cs.CV（2026-07-30）

📊 共 53 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (18 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (14 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (12 🔗2) 支柱一：机器人控制 (Robot Control) (4 🔗1) 支柱四：生成式动作 (Generative Motion) (2) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (18 篇)

#题目一句话要点标签🔗⭐
1 Capturing Token Tendencies for Training-Free Token Pruning in Multimodal Large Language Models 提出趋势感知修剪以解决多模态大语言模型的视觉令牌过滤问题 large language model multimodal
2 MMOOC: A Comprehensive Benchmark for Out-of-Context Evaluation in Multimodal Large Language Models 提出MMOOC基准以解决多模态大语言模型的上下文评估问题 large language model multimodal
3 Witness Evidence Portfolios: Single-Prefill Risk Detection for Closed Multimodal Answers 提出见证证据组合以解决闭合多模态答案的风险检测问题 large language model multimodal
4 MIND: Multimodal Intent-Driven Network via Diffusion Transformers for Medical Image Fusion 提出MIND以解决医学图像融合中的意图驱动问题 multimodal
5 Negative controls reveal volume-driven confounding in radiomics and imaging foundation model features 提出READII-2-ROQC以解决影像组学中的体积驱动混淆问题 foundation model
6 Beyond Classification: Pathology Foundation Models as Detection Encoders for Mitotic Figures 提出病理基础模型作为有丝分裂图像检测编码器 foundation model ✅
7 Benchmarking Foundation and Large Language Models for Few-Shot Medical Image Segmentation 提出FAME基准以评估少样本医学图像分割方法 large language model
8 DS@GT ARC at ImageCLEFmedical 2026: Architectural Diversity for Concept Detection and Foundation-Model Scaling for Caption Prediction in Medical Image Analysis 提出多样化架构以解决医学图像分析中的概念检测与标题生成问题 foundation model ✅
9 FiRE: Enhancing MLLMs with Fine-Grained Context Learning for Complex Image Retrieval 提出FiRE以解决复杂图像检索中的细粒度上下文建模问题 large language model multimodal
10 LAST: The Last Query Token Guides Visual Token Pruning for Edge-Cloud Collaborative MLLM Inference 提出LAST框架以解决边缘云协作MLLM推理中的视觉令牌修剪问题 foundation model multimodal
11 Thinking Once Is Enough: Intermediate-Layer Evidence Routing for High-Resolution VQA 提出Thinking-Once以解决高分辨率视觉问答中的证据获取问题 large language model multimodal
12 LoMeVQA: A Comprehensive Benchmark for Longitudinal Medical VQA 提出LoMeVQA以解决纵向医学视觉问答问题 large language model multimodal ✅
13 ROAD: Reciprocal-Objective Alignment of Discriminative Semantics for 3D Shape Generation 提出ROAD框架以降低3D形状生成的训练成本 foundation model ✅
14 ObjectStream: Latent Objects as Memory Anchors for Streaming Video Understanding 提出ObjectStream以解决流媒体视频理解中的记忆管理问题 large language model
15 Scaling Vision-Language Models Is Not Enough to Mitigate Bias 大规模研究揭示视觉-语言模型偏见问题的复杂性 multimodal
16 TongueReenact: Geometry-Anchored Tongue Synthesis for Face Reenactment 提出几何锚定的舌头合成方法以解决面部重演中的舌头动态问题 foundation model
17 MeshFM: 2D Features Are All You Need for 3D Shape Understanding 提出MeshFM以解决3D形状理解中的2D特征利用问题 foundation model ✅
18 Drawing-Recode: Annotation Grounding for Parametric CAD Code Generation from Raster 2D CAD Drawings 提出Drawing-Recode以解决光栅2D CAD图纸的参数化CAD代码生成问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (14 篇)

#题目一句话要点标签🔗⭐
19 OPLD: On-Policy Latent Distillation for Multimodal Reasoning 提出OPLD以解决多模态推理中的抽象思维问题 distillation multimodal chain-of-thought
20 VAD: Attributing Visual Evidence for Target Reconstruction in Multimodal On-Policy Distillation 提出视觉归因蒸馏以解决多模态知识转移问题 distillation multimodal
21 Theia: Large-Scale Multimodal Captioning and Automated Validation of the Incidents1M Dataset for Data-Free Distillation 提出一种新方法以构建和验证灾害响应多模态数据集 distillation multimodal
22 AuricularWorld: Hierarchical Action-Guided World Modeling for Fine-Grained Auricular Structure Segmentation from CT Scans 提出基于世界模型的分割框架以解决耳部CT图像细粒度分割问题 world model world models latent dynamics
23 RefineSVG: Visual Feedback-Driven Reinforcement Learning for Image-to-SVG Generation 提出RefineSVG以解决图像到SVG生成中的几何漂移问题 reinforcement learning large language model multimodal ✅
24 Learning to Understand Body Language from Flight through Robust 3D Avatar Placing 提出Drones2BodyLanguage数据集以解决无人机社交智能问题 world model world models monocular depth
25 PhiZero: A World Model Built Around Physical Language 提出PhiZero以解决物理世界建模中的隐式动态问题 world model world models
26 ShadowDancer: Teaching Video World Models Any Action by Learning Unified Dynamics Representations from a Video and Its Shadow 提出ShadowDancer以解决视频世界模型中的动作控制问题 world model world models
27 Beacon: Knowing When and How to Perform Agentic Visual Reasoning 提出Beacon模型以提升多模态大语言模型的视觉推理能力 reinforcement learning large language model multimodal
28 One Patch Is Enough: Reinforcement-Optimized Visual Token Grounding for MLLM-Based Scene Text Spotting 提出单补丁文本检测方法以解决场景文本定位精度问题 reinforcement learning large language model multimodal
29 SPFM-Net: Semantic-Prior-Guided Frequency-Constrained Mamba for Invisible Watermark Attack 提出SPFM-Net以解决隐形水印攻击效果与视觉保真度的权衡问题 Mamba masked autoencoder
30 ViP-Rig: Visual-Prompted Controllable Rigging 提出ViP-Rig以解决动画任务中骨骼控制不足的问题 VIP
31 Learning Color Grading, No Photo Sharing: Federated Aesthetic Preference Learning for Personalized Image Enhancement 提出FedPAIE框架以解决个性化图像增强中的隐私问题 preference learning
32 Temporal Concentration from Rollout Errors: Implicit Preference Optimization for Text-to-Video Diffusion 提出集中隐式偏好优化以解决视频生成中的时间稀疏伪影问题 DPO direct preference optimization

