CVPR 2025 Papers

Diffusion models (DMs) have demonstrated exceptional generative capabilities across various domains, including image, video, and so on. A key factor contributing to their effectiveness is the high quantity and quality of data used during training. However, mainstream DMs now consume increasingly large amounts of data. For example, training a Stable Diffusion model requires billions of image-text pairs. This enormous data requirement poses significant challenges for training large DMs due to high data acquisition costs and storage expenses. To alleviate this data burden, we propose a novel scenario: using existing DMs as data sources to train new DMs with any architecture. We refer to this scenario as Data-Free Knowledge Distillation for Diffusion Models (DKDM), where the generative ability of DMs is transferred to new ones in a data-free manner. To tackle this challenge, we make two main contributions. First, we introduce a DKDM objective that enables the training of new DMs via distillation, without requiring access to the data. Second, we develop a dynamic iterative distillation method that efficiently extracts time-domain knowledge from existing DMs, enabling direct retrieval of training data without the need for a prolonged generative process. To the best of our knowledge, we are the first to explore this scenario. Experimental results demonstrate that our data-free approach not only achieves competitive generative performance but also, in some instances, outperforms models trained with the entire dataset.

Lifelong person re-identification (LReID) aims to match the same person using sequentially collected data. However, due to the long-term nature of lifelong learning, the inevitable changes in human clothes prevent the model from relying on unified discriminative information (e.g., clothing style) to match the same person in the streaming data, demanding differentiated cloth-irrelevant information. Unfortunately, existing LReID methods typically fail to leverage such knowledge resulting in the exacerbation of catastrophic forgetting issues. Therefore, in this paper, we focus on a challenging practical task called Cloth-Hybrid Lifelong Person Re-identification (CH-LReID), which requires matching the same person wearing different clothes using sequentially collected data. A Differentiated Knowledge Consolidation (DKC) framework is designed to unify and balance distinct knowledge across streaming data. The core idea is to adaptively balance differentiated knowledge and compatibly consolidate cloth-relevant and cloth-irrelevant information. To this end, a Differentiated Knowledge Transfer (DKT) module and a Latent Knowledge Consolidation (LKC) module are designed to adaptively discover differentiated new knowledge, while eliminating the derived domain shift of old knowledge via reconstructing the old latent feature space, respectively. Then, to further alleviate the catastrophic conflict between differentiated new and old knowledge, we further propose a Dual-level Distribution Alignment (DDA) module to align the distribution of discriminative knowledge at both the instance level and the fine-grained level. Extensive experiments on multiple benchmarks demonstrate the superiority of our method against existing methods in both CH-LReID and traditional LReID tasks. The source code of this paper is available at https://github.com/PKU-ICST-MIPL/DKC-CVPR2025.

We present DIO, a flexible world model that can estimate the scene occupancy-flow from a sparse set of LiDAR observations, and decompose it into individual instances. DIO can not only complete instance shapes at the present time, but also forecast their occupancy-flow evolution over a future horizon. Thanks to its flexible prompt representation, DIO can take instance prompts from off-the-shelf models like 3D detectors, achieving state-of-the-art performance in the task of 4D semantic occupancy completion and forecasting on the Argoverse 2 dataset. Moreover, our world model can easily and effectively be transferred to downstream tasks like LiDAR point cloud forecasting, ranking first compared to all baselines in the Argoverse 4D occupancy forecasting challenge.

Self-supervised visual foundation models produce powerful embeddings that achieve remarkable performance on a wide range of downstream tasks. However, unlike vision-language models such as CLIP, self-supervised visual features are not readily aligned with language, hindering their adoption in open-vocabulary tasks. Our method unlocks this new ability for DINOv2, a widely used self-supervised visual encoder. We build upon the LiT training strategy, which trains a text encoder to align with a frozen vision model but leads to unsatisfactory results on dense tasks. We propose several key ingredients to improve performance on both global and dense tasks, such as concatenating the [CLS] token with the patch average to train the alignment and curating data using both text and image modalities. With these, we successfully train a CLIP-like model with only a fraction of the computational cost compared to CLIP while achieving state-of-the-art results in zero-shot classification and open-vocabulary semantic segmentation.

