Person Re-identification on MSMT17 v1 (test)
78.7mAPReText
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ReTextTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 78.7 | 91.4 | — | — | |
| DynaMixTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 76.3 | 90.4 | — | — | |
| TransReID-SSLBackbone=ViT1-B, Pre-training=LUPerson, Resolution=384x1282021.11 | 75 | 89.5 | — | — | |
| TransReID-SSLBackbone=ViT1-S, Pre-training=LUPerson, Resolution=384x1282021.11 | 70.5 | 87.8 | — | — | |
| DCACTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 70.1 | 88.3 | — | — | |
| TransReIDBackbone=ViT-B, Pre-training=ImageNet-21K, Resolution=384x1282021.11 | 69.4 | 89.2 | — | — | |
| TransReID-SSLBackbone=ViT1-S, Pre-training=LUPerson, Resolution=Standard2021.11 | 68.1 | 86.1 | — | — | |
| MoCoV2Backbone=MGN (ResNet-50), Pre-training=LUPerson, Resolution=384x1282021.11 | 65.7 | 85.5 | — | — | |
| ReMixTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 63.9 | 84.8 | — | — | |
| MGNBackbone=ResNet-50, Pre-training=ImageNet, Resolution=384x1282021.11 | 63.7 | 85.1 | — | — | |
| BOTBackbone=ResNet50-IBN-a, Pre-training=LUPerson, Resolution=Standard2021.11 | 60.8 | 81.4 | — | — | |
| ABDNetBackbone=ResNet-50, Pre-training=ImageNet, Resolution=384x1282021.11 | 60.8 | 82.3 | — | — | |
| SCSNBackbone=ResNet-50, Pre-training=ImageNet, Resolution=384x1282021.11 | 58.5 | 83.8 | — | — | |
| BOTBackbone=ResNet50-IBN-a, Pre-training=ImageNet, Resolution=Standard2021.11 | 57.8 | 80.7 | — | — | |
| PATTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 52 | 75.9 | — | — | |
| QAConv-GSTraining Dataset=MSMT17, Protocol=Protocol 12026.02 | 50.9 | 79.2 | — | — | |
| ReTextTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 43.1 | 70.1 | — | — | |
| ReTextTraining Dataset=RandPerson, Protocol=Protocol 12026.02 | 42.3 | 70.8 | — | — | |
| DynaMixTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 36.7 | 65.5 | — | — | |
| DynaMixTraining Dataset=RandPerson, Protocol=Protocol 12026.02 | 36.5 | 67.1 | — | — | |
| SDA + MMTclustering=k-means2020.03 | 30.3 | 59.6 | 71.7 | 76.2 | |
| SDA + MMTclustering=k-means2020.03 | 29 | 57 | 69.5 | 74.1 | |
| SpCLSource Dataset=Market-1501, Backbone=ResNet-502020.06 | 26.8 | 53.7 | 65 | 69.8 | |
| SDAclustering=DBSCAN2020.03 | 25.6 | 54.4 | 66.4 | 71.3 | |
| GPRadaptation_type=Pseudo-label-based2020.03 | 24.3 | 51.7 | 64 | 68.9 | |
| MMT-dbscanSource Dataset=Market-1501, Clustering=DBSCAN, Publication=ICLR'202020.06 | 24 | 50.1 | 63.5 | 69.3 | |
| DCACTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 23.4 | 52.1 | — | — | |
| MMTclustering=k-means2020.03 | 23.3 | 50.1 | 63.9 | 69.8 | |
| SDAclustering=DBSCAN2020.03 | 23.2 | 49.5 | 62.2 | 67.7 | |
| SDAclustering=k-means2020.03 | 23 | 51.7 | 64.2 | 69.6 | |
| MMT-kmeansSource Dataset=Market-1501, Clustering=k-means, Publication=ICLR'202020.06 | 22.9 | 49.2 | 63.1 | 68.8 | |
| MMTclustering=k-means2020.03 | 22.9 | 49.2 | 63.1 | 68.8 | |
| UDAPclustering=DBSCAN (modified)2020.03 | 22.8 | 49.3 | 61.7 | 67.2 | |
| DG-Net++Source Dataset=Market-1501, Publication=ECCV'202020.06 | 22.1 | 48.4 | 60.9 | 66.1 | |
| DG-Net++adaptation_type=Pseudo-label-based2020.03 | 22.1 | 48.4 | 60.9 | 66.1 | |
| DG-Net++adaptation_type=Pseudo-label-based2020.03 | 22.1 | 48.8 | 60.9 | 65.9 | |
| SDAclustering=k-means2020.03 | 20.6 | 46.8 | 59.9 | 65 | |
| NRMTadaptation_type=Pseudo-label-based2020.03 | 20.6 | 45.2 | 57.8 | 63.3 | |
| GPRSource Dataset=Market-1501, Publication=ECCV'202020.06 | 20.4 | 43.7 | 56.1 | 61.9 | |
