CVPR 2026 论文列表

Zero-shot (ZS) 3D anomaly detection is crucial for reliable industrial inspection, as it enables detecting and localizing defects without requiring any target-category training data. Existing approaches render 3D point clouds into 2D images and leverage pre-trained Vision-Language Models (VLMs) for anomaly detection. However, such strategies inevitably discard geometric details and exhibit limited sensitivity to local anomalies. In this paper, we revisit intrinsic 3D representations and explore the potential of pre-trained Point-Language Models (PLMs) for ZS 3D anomaly detection. We propose BTP (Back To Point), a novel framework that effectively aligns 3D point cloud and textual embeddings. Specifically, BTP aligns multi-granularity patch features with textual representations for localized anomaly detection, while incorporating geometric descriptors to enhance sensitivity to structural anomalies. Furthermore, we introduce a joint representation learning strategy that leverages auxiliary point cloud data to improve robustness and enrich anomaly semantics. Extensive experiments on Real3D-AD and Anomaly-ShapeNet demonstrate that BTP achieves superior performance in ZS 3D anomaly detection.

Recent unified multimodal models (UMMs) have achieved remarkable progress in visual comprehension and generation. However, existing datasets and benchmarks focus predominantly on single-turn interactions, overlooking the multi-turn, context-dependent nature of real-world image creation and editing. To bridge this gap, we introduce W E A V E, the first comprehensive suite for in-context interleaved cross-modality comprehension and generation, comprising two complementary components. W E A V E-100k is a large-scale dataset of 100,000 interleaved samples spanning over 370,000 dialogue turns and 500,000 images, encompassing comprehension, editing, and generation tasks that demand reasoning over prior context. W E A V E-Bench is a human-annotated benchmark of 100 items with 480 images, equipped with a hybrid VLM judge evaluation framework that jointly leverages reference images and original-image-instruction pairs to assess multi-turn generation, visual memory, and world-knowledge reasoning across diverse domains. Experiments show that training on W E A V E-100k substantially improves vision comprehension, image editing, and comprehension-generation collaboration, while further enabling the emergence of visual-memory capabilities in UMMs. Extensive evaluations on W E A V E-Bench reveal persistent limitations of current approaches in multi-turn, context-aware image generation and editing. We hope W E A V E provides both a perspective and a foundation for advancing in-context interleaved comprehension and generation in the multimodal community.

Vision-Language Models (VLMs) have shown remarkable progress in Vision-Language Navigation (VLN), offering new possibilities for navigation decision-making that could benefit both robotic platforms and human users. However, real-world navigation is inherently conditioned by the agent's mobility constraints. For example, a sweeping robot cannot traverse stairs, while a quadruped can. We introduce Capability-Conditioned Navigation (CapNav), a benchmark designed to evaluate how well VLMs can navigate complex indoor spaces given an agent's specific physical and operational capabilities. CapNav defines five representative human and robot agents, each described with physical dimensions, mobility capabilities, and environmental interaction abilities. CapNav provides 45 real-world indoor scenes, 473 navigation tasks, and 2365 QA pairs to test if VLMs can traverse indoor environments based on agent capabilities. We evaluate 13 modern VLMs and find that current VLM's navigation performance drops sharply as mobility constraints tighten, and that even state-of-the-art models struggle with obstacle types that require reasoning on spatial dimensions. We conclude by discussing the implications for capability-aware navigation and the opportunities for advancing embodied spatial reasoning in future VLMs. The benchmark is available at https://makeabilitylab.github.io/CapNav/

Large language models (LLMs) often generate self-contradictory outputs, which severely impacts their reliability and hinders their adoption in practical applications. In video-language models (Video-LLMs), this phenomenon recently draws the attention of researchers. Specifically, these models fail to provide logically consistent responses to rephrased questions based on their grounding outputs. However, the underlying causes of this phenomenon remain underexplored. In this work, we adopt an interpretability-driven approach to analyze, statistically summarize, and intervention the potential factors of the phenomenon. We find that one of the primary reasons for the inconsistency in responses lies in the inability of cross-modal attention heads to effectively distinguish video tokens across different timestamps. To address this, we propose an attention enhancement method called **T**emporally **C**onditioned **A**ttention **S**harpening (**TCAS**), which constructs an enhancement objective based on attention distinctions to enhance the model's temporal resolution capability, thereby improving its temporal understanding logic consistency. Experimental results demonstrate that our method significantly enhances the temporal logic consistency of Video-LLMs. Further analyses reveal that our method indeed improves the temporal discriminability of attention heads, validating our conclusions. Additionally, our method even achieves performance improvements in general video temporal grounding tasks, suggesting that temporal logic consistency is an important factor in temporal understanding.

