Sequential Visual Place Recognition on Oxford Hard
69.4Recall@1Adapt-STformer
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Adapt-STformerDescriptor Dimension=24576, Backbone=CCT384, Mem (MB)=180.39, Time (s)=0.01782025.10 | 69.4 | 77.06 | 79.69 | |
| Adapt-STformerDescriptor Dimension=4096 (PCA), Backbone=CCT384, Mem (MB)=180.39, Time (s)=0.01782025.10 | 68.74 | 76.59 | 79.18 | |
| Adapt-STformerDescriptor Dimension=512 (PCA), Backbone=CCT384, Mem (MB)=180.39, Time (s)=0.01782025.10 | 65.16 | 74.59 | 77.65 | |
| STformer pcDescriptor Dimension=24576, Backbone=CCT384, Mem (MB)=276.55, Time (s)=0.02772025.10 | 57.54 | 65.48 | 70.11 | |
| STformer pcDescriptor Dimension=4096 (PCA), Backbone=CCT384, Mem (MB)=276.55, Time (s)=0.02772025.10 | 56.48 | 64.53 | 69.68 | |
| STformer pcDescriptor Dimension=512 (PCA), Backbone=CCT384, Mem (MB)=276.55, Time (s)=0.02772025.10 | 53.77 | 61.98 | 67.2 | |
| SeqvladDescriptor Dimension=24576, Backbone=CCT384, Mem (MB)=177.66, Time (s)=0.01812025.10 | 47.09 | 56.01 | 61.51 | |
| SeqvladDescriptor Dimension=4096 (PCA), Backbone=CCT384, Mem (MB)=177.66, Time (s)=0.01812025.10 | 46.03 | 55.5 | 61.51 | |
| JIST pcDescriptor Dimension=512 (PCA), Backbone=ResNet, Mem (MB)=159.65, Time (s)=0.00532025.10 | 44.23 | 54.05 | 61.39 | |
| SeqvladDescriptor Dimension=512 (PCA), Backbone=CCT384, Mem (MB)=177.66, Time (s)=0.01812025.10 | 43.83 | 53.3 | 60.49 | |
| SeqnetDescriptor Dimension=4096 (PCA), Backbone=NetVLAD, Mem (MB)=676.09, Time (s)=0.00142025.10 | 30.64 | 45.05 | 52.28 |