CVPR 2025 Papers

We introduce the task of predicting functional 3D scene graphs for real-world indoor environments from posed RGB-D images. Unlike traditional 3D scene graphs that focus on spatial relationships of objects, functional 3D scene graphs capture objects, interactive elements, and their functional relationships. Due to the lack of training data, we leverage foundation models, including visual language models (VLMs) and large language models (LLMs), to encode functional knowledge. We evaluate our approach on an extended SceneFun3D dataset and a newly collected dataset, FunGraph3D, both annotated with functional 3D scene graphs. Our method significantly outperforms adapted baselines, including Open3DSG and ConceptGraph, demonstrating its effectiveness in modeling complex scene functionalities. We also demonstrate downstream applications such as 3D question answering and robotic manipulation using functional 3D scene graphs. See our project page at https://openfungraph.github.io

Estimating canopy height and its changes at meter resolution from satellite imagery is a significant challenge in computer vision with critical environmental applications. However, the lack of open-access datasets at this resolution hinders the reproducibility and evaluation of models. We introduce Open-Canopy, the first open-access, country-scale benchmark for very high-resolution (1.5 m) canopy height estimation, covering over 87,000 km2 across France with 1.5 m resolution satellite imagery and aerial LiDAR data. Additionally, we present Open-Canopy-, a benchmark for canopy height change detection between images from different years at tree level--a challenging task for current computer vision models. We evaluate state-of-the-art architectures on these benchmarks, highlighting significant challenges and opportunities for improvement. Our datasets and code are publicly available at https://github.com/fajwel/Open-Canopy.

Open set label shift (OSLS) occurs when label distributions change from a source to a target distribution, and the target distribution has an additional out-of-distribution (OOD) class.In this work, we build estimators for both source and target open set label distributions using a source domain in-distribution (ID) classifier and an ID/OOD classifier. With reasonable assumptions on the ID/OOD classifier, the estimators are assembled into a sequence of three stages: 1) an estimate of the source label distribution of the OOD class, 2) an EM algorithm for Maximum Likelihood estimates (MLE) of the target label distribution, and 3) an estimate of the target label distribution of OOD class under relaxed assumptions on the OOD classifier.The sampling errors of estimates in 1) and 3) are quantified with a concentration inequality.The estimation result allows us to correct the ID classifier trained on the source distribution to the target distribution without retraining.Experiments on a variety of open set label shift settings demonstrate the effectiveness of our model.Our code is available at https://github.com/ChangkunYe/OpenSetLabelShift.

Adaptive categorization of visual scenes is essential for AI agents to handle changing tasks. Unlike fixed common categories for plants or animals, ad-hoc categories, such as things to sell at a garage sale, are created dynamically to achieve specific tasks. We study open ad-hoc categorization, where the goal is to infer novel concepts and categorize images based on a given context, a small set of labeled exemplars, and some unlabeled data. We have two key insights: 1) recognizing ad-hoc categories relies on the same perceptual processes as common categories; 2) novel concepts can be discovered semantically by expanding contextual cues or visually by clustering similar patterns. We propose OAK, a simple model that introduces a single learnable context token into CLIP, trained with CLIP's objective of aligning visual and textual features and GCD's objective of clustering similar images. On Stanford and Clevr-4 datasets, OAK consistently achieves the state-of-art in accuracy and concept discovery across multiple categorizations, including 87.4% novel accuracy on Stanford Mood, surpassing CLIP and GCD by over 50%. Moreover, OAK generates interpretable saliency maps, focusing on hands for Action, faces for Mood, and backgrounds for Location, promoting transparency and trust while enabling accurate and flexible categorization.

Online 3D open-vocabulary segmentation of a progressively reconstructed scene is both a critical and challenging task for embodied applications. With the success of visual foundation models (VFMs) in the image domain, leveraging 2D priors to address 3D online segmentation has become a prominent research focus. Since segmentation results provided by 2D priors often require spatial consistency to be lifted into final 3D segmentation, an efficient method for identifying spatial overlap among 2D masks is essential--yet existing methods rarely achieve this in real time, mainly limiting its use to offline approaches. To address this, we propose an efficient method that lifts 2D masks generated by VFMs into a unified 3D instance using a hashing technique. By employing voxel hashing for efficient 3D scene querying, our approach reduces the time complexity of costly spatial overlap queries from O(n^2) to O(n). Accurate spatial associations further enable 3D merging of 2D masks through simple similarity-based filtering in a zero-shot manner, making our approach more robust to incomplete and noisy data. Evaluated on the ScanNet and SceneNN benchmarks, our approach achieves state-of-the-art performance in online, open-vocabulary 3D instance segmentation with leading efficiency.

