CVPR 2026 论文列表

Recently, unified multimodal models (UMMs) have made remarkable progress in integrating visual understanding and generation, demonstrating strong potential for complex text-to-image (T2I) tasks. Despite their theoretical promise, a persistent capability gap exists: UMMs typically exhibit superior visual understanding but comparatively weaker generative capabilities. This discrepancy arises largely from the intrinsic decoupling between the understanding and generation processes. While a UMM can accurately interpret fine-grained visual details, it often struggles to produce semantically coherent images from complex textual prompts. To address this challenge, we explore UMMs' internal understanding capability to enhance generation quality. We propose a token-level intrinsic text-image alignment reward mechanism, GvU, enabling the UMM to act simultaneously as teacher and student: it evaluates its own outputs using the understanding branch to guide the generations accordingly. Building upon this, we design a self-supervised reinforcement learning framework, allowing UMMs to iteratively improve their generation quality through understanding-based intrinsic reward signals--without reliance on external supervision. Experimental results show that our method substantially boosts UMMs' generation, which in turn strengthens their fine-grained visual understanding, narrowing the capability gap between UMMs' visual understanding and generation. The project page is https://matrix0721.github.io/gvu.github.io/.

Recent video diffusion models generate photorealistic, temporally coherent videos, yet they fall short as reliable world models for autonomous driving, where structured motion and physically consistent interactions are essential. Adapting these generalist video models to driving domains has shown promise but typically requires massive domain-specific data and costly fine-tuning. We propose an efficient adaptation framework that converts generalist video diffusion models into controllable driving world models with minimal supervision. The key idea is to decouple motion learning from appearance synthesis. First, the model is adapted to predict structured motion in a simplified form: videos of skeletonized agents and scene elements, focusing learning on physical and social plausibility. Then, the same backbone is reused to synthesize realistic RGB videos conditioned on these motion sequences, effectively "dressing" the motion with texture and lighting. This two-stage process mirrors a reasoning-rendering paradigm: first infer dynamics, then render appearance. Our experiments show this decoupled approach is exceptionally efficient: adapting SVD, we match prior SOTA models with less than 6% of their compute. Scaling to LTX, our MAD-LTX model outperforms all open-source competitors, and supports a comprehensive suite of text, ego, and object controls. Project page: \href https://vita-epfl.github.io/MAD-World-Model/ https://vita-epfl.github.io/MAD-World-Model/

Test-Time Training (TTT) has recently emerged as a promising direction for efficient sequence modeling. TTT reformulates attention operation as an online learning problem, constructing a compact inner model from key-value pairs at test time. This reformulation opens a rich and flexible design space while achieving linear computational complexity. However, crafting a powerful visual TTT design remains challenging: fundamental choices for the inner module and inner training lack comprehensive understanding and practical guidelines. To bridge this critical gap, in this paper, we present a systematic empirical study of TTT designs for visual sequence modeling. From a series of experiments and analyses, we distill six practical insights that establish design principles for effective visual TTT and illuminate paths for future improvement. These findings culminate in the Vision Test-Time Training (ViT^3) model, a pure TTT architecture that achieves linear complexity and parallelizable computation. We evaluate ViT^3 across diverse visual tasks, including image classification, image generation, object detection, and semantic segmentation. Results show that ViT^3 consistently matches or outperforms advanced linear-complexity models (e.g., Mamba and linear attention variants) and effectively narrows the gap to highly optimized vision Transformers. We hope this study and the ViT^3 baseline can facilitate future work on visual TTT models. Code: github.com/LeapLabTHU/ViTTT.

In this paper, we propose a cross-view fusion framework that enhances the robustness of 6-DoF grasp pose estimation in corner views.Our framework alleviates occlusion by incorporating an auxiliary view and avoids the time-consuming, task-agnostic multi-view reconstruction through a post-fusion strategy.To enable cross-view fusion, we propose a self-supervised contrastive learning strategy that leverages cross-view associations to regularize point cloud features.In brief, a cross-view point pair is considered a match if the two points correspond to the same 3D location, and a non-match if they represent distinct grasp directions.The learning strategy significantly enhances the spatial consistency and direction distinctiveness of point features, thereby facilitating cross-view fusion and improving estimation robustness.Furthermore, we propose a cross-view-aligned cylinder integration module to fuse grasp-relevant geometry into a comprehensive representation.Specifically, the module first aligns the cross-view points and features according to their similarity to enhance the robustness against noise.Subsequently, these points are registered into the cylindrical coordinate frame, emphasizing the rotation-symmetric geometry which is important for grasping.Finally, local self-attention and seed cross-attention layers are alternately employed, respectively enabling interactions within single views and across views, which supports fine-grained representation of grasp-relevant geometry.Our framework achieves strong performance on the GraspNet-1Billion benchmark and in real-world applications. Code is available at GitHub.

