cs.CV（2026-10-07）

📊 共 49 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) (13 🔗2) 支柱三：空间感知与语义 (Perception & Semantics) (12 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗3) 支柱一：机器人控制 (Robot Control) (6) 支柱七：动作重定向 (Motion Retargeting) (3) 支柱六：视频提取与匹配 (Video Extraction) (2 🔗1) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱四：生成式动作 (Generative Motion) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (13 篇)

#题目一句话要点标签🔗⭐
1 Explicit Geometric Chain-of-Thought for Vision-Language-Action in Autonomous Driving 提出GeoCoTDrive以解决VLA模型在自动驾驶中的几何信息不足问题 vision-language-action VLA foundation model
2 HarnessIR: Harnessing Multimodal Foundation Models for Universal Real-World Image Restoration 提出HarnessIR以解决复杂真实世界图像恢复问题 foundation model multimodal ✅
3 MemoCare: An Interactive Multimodal Mobile System for Automated Cognitive Screening 提出MemoCare以实现自动化多模态认知筛查 multimodal
4 DisParQ: Self-Supervised Part Concepts for Interpretable Vision Foundation Models 提出DisParQ以解决视觉模型可解释性不足的问题 foundation model
5 Mixture of Layers: Dynamic Layer Routing for Visual Reasoning 提出动态层路由方法以增强视觉推理能力 large language model multimodal visual grounding ✅
6 Quantifying Volumetric Risk: Class-Aware Asymmetric Weighted Conformal Prediction for 3D Medical Image Segmentation 提出类感知非对称加权保形预测以解决3D医学图像分割的不确定性问题 large language model foundation model multimodal
7 GRC-Net: Global Representation Consistency Network for Unsupervised Multimodal Anomaly Detection 提出GRC-Net以解决多模态异常检测中的重建不稳定问题 multimodal
8 SAREO-FM: Decoupled Semantic Supervision for SAR-EO Foundation Models 提出SAREO-FM以解决SAR与EO图像融合问题 foundation model
9 From Pixel to Coding: Evaluating the Figure Reproduction Capabilities of MLLMs 提出FigCodeBench以解决多模态大语言模型的图形重现评估问题 large language model multimodal
10 BagDINO: Multi-View Baggage Re-Identification with DINOv3 提出BagDINO以解决多视角行李重新识别问题 foundation model
11 Flow-of-Thought: A Framework for Visual Reasoning 提出Flow-of-Thought框架以解决视觉推理中的空间理解问题 large language model
12 Spatial Latent Reasoning for Embodied Reference Understanding 提出空间潜在推理框架以解决指向手势视觉定位问题 visual grounding
13 CRT-HMAR: Causal Requirement Tracing-Guided Hierarchical Multi-Agent Regulation for Open-Task-Aware Infrared-Visible Image Fusion 提出CRT-HMAR以解决开放任务场景下红外-可见光图像融合问题 multimodal

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

#题目一句话要点标签🔗⭐
14 DeltaSplat: Iterative Gaussian Refinement for Pose-Free Feed-Forward 3D Gaussian Splatting 提出DeltaSplat以解决无姿态3D高斯点云重建问题 3D gaussian splatting 3DGS gaussian splatting
15 TileSkipper: Region-Adaptive Tile Pruning for 3D Gaussian Splatting 提出TileSkipper以解决3D高斯点云渲染效率问题 3D gaussian splatting gaussian splatting splatting
16 ActiveLang: Active Open-Vocabulary 3D Mapping with Semantic-Uncertainty-Guided Exploration 提出ActiveLang以解决开放词汇3D映射中的语义不确定性问题 scene understanding open-vocabulary open vocabulary
17 RT-DETR-World: Transferring Rich LLM Semantics to Real-Time Open-Vocabulary Detection 提出RT-DETR-World以解决实时开放词汇检测问题 open-vocabulary open vocabulary
18 WAPR: A Foundation Model for Wide-Angle Refinement in Unseen Object Pose Estimation 提出WAPR以解决未见物体的6D姿态估计问题 6D pose estimation foundation model ✅
19 Gaussian Density Splatting Network 提出高斯密度溅射网络以解决人群计数问题 gaussian splatting splatting
20 Argos: Adapt Rich Geometric Priors for Generalizable Online Scene-Change-Detection 提出Argos以解决动态环境中的场景变化检测问题 3D reconstruction foundation model
21 Tetris3D: 3D Scene Generation With Objects That Fit Together 提出Tetris3D以解决3D场景重建中物体相互兼容性问题 scene reconstruction physically plausible
22 Hardware-aware Calibrated Clustered Attention for Efficient Visual Geometric Transformers 提出块状聚类注意力以解决VGGT的延迟瓶颈问题 scene reconstruction VGGT
23 KASALv2: Fully Automatic 3D Rotational Symmetry Classification and Axis Localization 提出KASALv2以解决3D物体旋转对称性自动分类问题 6D pose estimation ✅
24 OmniCam: Omni-Camera Trajectory Generation via Geometry-Grounded Pose Token Learning 提出OmniCam以解决全景图像生成相机轨迹问题 scene reconstruction
25 PPCAR-Net: Projection-Refined Parametric 3D Coronary Artery Reconstruction from Sparse X-ray Angiographic Views 提出PPCAR-Net以解决稀疏视图下冠状动脉重建问题 VGGT