Text-Image Person Re-Identification (TI-ReID) is widely deployed in intelligent surveillance. Built on deep neural networks and vision-language models, TI-ReID models inherit vulnerabilities to adversarial attacks, posing security risks. Yet its security remains less explored than retrieval accuracy, and its robustness to adversarial attacks is largely unexplored. To fill this gap, we propose Reconstruction-residual based Targeted and Untargeted Attack (R^2TUA), which generates perturbations from an image and adversarial prompt that mislead TI-ReID into matching the perturbed image with the adversarial identity. To precisely inject identity attributes into perturbations and achieve fine-grained targeted attack, R^2TUA proposes Transformer-based Gradual Multimodal Fusion (TGMF) that fuses image and adversarial prompt progressively across layers with tunable cross-modal weight. In addition, we propose a fully differentiable Soft Clamp Function (SCF), ensuring perturbations remain inconspicuous while avoiding local gradient vanishing effects that would trap training into suboptimal local minima. To further align perturbed images with the adversarial text descriptions while leading them to mismatch their original descriptions, R^2TUA employs (PPLs) and matching losses during training. Extensive evaluations across multiple datasets and models demonstrate the superior untargeted attack and targeted attack performance of R^2TUA. It also exhibits strong adaptability and transferability against black-box models, outperforming all related attacks across multiple tasks.

With the rapid advances of powerful multimodal models such as GPT-4o, Nano Banana, and Seedream 4.0 in Image Editing, the performance gap between closed-source and open-source models is widening, primarily due to the scarcity of large-scale, high-quality training data and comprehensive benchmarks capable of diagnosing model weaknesses across diverse editing behaviors. Existing data construction methods face a scale-quality trade-off: human annotations are high-quality but not scalable, while automated pipelines suffer from error propagation and noise. To address this, we introduce a lightweight data pipeline that replaces multi-toolchains with an end-to-end model and a unified post-verification stage. For scalable quality control, we train a 7B dual-task expert model, Qwen-Verify, for efficient failure detection and instruction recaptioning. This pipeline yields UnicEdit-10M, a 10M-scale dataset spanning diverse basic and complex editing tasks. We also propose UnicBench, a general benchmark that extends beyond basic edits to explicitly assess spatial and knowledge-driven reasoning. To enable fine-grained diagnosis, we introduce novel metrics, including Non-edit Consistency and Reasoning Accuracy. Our analysis of mainstream models on UnicBench reveals their limitations and provides clear directions for future research. The dataset, benchmark, and code will be released.

Existing concept customization methods have achieved remarkable outcomes in high-fidelity and multi-concept customization. However, they often neglect the influence on the original model's behavior and capabilities when learning new personalized concepts. To address this issue, we propose PureCC. PureCC introduces a novel decoupled learning objective for concept customization, which combines the implicit guidance of the target concept with the original conditional prediction. This separated form enables PureCC to substantially focus on the original model during training. Moreover, based on this objective, PureCC designs a dual-branch training pipeline that includes a frozen extractor providing purified target concept representations as implicit guidance and a trainable flow model producing the original conditional prediction, jointly achieving pure learning for personalized concepts. Furthermore, PureCC introduces a novel adaptive guidance scale l* to dynamically adjust the guidance strength of the target concept, balancing customization fidelity and model preservation. Extensive experiments show that PureCC achieves state-of-the-art performance in preserving the original behavior and capabilities while enabling high-fidelity concept customization.