Decoupled dataset distillation (DD) compresses large corpora into a few synthetic images by matching a frozen teacher's statistics. However, current residual-matching pipelines rely on static real patches, creating a fit-complexity gap and a pull-to-anchor effect that reduce intra-class diversity and hurt generalization. To address these issues, we introduce RETA--a Retrieval and Topology Alignment framework for decoupled DD. First, Dynamic Retrieval Connection (DRC) selects a real patch from a prebuilt pool by minimizing a fit-complexity score in teacher feature space; the chosen patch is injected via a residual connection to tighten feature fit while controlling injected complexity. Second, Persistent Topology Alignment (PTA) regularizes synthesis with persistent homology: we build a mutual k-NN feature graph, compute persistence images of components and loops, and penalize topology discrepancies between real and synthetic sets, mitigating pull-to-anchor effect. Across CIFAR-100, Tiny-ImageNet, ImageNet-1K, and multiple ImageNet subsets, RETA consistently outperforms various baselines under comparable time and memory, especially reaching 64.3% top-1 accuracy on ImageNet-1K with ResNet-18 at 50 images per class, +3.1% over the best prior.

Masked Image Modeling has been one of the most popular self-supervised learning paradigms to learn representations from large-scale, unlabeled Earth Observation images. While incorporating multi-modal and multi-temporal Earth Observation data into Masked Image Modeling has been widely explored, the spatial dependencies between images captured from neighboring areas remains largely overlooked. Since the Earth's surface is continuous, neighboring images are highly related and offer rich contextual information for self-supervised learning. To close this gap, we propose NeighborMAE, which learns spatial dependencies by joint reconstruction of neighboring Earth Observation images. To ensure that the reconstruction remains challenging, we leverage a heuristic strategy to dynamically adjust the mask ratio and the pixel-level loss weight. Experimental results across various pretraining datasets and downstream tasks show that NeighborMAE significantly outperforms existing baselines, underscoring the value of neighboring images in Masked Image Modeling for Earth Observation and the efficacy of our designs. Our code is available on https://github.com/LeungTsang/NeighborMAE.

In hematoxylin-eosin (H&E) to virtual immunohistochemistry (IHC) staining, paired images enable supervised learning but suffer from inherent spatial dislocation, limiting pixel-level constraints. Thus, auxiliary tasks have been increasingly employed with paired data to provide complementary supervision. However, existing methods largely overlook the rich semantic information embedded in auxiliary task models. This paper proposes a novel framework for virtual IHC staining guided by dual-aligned multi-task features, which fully explores semantic cues from auxiliary tasks. To realize effective guidance, we address two obstacles: (1) the spatial mismatch between paired H&E and IHC feature representations; (2) the task gap between auxiliary task features and virtual staining features. To resolve the spatial mismatch, we generate an alignment matrix that aligns H&E and IHC features. Specifically, we first introduce structure-enhanced learning to restore semantic consistency in regions affected by inaccurate staining in virtual IHC images. Then, we separately cluster features from virtual IHC and real IHC images, and establish semantic correspondences using an active-passive matching mechanism. This ensures that only semantically aligned regions are matched, reducing the impact of staining variability on the alignment matrix. To bridge the task gap, we introduce a task-gap alignment module trained under the principle that auxiliary features are considered aligned if they improve the performance of the virtual IHC staining model. Extensive experiments on two public datasets with four biomarkers demonstrate the effectiveness of our framework. Code is available at https://github.com/U-RBook/VSMT.

