Point cloud place recognition on Oxford RobotCar (test)
86.4Average Recall @ 1%PCAN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PCANTraining Protocol=Refined (Oxford, U.S., R.A.)2019.04 | 86.4 | 70.72 | |
| PCANTraining Dataset=Oxford, Output Vector Dimension=256-dim2019.04 | 83.81 | — | |
| PN_VLADTraining Dataset=Oxford, Output Vector Dimension=256-dim2019.04 | 81.01 | — | |
| PN_VLADTraining Protocol=Refined (Oxford, U.S., R.A.)2019.04 | 80.7 | 63.33 | |
| PointNetVLAD (PN_VLAD)Clusters (K)=64, Loss=lazy quadruplet loss, Output Dimension=2562018.04 | 80.31 | — | |
| PN_VLADArchitecture=PointNetVLAD, Training Datasets=Oxford, U.S., R.A.2018.04 | 80.09 | 63.33 | |
| PN_MAXArchitecture=PointNet + Max Pooling, Training Datasets=Oxford, U.S., R.A.2018.04 | 73.87 | 54.16 | |
| PN_MAXLoss=lazy quadruplet loss, Output Dimension=2562018.04 | 73.44 | — | |
| PN_MAXTraining Dataset=Oxford, Output Vector Dimension=256-dim2019.04 | 73.44 | — | |
| PN_STDOutput Dimension=2562018.04 | 46.52 | — | |
| PN_STDArchitecture=PointNet + Standard Deviation, Training Datasets=Oxford, U.S., R.A.2018.04 | 46.52 | 31.87 | |
| PN_STDTraining Dataset=Oxford, Output Vector Dimension=256-dim2019.04 | 46.52 | — |