Place Recognition on Oxford Robotcar (test)
96.4Avg Recall @1%SOE-Net
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SOE-NetDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 96.4 | — | |
| CORAL-VLADFusion strategy=element-wise concatenation in four residual blocks, Alternative Name=Con-Four2020.11 | 96.13 | 88.93 | |
| FN-SF-VLADaggregation_structure=SF, representing=LPD-Net2018.12 | 94.92 | 86.28 | |
| LPD-NetDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 94.92 | — | |
| Sum-FourFusion strategy=element-wise summarization in four residual blocks2020.11 | 94.43 | 86.23 | |
| LPD-Net2020.11 | 94.42 | 86.28 | |
| Con-FirstFusion strategy=element-wise concatenation in the first residual block2020.11 | 93.62 | 84.82 | |
| Ele-VLADStream=Elevation image NetVLAD2020.11 | 93.61 | 82.49 | |
| Sum-FirstFusion strategy=element-wise summarization in the first residual block2020.11 | 92.71 | 82.44 | |
| FN-PC-VLADaggregation_structure=PC2018.12 | 92.27 | 81.41 | |
| FN-DG-VLADgraph_type=Dynamic Graph2018.12 | 91.44 | 80.14 | |
| Aug-Net2020.11 | 91.24 | 79.47 | |
| FN-PM-VLADaggregation_structure=PM2018.12 | 91.2 | 78.77 | |
| FN-NG-VLADgraph_type=Neighbor Graph2018.12 | 90.38 | 77.74 | |
| FN-VLADmodule=without graph-based neighborhood aggregation2018.12 | 89.77 | 75.79 | |
| DAGCDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 87.49 | — | |
| Img-VLAD2020.11 | 85.24 | 64.47 | |
| DH3DDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 84.26 | — | |
| PCANDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 83.81 | — | |
| Vis-VLADStream=Vision NetVLAD2020.11 | 83.05 | 57.62 | |
| PN-VLAD baseline2018.12 | 81.01 | 62.76 | |
| PN-VLAD2020.11 | 81.01 | 67.94 | |
| PN-VLAD refine2018.12 | 80.71 | 63.33 | |
| PointNetVLADDescriptor Dimension=256-dim, Training Data=Oxford RobotCar2020.11 | 80.31 | — | |
| NN-VLADclustering=kNN (k=20)2018.12 | 79.21 | 61.96 | |
| PN MAX2018.12 | 73.87 | 54.16 | |
| PN STD2018.12 | 46.52 | 31.87 |