CVPR 2026 论文列表

Closed-set object detection in remote sensing imagery has made significant progress, but achieving high detection accuracy remains challenging. Vision-Language Models (VLMs), which possess rich prior knowledge, offer a promising solution to this challenge. However, most existing VLMs are designed for open-vocabulary tasks and exhibit inherent limitations when directly applied to closed-set scenarios, such as notable accuracy degradation and high deployment costs. To address these issues, we propose VLM4RSDet, a novel collaborative training framework that leverages vision-language model to enhance the performance of conventional closed-set remote sensing object detectors. Notably, during inference, VLM4RSDet only retains the standard object detection architecture, thus avoiding any additional deployment overhead. Furthermore, we introduce a Global-Local Cross-Attention (GLCA) module and a Learnable Hierarchical Prediction Strategy (LHPS) to further improve collaborative training performance. Extensive experiments on five benchmark datasets demonstrate the effectiveness and robustness of our approach. In particular, our method outperforms the state-of-the-art by 7.5% in mAP_ 0.5:0.95 on the VisDrone2019 dataset. Our code is available at \href https://github.com/cszzshi/VLM4RSDet https://github.com/cszzshi/VLM4RSDet .

Vision-Language-Action (VLA) models have recently enabled robotic manipulation by grounding visual and linguistic cues into actions. However, most VLAs assume the Markov property, relying only on the current observation and thus suffering from temporal myopia that degrades long-horizon coherence. In this work, we view motion as a more compact and informative representation of temporal context and world dynamics, capturing inter-state changes while filtering static pixel-level noise. From this perspective, HiF-VLA equips a motion-centric world model for the VLA, enabling agents to reason about temporal dynamics for future evolution during action generation. Building on this idea, we propose HiF-VLA (Hindsight, Insight, and Foresight for VLAs), a unified framework that leverages motion for bidirectional temporal reasoning. HiF-VLA encodes past dynamics through hindsight priors, anticipates future motion via foresight reasoning, and integrates both through a hindsight-modulated joint expert to enable a "think-while-acting" paradigm for long-horizon manipulation. As a result, HiF-VLA surpasses strong baselines on LIBERO-Long and CALVIN ABC-D benchmarks, while incurring negligible additional inference latency. Furthermore, HiF-VLA achieves substantial improvements in real-world long-horizon manipulation tasks, demonstrating its broad effectiveness in practical robotic settings.

The boundary representation (B-Rep) models a 3D solid as its explicit boundaries: trimmed corners, edges, and faces. Recovering B-Rep representation from unstructured data is a challenging and valuable task of computer vision and graphics. Recent advances in deep learning have greatly improved the recovery of 3D shape geometry, but still depend on dense and clean point clouds and struggle to generalize to novel shapes. We propose B-Rep Gaussian Splatting (BrepGaussian), a novel framework that learns 3D parametric representations from 2D images. We employ a Gaussian Splatting renderer with learnable features, followed by a specific fitting strategy. To disentangle geometry reconstruction and feature learning, we introduce a two-stage learning framework that first captures geometry and edges and then refines patch features to achieve clean geometry and coherent instance representations. Extensive experiments demonstrate the superior performance of our approach to state-of-the-art methods.

Vision-language-action (VLA) models unify perception, language, and control for embodied agents but face significant challenges in practical deployment due to rapidly increasing compute and memory demands, especially as models scale to longer horizons and larger backbones. To address these bottlenecks, we introduce QuantVLA, a training-free post-training quantization (PTQ) framework that, to our knowledge, is the first PTQ approach for VLA systems and the first to successfully quantize a diffusion transformer (DiT) action head. QuantVLA incorporates three scale-calibrated components: (1) a selective quantization layout that integerizes all linear layers in both the language backbone and the DiT while keeping attention projections in floating point to preserve the original operator schedule; (2) attention temperature matching, a lightweight per-head scaling mechanism that stabilizes attention logits and is folded into the dequantization scales at inference; and (3) output head balancing, a per-layer residual interface calibration that mitigates post-projection energy drift. The framework requires no additional training, uses only a small unlabeled calibration buffer, and supports integer kernels for low-bit weights and activations while leaving the architecture unchanged. Across representative VLA models on LIBERO, QuantVLA exceeds the task success rates of full-precision baselines, achieves about 70% relative memory savings on the quantized components, providing a practical pathway toward scalable low-bit embodied intelligence under strict compute, memory, and power constraints.

