with recent advancements leveraging transformer architectures to improve both visual and artistic quality. However, Diffusion Transformers (DiTs) continue to encounter challenges related to low inference speed, primarily due to the iterative denoising process. To address this issue, we propose BlockDance, a training-free approach that explores feature similarities at adjacent time steps to accelerate DiTs. Unlike previous feature-reuse methods that lack tailored reuse strategies for features at different scales, BlockDance prioritizes the identification of the most structurally similar features, referred to as Structurally Similar Spatio-Temporal (STSS) features. These features are primarily located within the structure-focused blocks of the transformer during the later stages of denoising. BlockDance caches and reuses these highly similar features to mitigate redundant computation, thereby accelerating DiTs while maximizing consistency with the generated results of the original model. Furthermore, considering the diversity of generated content and the varying distributions of redundant features, we introduce BlockDance-Ada, a lightweight decision-making network tailored for instance-specific acceleration. BlockDance-Ada dynamically allocates resources and provides superior content quality. Both BlockDance and BlockDance-Ada have proven effective across various generation tasks and models, achieving accelerations between 25% and 50% while maintaining generation quality.