Open-vocabulary change detection (OVCD) seeks to recognize arbitrary changes of interest by enabling generalization beyond a fixed set of predefined classes. We reformulate OVCD as a two-stage pipeline: first generate class-agnostic change proposals using visual foundation models (VFMs) such as SAM and DINOv2, and then perform category identification with vision-language models (VLMs) such as CLIP. We reveal that category identification errors are the primary bottleneck of OVCD, mainly due to the limited ability of VLMs based on image-text matching to represent fine-grained land-cover categories. To address this, we propose OpenDPR, a training-free vision-centric diffusion-guided prototype retrieval framework. OpenDPR leverages diffusion models to construct diverse prototypes for target categories offline, and to perform similarity retrieval with change proposals in the visual space during inference. The secondary bottleneck lies in change localization, due to the inherent lack of change priors in VFMs. To bridge this gap, we design a spatial-to-change weakly supervised change detection module named S2C to adapt their strong spatial modeling capabilities for change localization. Integrating the pretrained S2C into OpenDPR leads to an optional weakly supervised variant named OpenDPR-W, which further improves OVCD with minimal supervision. Experimental results on four benchmark datasets demonstrate that the proposed methods achieve state-of-the-art performance under both supervision modes. Code is available at https://github.com/guoqi2002/OpenDPR.

Recent work in 3D scene understanding is moving beyond purely spatial analysis toward functional scene understanding. However, existing methods often consider functional relationships between object pairs in isolation, failing to capture the scene-wide interdependence that humans use to resolve ambiguity. We introduce FunFact, a framework for constructing probabilistic open-vocabulary functional 3D scene graphs from posed RGB-D images. FunFact first builds an object- and part-centric 3D map and uses foundation models to propose semantically plausible functional relations. These candidates are converted into factor graph variables and constrained by both LLM-derived common-sense priors and geometric priors. This formulation enables joint probabilistic inference over all functional edges and their marginals, yielding substantially better calibrated confidence scores. To benchmark this setting, we introduce FunThor, a synthetic dataset based on AI2-THOR with part-level geometry and rule-based functional annotations. Experiments on SceneFun3D, FunGraph3D, and FunThor show that FunFact improves node and relation discovery recall and significantly reduces calibration error for ambiguous relations, highlighting the benefits of holistic probabilistic modeling for functional scene understanding. See our project page at https://funfact-scenegraph.github.io/.

Visual tokenizers play a crucial role in diffusion models. The dimensionality of latent space governs both reconstruction fidelity and the semantic expressiveness of the latent feature. However, a fundamental trade-off is inherent between dimensionality and generation quality, constraining existing methods to low-dimensional latent spaces. Although recent works have leveraged vision foundation models (VFMs) to enrich the semantics of visual tokenizers and accelerate convergence, high-dimensional tokenizers still underperform their low-dimensional counterparts. In this work, we propose RecTok, which overcomes the limitations of high-dimensional visual tokenizers through two key innovations: flow semantic distillation and reconstruction-alignment distillation. Our key insight is to make the forward flow in flow matching semantically rich, which serves as the training space of diffusion transformers, rather than focusing on the latent space as in previous works. Specifically, our method distill the semantic information in VFMs into the forward flow trajectories in flow matching. And we further enhance the semantics by introducing a masked feature reconstruction loss. Our RecTok achieves superior image reconstruction, generation quality, and discriminative performance. It achieves state-of-the-art results on the gFID-50K under both with and without classifier-free guidance settings, while maintaining a semantically rich latent space structure. Furthermore, as the latent dimensionality increases, we observe consistent improvements. Code and model will be publicly available.

Vision-Language Models (VLMs) still lack robustness in spatial intelligence, demonstrating poor performance on spatial understanding and reasoning tasks. We attribute this gap to the absence of a visual geometry learning process capable of reconstructing 3D space from 2D images. We present G^2VLM, a geometry grounded vision-language model that bridges two fundamental aspects of spatial intelligence: spatial 3D reconstruction and spatial understanding. G^2VLM natively leverages learned 3D visual geometry features to directly predict 3D attributes and enhance spatial reasoning tasks via in-context learning and interleaved reasoning. Our unified design is highly scalable for spatial understanding: it trains on abundant multi-view image and video data, while simultaneously leveraging the benefits of 3D visual priors that are typically only derived from hard-to-collect annotations.Experimental results demonstrate G^2VLM is proficient in both tasks, achieving comparable results to state-of-the-art feed-forward 3D reconstruction models and achieving better or competitive results across spatial understanding and reasoning tasks.By unifying a semantically strong VLM with low-level 3D vision tasks, we hope G^2VLM can serve as a strong baseline for the community and unlock more future applications, such as 3D scene editing.