Clothes-changing person re-identification (CC-ReID) aims to recognize individuals under different clothing scenarios. Current CC-ReID approaches either concentrate on modeling body shape using additional modalities including silhouette, pose, and body mesh, potentially causing the model to overlook other critical biometric traits such as gender, age, and style, or they incorporate supervision through additional labels that the model tries to disregard or emphasize, such as clothing or personal attributes. However, these annotations are discrete in nature and do not capture comprehensive descriptions. In this work, we propose DIFFER: Disentangle Identity Features From Entangled Representations, a novel adversarial learning method that leverages textual descriptions to disentangle identity features. Recognizing that image features inherently mix inseparable information, DIFFER introduces NBDetach, a mechanism designed for feature disentanglement by leveraging the separable nature of text descriptions as supervision. It partitions the feature space into distinct subspaces and, through gradient reversal layers, effectively separates identity-related features from non-biometric features. We evaluate DIFFER on 4 different benchmark datasets (LTCC, PRCC, CelebreID-Light, and CCVID) to demonstrate its effectiveness and provide state-of-the-art performance across all the benchmarks. DIFFER consistently outperforms the baseline method, with improvements in top-1 accuracy of 3.6% on LTCC, 3.4% on PRCC, 2.5% on CelebReID-Light, and 1% on CCVID. Our code can be found at https://github.com/xliangp/DIFFER.git.

Procedural Content Generation (PCG) is powerful in creating high-quality 3D contents, yet controlling it to produce desired shapes is difficult and often requires extensive parameter tuning. Inverse Procedural Content Generation aims to automatically find the best parameters under the input condition. However, existing sampling-based and neural network-based methods still suffer from numerous sample iterations or limited controllability. In this work, we present DI-PCG, a novel and efficient method for Inverse PCG from general image conditions. At its core is a lightweight diffusion transformer model, where PCG parameters are directly treated as the denoising target and the observed images as conditions to control parameter generation. DI-PCG is efficient and effective. With only 7.6M network parameters and 30 GPU hours to train, it demonstrates superior performance in recovering parameters accurately, and generalizing well to in-the-wild images. Quantitative and qualitative experiment results validate the effectiveness of DI-PCG in inverse PCG and image-to-3D generation tasks. DI-PCG offers a promising approach for efficient inverse PCG and represents a valuable exploration step towards a 3D generation path that models how to construct a 3D asset using parametric models.

Vision-Language (VL) alignment across image and text modalities is a challenging task due to the inherent semantic ambiguity of data with multiple possible meanings. Existing methods typically solve it by learning multiple sub-representation spaces to encode each input data as a set of embeddings, and constraining diversity between whole subspaces to capture diverse semantics for accurate VL alignment. Despite their promising outcomes, existing methods suffer two imperfections: 1) actually, specific semantics is mainly expressed by some local dimensions within the subspace. Ignoring this intrinsic property, existing diversity constraints imposed on the whole subspace may impair diverse embedding learning; 2) multiple embeddings are inevitably introduced, sacrificing computational and storage efficiency. In this paper, we propose a simple yet effective Diverse and Hybrid Set-embeddings learning framework (DH-Set), which is distinct from prior work in three aspects. DH-Set 1) devises a novel semantic importance dissecting method to focus on key local dimensions within each subspace; and thereby 2) not only imposes finer-grained diversity constraint to improve the accuracy of diverse embedding learning, 3) but also mixes key dimensions of all subspaces into the single hybrid embedding to boost inference efficiency. Extensive experiments on various benchmarks and model backbones show the superiority of DH-Set over state-of-the-art methods, achieving substantial 2.3%-14.7% rSum improvements while lowering computational and storage complexity.

