PointNetVLAD: Deep Point Cloud Based Retrieval for Large-Scale Place Recognition

Benchmark Model Rank Results
3d-place-recognition-on-cs-campus3dPointNetVLAD#7AR@1%: 41.46AR@1: 35.57AR@1% cross-source: 43.53
3d-place-recognition-on-oxford-robotcarpointnetvlad#8AR@1%: 80.3
visual-localization-on-oxford-radar-robotcarPointNetVLAD#15Mean Translation Error: 28.48
visual-place-recognition-on-kitti360posePointNetVLAD#3Localization Recall@1: 0.21