Person Re-identification on Market1501 (test)
98.1Rank-1 Accuracyst-ReID
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| st-ReIDRandom Erasing (RE)=true2018.12 | 98.1 | 87.6 | 99.3 | 99.6 | — | |
| st-ReIDRandom Erasing (RE)=true, re-ranking=true2018.12 | 98 | 95.5 | 98.9 | 99.1 | — | |
| SOTARetrieval=Instance-level with spatial/temporal constraints2021.04 | 98 | 95.5 | 98.9 | 99.1 | — | |
| CTL-S-R50Backbone=ResNet-50, Inference Mode=Centroid-based2021.04 | 98 | 98.3 | 98.6 | 99.5 | — | |
| MoSCE-ReIDMethod Category=Attribute-based, Re-ranking=without2025.12 | 97.9 | 94.2 | — | — | — | |
| MoSAIC-ReIDMethod Category=General, Re-ranking=without, Backbone=ViT-B-16, LoRA rank=162025.12 | 97.9 | 95.5 | — | — | — | |
| Baseline+InSTDCategory=spatial-temporal2021.07 | 97.6 | 90.8 | 99.5 | 99.7 | — | |
| PLIPBackbone=Swin-B, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 97.3 | 93.2 | — | — | — | |
| CIONBackbone=R50-IBN, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 97.3 | 93.3 | — | — | — | |
| st-ReID2018.12 | 97.2 | 86.7 | 99.3 | 99.5 | — | |
| MGNQuery Setting=Multiple Query, Re-ranking (RK)=true2018.04 | 97.1 | 95.9 | — | — | — | |
| UPREIDBackbone=RN50, Pretrain=LUP, Protocol=fully supervised2023.05 | 97.1 | 91.1 | — | — | — | |
| PLIPBackbone=RN152, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 97.1 | 92.6 | — | — | — | |
| UPReIDBackbone=R50, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 97.1 | 91.1 | — | — | — | |
| CIONBackbone=R50, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 97.1 | 92.3 | — | — | — | |
| Ours(R101)+MGNBackbone=ResNet-101, Pre-training=LUPerson, Base Architecture=MGN2020.12 | 97 | 92 | — | — | — | |
| PromptSGMethod Category=General, Re-ranking=without2025.12 | 97 | 94.6 | — | — | — | |
| MGNQuery Setting=Multiple Query, Re-ranking (RK)=false2018.04 | 96.9 | 90.7 | — | — | — | |
| Baseline+st-ReIDCategory=spatial-temporal2021.07 | 96.9 | 88.6 | 99.2 | 99.5 | — | |
| PLIPBackbone=RN101, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 96.9 | 92 | — | — | — | |
| ISRBackbone=R50-IBN, Pre-training data=Large-scale Person Images, Number of pre-training images=47.8M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 96.9 | 92.3 | — | — | — | |
| CIONBackbone=ViT-B, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M2024.09 | 96.9 | 93.1 | — | — | — | |
| SOLIDERMethod Category=General, Re-ranking=without2025.12 | 96.9 | 93.9 | — | — | — | |
| PASSBackbone=ViT-B, Pretrain=LUP, Protocol=fully supervised2023.05 | 96.8 | 93 | — | — | — | |
| PLIPBackbone=RN50, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 96.8 | 91.4 | — | — | — | |
| PASSBackbone=ViT-B, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M2024.09 | 96.8 | 93 | — | — | — | |
| CIONBackbone=Swin-T, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M, Input size=384 x 1282024.09 | 96.8 | 92.2 | — | — | — | |
| PASS*Pre-training=4 million pedestrian images, Input image size=256x1282023.06 | 96.8 | 93 | — | — | — | |
| CLIMB-ReIDMethod Category=General, Re-ranking=without2025.12 | 96.8 | 92.6 | — | — | — | |