Multimodal Large Language Models (MLLMs) have significantly progressed in offline video understanding. However, applying these models to real-world scenarios, such as autonomous driving and human-computer interaction, presents unique challenges due to the need for real-time processing of continuous online video streams. To this end, this paper presents systematic efforts from three perspectives: evaluation benchmark, model architecture, and training strategy. First, we introduce OVBench, a comprehensive question-answering benchmark designed to evaluate models' ability to perceive, memorize, and reason within online video contexts. It features 6 core task types across three temporal contexts--past, current, and future--forming 16 subtasks from diverse datasets. Second, we propose a new Pyramid Memory Bank (PMB) that effectively retains key spatiotemporal information in video streams. Third, we proposed an offline-to-online learning paradigm, designing an interleaved dialogue format for online video data and constructing an instruction-tuning dataset tailored for online video training. This framework led to the development of VideoChat-Online, a robust and efficient model for online video understanding. Despite the lower computational cost and higher efficiency, VideoChat-Online outperforms existing state-of-the-art offline and online models across popular offline video benchmarks and OVBench, demonstrating the effectiveness of our model architecture and training strategy.

Recent continuous learning (CL) research primarily addresses catastrophic forgetting within a straightforward learning framework where class and task information are predefined. However, in the Task-Free Continual Learning (TFCL), representing a more realistic and challenging CL scenarios, such information is typically absent. In this paper, we address the online TFCL by introducing an innovative memory management approach, by incorporating a dynamic memory system for storing selected data representatives from evolving distributions while a dynamically expandable memory system enables the retention of essential long-term knowledge. The proposed dynamic expandable memory system manages a series of memory distributions, each designed to represent the information from a distinct data category. A new memory expansion mechanism that assesses the proximity between incoming samples and existing memory distributions is proposed for evaluating when to add new memory distributions into the system. Additionally, a novel memory distribution augmentation technique is proposed for selectively gathering suitable samples for each memory distribution, enhancing the statistical robustness over time. To prevent memory saturation before the training phase, we introduce a memory distribution reduction strategy that automatically eliminates overlapping memory distributions, ensuring adequate capacity for accommodating new information in subsequent learning episodes. We conduct a series of experiments demonstrating that our proposed approach attains state-of-the-art performance in both supervised and unsupervised learning contexts. The source code is available at https://github.com/dtuzi123/DEMD.

6D object pose estimation remains challenging for many applications due to dependencies on complete 3D models, multi-view images, or training limited to specific object categories. These requirements make generalization to novel objects difficult for which neither 3D models nor multi-view images may be available. To address this, we propose a novel method One2Any that estimates the relative 6-degrees of freedom (DOF) object pose using only a single reference-single query RGB-D image, without prior knowledge of its 3D model, multi-view data, or category constraints. We treat object pose estimation as an encoding-decoding process: first, we obtain a comprehensive Reference Object Pose Embedding (ROPE) that encodes an object's shape, orientation, and texture from a single reference view. Using this embedding, a U-Net-based pose decoding module produces Reference Object Coordinate (ROC) for new views, enabling fast and accurate pose estimation. This simple encoding-decoding framework allows our model to be trained on any pair-wise pose data, enabling large-scale training and demonstrating great scalability. Experiments on multiple benchmark datasets demonstrate that our model generalizes well to novel objects, achieving state-of-the-art accuracy and robustness even rivaling methods that require multi-view or CAD inputs, at a fraction of compute.

3D object canonicalization is a fundamental task, essential for various downstream tasks. Existing methods rely on either cumbersome manual processes or priors learned from extensive, per-category training samples. Real-world datasets, however, often exhibit long-tail distributions, challenging existing learning-based methods, especially in categories with limited samples. We address this by introducing the first one-shot category-level object canonicalization framework that operates under arbitrary poses, requiring only a single canonical model as a reference (the "prior model") for each category. To canonicalize any object, our framework first extracts semantic cues with large language models (LLMs) and vision-language models (VLMs) to establish correspondences with the prior model. We introduce a novel joint energy function to enforce geometric and semantic consistency, aligning object orientations precisely despite significant shape variations. Moreover, we adopt a support-plane strategy to reduce search space for initial poses and utilize a semantic relationship map to select the canonical pose from multiple hypotheses. Extensive experiments on multiple datasets demonstrate that our framework achieves state-of-the-art performance and validates key design choices. Using our framework, we create the Canonical Objaverse Dataset (COD), canonicalizing 32K samples in the Objaverse-LVIS dataset, underscoring the effectiveness of our framework on handling large-scale datasets. Project page at https://jinli998.github.io/One-shot_3D_Object_Canonicalization/