🔬 支柱三：空间感知与语义 (Perception & Semantics) (12 篇)

#题目一句话要点标签🔗⭐
33 Beyond Visual Ambiguity: Guiding Robust Monocular Depth Estimation in Challenging Scenarios via Detailed Long Captions 提出CapDepth以解决单目深度估计中的视觉歧义问题 depth estimation monocular depth spatial relationship
34 ViewMind3D: Modular View-Aware Inference for Training-Free 3D-QA 提出ViewMind3D以解决训练依赖的3D问答问题 3D reconstruction embodied AI large language model
35 MonoVoc: Decoupling Geometry and Semantics for Lightweight Monocular Open-Vocabulary 3D Gaussians 提出MonoVoc以解决轻量级单目开放词汇3D场景理解问题 scene understanding open-vocabulary open vocabulary
36 Endo-NeRF++: Uncertainty-Aware Neural Rendering with Multi-Resolution Hash Encoding for Dynamic Surgical Scene Reconstruction 提出Endo-NeRF++以解决动态外科场景重建中的不确定性问题 NeRF scene reconstruction
37 4DHumanDiff: Direct Text-to-4DGS Generation for Consistent 360-Degree Dynamic Humans 提出4DHumanDiff以解决360度动态人类生成问题 gaussian splatting splatting scene reconstruction
38 S-Avatar: Diffusion-Guided Gaussian Head Avatars from a Single Image 提出S-Avatar以解决单图生成3D头像一致性问题 3D gaussian splatting 3DGS gaussian splatting ✅
39 AdaAnchor4D: Anchor-Conditioned Spatiotemporal Feature Aggregation for Monocular UAV 4D Reconstruction 提出AdaAnchor4D以解决单目无人机动态场景重建问题 scene reconstruction spatiotemporal
40 Convolutional Neural Shading for High-Quality 3D Reconstruction from Multi-View Images 提出卷积神经着色以解决多视图图像高质量3D重建问题 3D reconstruction neural radiance field
41 Filling the Pareto-Optimal Front for Affordance Segmentation on Embedded Devices Using RGB-D Cameras 提出深度信息融合方法以优化可穿戴机器人中的可供性分割 affordance
42 Split and Drive: Dual-Axis Disentanglement for Real-Time Gaussian Head Avatars 提出SpiD框架以解决单图像生成头部头像问题 3D gaussian splatting gaussian splatting splatting
43 TARS: Timestep-Aware Data Scaling for 3D-Free Video Re-Shooting 提出TARS以解决3D视频重拍中的视角控制问题 3D reconstruction spatiotemporal ✅
44 TSOG: A Format For Temporally And Spatially Ordered Gaussians 提出TSOG格式以高效表示动态4D高斯内容 gaussian splatting splatting

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
45 ACE-Data-0: Human-Centric Ambient Capture as Embodied Data Engine 提出ACE-Data-0以解决人类中心数据捕获的瓶颈问题 locomotion manipulation dexterous manipulation
46 EgoGenesis: Egocentric World-Action Modeling with Online Anchored Projective Memory and Action-3D RoPE 提出EgoGenesis以解决稀缺的自我中心数据问题 manipulation dual-arm egocentric
47 FaithEyes: Towards Faithful Tool Use via Multi-Agent Process-Image Verification 提出FaithEyes以解决工具使用不可靠问题 manipulation multimodal ✅
48 Can Vision-Language Models Reason about AI Edits in Images? 提出基于强化学习的VLM框架以检测AI篡改图像 manipulation reinforcement learning

🔬 支柱四：生成式动作 (Generative Motion) (2 篇)

#题目一句话要点标签🔗⭐
49 MPIE-Bench: Benchmarking Anatomically Plausible Multi-Person Interaction Editing 提出MPIE-Bench以解决多人物交互编辑中的解剖学问题 penetration multi-person interaction
50 Articulated Object Reconstruction from Rest-State Observation 提出静态状态重建方法以解决关节物体重建问题 physically plausible geometric consistency

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
51 Hand-Object Interaction in the Age of Large Foundation Models:Reconstruction, Generation, and Embodied Transfer 提出系统性框架以提升手-物交互建模能力 HOI human-to-robot foundation model

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
52 ENCORE: Event-Assisted Complementary Motion Refinement for Learned Video Compression 提出ENCORE框架以解决学习视频压缩中的运动建模问题 motion representation TAMP

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
53 Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformers 提出Chimera以解决高分辨率图像和长视频生成问题 spatiotemporal multimodal

⬅️ 返回 cs.CV 首页 · 🏠 返回主页