Visual Place Recognition on Oxford RobotCar Day vs Night
86.45Recall@1Vanilla SP + SuperGlue (SG)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Vanilla SP + SuperGlue (SG)Base Architecture=SuperPoint + SuperGlue2021.07 | 86.45 | 1 | 1 | |
| ES UtilityBase Architecture=SuperPoint + SuperGlue, Selection Utility Method=ES2021.07 | 86.06 | 0.53 | 0.76 | |
| ES + PS UtilityBase Architecture=SuperPoint + SuperGlue, Selection Utility Method=ES + PS2021.07 | 86.06 | 0.48 | 0.74 | |
| Vanilla HF-NetBase Architecture=HF-Net2021.07 | 86 | 1 | 1 | |
| ES + PS UtilityBase Architecture=HF-Net, Selection Utility Method=ES + PS2021.07 | 85.66 | 0.47 | 0.75 | |
| ES UtilityBase Architecture=HF-Net, Selection Utility Method=ES2021.07 | 84.46 | 0.52 | 0.77 | |
| PS UtilityBase Architecture=SuperPoint + SuperGlue, Selection Utility Method=PS2021.07 | 82.86 | 0.37 | 0.71 | |
| PS UtilityBase Architecture=HF-Net, Selection Utility Method=PS2021.07 | 81.67 | 0.47 | 0.75 | |
| ES UtilityBase Architecture=SuperPoint, Selection Utility Method=ES2021.07 | 74.1 | 0.53 | 0.76 | |
| ES + PS UtilityBase Architecture=SuperPoint, Selection Utility Method=ES + PS2021.07 | 74.1 | 0.48 | 0.74 | |
| Vanilla SuperPoint (SP)Base Architecture=SuperPoint2021.07 | 72.11 | 1 | 1 | |
| PS UtilityBase Architecture=SuperPoint, Selection Utility Method=PS2021.07 | 69.32 | 0.37 | 0.71 | |
| Semantic ConsistencyBase Architecture=SuperPoint2021.07 | 64.14 | 0.81 | 0.9 | |
| HF-Net's MobileNetVLADBase Architecture=HF-Net2021.07 | 60.55 | — | — | |
| Cluster ConsistencyBase Architecture=SuperPoint2021.07 | 58.96 | 0.62 | 0.8 | |
| NetVLADBase Architecture=NetVLAD2021.07 | 46.61 | — | — |