Video-to-Video Person Re-identification on MARS (test)
90Top-1 AccuracyNVAN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NVANProtocol=V2V2020.07 | 90 | — | — | 82.8 | |
| ResVKD-50bamBackbone=ResNet-50, Spatial Attention=bam, Protocol=V2V2020.07 | 89.4 | 96.8 | — | 83.1 | |
| STE-NVANProtocol=V2V2020.07 | 88.9 | — | — | 81.2 | |
| ResVKD-50Backbone=ResNet-50, Protocol=V2V2020.07 | 88.7 | 96.1 | — | 82.2 | |
| CSACSE+OFProtocol=V2V2020.07 | 86.3 | 94.7 | — | 76.1 | |
| STAProtocol=V2V2020.07 | 86.3 | 95.7 | — | 80.8 | |
| OSM+CAABackbone=GoogLeNet V22019.05 | 84.7 | — | — | 72.4 | |
| DMBackbone=GoogLeNet V2, Note=Our Trained Results2019.05 | 84.3 | — | — | 72.8 | |
| TKP2019.08 | 84 | 93.7 | 95.7 | 73.3 | |
| TKPProtocol=V2V2020.07 | 84 | 93.7 | — | 73.3 | |
| CAEBackbone=ResNet-502019.05 | 82.4 | — | — | 67.5 | |
| DRSA2019.08 | 82.3 | — | — | 65.8 | |
| DRSABackbone=ResNet-502019.05 | 82.3 | — | — | 65.8 | |
| multi-snippetsoptical flow=false2019.08 | 81.2 | 92.1 | — | 69.4 | |
| DuATNProtocol=V2V2020.07 | 81.2 | 92.5 | — | 67.7 | |
| DuATM2019.08 | 78.7 | 90.9 | — | 62.3 | |
| MGCAM2019.08 | 77.2 | — | — | 71.2 | |
| CCEBackbone=GoogLeNet V2, Note=Our Trained Results2019.05 | 73.8 | — | — | 58.1 | |
| SDM2019.08 | 71.2 | 85.7 | 91.8 | — | |
| GCEBackbone=GoogLeNet V2, Note=Our Trained Results2019.05 | 51.5 | — | — | 31.6 | |
| MSEBackbone=GoogLeNet V2, Note=Our Trained Results2019.05 | 39.3 | — | — | 19.6 | |
| MAEBackbone=GoogLeNet V2, Note=Our Trained Results2019.05 | 26 | — | — | 12 |