Generating new Low-Rank Adaptation (LoRA) weights from pre-trained LoRAs has demonstrated strong generalization capabilities across various tasks, enabling the efficient transfer of AI models, particularly on resource-constrained edges. However, previous studies either merge base LoRAs via weighting coefficients or train a generative model under the closed-world assumption, limiting their efficiency and flexibility in complex edge user cases. This challenge may further increase when there are significant domain shifts between training and deployment. To this end, we propose Semantic-Guided LoRA Parameter Generation (SG-LoRA), a tuning-free generative framework to efficiently produce task-specific parameters for unseen tasks in a semantic-to-LoRA pipeline. Concretely, SG-LoRA uses task descriptions as the semantic bridge, measuring their proximity to a set of known expert tasks in a shared embedding space. Based on this semantic guidance, it models the target task's LoRA parameter distribution to generate high-performing parameters for novel tasks. SG-LoRA enables the real-time construction of LoRA models aligned with individual intents by distilling knowledge from prominent LoRA experts, while also offering a privacy-preserving solution for personalized model adaptation in a novel zero-shot open-world setting proposed in this work. Extensive experiments on multiple challenging tasks confirm the superior performance and remarkable adaptability of SG-LoRA. The code is attached in the supplementary material.

Multimodal Large Language Models (MLLMs) have demonstrated promising advancements in augmenting the capabilities of LLMs to comprehend visual input. However, modality misalignment between vision and text remains a key challenge in MLLM, which can be attributed to two aspects: misalignment of modality-specific representations and depletion of modality-specific details. To address the issue of modality misalignment, we propose DeepAlign, a novel multimodal alignment framework to mitigate modality conflict, which employs representation intervention and structure-induced knowledge distillation to prevent the misalignment and depletion of modality-specific information. Extensive experiments demonstrate that DeepAlign significantly mitigates modality conflicts, leading to substantial performance improvements compared to backbone models across multiple vision-language tasks. It also stimulates some emergent abilities in MLLMs, such as multimodal in-context learning on interleaved text-image sequences.

Domain adaptive object detection (DAOD) aims to generalize an object detector trained on a source domain to a target domain, where the domain gap degrades the adaptability. Recently, large-scale vision foundation models (VFMs), pretrained on web-scale datasets, exhibit such powerful generalization capabilities that many approaches leverage them to bridge the domain gap. However, their generalized knowledge is not tailored to the specific domain, which makes it difficult to offer precise guidance in the target domain. In this paper, we propose an Expert-Teacher-Student collaborative learning (ETS) framework to synergize the generalized knowledge from VFMs with the domain-specific knowledge from the teacher model. Concretely, we first design an Expert-Teacher Collaborative Teaching (ETCT) module, which leverages the complementary knowledge of expert and teacher models to collaboratively generate high-quality pseudo labels for supervising student model learning. Second, we devise an Expert-Teacher Joint Consolidating (ETJC) module, which introduces class-wise prototype alignment among expert, teacher, and student models, to jointly consolidate generalized and domain-specific knowledge within the student model. ETS leverages VFMs as the expert model in a free lunch manner, thus avoiding significant additional training costs. Extensive experiments exhibit that our method outperforms the existing SOTA methods on three benchmarks.Our code is available in the supplementary materials.

Model merging aims to integrate multiple task-specific fine-tuned models derived from a shared pre-trained checkpoint into a single multi-task model without additional training. Despite extensive research, task interference remains a major obstacle that often undermines the performance of merged models. In this paper, we propose ESM (Essential Subspace Merging) , a robust framework for effective model merging. We begin by performing Principal Component Analysis (PCA) on feature shifts induced by parameter updates. The resulting principal directions span an essential subspace that dominantly influences feature representations. Each task's parameter update matrix is projected onto its respective essential subspace for low-rank decomposition before merging. This methodology mitigates inter-task interference while preserving core task-specific functionality. Furthermore, we introduce a multi-level polarized scaling strategy that amplifies parameters containing critical knowledge and suppresses redundant ones, preventing essential knowledge from being overwhelmed during fusion. Extensive experiments across multiple task sets and model scales demonstrate that our method achieves state-of-the-art performance in multi-task model merging.