Recent advances in scene understanding benefit a lot from depth maps because of the 3D geometry information, especially in complex conditions (e.g., low light and overexposed). Existing approaches encode depth maps along with RGB images and perform feature fusion between them to enable more robust predictions. Taking into account that depth can be regarded as a geometry supplement for RGB images, a straightforward question arises: Do we really need to explicitly encode depth information with neural networks as done for RGB images? Based on this insight, in this paper, we investigate a new way to learn RGBD feature representations and present DFormerv2, a strong RGBD encoder that explicitly uses depth maps as geometry priors rather than encoding depth information with neural networks. Our goal is to extract the geometry clues from the depth and spatial distances among all the image patch tokens, which will then be used as geometry priors to allocate attention weights in self-attention. Extensive experiments demonstrate that DFormerv2 exhibits exceptional performance in various RGBD semantic segmentation benchmarks. Code is available at: https://github.com/VCIP-RGBD/DFormer.

Feature matching methods for unsupervised anomaly detection have demonstrated impressive performance. Existing methods primarily rely on self-supervised training and handcrafted matching schemes for task adaptation. However, they can only achieve an inferior feature representation for anomaly detection because the feature extraction and matching modules are separately trained. To address these issues, we propose a Differentiable Feature Matching (DFM) framework for joint optimization of the feature extractor and the matching head. DFM transforms nearest-neighbor matching into a pooling-based module and embeds it within a Feature Matching Network (FMN). This design enables end-to-end feature extraction and feature matching module training, thus providing better feature representation for anomaly detection tasks. DFM is generic and can be incorporated into existing feature-matching methods. We implement DFM with various backbones and conduct extensive experiments across various tasks and datasets, demonstrating its effectiveness. Notably, we achieve state-of-the-art results in the continual anomaly detection task with instance-AUROC improvement of up to 3.9% and pixel-AP improvement of up to 5.5%.

Recent advances in dataset distillation have led to solutions in two main directions. The conventional batch-to-batch matching mechanism is ideal for small-scale datasets and includes bi-level optimization methods on models and syntheses, such as FRePo, RCIG, and RaT-BPTT, as well as other methods like distribution matching, gradient matching, and weight trajectory matching. Conversely, batch-to-global matching typifies decoupled methods, which are particularly advantageous for large-scale datasets. This approach has garnered substantial interest within the community, as seen in SRe^2L, G-VBSM, WMDD, and CDA. A primary challenge with the second approach is the lack of diversity among syntheses within each class since samples are optimized independently and the same global supervision signals are reused across different synthetic images. In this study, we propose a new Diversity-driven EarlyLate Training (DELT) scheme to enhance the diversity of images in batch-to-global matching with less computation. Our approach is conceptually simple yet effective, it partitions predefined IPC samples into smaller subtasks and employs local optimizations to distill each subset into distributions from distinct phases, reducing the uniformity induced by the unified optimization process. These distilled images from the subtasks demonstrate effective generalization when applied to the entire task. We conduct extensive experiments on CIFAR, Tiny-ImageNet, ImageNet-1K, and its sub-datasets. Our approach outperforms the previous state-of-the-art by 2~5% on average across different datasets and IPCs (images per class), increasing diversity per class by more than 5% while reducing synthesis time by up to 39.3% for enhancing the training efficiency.

We introduce DEIM, an innovative and efficient training framework designed to accelerate convergence in real-time object detection with Transformer-based architectures (DETR). To mitigate the sparse supervision inherent in one-to-one (O2O) matching in DETR models, DEIM employs a Dense O2O matching strategy. This approach increases the number of positive samples per image by incorporating additional targets, using standard data augmentation techniques. While Dense O2O matching speeds up convergence, it also introduces numerous low-quality matches that could affect performance. To address this, we propose the Matchability-Aware Loss (MAL), a novel loss function that optimizes matches across various quality levels, enhancing the effectiveness of Dense O2O. Extensive experiments on the COCO dataset validate the efficacy of DEIM. When integrated with RT-DETR and D-FINE, it consistently boosts performance while reducing training time by 50%. Notably, paired with RT-DETRv2, DEIM achieves 53.2% AP in a single day of training on an NVIDIA 4090 GPU. Additionally, DEIM-trained real-time models outperform leading real-time object detectors, with DEIM-D-FINE-L and DEIM-D-FINE-X achieving 54.7% and 56.4% AP at 124 and 78 FPS on an NVIDIA T4 GPU, respectively, without the need for additional data. We believe DEIM sets a new baseline for advancements in real-time object detection. Our code and pre-trained models are available at https://www.shihuahuang.cn/DEIM/.