Large Vision-Language Models (VLMs) often answer classic visual illusions "correctly" on original images, yet persist with the same responses when illusion factors are inverted, even though the visual change is obvious to humans. This raises a fundamental question: do VLMs perceive visual changes or merely recall memorized patterns? While several studies have noted this phenomenon, the underlying causes remain unclear. To move from observations to systematic understanding, this paper introduces VI-Probe, a controllable visual-illusion framework with graded perturbations and matched visual controls (without illusion inducer) that disentangles visually grounded perception from language-driven recall. Unlike prior work that focuses on averaged accuracy, we measure stability and sensitivity using Polarity-Flip Consistency, Template Fixation Index, and an illusion multiplier normalized against matched controls. Experiments across different families reveal that response persistence arises from heterogeneous causes rather than a single mechanism. For instance, GPT-5 exhibits memory override, Claude-Opus-4.1 shows perception-memory competition, while Qwen variants suggest visual-processing limits. Our findings challenge single-cause views and motivate probing-based evaluation that measures both knowledge and sensitivity to controlled visual change. Data and code are available at https://sites.google.com/view/vi-probe/.

We address hallucination detection in Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) by framing the problem through the lens of dynamical systems stability theory. Rather than treating hallucination as a straightforward classification task, we conceptualize (M)LLMs as dynamical systems, where factual knowledge is represented by stable equilibrium points within the representation space. Our main insight is that hallucinations tend to arise at the boundaries of knowledge--transition regions separating stable and unstable zones. To capture this phenomenon, we propose Lyapunov Probes: lightweight networks trained with derivative-based stability constraints that enforce a monotonic decay in confidence under input perturbations. By performing systematic perturbation analysis and applying a two-stage training process, these probes reliably distinguish between stable factual regions and unstable, hallucination-prone areas. Experiments on diverse datasets and models demonstrate consistent improvements over existing baselines.

Abstract reasoning from minimal examples remains a core unsolved problem for frontier foundation models such as GPT-5. These models still fail to infer structured transformation rules from a handful of examples, which is a key hallmark of human intelligence. The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) provides a rigorous testbed for this capability, demanding conceptual rule induction and transfer to novel tasks. Most existing methods treat ARC-AGI as a purely textual reasoning task, overlooking the fact that humans rely heavily on visual abstraction when solving such puzzles. However, our experiments reveal a paradox: naively rendering ARC-AGI grids as images degrades performance due to imprecise rule execution. This leads to our central hypothesis that vision and language possess complementary strengths across distinct reasoning stages: vision supports global pattern abstraction and verification, whereas language specializes in symbolic rule formulation and precise execution. Building on this insight, we introduce two synergistic strategies: (1) Vision-Language Synergy Reasoning (VLSR), which decomposes ARC-AGI into modality-aligned subtasks; and (2) Modality-Switch Self-Correction (MSSC), which leverages vision to verify text-based reasoning for intrinsic error correction. Extensive experiments demonstrate that our approach yields up to a 4.33% improvement over text-only baselines across diverse flagship models and multiple ARC-AGI tasks. Our findings suggest that unifying visual abstraction with linguistic reasoning is a crucial step toward achieving generalizable, human-like intelligence.

We present AutoTraces, an autoregressive vision-language-trajectory model for robot trajectory forecasting in humam-populated environments, which harnesses the inherent reasoning capabilities of large language models (LLMs) to model complex human behaviors. In contrast to prior works that rely solely on textual representations, our key innovation lies in a novel trajectory tokenization scheme, which represents physical waypoints as discrete tokens, seamlessly integrated into the LLM's space through a lightweight encoder-decoder architecture. This design preserves the LLM's native autoregressive generation mechanism while extending it to physical coordinate spaces, facilitates modeling of long-term interactions in trajectory data. We further introduce an automated chain-of-thought (CoT) generation mechanism that leverages a multimodal LLM to infer spatio-temporal relationships from visual observations and trajectory data, eliminating reliance on manual annotation. Through a two-stage training strategy, our AutoTraces achieves SOTA forecasting accuracy, particularly in long-horizon prediction, while exhibiting strong cross-scene generalization and supporting flexible-length forecasting.

In recent years, there has been a growing trend in computer vision towards exploiting RAW sensor data, which preserves richer information compared to conventional low-bit RGB images. Early studies mainly focused on enhancing visual quality, while more recent efforts aim to leverage the abundant information in RAW data to improve the performance of visual perception tasks such as object detection and segmentation. However, existing approaches still face two key limitations: large-scale ISP networks impose heavy computational overhead, while methods based on tuning traditional ISP pipelines are restricted by limited representational capacity.To address these issues, we propose Task-Aware Image Signal Processor (TA-ISP), a compact RAW-to-RGB framework that produces task-oriented representations for pretrained vision models. Instead of heavy dense convolutional pipelines, TA-ISP predicts a small set of lightweight, multi-scale modulation operators that act at global, regional, and pixel scales to reshape image statistics across different spatial extents. This factorized control significantly expands the range of spatially varying transforms that can be represented while keeping memory usage, computation, and latency tightly constrained. Evaluated on several RAW-domain detection and segmentation benchmarks under both daytime and nighttime conditions, TA-ISP consistently improves downstream accuracy while markedly reducing parameter count and inference time, making it well suited for deployment on resource-constrained devices.

