Person Re-identification on MSMT17 to Market-1501
79.4mAPAWB
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| AWBAdaptation Stage=Post-A, Attention Module=Non-local2020.06 | 79.4 | 92.6 | 97.1 | 98.2 | |
| Non-local2020.06 | 77.1 | 90.4 | 96.6 | 97.9 | |
| AWBAdaptation Stage=Pre-A, Attention Module=I-CBAM2020.06 | 77.1 | 91.2 | 96.7 | 97.8 | |
| WaveBlock2020.06 | 75.8 | 90.6 | 96.4 | 97.9 | |
| I-CBAM2020.06 | 71.6 | 87.5 | 94.9 | 97.1 | |
| MMTnote=implementation based on original code2020.06 | 69.4 | 85.9 | 94.4 | 96.2 | |
| MUVAType=VLM-based2026.03 | 59.5 | 82.7 | — | — | |
| DMRLType=VLM-based, Uses extra datasets for pre-training=true2026.03 | 55.1 | 81.3 | — | — | |
| QAConv-MSType=Image Matching, Input image size larger than 256x128=true2026.03 | 54.5 | 83 | — | — | |
| STLType=Normalization2026.03 | 54 | 82.5 | — | — | |
| IGMGType=Normalization, Input image size larger than 256x128=true2026.03 | 53.5 | 80.9 | — | — | |
| MDAType=Meta-Learning, Uses target domain data for test-time updating=true2026.03 | 53 | 79.7 | — | — | |
| DCACType=VLM-based2026.03 | 52.1 | 77.9 | — | — | |
| TransMatcherType=Image Matching, Input image size larger than 256x128=true2026.03 | 52 | 80.1 | — | — | |
| QAConv-GSType=Image Matching, Input image size larger than 256x128=true2026.03 | 49.5 | 79.1 | — | — | |
| MSINet-SAMParams=2.4M, Inference Time=0.71×, Pre-training=ImageNet2023.03 | 48.4 | 76.3 | — | — | |
| PATType=Part-based2026.03 | 47.3 | 72.2 | — | — | |
| MSINetParams=2.3M, Inference Time=0.71×, Pre-training=ImageNet2023.03 | 46.2 | 74.9 | — | — | |
| Unsupervised and self-adaptative ReIDRe-Ranking=true, Source pre-training=true2021.03 | 45.2 | 68.1 | 76 | 79.2 | |
| CBNType=Normalization2026.03 | 45 | 73.7 | — | — | |
| LDUType=Normalization2026.03 | 44.8 | 74.6 | — | — | |
| CLIP-FGDIType=VLM-based, Input image size larger than 256x128=true2026.03 | 44.7 | 73 | — | — | |
| QAConvType=Image Matching, Input image size larger than 256x128=true2026.03 | 43.1 | 72.6 | — | — | |
| MTIType=Normalization2026.03 | 42.7 | 72.9 | — | — | |
| MMETType=VLM-based, Uses extra datasets for pre-training=true2026.03 | 40.6 | 76.3 | — | — | |
| PAUL2020.06 | 40.1 | 68.5 | 82.4 | 87.4 | |
| MetaBINType=Meta-Learning2026.03 | 39.9 | 70.4 | — | — | |
| SpCLSource pre-training=true2021.03 | 37.9 | 58.1 | 69.6 | 74.1 | |
| OSNetParams=2.2M, Inference Time=0.79×, Pre-training=ImageNet2023.03 | 37.5 | 66.6 | — | — | |
| GN+SNRType=Normalization2026.03 | 37.5 | 68 | — | — | |
| Unsupervised and self-adaptative ReIDRe-Ranking=false, Source pre-training=true2021.03 | 33.2 | 62.3 | 74.1 | 78.5 | |
| ResNet50Params=~24M, Inference Time=1×, Pre-training=ImageNet2023.03 | 31.8 | 58.8 | — | — | |
| ABMTSource pre-training=true2021.03 | 27.8 | 55.5 | — | — | |
| CycAsSource pre-training=false2021.03 | 26.7 | 50.1 | — | — | |
| Dual-RefinementSource pre-training=true2021.03 | 25.1 | 53.3 | 66.1 | 71.5 | |
| SSKDSource pre-training=true2021.03 | 23.8 | 49.6 | 63.1 | 68.8 | |
| MMTSource pre-training=true2021.03 | 22.9 | 49.2 | 63.1 | 68.8 | |
| DG-Net++Source pre-training=true2021.03 | 22.1 | 48.4 | 60.9 | 66.1 | |
| SGG++Source Dataset=Market1501, Protocol=Semi-supervised Domain Adaptation2018.11 | 16.6 | 37.6 | — | 57.2 | |
| ECN-GPPSource pre-training=true2021.03 | 15.2 | 40.4 | 53.1 | 58.7 | |
| MMCLSource pre-training=true2021.03 | 15.1 | 40.8 | 51.8 | 56.7 | |
| SSGSource pre-training=true2021.03 | 13.2 | 31.6 | — | 49.6 | |
| SSGSource Dataset=Market1501, Protocol=Unsupervised Domain Adaptation2018.11 | 13.2 | 31.6 | — | 49.6 | |
| CCSESource pre-training=false2021.03 | 9.9 | 31.4 | 41.4 | 45.7 | |
| ECNSource pre-training=true2021.03 | 8.5 | 25.3 | 36.3 | 42.1 | |
| PTGANSource Dataset=Market1501, Protocol=Unsupervised Domain Adaptation2018.11 | 2.9 | 10.2 | — | 24.4 |