cs.CV（2026-09-04）

📊 共 30 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱三：空间感知与语义 (Perception & Semantics) (8 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗5) 支柱九：具身大模型 (Embodied Foundation Models) (6) 支柱一：机器人控制 (Robot Control) (4) 支柱六：视频提取与匹配 (Video Extraction) (3) 支柱八：物理动画 (Physics-based Animation) (2 🔗1)

🔬 支柱三：空间感知与语义 (Perception & Semantics) (8 篇)

#题目一句话要点标签🔗⭐
1 Compact Neural Appearance Models for Efficient Gaussian Splatting 提出紧凑神经外观模型以提高高斯点云渲染效率 3D gaussian splatting gaussian splatting splatting
2 Temporal Residual Neural Radiance Fields for Monocular Video Dynamic Human Body Reconstruction 提出时序残差神经辐射场以解决动态人类体重建问题 NeRF neural radiance field scene reconstruction
3 CrossDepth: Geometry-Constrained Attention for Generalizable Multi-View Surround Depth Estimation 提出CrossDepth以解决多视角深度估计中的跨图像不一致问题 depth estimation ✅
4 BLASt3R: Bundle Adjustment of Any Image Set with Multi-View Matching and Monocular Priors 提出BLASt3R以解决图像集的束调整问题 visual SLAM 3D reconstruction
5 WorldSculpt: Generating Compositional Worlds from Grounded Videos 提出WorldSculpt以解决复杂场景生成问题 3DGS
6 VoxelFix: Post-Hoc Semantic Correction of Completed 3D Voxel Maps 提出VoxelFix以解决3D体素地图语义修正问题 semantic map
7 CLON: Cue-Calibrated Linguistic Object Onboarding for Zero-Shot 6D Pose Front-Ends 提出CLON以解决零-shot 6D姿态估计中的前端问题 6D pose estimation
8 HiSfM: Disambiguating Structure-from-Motion via Scaffold-Anchored Hierarchical Reconstruction 提出HiSfM以解决结构光重建中的视觉歧义问题 3D reconstruction ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#题目一句话要点标签🔗⭐
9 Weather-Conditioned Depth Anything 提出Weather-Conditioned Depth Anything以解决恶劣天气下深度估计问题 distillation depth estimation monocular depth ✅
10 MEOX: Compact Multimodal Mixture-of-Experts for Earth Observation 提出MEOX以解决地球观测中的多模态学习问题 representation learning masked autoencoder multimodal
11 Learning 3D Editing without Paired Supervision via Generative Prior Distillation 提出无配对监督的3D编辑学习方法以解决数据稀缺问题 distillation foundation model instruction following ✅
12 First Things First: Teaching LLM-Based Agents to Prioritize Must-Haves before Nice-to-Haves 提出FTF-rl以解决多优先级用户需求的推理问题 reinforcement learning large language model multimodal ✅
13 Importance-Aware Low-Rank Distillation of Diffusion Transformers 提出SVDtrunc以解决扩散变换器的参数压缩问题 distillation large language model ✅
14 BEAM3R: Beam's-eye-view architecture with Mamba-3 for implicit dose reconstruction 提出BEAM3R以解决剂量重建中的高效性与准确性问题 Mamba MAE
15 SeamFlow: Structure-Aware Flow Matching on Edge Probabilities for Artist-Like UV Unwrapping 提出SeamFlow以解决3D表面切割与UV展开问题 flow matching ✅

🔬 支柱九：具身大模型 (Embodied Foundation Models) (6 篇)

#题目一句话要点标签🔗⭐
16 MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression 提出MCPO以解决多模态推理链压缩问题 multimodal chain-of-thought
17 Conserved Immune Topology Improves Pathology Foundation Model Generalization for Cross-Cancer MSI-H Prediction 提出保守免疫拓扑以解决跨癌症MSI-H预测问题 foundation model
18 SMILE: Self-Explainable Multimodal Information Bottleneck for Medical Diagnosis 提出自解释多模态信息瓶颈以解决医疗诊断中的可解释性问题 multimodal
19 WeAgent-MMGenEdit: A Full-Stack Recipe for Multimodal Agentic Image Generation and Editing 提出WeAgent-MMGenEdit以解决多模态图像生成与编辑中的知识依赖问题 multimodal
20 Where to Look Matters: Learning Influential Views for VLM-based 3D Visual Grounding 提出IVSGround以解决3D视觉定位中的视角选择问题 visual grounding
21 LetOccVote: Learning Weakly Supervised 3D Occupancy through Consensus 提出LetOccVote以解决弱监督3D占用预测问题 foundation model

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
22 TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single Image 提出TourPhysics以解决长视频合成中的物理一致性问题 manipulation world model world models
23 UniMate: One Unified Model to Animate Diverse Skeletons 提出UniMate以解决多样化骨骼动画生成问题 quadruped bipedal biped
24 An Evaluation Framework for Generating Multi-View Images of a Person in a Scene 提出HSRD度量以解决多视角图像生成中的一致性问题 manipulation foundation model
25 LensStyle: Learning the Optical Aesthetics for Controllable Stylized Lens Effect Rendering 提出LensStyle以解决镜头效果渲染的多样性与可控性问题 manipulation

🔬 支柱六：视频提取与匹配 (Video Extraction) (3 篇)

#题目一句话要点标签🔗⭐
26 MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision 提出MINT模型以解决世界坐标下相机与手部运动估计问题 egocentric hand reconstruction motion estimation
27 Linguistic Trajectory Encoding for Efficient Long-Horizon Spatial Memory in Embodied Agents 提出语言轨迹编码以解决长时间空间记忆问题 Ego4D
28 Hidden In Plain Gaze: Gaze Representations as Privacy Controls for Utility and Re-identification Risk in XR 提出基于注视表示的隐私控制以降低XR中的身份重识别风险 egocentric

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
29 Enhancing Multimodal Emotion Recognition via Multi-Feature Encoding and Attention-Based Fusion 提出多特征编码与注意力融合以增强多模态情感识别 spatiotemporal multimodal
30 Sound-based Multi-Person 3D Pose Estimation 提出SoundMHPE以解决声学信号下的多人体姿态估计问题 AMP ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页