cs.CV（2026-10-23）

📊 共 28 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (11 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (10) 支柱八：物理动画 (Physics-based Animation) (2) 支柱三：空间感知与语义 (Perception & Semantics) (2 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (1 🔗1) 支柱一：机器人控制 (Robot Control) (1) 支柱七：动作重定向 (Motion Retargeting) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (11 篇)

#题目一句话要点标签🔗⭐
1 Multimodal Generative AI Assists Comprehensive Echocardiographic Interpretation 提出EchoChat以解决心脏超声解读专家短缺问题 foundation model multimodal chain-of-thought
2 LapaFM: A Laparoscopic Foundation Model via Multi-Stage Continual Learning 提出LapaFM以解决腹腔镜手术视频分析中的多任务泛化问题 foundation model multimodal
3 PROVE: Probabilistic Visual Grounding via Patch-Level Evidence for Vision-Language Models 提出PROVE以解决视觉语言模型中的不确定性问题 visual grounding
4 An Error Budget for Shuffled Multimodal Windows in the 3rd WEAR Challenge: Stacking, Count Constraints and Seam Matching 提出基于堆叠分类器的多模态窗口活动识别方法 multimodal
5 Video-MemComp: Taylor-Guided Sensitivity-Aware Memory Management for Streaming Video Understanding 提出Video-MemComp以解决长视频理解中的内存瓶颈问题 large language model multimodal
6 The Answer Knows Where to Look: Zooming before MLLMs Answer under Visual-Token Budgets 提出一种新方法以优化多模态大语言模型的视觉信息获取 large language model multimodal
7 SPOOK: Sparse Pixel Optimization for Off-Manifold Knowledge in Vision Encoders 提出稀疏像素优化方法以探测视觉编码器的非流形知识 multimodal
8 VERA-Hand: A Verified Multidimensional Benchmark for 2D Hand Pose Estimation in Third-Person Dynamic Gesture Videos 提出VERA-Hand以解决手势视频中手部姿态估计评估不足问题 foundation model ✅
9 ReLEAP: Upcycling Vision Datasets through Cross-Task Unification and Semi-Supervised Annotation 提出ReLEAP以解决视觉数据集整合与标注不足问题 foundation model
10 Separating Angular Alignment from Radial Ordering in Hyperbolic Vision-Language Models 提出AURORA框架以解决超曲面视觉语言模型的目标耦合问题 multimodal
11 Guided Zero-Shot Anomaly Explanation for Anonymized Crowd Video Monitoring: Context, Categorization, and Spatial Localization 提出XAVAD以解决视频异常检测中的可解释性与隐私保护问题 multimodal ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
12 Past as State, Present as Attention: Persistent-State Blockwise Flow Matching for Long-Horizon Co-Speech Motion Generation 提出PASPA以解决长时间段共语动作生成问题 flow matching motion generation classifier-free guidance
13 Learning Fine-Grained Visual Classification with Text as Privileged Information via Multimodal Distillation 提出FG-LUPI框架以利用文本信息提升细粒度视觉分类 distillation privileged information multimodal
14 Reuse, Don't Retrain: Efficient, Modality-Versatile Geospatial Foundation Models through Feature Distillation 通过特征蒸馏提高地理基础模型的效率与多模态适应性 distillation feature matching foundation model
15 GeoLEON-CLIP: Graph-Anchored Multimodal Contrastive Learning for Geographic Region Representation 提出GeoLEON-CLIP以解决地理区域多模态表示问题 contrastive learning multimodal
16 UniAction: Unified Action Conditioning for World Models 提出UniAction以解决异构动作条件化世界模型问题 world model world models spatiotemporal
17 UniWorld: A Unified World Model for Real-Time Visual World Simulation 提出UniWorld以解决实时视觉世界模拟问题 world model world models
18 Representation Learning via Composite Subspace Flow Matching 提出复合子空间流匹配方法以解决图像表示学习挑战 flow matching representation learning
19 ACWM-RM: A Reference-Free Reward Model from Complementary Mismatch Preferences 提出ACWM-RM以解决视频质量与动作跟随评估的奖励模型问题 reinforcement learning flow matching world model
20 TileShare: Training-Free Acceleration of Overlapping Tiled ViT Decoding 提出TileShare以加速重叠切片ViT解码，提升效率 distillation open-vocabulary open vocabulary
21 LoCA: Enabling Multi-Modality in Geospatial Location Encoders 提出LoCA以解决多模态地理位置编码问题 contrastive learning multimodal

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

#题目一句话要点标签🔗⭐
22 4DMemory: Benchmarking Long-Horizon Spatiotemporal Memory with Harness-Augmented Agents 提出4DMemory以解决长时空记忆评估问题 spatiotemporal foundation model multimodal
23 Distortion-Aware Spatiotemporal Learning for Real-Time Wide-Angle Video Rectification 提出FastVC以解决广角视频畸变问题 spatiotemporal

🔬 支柱三：空间感知与语义 (Perception & Semantics) (2 篇)

#题目一句话要点标签🔗⭐
24 CompMVR: Compression-Aware Multi-View Restoration Using Diffusion Models for Geometrically Consistent 3D Reconstruction 提出CompMVR以解决压缩导致的多视图几何不一致问题 3D reconstruction feature matching ✅
25 Easy-3Dfy: Object-Level Geometric Composition for Single-Image 3D Scene Construction 提出Easy-3Dfy以解决单图像3D场景构建问题 sam 3D SAM 3D foundation model

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
26 FlexGlue: Sparse-to-Dense Feature Matching and Visual Disambiguation 提出FlexGlue以解决图像匹配与视觉歧义问题 feature matching ✅

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
27 MiD-VLM: Detecting and Grounding Global Semantic Misinformation in Image-Caption Pairs 提出MiD-VLM以解决图像-标题对中的语义错误问题 manipulation multimodal

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

#题目一句话要点标签🔗⭐
28 CityLift: Compositional 3D Urban Generation from a Single Image 提出CityLift框架以解决单图生成3D城市场景问题 spatial relationship ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页