cs.CV（2026-07-06）

📊 共 37 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) (12 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (11 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗1) 支柱一：机器人控制 (Robot Control) (2 🔗1) 支柱四：生成式动作 (Generative Motion) (1) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (12 篇)

#题目一句话要点标签🔗⭐
1 Green for Go, Red for No: Visual Grounding via Semantic Segmentation for VLA Navigation Policies 提出基于语义分割的视觉定位方法以提升VLA导航策略 vision-language-action VLA visual grounding
2 Deep Learning for Semen Analysis in Male Infertility: Computer Vision, Multimodal Fusion, and Clinical Translation 提出深度学习方法以解决男性不育的精液分析问题 multimodal
3 A Multimodal Reasoning Typology for Grounding Chart-Image Coherence in Science Communication 提出多模态推理分类法以解决科学传播中的图表图像一致性问题 multimodal
4 EventCoT: Event-centric Video Chain-of-thought for Reasoning Temporal Localization 提出EventCoT框架以解决视频时间定位问题 chain-of-thought
5 Enhancing Large Multimodal Models in Key Information Extraction via Scene-Aware Document Synthesis 提出SAYRE框架以解决KIE训练数据生成问题 multimodal
6 DGSeg: Dynamic Gating of Semantic-Spatial Guided Predictions for Reasoning Segmentation 提出DGSeg以解决推理分割中的模糊性问题 large language model multimodal ✅
7 DiCE-CIR: Direct Composition Learning for Efficient Zero-Shot Composed Image Retrieval 提出DiCE-CIR以解决零-shot复合图像检索效率问题 large language model multimodal
8 Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation 提出SearchGen框架以解决视觉生成中的知识边界问题 multimodal
9 ChatImage: Navigating Long-Form LLM Answers through Interactive Images 提出ChatImage以解决长文本LLM答案交互性不足问题 large language model
10 Probe-EM: Targeted Neuron Tracing via Training-Free Semantic Verification 提出Probe-EM框架以解决神经元追踪中的数据依赖问题 foundation model ✅
11 Hierarchical Evidence-Driven Reasoning for Long Document Understanding 提出HIEVI-RAG以解决长文档理解中的检索与生成问题 multimodal
12 QSVideo: Query-Conditioned Semantic Temporal Retrieval for Video Understanding 提出QSVideo以解决视频理解中的多模态检索问题 multimodal ✅

🔬 支柱三：空间感知与语义 (Perception & Semantics) (11 篇)

#题目一句话要点标签🔗⭐
13 DeGenseGS: Geometrically and Semantically Decoupled Surgical Scene Understanding in 4D Gaussian Splatting 提出DeGenseGS以解决手术场景理解中的几何与语义耦合问题 gaussian splatting splatting scene understanding
14 Beyond Isolated Objects: Relationship-aware Open Vocabulary Scene Understanding via 3D Scene Graph Analysis 提出RelGraphOV框架以解决开放词汇3D场景理解中的关系建模问题 scene understanding open-vocabulary open vocabulary ✅
15 WildSplat: Feedforward Gaussian Splatting from Unposed In-the-Wild Images 提出WildSplat以解决不一致光照下的3D重建问题 3D gaussian splatting 3D reconstruction gaussian splatting
16 GUSH3R: Everyone Everywhere All at Once as Gaussians 提出GUSH3R以解决动态人场景重建问题 3D gaussian splatting 3DGS 3D reconstruction
17 ReCal3R: Reliability-Calibrated Learning Rates for Streaming 3D Reconstruction 提出ReCal3R以解决流式3D重建中的可靠性问题 3D reconstruction ✅
18 Targeted Structure Completion for Sparse-View 3D Reconstruction in Autonomous Driving 提出FocusGS以解决稀疏视图3D重建中的结构完整性问题 3D reconstruction
19 PixWorld: Unifying 3D Scene Generation and Reconstruction in Pixel Space 提出PixWorld以统一3D场景生成与重建问题 3D reconstruction foundation model
20 Vision Pretraining for Dense Spatial Perception 提出边界中心的视觉预训练以提升密集空间感知能力 depth estimation foundation model
21 Reference-Induced Consensus for Selective Posed-Reference Visual Localization 提出RIC-Loc以解决无场景训练的定位问题 VGGT ✅
22 TubeLite: Lightweight Multi-Actor Spatio-Temporal Action Detection 提出TubeLite以解决多演员时空动作检测中的稳定性问题 optical flow
23 Enhancing Video Physical Consistency via Role-aware Joint Training and Modality-decoupled Denoising 提出VPT框架以解决视频物理一致性问题 optical flow ✅

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

#题目一句话要点标签🔗⭐
24 PixelPilot: Scalable Vision-Language-Action Models for End-to-End Autonomous Driving 提出PixelPilot以解决自主驾驶中的2D到3D预测问题 policy learning scene understanding vision-language-action
25 TGRIP: A Text-Guided Approach to Vehicle Instance Prediction in Autonomous Driving 提出TGRIP以解决自动驾驶中车辆实例预测的语义不足问题 teacher-student optical flow motion prediction ✅
26 Multiplayer Interactive World Models with Representation Autoencoders 提出多玩家交互世界模型以解决动态环境中的复杂物理交互问题 world model world models
27 UNIVERSE: Unified Video Action Models for Autonomous Driving with Flexible Mask-Modulated Modality Generation 提出UNIVERSE以解决自主驾驶中的视频与动作模型融合问题 world action model world action models mutual attention
28 Geometry-aware Depth-guided Representation Learning for Structure-preserving Low-light Image Enhancement 提出DMSA-Net以解决低光照图像增强中的结构一致性问题 representation learning scene understanding
29 Claim-Level Rubric Rewards for Video Caption Reinforcement Learning 提出Claim-Level Rubric Rewards以解决视频字幕强化学习中的奖励设计瓶颈 reinforcement learning reward design
30 Graph Representation Learning of Longitudinal Medical Imaging Trajectories for Treatment Response Prediction 提出基于图表示学习的框架以预测乳腺癌治疗反应 predictive model representation learning
31 Steering Optimisation Trajectories in Diffusion Representation Learning 提出SteeringDRL以优化扩散表示学习中的轨迹 representation learning
32 TimeThink: Reasoning with Time for Video LLMs 提出TimeThink框架以解决视频推理中的时间证据发现问题 reinforcement learning large language model
33 MV-Forcing: Long Multi-View Video Generation via 4D-Grounded Spatio-Temporal Self-Forcing 提出MV-Forcing以解决长多视角视频生成问题 distillation 3D reconstruction

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
34 From Fixed to Free Cameras: Calibration-Free View-Robust Vision-Language-Action Model 提出Camera-Centric VLA以解决相机校准问题 manipulation vision-language-action VLA ✅
35 Unsupervised Pixel-Level Semantic Left-Right Understanding of In-the-Wild Images 提出无监督学习框架以解决野外图像的像素级左右理解问题 quadruped

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
36 When Does High-CFG Diffusion Inversion Fail? A Controlled Study of Prompt--Latent Interactions 提出高-CFG扩散反演失败的控制研究以解决图像编辑问题 classifier-free guidance

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

#题目一句话要点标签🔗⭐
37 Hybrid Deep Learning for Traceability and Classification of Industrial Slate Tiles 提出混合深度学习方法以解决工业板岩瓷砖追溯与分类问题 feature matching

⬅️ 返回 cs.CV 首页 · 🏠 返回主页