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
26 Controllable Crowd Generation through World-Model Planning 提出Ctrl-CWM以解决灵活控制人群生成问题 world model world models human motion ✅
27 UltraWorld: Learning Interactive Ultrasound World Models from Untracked Clinical Videos with Acoustic Sampling Map 提出UltraWorld以解决无追踪临床视频中的超声世界模型学习问题 world model world models distillation ✅
28 FedSSMCoOp: SSM Encoders for light-weight Federated Prompt Learning for Few-shot Classification 提出FedSSMCoOp以解决生物医学领域的联邦少样本分类问题 Mamba SSM large language model
29 Event-Aligned Visual Action Reasoning for World Action Models 提出事件对齐视觉动作推理框架以优化世界动作模型 world action model world action models
30 $Δ$Representation: Geometry Supervised Representation Learning of Phenotypes via Counterfactual Reasoning for Medical VLMs 提出$Δ$Representation以解决医学VLM中病理表型表示不足问题 representation learning spatial relationship
31 MORCA: Offline-to-Online Reinforcement Learning for Adaptive Cache Reuse in Video Diffusion Acceleration 提出MORCA以解决视频扩散加速中的缓存重用问题 reinforcement learning ✅
32 VideoEvolve: Co-Evolving Memory and Retrieval for Long Video Understanding 提出VideoEvolve以解决长视频理解中的记忆与检索不匹配问题 reinforcement learning multimodal
33 Global Average Precision for Representation Learning 提出gSAP以解决现有信息检索指标的局限性 representation learning
34 MeshCarve: Artisan Mesh Generation with Flow Matching in Compact Latent Spaces 提出MeshCarve以解决艺术网格生成中的效率与质量问题 flow matching
35 TiTok: Audio-Visual LLM for Multi-Segment Temporal Grounding 提出TiTok以解决音视频多段时间定位问题 reinforcement learning large language model
36 SkillCycle: Co-Evolving Agent Policies and Skill Banks 提出SkillCycle框架以解决技能内化与规则修订的耦合问题 policy learning distillation

🔬 支柱一：机器人控制 (Robot Control) (6 篇)

#题目一句话要点标签🔗⭐
37 Video Prediction Policy 2: Predict Better, Act Better 提出Video Prediction Policy 2以解决现有WAMs在开放环境中的预测不足问题 manipulation world action model world action models
38 Self-correction Optimization for Interleaved Multimodal Generation 提出自我修正优化方法以解决多模态生成中的一致性问题 manipulation classifier-free guidance large language model
39 From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators 提出DIGHT框架以解决人形机器人交互生成的物理可执行性问题 humanoid DPO direct preference optimization
40 STRIKE: Learning Visual State Transitions for Physical World Modeling 提出STRIKE框架以解决物理世界建模中的状态转移问题 manipulation world model world models
41 TouchScale: 500 Hours of Human Vision and Touch for Visual-Tactile Learning 提出TouchScale以解决视觉-触觉学习中的数据不足问题 manipulation bi-manual egocentric
42 PalmSpace: Towards a Versatile On-Palm Interaction Space through Unified Touch Modeling 提出PalmSpace以解决智能眼镜输入挑战 manipulation

🔬 支柱七：动作重定向 (Motion Retargeting) (3 篇)

#题目一句话要点标签🔗⭐
43 Video-Conditioned Generative Joint 2D-3D Hand Motion Recovery 提出JoHan框架以解决视频中3D手部运动恢复问题 motion recovery motion reconstruction
44 Scalable Patch-Level Self-Supervised Learning 提出JEM算法以解决大规模自监督学习的稳定性问题 structure preservation
45 DynStream: Online Streaming 4D Gaussian Reconstruction of Dynamic Worlds from Unposed Video 提出DynStream以解决动态场景在线重建问题 geometric consistency

🔬 支柱六：视频提取与匹配 (Video Extraction) (2 篇)

#题目一句话要点标签🔗⭐
46 RLHND: Video Foundation Models as Physically Grounded Hand Trackers for Robot Learning 提出RLHND以解决机器人学习中的手部追踪问题 egocentric foundation model ✅
47 Never Look Back: Understanding Persistence in 3D Object Memory from Egocentric Videos 提出Ledger以解决3D物体记忆的持久性问题 egocentric Ego4D

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

#题目一句话要点标签🔗⭐
48 ECHO: Embodied Camera Observations of Human Object Carrying 提出ECHO数据集以解决人类物体携带场景理解问题 human-object interaction

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
49 VolCo: Volumetric Contact for High-Fidelity Human Grasp Generation 提出VolCo以解决高保真度人手抓取生成问题 penetration ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页