Diffusion-based approaches have recently driven remarkable progress in real-world image super-resolution (SR). However, existing methods still struggle to simultaneously preserve fine details and ensure high-fidelity reconstruction, often resulting in suboptimal visual quality. In this paper, we propose FiDeSR, a high-fidelity and detail-preserving one-step diffusion super-resolution framework. During training, we introduce a detail-aware weighting strategy that adaptively emphasizes regions where the model exhibits higher prediction errors. During inference, low- and high-frequency adaptive enhancers further refine the reconstruction without requiring model retraining, enabling flexible enhancement control. To further improve the reconstruction accuracy, FiDeSR incorporates a latent residual refinement, which corrects prediction errors in the diffusion noise and enhances fine detail recovery. FiDeSR achieves superior real-world SR performance compared to existing diffusion-based methods, producing outputs with both high perceptual quality and faithful content restoration. The source code will be released at: https://github.com/Ar0Kim/FiDeSR.

Despite advances in the application of MLLMs for various video tasks, video event prediction (VEP) remains relatively underexplored. VEP requires the model to perform fine-grained temporal modeling of videos and establish logical relationships between videos and future events, which current MLLMs still struggle with. In this work, we first present a comprehensive evaluation of current leading MLLMs on the VEP task, revealing the reasons behind their inaccurate predictions, including lack of logical reasoning ability for future events prediction and insufficient utilization of visual information. To address these challenges, we propose **C**hain **o**f **E**vents (**CoE**) paradigm, which constructs temporal event chains to implicitly enforce MLLM focusing on the visual content and the logical connections between videos and future events, incentivizing model's reasoning capability with multiple training protocols. Experimental results on public benchmarks demonstrate that our method outperforms both leading open-source and commercial MLLMs, establishing a new state-of-the-art on the VEP task.

Latent-space modeling has been the standard for Diffusion Transformers (DiTs). However, it relies on a two-stage pipeline where the pretrained autoencoder introduces lossy reconstruction, leading to error accumulation while hindering joint optimization. To address these issues, we propose PixelDiT, a single-stage, end-to-end model that eliminates the need for the autoencoder and learns the diffusion process directly in the pixel space. PixelDiT adopts a fully transformer-based architecture shaped by a dual-level design: a patch-level DiT that captures global semantics and a pixel-level DiT that refines texture details, enabling efficient training of a pixel-space diffusion model while preserving fine details. PixelDiT achieves 1.61 FID on ImageNet 256 and 1.81 FID on ImageNet 512, surpassing existing pixel generative models by a large margin. We further extend PixelDiT to text-to-image generation and pretrain it at the 1024^ 2 resolution in pixel space. It achieves 0.74 on GenEval and 83.5 on DPG-bench, approaching the best latent diffusion models.

Recent studies have demonstrated that incorporating trainable prompts into pretrained models enables effective incremental learning. However, the application of prompts in incremental object detection (IOD) remains underexplored. Our study reveals that existing prompts-pool-based approaches assume disjoint class sets across incremental tasks, which are unsuitable for IOD as they overlook the inherent co-occurrence phenomenon in detection. In co-occurring scenarios, unlabeled objects from previous tasks may appear in current task images, leading to confusion in prompts pool. In this paper, we hold that prompt structures should exhibit adaptive consolidation properties across tasks, with constrained updates to prevent confusion and catastrophic forgetting. Motivated by this, we introduce Parameterized Prompts for Incremental Object Detection (P^2IOD). Leveraging neural networks global evolution properties, P^2IOD employs networks as the parameterized prompts to adaptively consolidate knowledge across tasks. To constrain prompts structure updates, P^2IOD further engages a parameterized prompts fusion strategy. Extensive experiments on PASCAL VOC2007 and MS COCO datasets demonstrate that P^2IOD's effectiveness in IOD and achieves the state-of-the-art performance among existing baselines.