Despite Multimodal Large Language Models (MLLMs) having shown impressive capabilities, they may suffer from hallucinations. Empirically, we find that MLLMs attend disproportionately to task-irrelevant background regions compared with text-only LLMs, implying spurious background-answer correlations. We claim and analyze that (i) outcome-based rewards can be an important factor leading to spurious correlations, and (ii) spurious correlations can be an important factor leading to hallucinations. Based on these results, we propose Causal-Oriented Policy Optimization (COPO) to mitigate these spurious correlations, thus addressing the issue of hallucinations. It imposes token-level sufficiency and necessity constraints to measure each inference token's causal contribution, thus ensuring correct and evidence-grounded output. Specifically, we first evaluate each token's causal contribution via a newly proposed causal completeness reward. This reward is then used to construct a causally informed advantage function within the GRPO optimization framework, encouraging the model to focus on tokens that are causally sufficient and necessary for accurate generation. Experimental results across various benchmarks demonstrate the advantages of COPO.

Multimodal large language models (MLLMs) are increasingly deployed in real-world, agentic settings where outputs must not only be correct, but also conform to pre-defined data schemas. Despite recent progress in structured generation in textual domain, there is still no benchmark that systematically evaluates schema-grounded information extraction and reasoning over visual inputs. In this work, we conduct a comprehensive study of visual structural output capabilities for MLLMs with our carefully designed SO-BENCH benchmark. Covering four visual domains, including UI screens, natural images, documents, and charts, SO-BENCH is built from over 6.5K diverse JSON schemas and 1.8K curated image-schema pairs with human-verified quality. Benchmarking experiments on open-sourced andfrontier proprietary models reveal persistent gaps in predicting accurate, schema compliant outputs, highlighting the need for better multimodal structured reasoning. Beyond benchmarking, we further conduct training experiments to largely improve the model's structured output capability. We make the benchmark publicly available at https://github.com/apple/ml-sobench.

Transformers rely on explicit positional encoding to model structure in data. WhileRotary Position Embedding (RoPE) excels in 1D domains, its application to image generation reveals significant limitations such as fine-grained spatial relationmodeling, color cues, and object counting. This paper identifies key limitationsof standard multi-dimensional RoPE--rigid frequency allocation, axis-wise independence, and uniform head treatment--in capturing the complex structural biasesrequired for fine-grained image generation. We propose HARoPE, a head-wiseadaptive extension that inserts a learnable linear transformation parameterized viasingular value decomposition (SVD) before the rotary mapping. This lightweightmodification enables dynamic frequency reallocation, semantic alignment of rotaryplanes, and head-specific positional receptive fields while rigorously preservingRoPE's relative-position property. Extensive experiments on class-conditional ImageNet and text-to-image generation (Flux and MMDiT) demonstrate that HARoPEconsistently improves performance over strong RoPE baselines and other extensions. The method serves as an effective drop-in replacement, offering a principledand adaptable solution for enhancing positional awareness in transformer-basedimage generative models.

Document layout understanding remains data-intensive despite advances in semi-supervised learning. We present a framework that enhances semi-supervised detection by fusing visual predictions with structural priors from text-pretrained LLMs via principled probabilistic weighting. Given unlabeled documents, an OCR-LLM pipeline infers hierarchical regions which are combined with teacher detector outputs through inverse-variance fusion to generate refined pseudo-labels.Our method demonstrates consistent gains across model scales. With a lightweight SwiftFormer backbone (26M params), we achieve 88.2\pm0.3 AP using only 5% labels on PubLayNet. When applied to document-pretrained LayoutLMv3 (133M params), our fusion framework reaches 89.7\pm0.4 AP, surpassing both LayoutLMv3 with standard semi-supervised learning (89.1\pm0.4 AP, p=0.02) and matching UDOP [??] (89.8 AP) which requires 100M+ pages of multimodal pretraining. This demonstrates that LLM structural priors are complementary to both lightweight and pretrained architectures. Key findings include: (1) learned instance-adaptive gating improves over fixed weights by +0.9 AP with data-dependent PAC bounds correctly predicting convergence; (2) open-source LLMs enable privacy-preserving deployment with minimal loss (Llama-3-70B: 87.1 AP lightweight, 89.4 AP with LayoutLMv3); (3) LLMs provide targeted semantic disambiguation (18.7% of cases, +3.8 AP gain) beyond simple text heuristics.Total system cost includes \12 for GPT-4o-mini API or 17 GPU-hours for local Llama-3-70B per 50K pages, amortized across training runs.