Stereo matching is a key technique for metric depth estimation in computer vision and robotics. Real-world challenges like occlusion and non-texture hinder accurate disparity estimation from binocular matching cues. Recently, monocular relative depth estimation has shown remarkable generalization using vision foundation models. Thus, to facilitate robust stereo matching with monocular depth cues, we incorporate a robust monocular relative depth model into the recurrent stereo-matching framework, building a new framework for depth foundation model-based stereo-matching, DEFOM-Stereo. In the feature extraction stage, we construct the combined context and matching feature encoder by integrating features from conventional CNNs and DEFOM. In the update stage, we use the depth predicted by DEFOM to initialize the recurrent disparity and introduce a scale update module to refine the disparity at the correct scale. DEFOM-Stereo is verified to have much stronger zero-shot generalization compared with SOTA methods. Moreover, DEFOM-Stereo achieves top performance on the KITTI 2012, KITTI 2015, Middlebury, and ETH3D benchmarks, ranking 1^ st on many metrics. In the joint evaluation under the robust vision challenge, our model simultaneously outperforms previous models on the individual benchmarks, further demonstrating its outstanding capabilities.

Thermal imaging is often compromised by dynamic, complex degradations caused by hardware limitations and unpredictable environmental factors. The scarcity of high-quality infrared data, coupled with the challenges of dynamic, intricate degradations, makes it difficult to recover details using existing methods. In this paper, we introduce thermal degradation simulation integrated into the training process via a mini-max optimization, by modeling these degraded factors as adversarial attacks on thermal images. The simulation is dynamic to maximize objective functions, thus capturing a broad spectrum of degraded data distributions. This approach enables training with limited data, thereby improving model performance. Additionally, we introduce a dual-interaction network that combines the benefits of spiking neural networks with scale transformation to capture degraded features with sharp spike signal intensities. This architecture ensures compact model parameters while preserving efficient feature representation. Extensive experiments demonstrate that our method not only achieves superior visual quality under diverse single and composited degradation, but also delivers a significant reduction in processing when trained on only fifty clear images, outperforming existing techniques in efficiency and accuracy. The source code will be available at https://github.com/LiuZhu-CV/DEAL.

Infrared and visible image fusion integrates information from distinct spectral bands to enhance image quality by leveraging the strengths and mitigating the limitations of each modality. Existing approaches typically treat image fusion and subsequent high-level tasks as separate processes, resulting in fused images that offer only marginal gains in task performance and fail to provide constructive feedback for optimizing the fusion process. To overcome these limitations, we propose a Discriminative Cross-Dimension Evolutionary Learning Framework, termed DCEvo, which simultaneously enhances visual quality and perception accuracy. Leveraging the robust search capabilities of Evolutionary Learning, our approach formulates the optimization of dual tasks as a multi-objective problem by employing an Evolutionary Algorithm (EA) to dynamically balance loss function parameters. Inspired by visual neuroscience, we integrate a Discriminative Enhancer (DE) within both the encoder and decoder, enabling the effective learning of complementary features from different modalities. Additionally, our Cross-Dimensional Embedding (CDE) block facilitates mutual enhancement between high-dimensional task features and low-dimensional fusion features, ensuring a cohesive and efficient feature integration process. Experimental results on three benchmarks demonstrate that our method significantly outperforms state-of-the-art approaches, achieving an average improvement of 9.32% in visual quality while also enhancing subsequent high-level tasks. The code is available at https://github.com/Beate-Suy-Zhang/DCEvo.