The heterogeneity between high-level vision-language understanding and low-level action control remains a fundamental challenge in robotic manipulation. Although recent methods have advanced task-specific action alignment, they often struggle to generate robust and accurate actions for novel or semantically related tasks. To address this, we propose the Language-Grounded Decoupled Action Representation (LaDA) framework, which leverages natural language as a semantic bridge to connect perception and control. LaDA introduces a fine-grained intermediate layer of three interpretable action primitives--translation, rotation, and gripper control--providing explicit semantic structure for low-level actions. It further employs a semantic-guided soft-label contrastive learning objective to align similar action primitives across tasks, enhancing generalization and motion consistency. An adaptive weighting strategy, inspired by curriculum learning, dynamically balances contrastive and imitation objectives for stable and effective training. Extensive experiments on simulated benchmarks (LIBERO and MimicGen) and real-world demonstrations validate that LaDA achieves strong performance and generalizes effectively to unseen or related tasks.

Two-hand reconstruction from monocular images is hampered by complex poses and severe occlusions, which often cause interaction misalignment and two-hand penetration. We address this by decoupling the problem into 2D structural alignment and 3D spatial interaction alignment, each handled by a tailored component. For 2D alignment, we pioneer the attempt to unify heterogeneous structural priors (keypoints, segmentation, and depth) from vision foundation models as complementary structured guidance for two-hand recovery. Instead of extracting priors prediction as explicit inputs, we propose a fusion-alignment encoder that absorbs their structural knowledge implicitly, achieving foundation-level guidance without foundation-level cost. For 3D spatial alignment, we propose a two-hand penetration-free diffusion model that learns a generative mapping from interpenetrated poses to realistic, collision-free configurations. Guided by collision gradients during denoising, the model converges toward the manifold of valid two-hand interactions, preserving geometric and kinematic coherence. This generative formulation approach enables physically credible reconstructions even under occlusion or ambiguous visual input. Extensive experiments on InterHand2.6M and HIC show state-of-the-art or leading performance in interaction alignment and penetration suppression. Project: https://gaogehan.github.io/A2P/

Affordance learning is a complex challenge in many applications, where existing approaches primarily focus on the geometric structures, visual knowledge, and affordance labels of objects to determine interactable regions. However, extending this learning capability to a scene is significantly more complicated, as incorporating object- and scene-level semantics is not straightforward; for example, 3D instance identification often struggles with small, interactable, functional parts (i.e., knobs, handles, etc.). In this work, we introduce AffordBridge, a large-scale dataset with 291,637 functional interaction annotations across 685 high-resolution indoor scenes in the form of point clouds. Our affordance annotations are complemented by RGB images that are linked to the same instances within scenes. Building upon our dataset, we propose AffordMatcher, an affordance learning method that establishes coherent semantic correspondences between image-based and point cloud-based instances for keypoint matching, enabling a more precise identification of affordance regions based on cues, so-called visual signifiers. Experimental results on our dataset demonstrate the effectiveness of our approach against other methods. Our code and dataset will be made publicly available.

Visual navigation requires the robot to reach a specified goal such as an image, based on a sequence of first-person visual observations. While recent learning-based approaches have made significant progress, they often focus on improving policy heads or decision strategies while relying on simplistic feature encoders and temporal pooling to represent visual input. This leads to the loss of fine-grained spatial and temporal structure, ultimately limiting accurate action prediction and progress estimation. In this paper, we propose a unified spatio-temporal representation framework that enhances visual encoding for robotic navigation. Our approach extracts features from both image sequences and goal observations, and fuses them using the designed spatio-temporal fusion module. This module performs spatial graph reasoning within each frame and models temporal dynamics using a hybrid temporal shift module combined with multi-resolution difference-aware convolution. Experimental results demonstrate that our approach consistently improves navigation performance and offers a generalizable visual backbone for goal-conditioned control. Code is available at https://github.com/hren20/STRNet.

