FlashMix: Fast Map-Free LiDAR Localization Via Feature Mixing and Contrastive-Constrained Accelerated Training
IEEE/CVF Winter Conference on Applications of Computer Vision（2025）
关键词
LiDAR Localization,Training Time,Localization Accuracy,Point Cloud,Real-world Scenarios,Local Descriptors,Contrastive Loss,Metric Learning,Functional Need,Raw Point Cloud,Representation Learning,Global Pooling,Pose Estimation,Rigid Transformation,Global Average Pooling,Linear Layer,Direct Prediction,Self-supervised Learning,Storage Requirements,Global Descriptors,Pose Prediction,Translation Error,Random Sample Consensus,Rotation Error,Place Recognition,Triplet Loss,Feature Encoder,Global Relations,Input Point Cloud
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要