The automatic generation of radiology reports has emerged as a promising solution to reduce a time-consuming task and accurately capture critical disease-relevant findings in X-ray images. Previous approaches for radiology report generation have shown impressive performance. However, there remains significant potential to improve accuracy by ensuring that retrieved reports contain disease-relevant findings similar to those in the X-ray images and by refining generated reports. In this study, we propose a Disease-aware image-text Alignment and self-correcting Re-alignment for Trustworthy radiology report generation (DART) framework. In the first stage, we generate initial reports based on image-to-text retrieval with disease-matching, embedding both images and texts in a shared embedding space through contrastive learning. This approach ensures the retrieval of reports with similar disease-relevant findings that closely align with the input X-ray images. In the second stage, we further enhance the initial reports by introducing a self-correction module that re-aligns them with the X-ray images. Our proposed framework achieves state-of-the-art results on two widely used benchmarks, surpassing previous approaches in both report generation and clinical efficacy metrics, thereby enhancing the trustworthiness of radiology reports.

Text-driven avatar generation has gained significant attention owing to its convenience. However, existing methods typically model the human body with all garments as a single 3D model, limiting its usability, such as clothing replacement, and reducing user control over the generation process. To overcome the limitations above, we propose DAGSM, a novel pipeline that generates disentangled human bodies and garments from the given text prompts. Specifically, we model each part (e.g., body, upper/lower clothes) of the clothed human as one GS-enhanced mesh (GSM), which is a traditional mesh attached with 2D Gaussians to better handle complicated textures (e.g., woolen, translucent clothes) and produce realistic cloth animations. During the generation, we first create the unclothed body, followed by a sequence of individual cloth generation based on the body, where we introduce a semantic-based algorithm to achieve better human-cloth and garment-garment separation. To improve texture quality, we propose a view-consistent texture refinement module, including a cross-view attention mechanism for texture style consistency and an incident-angle-weighted denoising (IAW-DE) strategy to update the appearance. Extensive experiments have demonstrated that DAGSM generates high-quality disentangled avatars, supports clothing replacement and realistic animation, and outperforms the baselines in visual quality.

Visual Prompt Tuning (VPT) has become a promising solution for Parameter-Efficient Fine-Tuning (PEFT) approach for Vision Transformer (ViT) models by partially fine-tuning learnable tokens while keeping most model parameters frozen. Recent research has explored modifying the connection structures of the prompts. However, the fundamental correlation and distribution between the prompts and image tokens remain unexplored. In this paper, we leverage metric learning techniques to investigate how the distribution of prompts affects fine-tuning performance. Specifically, we propose a novel framework, Distribution Aware Visual Prompt Tuning (DA-VPT), to guide the distributions of the prompts by learning the distance metric from their class-related semantic data. Our method demonstrates that the prompts can serve as an effective bridge to share semantic information between image patches and the class token. We extensively evaluated our approach on popular benchmarks in both recognition and segmentation tasks. The results demonstrate that our approach enables more effective and efficient fine-tuning of ViT models by leveraging semantic information to guide the learning of the prompts, leading to improved performance on various downstream vision tasks.

The current advancements in Dynamic Facial Expression Recognition (DFER) methods mainly focus on better capturing the spatial and temporal features of facial expressions. However, DFER datasets contain a substantial amount of noisy samples, and few have addressed the issue of handling this noise. We identified two types of noise: one is caused by low-quality data resulting from factors such as occlusion, dim lighting, and blurriness; the other arises from mislabeled data due to annotation bias by annotators. Addressing the two types of noise, we have meticulously crafted a Dynamic Dual-Stage Purification (D2SP) Framework. This initiative aims to dynamically purify the DFER datasets of these two types of noise, ensuring that only high-quality and correctly labeled data is used in the training process. To mitigate low-quality samples, we introduce the Coarse-Grained Pruning (CGP) stage, which computes sample weights and prunes those low-weight samples. After CGP, the Fine-Grained Correction (FGC) stage evaluates prediction stability to correct mislabeled data. Moreover, D2SP is conceived as a general and plug-and-play framework, tailored to integrate seamlessly with prevailing DFER methods. Extensive experiments covering prevalent DFER datasets and deploying multiple benchmark methods have substantiated D2SP's ability to significantly enhance performance metrics.