Large Multimodal Models (LMMs) that process 3D data typically rely on heavy, pretrained visual encoders to extract geometric features. While recent 2D LMMs have begun to eliminate such encoders for efficiency and scalability, extending this paradigm to 3D remains challenging due to the unordered and large-scale nature of point clouds. This leaves a critical unanswered question: How can we design an LMM that tokenizes unordered 3D data effectively and efficiently without a cumbersome encoder? We propose Fase3D, the first efficient encoder-free Fourier-based 3D scene LMM. Fase3D tackles the challenges of scalability and permutation invariance with a novel tokenizer that combines point cloud serialization and the Fast Fourier Transform (FFT) to approximate self-attention. This design enables an effective and computationally minimal architecture, built upon three key innovations: First, we represent large scenes compactly via structured superpoints. Second, our space-filling curve serialization followed by an FFT enables efficient global context modeling and graph-based token merging. Lastly, our Fourier-augmented LoRA adapters inject global frequency-aware interactions into LLM backbones at a negligible cost. Fase3D achieves performance comparable to encoder-based 3D LMMs while being significantly more efficient in computation and parameters.

Humans naturally allocate more time before acting when handling complex tasks in the physical world. This paradigm has recently led to remarkable advances in boosting Large Language Models (LLMs) on complex tasks in digital domains. However, the potential of test-time computing remains largely unexplored for robotic foundation models that interact with the physical world. In this work, we propose FM-Steer, a test-time computing framework that augments flow-based Vision-Language-Action (VLA) generalist policies with value-guided sampling and cascaded action denoising, enabling stronger control performance and real-time action rates for dexterous robot manipulation. FM-Steer first introduces an intermediate flow verifier to estimate state-action values for candidate actions. At test time, the policy iteratively samples multiple noisy action proposals and retains the one with the highest predicted value, yielding value-aligned, high-quality actions without retraining. To satisfy the stringent frequency demands of robot control, FM-Steer further introduces cascaded action denoising, decoupling expensive value-guided sampling from fast action refinement. A lightweight Lite-Flow denoiser asynchronously takes the selected high-value noisy action and rapidly denoises it to produce the final control signal, enabling fluid, high-rate execution. During deployment, the intermediate verifier operates at a low frequency to provide value-guided sampling, while the Lite-Flow denoiser continually processes selected candidates to maintain real-time control. Extensive experiments demonstrate that FM-Steer scales flow-based VLA models effectively at test time and achieves state-of-the-art performance across diverse simulation benchmarks and real-world dexterous robotic tasks. The source code is available on the project page: https://hume-vla.github.io.

Few-Shot Class-Incremental Learning (FSCIL) poses a critical challenge in machine learning, requiring models to continuously integrate novel classes with limited samples while preserving knowledge of previously seen classes. While existing FSCIL approaches have demonstrated promising results, they still suffer from catastrophic forgetting and few-shot overfitting due to the challenge of balancing old knowledge retention with new knowledge acquisition. To address these challenges, we propose an innovative Semantic-Guided Global-Local Collaborative Prompt Learning (SGLC) framework. Built upon powerful pre-trained Vision-Language Models (VLMs), the framework first introduces a dual-alignment mechanism: globally aligning visual features with visual-textual prototypes and locally aligning multi-view visual features with local textual attribute features, which facilitates effective knowledge learning while preserving existing knowledge via frozen prototypes of previous classes. Furthermore, to alleviate overfitting, we incorporate Large Language Models (LLMs) to generate semantically rich textual descriptions, which simultaneously guide both global and local prompt learning through knowledge distillation. Extensive experiments on the miniImageNet, CIFAR-100, and CUB200 datasets demonstrate that SGLC performs favorably against the state-of-the-art methods.