Is basic visual understanding really solved in state-of-the-art VLMs? We present VisualOverload, a slightly different visual question answering (VQA) benchmark comprising 2,720 question-answer pairs, with privately held ground-truth responses. Unlike prior VQA datasets that typically focus on near global image understanding, VisualOverload challenges models to perform simple, knowledge-free vision tasks in densely populated (or, overloaded) scenes. Our dataset consists of high-resolution scans of public-domain paintings that are populated with multiple figures, actions, and unfolding subplots set against elaborately detailed backdrops. We manually annotated these images with questions across six task categories to probe for a thorough understanding of the scene. We hypothesize that current benchmarks overestimate the performance of VLMs, and encoding and reasoning over details is still a challenging task for them, especially if they are confronted with densely populated scenes. Indeed, we observe that even the best model (o3) out of 37 tested models only achieves 19.6% accuracy on our hardest test split and overall 69.5% accuracy on all questions. Beyond a thorough evaluation, we complement our benchmark with an error analysis that reveals multiple failure modes, including a lack of counting skills, failure in OCR, and striking logical inconsistencies under complex tasks. Altogether, VisualOverload exposes a critical gap in current vision models and offers a crucial resource for the community to develop better models.

Test-Time Adaptation (TTA) is essential for enabling deep learning models to handle real-world data distribution shifts. However, current approaches face significant limitations: backpropagation-based methods are not suitable for low-end deployment devices, due to their high computation and memory requirements, as well as their tendency to modify model weights during adaptation; while traditional backpropagation-free techniques exhibit constrained adaptation capabilities. In this work, we propose Forward-Only Zeroth-Order Optimization (FOZO), a novel and practical backpropagation-free paradigm for TTA. FOZO leverages a memory-efficient zeroth-order prompt optimization, which is led by objectives optimizing both intermediate feature statistics and prediction entropy. To ensure efficient and stable adaptation over the out-of-distribution data stream, we introduce a dynamically decaying perturbation scale during zeroth-order gradient estimation and theoretically prove its convergence under the TTA data stream assumption. Extensive continual adaptation experiments on ImageNet-C, ImageNet-R, and ImageNet-Sketch demonstrate FOZO's superior performance, achieving 59.52% Top-1 accuracy on ImageNet-C (5K, level 5) and outperforming main gradient-based methods and SOTA forward-only FOA (58.13%). Furthermore, FOZO exhibits strong generalization on quantized (INT8) models. These findings demonstrate that FOZO is a highly competitive solution for TTA deployment in resource-limited scenarios. Code: https://github.com/eVI-group-SCU/FOZO

MeanFlow promises high-quality generative modeling in few steps, by jointly learning instantaneous and average velocity fields. Yet, the underlying training dynamics remain unclear. We analyze the interaction between the two velocities and find: (i) well-established instantaneous velocity is a prerequisite for learning average velocity; (ii) learning of instantaneous velocity benefits from average velocity when the temporal gap is small, but degrades as the gap increases; and (iii) task-affinity analysis indicates that smooth learning of large-gap average velocities, essential for one-step generation, depends on the prior formation of accurate instantaneous and small-gap average velocities. Guided by these observations, we design an effective training scheme that accelerates the formation of instantaneous velocity, then shifts emphasis from short- to long-interval average velocity. Our enhanced MeanFlow training yields faster convergence and significantly better few-step generation: With the same DiT-XL backbone, our method reaches an impressive FID of 2.87 on 1-NFE ImageNet 256x256, compared to 3.43 for the conventional MeanFlow baseline. Alternatively, our method matches the performance of the MeanFlow baseline with 2.5x shorter training time, or with a smaller DiT-L backbone.