| VA-reIDreranking=true2019.12 | 96.79 | 95.43 | 98.31 | — | — | |
| PLIPBackbone=R50, Pre-training data=Large-scale Person Images, Number of pre-training images=4.8M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 96.7 | 91.3 | — | — | — | |
| TransReID-SSL + DenoiseRepBackbone=ViT-small2024.06 | 96.68 | 92.05 | — | — | — | |
| MGNQuery Setting=Single Query, Re-ranking (RK)=true2018.04 | 96.6 | 94.2 | — | — | — | |
| LUP-NLBackbone=RN50, Pretrain=LUP-NL, Protocol=fully supervised2023.05 | 96.6 | 91.9 | — | — | — | |
| LUP NLBackbone=R50, Pre-training data=Large-scale Person Images, Number of pre-training images=10.7M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 96.6 | 91.9 | — | — | — | |
| SOLIDERBackbone=Swin-S, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M, Input size=384 x 1282024.09 | 96.6 | 93.3 | — | — | — | |
| CIONBackbone=Swin-S, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M, Input size=384 x 1282024.09 | 96.6 | 93.4 | — | — | — | |
| TransReID-SSL + DenoiseRepBackbone=ViT-base2024.06 | 96.55 | 93.26 | — | — | — | |
| TransReID-SSLBackbone=ViT-base2024.06 | 96.52 | 93.1 | — | — | — | |
| M³ + HA-CNNPublication=CVPR20, Re-ranking=false2021.11 | 96.5 | 85.2 | — | — | — | |
| CIONBackbone=ViT-S, Pre-training data=Large-scale Person Images, Number of pre-training images=3.9M2024.09 | 96.5 | 92.8 | — | — | — | |
| IRM (MTL)Training Protocol=Multi-task Learning, Input image size=256x1282023.06 | 96.5 | 93.5 | — | — | — | |
| Ours(R50)+MGNBackbone=ResNet-50, Pre-training=LUPerson, Base Architecture=MGN2020.12 | 96.4 | 91 | — | — | — | |
| LUPBackbone=RN50, Pretrain=LUP, Protocol=fully supervised2023.05 | 96.4 | 91 | — | — | — | |
| LUPBackbone=R50, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M, Fine-tuning protocol=MGN, Input size=384 x 1282024.09 | 96.4 | 91 | — | — | — | |
| PASSBackbone=ViT-S, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M2024.09 | 96.3 | 92.2 | — | — | — | |
| VA-reID2019.12 | 96.23 | 91.7 | 98.69 | — | — | |
| DSFLPublication=ACM MM20, Re-ranking=false2021.11 | 96.2 | 89.9 | — | — | — | |
| BoT + CAJRe-ranking=CA-Jaccard2023.11 | 96.2 | 94.5 | 98.1 | — | — | |
| IRM (STL)Training Protocol=Single-task Learning, Input image size=256x1282023.06 | 96.2 | 92.3 | — | — | — | |
| LDSRe-ranking=true2021.11 | 96.17 | 94.89 | — | — | — | |
| GNN-based re-ranking2020.12 | 96.11 | 94.65 | — | — | — | |
| PyrNetre-ranking=true2020.10 | 96.1 | 94 | — | — | — | |
| RGA-SCCategory=Attention2020.09 | 96.1 | 88.4 | — | — | — | |
| RGA-SCBackbone=Resnet 50, Feature Type=Global2021.02 | 96.1 | 88.4 | — | — | — | |
| SANBackbone=ResNet-502020.12 | 96.1 | 88 | — | — | — | |
| MGN+circleloss2021.08 | 96.1 | 87.4 | — | — | — | |
| CARPublication=ICCV19, Re-ranking=false2021.11 | 96.1 | 84.7 | — | — | — | |
| RGA-SCPublication=CVPR20, Re-ranking=false2021.11 | 96.1 | 88.4 | — | — | — | |
| SANPublication=AAAI20, Re-ranking=false2021.11 | 96.1 | 88 | — | — | — | |
| SANInput size=256x1282022.05 | 96.1 | 88 | — | — | — | |
| RGA-SCBackbone=ResNet-50, Input size=256×1282023.03 | 96.1 | 88.4 | — | — | — | |
