cs.CV（2025-11-08）

📊 共 12 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱三：空间感知 (Perception & SLAM) (5 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗2) 支柱一：机器人控制 (Robot Control) (2 🔗2) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱三：空间感知 (Perception & SLAM) (5 篇)

#	题目	一句话要点	标签	🔗	⭐
1	StreamSTGS: Streaming Spatial and Temporal Gaussian Grids for Real-Time Free-Viewpoint Video	提出StreamSTGS，用于实时自由视点视频的流式传输，兼顾性能与效率。	3D gaussian splatting 3DGS gaussian splatting	✅
2	Light-Field Dataset for Disparity Based Depth Estimation	提出用于视差深度估计的光场数据集，解决现有数据集的局限性。	depth estimation	✅
3	Point Cloud Segmentation of Integrated Circuits Package Substrates Surface Defects Using Causal Inference: Dataset Construction and Methodology	针对集成电路封装基板表面缺陷，提出基于因果推理的点云分割方法CINet。	point cloud
4	MiVID: Multi-Strategic Self-Supervision for Video Frame Interpolation using Diffusion Model	MiVID：基于扩散模型的多策略自监督视频帧插值	optical flow
5	Open-World 3D Scene Graph Generation for Retrieval-Augmented Reasoning	提出基于检索增强推理的开放世界3D场景图生成框架，用于通用和交互式3D场景理解。	scene understanding

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#	题目	一句话要点	标签	🔗	⭐
6	Adaptive Agent Selection and Interaction Network for Image-to-point cloud Registration	提出自适应Agent选择与交互网络，用于图像到点云的精确配准	reinforcement learning point cloud
7	MoEGCL: Mixture of Ego-Graphs Contrastive Representation Learning for Multi-View Clustering	提出MoEGCL，通过混合自 Ego 图对比学习提升多视图聚类性能	representation learning contrastive learning	✅
8	Latent Refinement via Flow Matching for Training-free Linear Inverse Problem Solving	提出LFlow：基于Flow Matching的免训练线性逆问题隐空间优化方法	flow matching	✅
9	CoMA: Complementary Masking and Hierarchical Dynamic Multi-Window Self-Attention in a Unified Pre-training Framework	CoMA：互补掩码与分层动态多窗口自注意力，提升MAE预训练效率。	masked autoencoder MAE

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
10	Exploring Category-level Articulated Object Pose Tracking on SE(3) Manifolds	提出PPF-Tracker以解决关节物体姿态跟踪问题	manipulation point cloud	✅
11	Reperio-rPPG: Relational Temporal Graph Neural Networks for Periodicity Learning in Remote Physiological Measurement	提出Reperio-rPPG，利用关系时序图神经网络学习远程生理信号的周期性，实现更鲁棒的心率估计。	walking	✅

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

#	题目	一句话要点	标签	🔗	⭐
12	DiLO: Disentangled Latent Optimization for Learning Shape and Deformation in Grouped Deforming 3D Objects	DiLO：解耦潜在空间优化，用于学习分组形变3D对象的形状和形变	latent optimization

⬅️ 返回 cs.CV 首页 · 🏠 返回主页