Person Re-identification on MSMT17 MS
62.7mAPCION
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CIONBackbone=ResNet50-IBN, Pre-training dataset=Large-scale Person Images2024.09 | 62.7 | 84.3 | |
| CIONBackbone=ViT-S, Pre-training dataset=Large-scale Person Images2024.09 | 58 | 79.9 | |
| TranSSLBackbone=ViT-S, Pre-training dataset=Large-scale Person Images, add CFS=true2024.09 | 57.8 | 79.5 | |
| CIONBackbone=ResNet-50, Pre-training dataset=Large-scale Person Images2024.09 | 57.4 | 81.3 | |
| TranSSLBackbone=ViT-S, Pre-training dataset=Large-scale Person Images2024.09 | 55 | 77.9 | |
| PASSBackbone=ViT-S, Pre-training dataset=Large-scale Person Images2024.09 | 49.1 | 72.7 | |
| LUPBackbone=ResNet50-IBN, Pre-training dataset=Large-scale Person Images2024.09 | 42.6 | 69.1 | |
| MUVAType=VLM-based2026.03 | 34.5 | 63.1 | |
| C-ContrastBackbone=ResNet-50, Pre-training dataset=ImageNet1K-1.3M2024.09 | 33.4 | 60.5 | |
| MCRNBackbone=ResNet-50, Pre-training dataset=ImageNet1K-1.3M2024.09 | 32.8 | 64.4 | |
| ISRType=Pre-training, Input image size=> 256x128, Extra datasets for pre-training=true2026.03 | 30.3 | 56.4 | |
| CLIP-FGDIType=VLM-based, Input image size=> 256x1282026.03 | 28.8 | 56.7 | |
| LUPBackbone=ResNet-50, Pre-training dataset=Large-scale Person Images2024.09 | 28.3 | 53.8 | |
| DCAC2025.02 | 27.5 | 56.7 | |
| DCACType=VLM-based2026.03 | 27.5 | 56.7 | |
| SpCLBackbone=ResNet-50, Pre-training dataset=ImageNet1K-1.3M2024.09 | 26.8 | 53.7 | |
| QAConv-MSType=Image Matching, Input image size=> 256x1282026.03 | 25.9 | 57.7 | |
| MMTBackbone=ResNet-50, Pre-training dataset=ImageNet1K-1.3M2024.09 | 24 | 50.1 | |
| DG-Net++Backbone=ResNet-50, Pre-training dataset=ImageNet1K-1.3M2024.09 | 22.1 | 48.4 | |
| PAT2025.02 | 21.6 | 45.6 | |
| PATType=Part-based2026.03 | 21.6 | 45.6 | |
| STLType=Normalization2026.03 | 21.3 | 45.1 | |
| RILLType=Normalization, Test-time updating=true2026.03 | 20.7 | 45.8 | |
| DCCLType=Normalization2026.03 | 20.5 | 47.4 | |
| SuA-SpMLType=Meta-Learning2026.03 | 20.4 | 47.1 | |
| UDSX2025.02 | 20.2 | 47.6 | |
| UDSXType=Embedding Expansion2026.03 | 20.2 | 47.6 | |
| SVILType=Normalization2026.03 | 19.9 | 44.6 | |
| IGMGType=Normalization, Input image size=> 256x1282026.03 | 19.7 | 41.6 | |
| DEX2025.02 | 18.7 | 43.5 | |
| DEXType=Embedding Expansion2026.03 | 18.7 | 43.5 | |
| SALDG2025.02 | 18.1 | 46.5 | |
| SALDGType=Mixture-of-Experts2026.03 | 18.1 | 46.5 | |
| M3LIBN2025.02 | 14.7 | 36.9 | |
| RaMoE2025.02 | 13.5 | 34.1 | |
| RaMoEType=Mixture-of-Experts2026.03 | 13.5 | 34.1 | |
| M3L2025.02 | 12.9 | 33 | |
| M3LType=Meta-Learning2026.03 | 12.9 | 33 | |
| QAConv502025.02 | 10 | 29.9 |