| SANBackbone=CNN, Supervision=Fully supervised, Resolution=256x1282023.10 | 96.1 | 88 | — | — | — | |
| PCL-CLIP LpclBackbone=ViT, Supervision=Fully supervised, Resolution=256x1282023.10 | 96.1 | 91 | 98.8 | 99.2 | — | |
| CIONBackbone=R50-IBN, Pre-training=Large-scale Person Images2024.09 | 96.1 | 91.4 | — | — | — | |
| SANBackbone=R50, Pre-training data=ImageNet1K-1.3M2024.09 | 96.1 | 88 | — | — | — | |
| SOLIDERBackbone=Swin-T, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M, Input size=384 x 1282024.09 | 96.1 | 91.6 | — | — | — | |
| CIONBackbone=R50-IBN, Pre-training=Large-scale Person Images2024.09 | 96.1 | 91.4 | — | — | — | |
| CIONBackbone=R50-IBN, Pre-training=Large-scale Person Images2024.09 | 96.1 | 91.4 | — | — | — | |
| RGA-SCBackbone=ResNet-502024.06 | 96.1 | 88.4 | — | — | — | |
| SANInput image size=256x1282023.06 | 96.1 | 88 | — | — | — | |
| CerberusMethod Category=Attribute-based, Re-ranking=without2025.12 | 96.1 | 89.8 | — | — | — | |
| alpha-QE2020.12 | 96.08 | 93.51 | — | — | — | |
| SCA2020.12 | 96.08 | 94.14 | — | — | — | |
| CACE-NetCategory=CACE-Net2020.09 | 95.96 | 90.3 | — | — | — | |
| Local CNN2019.08 | 95.9 | 87.4 | — | — | — | |
| Local CNNSupervision=Fully supervised, Note=reproduction result2019.04 | 95.9 | — | — | — | — | |
| k-reciprocal2020.12 | 95.9 | 94.38 | — | — | — | |
| PCL-CLIP Lpcl+LidBackbone=ViT, Supervision=Fully supervised, Resolution=256x1282023.10 | 95.9 | 91.4 | 98.5 | 99.2 | — | |
| BoT + ECNRe-ranking=ECN2023.11 | 95.9 | 94.4 | 97.8 | — | — | |
| TranSSLBackbone=ViT-S, Pre-training data=Large-scale Person Images, Number of pre-training images=4.2M2024.09 | 95.9 | 91.1 | — | — | — | |
| LDSRe-ranking=false2021.11 | 95.84 | 90.37 | — | — | — | |
| MGNSupervision=Fully supervised, Note=reproduction result2019.04 | 95.8 | — | — | — | — | |
| SCALBackbone=ResNet-502020.12 | 95.8 | 89.3 | — | — | — | |
| MGN(R101)Backbone=ResNet-1012020.12 | 95.8 | 89 | — | — | — | |
| SCAL2021.08 | 95.8 | 89.3 | 98.7 | — | — | |
| HAAPublication=ACM MM20, Re-ranking=false2021.11 | 95.8 | 89.5 | — | — | — | |
| HBFP-NetPublication=ACM MM20, Re-ranking=false2021.11 | 95.8 | 89.8 | — | — | — | |
| PATHBackbone=ViT-B, Pre-training data=Large-scale Person Images, Number of pre-training images=11M2024.09 | 95.8 | 89.5 | — | — | — | |
| TransReID-SSLBackbone=ViT-small2024.06 | 95.8 | 91.2 | — | — | — | |
| CLIP-REID + DenoiseRepBackbone=ViT-base2024.06 | 95.8 | 91.1 | — | — | — | |
| AGCLMethod Category=Attribute-based, Re-ranking=without2025.12 | 95.8 | 88.1 | — | — | — | |
| MGNQuery Setting=Single Query, Re-ranking (RK)=false2018.04 | 95.7 | 86.9 | — | — | — | |
| Pyramid-ours2018.10 | 95.7 | 88.2 | 98.4 | 99 | — | |
| MGNFeature Type=Part-based2018.10 | 95.7 | 86.9 | — | — | — | |
| PyramidType=Stripe-based, Nf=212019.03 | 95.7 | 88.2 | — | — | — | |
| MGNofficial_codes_released=false2019.08 | 95.7 | 86.9 | — | — | — | |
| MGNF-dim.=2,0482019.11 | 95.7 | 86.9 | — | — | — | |
| MGNCategory=stripe based2019.12 | 95.7 | 86.9 | — | — | — | |
| DSA-reIDCategory=human parsing2019.12 | 95.7 | 87.6 | 98.4 | — | — |