With the rise of generative models, there is a growing interest in unifying all tasks within a generative framework. Anomaly detection methods also fall into this scope and utilize diffusion models to generate or reconstruct normal samples when given arbitrary anomaly images. However, our study found that the diffusion model suffers from severe faithfulness hallucination and catastrophic forgetting, which can't meet the unpredictable pattern increments. To mitigate the above problems, we propose a continual diffusion model that uses gradient projection to achieve stable continual learning. Gradient projection deploys a regularization on the model updating by modifying the gradient towards the direction protecting the learned knowledge. But as a double-edged sword, it also requires huge memory costs brought by the Markov process. Hence, we propose an iterative singular value decomposition method based on the transitive property of linear representation, which consumes tiny memory and incurs almost no performance loss. Finally, considering the risk of over-fitting to normal images of the diffusion model, we propose an anomaly-masked network to enhance the condition mechanism of the diffusion model. For continual anomaly detection, ours achieves first place in 17/18 settings on MVTec and VisA.

Text-to-Image (T2I) diffusion models have made remarkable advancements in generative modeling; however, they face a trade-off between inference speed and image quality, posing challenges for efficient deployment. Existing distilled T2I models can generate high-fidelity images with fewer sampling steps, but often struggle with diversity and quality, especially in one-step models. From our analysis, we observe redundant computations in the UNet encoders. Our findings suggest that, for T2I diffusion models, decoders are more adept at capturing richer and more explicit semantic information, while encoders can be effectively shared across decoders from diverse time steps.Based on these observations, we introduce the first Time-independent Unified Encoder (TiUE) for the student model UNet architecture, which is a loop-free image generation approach for distilling T2I diffusion models. Using a one-pass scheme, TiUE shares encoder features across multiple decoder time steps, enabling parallel sampling and significantly reducing inference time complexity. In addition, we incorporate a KL divergence term to regularize noise prediction, which enhances the perceptual realism and diversity of the generated images. Experimental results demonstrate that TiUE outperforms state-of-the-art methods, including LCM, SD-Turbo, and SwiftBrushv2, producing more diverse and realistic results while maintaining the computational efficiency.

High-speed autofocus in extreme scenes remains a significant challenge. Traditional methods rely on repeated sampling around the focus position, resulting in "focus hunting". Event-driven methods have advanced focusing speed and improved performance in low-light conditions; however, current approaches still require at least one lengthy round of "focus hunting", involving the collection of a complete focus stack. We introduce the Event Laplacian Product (ELP) focus detection function, which combines event data with grayscale Laplacian information, redefining focus search as a detection task. This innovation enables the first one-step event-driven autofocus, cutting focusing time by up to two-thirds and reducing focusing error by 24 times on the DAVIS346 dataset and 22 times on the EVK4 dataset. Additionally, we present an autofocus pipeline tailored for event-only cameras, achieving accurate results across a range of challenging motion and lighting conditions. All datasets and code will be made publicly available.

Transformers today still struggle to generate one-minute videos because self-attention layers are inefficient for long context. Alternatives such as Mamba layers struggle to produce coherent scenes because their hidden states are small and less expressive. We experiment with Test-Time Training (TTT) layers, whose hidden states themselves can be neural networks, therefore larger and more expressive. Adding TTT layers into a pre-trained Transformer enables it to generate one-minute videos from text storyboards. We curate a dataset based on Tom and Jerry cartoons as a proof-of-concept benchmark. Compared to baselines such as Mamba 2, Gated DeltaNet, and sliding-window attention layers, TTT layers generate much more coherent videos that tell complete stories, leading by 34 Elo points in a human evaluation of 100 videos per method. Although promising, our results are still limited in physical realism, and the efficiency of our implementation can be further improved. Sample videos, code and annotations are available at: https://test-time-training.github.io/video-dit

Collaborative perception in autonomous driving significantly enhances the perception capabilities of individual agents. Immutable heterogeneity in collaborative perception, where agents have different and fixed perception networks, presents a major challenge due to the semantic gap in their exchanged intermediate features without modifying the perception networks. Most existing methods bridge the semantic gap through interpreters. However, they either require training a new interpreter for each new agent type, limiting extensibility, or rely on a two-stage interpretation via an intermediate standardized semantic space, causing cumulative semantic loss. To achieve both extensibility in immutable heterogeneous scenarios and low-loss feature interpretation, we propose PolyInter, a polymorphic feature interpreter. It contains an extension point through which emerging new agents can seamlessly integrate by overriding only their specific prompts, which are learnable parameters intended to guide the interpretation, while reusing PolyInter's remaining parameters. By leveraging polymorphism, our design ensures that single interpreter is sufficient to accommodate diverse agents and interpret their features into the ego agent's semantic space. Experiments conducted on OPV2V dataset demonstrate that PolyInter improves collaborative perception precision by up to 11.1% compared to SOTA interpreters, while comparable results can be achieved by training only 1.4% of PolyInter's parameters when adapting to new agents.