| GPRadaptation_type=Pseudo-label-based2020.03 | 20.4 | 43.7 | 56.1 | 61.9 | |
| JVTCSource Dataset=Market-1501, Publication=ECCV'202020.06 | 20.3 | 45.4 | 58.4 | 64.3 | |
| JVTCadaptation_type=Pseudo-label-based2020.03 | 20.3 | 45.4 | 58.4 | 64.3 | |
| UDAPclustering=DBSCAN (modified)2020.03 | 20.3 | 45.7 | 58.6 | 64.5 | |
| NRMTSource Dataset=Market-1501, Publication=ECCV'202020.06 | 19.8 | 43.7 | 56.5 | 62.2 | |
| NRMTadaptation_type=Pseudo-label-based2020.03 | 19.8 | 43.7 | 56.5 | 62.2 | |
| PATTraining Dataset=RandPerson, Protocol=Protocol 12026.02 | 19.4 | 45.5 | — | — | |
| JVTCadaptation_type=Pseudo-label-based2020.03 | 19 | 42.1 | 53.4 | 58.9 | |
| TransMatcherTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 18.4 | 47.3 | — | — | |
| PATTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 18.2 | 42.8 | — | — | |
| TransMatcherTraining Dataset=RandPerson, Protocol=Protocol 12026.02 | 17.7 | 48.3 | — | — | |
| QAConv-GSTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 17.2 | 45.9 | — | — | |
| MMCLadaptation_type=Pseudo-label-based2020.03 | 16.2 | 43.6 | 54.3 | 58.9 | |
| ECN++adaptation_type=Pseudo-label-based2020.03 | 16 | 42.5 | 55.9 | 61.5 | |
| QAConv-GSTraining Dataset=RandPerson, Protocol=Protocol 12026.02 | 15.5 | 45.1 | — | — | |
| M³LSource=D+M+C-NP, IDs=2,220, Images=36,823, Backbone=IBN-Net50, evaluation_subset=detected2020.12 | 15.4 | 37.1 | — | — | |
| D-MMDadaptation_type=Pseudo-label-based2020.03 | 15.3 | 34.4 | 51.1 | 58.5 | |
| ECN++Source Dataset=Market-1501, Publication=TPAMI'202020.06 | 15.2 | 40.4 | 53.1 | 58.7 | |
| ECN++adaptation_type=Pseudo-label-based2020.03 | 15.2 | 40.4 | 53.1 | 58.7 | |
| MMCLSource Dataset=Market-1501, Publication=CVPR'202020.06 | 15.1 | 40.8 | 51.8 | 56.7 | |
| MMCLadaptation_type=Pseudo-label-based2020.03 | 15.1 | 40.8 | 51.8 | 56.7 | |
| M³LSource=D+M+C, IDs=2,820, Images=55,748, Backbone=IBN-Net502020.12 | 14.7 | 36.9 | — | — | |
| D-MMDSource Dataset=Market-1501, Publication=ECCV'202020.06 | 13.5 | 29.1 | 46.3 | 54.1 | |
| D-MMDadaptation_type=Pseudo-label-based2020.03 | 13.5 | 29.1 | 46.3 | 54.1 | |
| LDUTraining Dataset=Market-1501, Protocol=Protocol 12026.02 | 13.5 | 35.7 | — | — | |
| SSGadaptation_type=Pseudo-label-based2020.03 | 13.3 | 32.2 | — | 51.2 | |
| SSGSource Dataset=Market-1501, Publication=ICCV'192020.06 | 13.2 | 31.6 | — | 49.6 | |
| SSGadaptation_type=Pseudo-label-based2020.03 | 13.2 | 31.6 | — | 49.6 | |
| M³LSource=D+M+C-NP, IDs=2,220, Images=36,823, Backbone=ResNet-50, evaluation_subset=detected2020.12 | 13.1 | 32 | — | — | |
| M³LSource=D+M+C, IDs=2,820, Images=55,748, Backbone=ResNet-502020.12 | 12.9 | 33 | — | — | |
| ECNadaptation_type=Pseudo-label-based2020.03 | 10.2 | 30.2 | 41.5 | 46.8 | |
| QAConv50Source=D+M+C-NP, IDs=2,220, Images=36,823, reimplemented=true, evaluation_subset=detected2020.12 | 10 | 29.9 | — | — | |
| QAConv50Source=D, IDs=702, Images=16,5222020.12 | 8.9 | 29 | — | — | |
| ECNSource Dataset=Market-1501, Publication=CVPR'192020.06 | 8.5 | 25.3 | 36.3 | 42.1 | |
| ECNadaptation_type=Pseudo-label-based2020.03 | 8.5 | 25.3 | 36.3 | 42.1 | |
| QAConv50Source=D+M+C, IDs=2,820, Images=55,748, reimplemented=true2020.12 | 7.5 | 24.3 | — | — | |
| PTGANadaptation_type=Domain translation-based2020.03 | 3.3 | 11.8 | — | 27.4 | |
| PTGANSource Dataset=Market-1501, Publication=CVPR'182020.06 | 2.9 | 10.2 | — | 24.4 | |
| PTGANadaptation_type=Domain translation-based2020.03 | 2.9 | 10.2 | — | 24.4 |