One of the major differentiators unlocked by learned codecs relative to their hard-coded traditional counterparts is their ability to be optimized directly to appeal to the human visual system. Despite this potential, a perceptual yet practical image codec is yet to be proposed.In this work, we aim to close this gap. We conduct a comprehensive study of the key modeling choices that govern the design of a practical learned image codec, jointly optimized for perceptual quality and runtime -- including within the ablations several novel techniques. We then perform performance-aware neural architecture search over millions of backbone configurations to identify models that achieve the target on-device runtime while maximizing compression performance as captured by perceptual metrics. We combine the various optimizations to construct a new codec that achieves a significantly improved tradeoff between speed and perceptual quality. Based on rigorous subjective user studies, it provides 2.3-3x bitrate savings against AV1, AV2, VVC, ECM and JPEG-AI, and 20-40% bitrate savings against the best learned codec alternatives. At the same time, on an iPhone 17 Pro Max, it encodes 12MP images as fast as 230ms, and decodes them in 150ms -- faster than most top ML-based codecs run on a V100 GPU.

Prompt learning has emerged as an effective technique for fine-tuning large-scale foundation models for downstream tasks. However, conventional prompt learning methods are prone to overfitting and can struggle with out-of-distribution generalization. To address these limitations, Bayesian prompt learning has been proposed, which frames prompt optimization as a Bayesian inference problem to enhance robustness. This paper introduces Repulsive Bayesian Prompt Learning (ReBaPL), a novel method for Bayesian prompt learning, designed to efficiently explore the complex and often multimodal posterior landscape of prompts. Our method integrates a cyclical step-size schedule with a stochastic gradient Hamiltonian Monte Carlo (SGHMC) algorithm, enabling alternating phases of exploration to discover new modes, and exploitation to refine existing modes. Furthermore, we introduce a repulsive force derived from a potential function over probability metrics (including Maximum Mean Discrepancy and Wasserstein distance) computed on the distributions of representations produced by different prompts. This representation-space repulsion diversifies exploration and prevents premature collapse to a single mode. Our approach allows for a more comprehensive characterization of the prompt posterior distribution, leading to improved generalization. In contrast to prior Bayesian prompt learning methods, our method provides a modular plug-and-play Bayesian extension of any existing prompt learning method based on maximum likelihood estimation. We demonstrate the efficacy of ReBaPL on several benchmark datasets, showing superior performance over state-of-the-art prompt learning methods.

A sliding-window inference strategy is commonly adopted in recent training-free open-vocabulary semantic segmentation methods to overcome limitation of the CLIP in processing high-resolution images. However, this approach introduces a new challenge: each window is processed independently, leading to semantic discrepancy across windows. To address this issue, we propose Global-Local Aligned CLIP (GLA-CLIP), a framework that facilitates comprehensive information exchange across windows. Rather than limiting attention to tokens within individual windows, GLA-CLIP extends key-value tokens to incorporate contextual cues from all windows. Nevertheless, we observe a window bias: outer-window tokens are less likely to be attended, since query features are produced through interactions within the inner window patches, thereby lacking semantic grounding beyond their local context. To mitigate this, we introduce a proxy anchor, constructed by aggregating tokens highly similar to the given query from all windows, which provides a unified semantic reference for measuring similarity across both inner- and outer-window patches. Furthermore, we propose a dynamic normalization scheme that adjusts attention strength according to object scale by dynamically scaling and thresholding the attention map to cope with small-object scenarios. Moreover, GLA-CLIP can be equipped on existing methods and broad their receptive field. Extensive experiments validate the effectiveness of GLA-CLIP in enhancing training-free open-vocabulary semantic segmentation performance. Codes are available at github.com/2btlFe/GLA-CLIP

Brain visual decoding aims to recognize and reconstruct perceptual visual content from brain activity, providing a promising potential for the development of brain-computer interfaces and brain-inspired intelligence. However, this task faces a fundamental challenge of information asymmetry: while natural images contain complex visual scenes with objects and backgrounds, the corresponding brain signals reflect focused attention on central objects while being contaminated by various neural noise. Previous methods that directly align visual and brain representations often overlook this inherent asymmetry, resulting in suboptimal decoding performance. To address this, we propose linguistic-prior-guided visual decoupling method, which introduces object-oriented textual descriptions as semantic guidance to explicitly decouple foreground objects from complex backgrounds in natural images. This design enables the model to automatically focus on task-relevant visual concepts while effectively filtering out irrelevant neural noise in brain signals, achieving a transition from asymmetric vision-brain alignment to semantic symmetric alignment. Extensive experiments on the THINGS-EEG and THINGS-MEG datasets demonstrate that our method achieves new state-of-the-art performance in the zero-shot brain-to-image retrieval task. The source code is available at https://github.com/TKQXX/BVSA.