Advanced image fusion methods mostly prioritise high-level missions, where task interaction struggles with semantic gaps, requiring complex bridging mechanisms. In contrast, we propose to leverage low-level vision tasks from digital photography fusion, allowing for effective feature interaction through pixel-level supervision. This new paradigm provides strong guidance for unsupervised multimodal fusion without relying on abstract semantics, enhancing task-shared feature learning for broader applicability. Owning to the hybrid image features and enhanced universal representations, the proposed GIFNet supports diverse fusion tasks, achieving high performance across both seen and unseen scenarios with a single model. Uniquely, experimental results reveal that our framework also supports single-modality enhancement, offering superior flexibility for practical applications. Our code will be available at https://github.com/AWCXV/GIFNet.

We introduce \texttt OneDiffusion - a single large-scale diffusion model designed to tackle a wide range of image synthesis and understanding tasks. It can generate images conditioned on text, depth, pose, layout, or semantic maps. It also handles super-resolution, multi-view generation, instant personalization, and text-guided image editing. Additionally, the same model is also effective for image understanding tasks such as depth estimation, pose estimation, open-vocab segmentation and camera pose estimation, etc.Our unified model is trained with simple but effective recipe: we casting all tasks as modeling a sequence of frames where we inject different noise scales during training to each frame. During inference, we can set any frame to a clean image for tasks like conditional image generation, camera pose estimation, or generating paired depth and images from a caption. Experiment results shows our proposed method achieve competitive performance on wide range of tasks. By eliminating the need for specialized architectures or finetuning, OneDiffusion provides enhanced flexibility and scalability, reflecting the emerging trend towards general-purpose diffusion models in the field.

Vision-language model (VLM) is one of the most important models for multi-modal tasks. Real industrial applications often meet the challenge of adapting VLMs to different scenarios, such as varying hardware platforms or performance requirements. Traditional methods involve training or fine-tuning to adapt multiple unique VLMs or using model compression techniques to create multiple compact models. These approaches are complex and resource-intensive. This paper introduces a novel paradigm called Once-Tuning-Multiple-Variants (OTMV). OTMV requires only a single tuning process to inject dynamic weight expansion capacity into the VLM with dynamic expansion capacity. This tuned VLM can then be expanded into multiple variants tailored for different scenarios in inference. The tuning mechanism of OTMV is inspired by the mathematical series expansion theorem, which helps to reduce the parameter size and memory requirements while maintaining accuracy for VLM. Experiment results show that OTMV-tuned models achieve comparable accuracy to baseline VLMs across various visual-language tasks. The experiments also demonstrate the dynamic expansion capability of OTMV-tuned VLMs, outperforming traditional model compression and adaptation techniques in terms of accuracy and efficiency.

Event cameras provide low-latency perception for only milliwatts of power. This makes them highly suitable for resource-restricted, agile robots such as small flying drones. Self-supervised learning based on contrast maximization holds great potential for event-based robot vision, as it foregoes the need for high-frequency ground truth and allows for online learning in the robot's operational environment. However, online, on-board learning raises the major challenge of achieving sufficient computational efficiency for real-time learning, while maintaining competitive visual perception performance. In this work, we improve the time and memory efficiency of the contrast maximization pipeline, making on-device learning of low-latency monocular depth possible. We demonstrate that online learning on board a small drone yields more accurate depth estimates and more successful obstacle avoidance behavior compared to only pre-training. Benchmarking experiments show that the proposed pipeline is not only efficient, but also achieves state-of-the-art depth estimation performance among self-supervised approaches. Our work taps into the unused potential of online, on-device robot learning, promising smaller reality gaps and better performance.

Pre-trained Vision-Language Models (VLMs) like CLIP, have demonstrated strong zero-shot generalization capabilities. Despite their effectiveness on various downstream tasks, they remain vulnerable to adversarial samples. Existing methods fine-tune VLMs to improve their performance via performing adversarial training on a certain dataset. However, this can lead to model overfitting and is not a true zero-shot scenario. In this paper, we propose a truly zero-shot and training-free approach that can significantly improve the VLM's zero-shot adversarial robustness. Specifically, we first discover that simply adding Gaussian noise greatly enhances the VLM's zero-shot performance. Then, we treat the adversarial examples with added Gaussian noise as anchors and strive to find a path in the embedding space that leads from the adversarial examples to the cleaner samples. We improve the VLMs' generalization abilities in a truly zero-shot and training-free manner compared to previous methods. Extensive experiments on 16 datasets demonstrate that our method can achieve state-of-the-art zero-shot robust performance, improving the top-1 robust accuracy by an average of 9.77%. The code will be publicly available.