Specular highlights distort appearance, obscure texture, and hinder geometric reasoning in both natural and surgical imagery. We present UnReflectAnything, an RGB-only framework that removes highlights from a single image by predicting a highlight map together with a reflection-free diffuse reconstruction. The model uses a frozen vision transformer encoder to extract multi-scale features, a lightweight head to localize specular regions, and a token-level inpainting module that restores corrupted feature patches before producing the final diffuse image. To overcome the lack of paired supervision, we introduce a Virtual Highlight Synthesis pipeline that renders physically plausible specularities using monocular geometry, Fresnel-aware shading, and randomized lighting which enables training on arbitrary RGB images with correct geometric structure. UnReflectAnything generalizes across natural and surgical domains where non-Lambertian surfaces and non-uniform lighting create severe highlights and it achieves competitive performance with state-of-the-art results on several benchmarks. Project Page: alberto-rota.github.io/UnReflectAnything

Most instruction-driven 3D editing methods rely on 2D models to guide the explicit and iterative optimization of 3D representations. This paradigm, however, suffers from two primary drawbacks. First, it lacks a universal design of different 3D editing tasks because the explicit manipulation of 3D geometry necessitates task-dependent rules, e.g., 3D appearance editing demands inherent source 3D geometry, while 3D removal alters source geometry. Second, the iterative optimization process is highly time-consuming, often requiring thousands of invocations of 2D/3D updating. We present Omni-3DEdit, a unified, learning-based model that generalizes various 3D editing tasks implicitly. One key challenge to achieve our goal is the scarcity of paired source-edited multi-view assets for training. To address this issue, we construct a data pipeline, synthesizing a relatively rich number of high-quality paired multi-view editing samples. Subsequently, we adapt the pre-trained generative model SEVA as our backbone by concatenating source view latents along with conditional tokens in sequence space. A dual-stream LoRA module is proposed to disentangle different view cues, largely enhancing our model's representational learning capability. As a learning-based model, our model is free of the time-consuming online optimization, and it can complete various 3D editing tasks in one forward pass, reducing the inference time from tens of minutes to approximately two minutes. Extensive experiments demonstrate the effectiveness and efficiency of Omni-3DEdit.

The transition from image to video understanding requires vision-language models (VLMs) to shift from recognizing static patterns to reasoning over temporal dynamics such as motion trajectories, speed changes, and state transitions. Yet current post-training methods fall short due to two critical limitations: (1) existing datasets often lack temporal-centricity, where answers can be inferred from isolated keyframes rather than requiring holistic temporal integration; (2) training data generated by proprietary models contains systematic errors in fundamental temporal perception, such as confusing motion directions or misjudging speeds. We introduce SynRL, a post-training framework that teaches models temporal primitives, the fundamental building blocks of temporal understanding including direction, speed, and state tracking. Our key insight is that these abstract primitives, learned from programmatically generated synthetic videos, transfer effectively to real-world scenarios. We decompose temporal understanding into short-term perceptual primitives (speed, direction) and long-term cognitive primitives (state tracking, retrodictive inference), constructing 7.7K CoT and 7K RL samples with ground-truth frame-level annotations through code-based video generation. Despite training on simple geometric shapes, SynRL achieves substantial improvements across 15 benchmarks spanning temporal grounding, complex reasoning, and general video understanding. Remarkably, our 7.7K synthetic CoT samples outperform Video-R1's 165K real-world samples. We attribute this to fundamental temporal skills--such as tracking frame-by-frame changes and comparing velocity--that transfer effectively from abstract synthetic patterns to complex real-world scenarios. This establishes a new paradigm for video post-training: video temporal learning through carefully designed synthetic data provides a more cost-efficient scaling path. Code is released on https://github.com/jiangsongtao/Synthetic-Video.

Conventional knowledge distillation, designed for model compression, fails on long-tailed distributions because the teacher model tends to be biased toward head classes and provides limited supervision for tail classes. We propose Long-Tailed Knowledge Distillation (LTKD), a novel framework that reformulates the conventional objective into two components: a cross-group loss, capturing mismatches in prediction distributions across class groups (head, medium, and tail), and a within-group loss, capturing discrepancies within each group's distribution. This decomposition reveals the specific sources of the teacher's bias. To mitigate the inherited bias, LTKD introduces (1) a rebalanced cross-group loss that calibrates the teacher's group-level predictions and (2) a reweighted within-group loss that ensures equal contribution from all groups. Extensive experiments on CIFAR-100-LT, TinyImageNet-LT, and ImageNet-LT demonstrate that LTKD significantly outperforms existing methods in both overall and tail-class accuracy, thereby showing its ability to distill balanced knowledge from a biased teacher for real-world applications.