Person Re-Identification on DukeMTMC (test)
87.3mAPMoSCE-ReID
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MoSCE-ReIDMethod Category=Attribute-based, Re-ranking=without2025.12 | 87.3 | 94.6 | |
| MoSAIC-ReIDMethod Category=General, Re-ranking=without, Backbone=ViT-B-16, LoRA rank=162025.12 | 85.7 | 93.3 | |
| PLIPBackbone=Swin-B, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 84.4 | 92.2 | |
| TransReID-SSL + DenoiseRepBackbone=ViT-base2024.06 | 84.31 | 92.9 | |
| LUP-NLBackbone=RN50, Pretrain=LUP-NL, Protocol=fully supervised2023.05 | 84.3 | 92 | |
| Ours(R101)+MGNBackbone=ResNet-101, Pre-training=LUPerson, Base Architecture=MGN2020.12 | 84.1 | 91.9 | |
| TransReID-SSLBackbone=ViT-base2024.06 | 84.1 | 92.6 | |
| CLIP-REID + DenoiseRepBackbone=ViT-base2024.06 | 83.7 | 91.6 | |
| NformerBackbone=ResNet-50, Input size=256×128, epochs=1602023.03 | 83.5 | 89.4 | |
| NformerBackbone=ResNet-502024.06 | 83.5 | 89.4 | |
| PLIPBackbone=RN152, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 83.1 | 92.1 | |
| CLIP-REIDBackbone=ViT-base2024.06 | 83.1 | 90.8 | |
| AdaSPBackbone=ResNet-50, Input size=384×1282023.03 | 83 | 91.7 | |
| CLIP-ReIDMethod Category=General, Re-ranking=without, Note=baseline2025.12 | 82.5 | 90 | |
| TransReID + DenoiseRepBackbone=ViT-base-ics2024.06 | 82.35 | 90.87 | |
| PLIPBackbone=RN101, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 82.3 | 91.8 | |
| TransReIDBackbone=ViT-base-ics2024.06 | 82.2 | 90.7 | |
| Ours(R50)+MGNBackbone=ResNet-50, Pre-training=LUPerson, Base Architecture=MGN2020.12 | 82.1 | 91 | |
| LUPBackbone=RN50, Pretrain=LUP, Protocol=fully supervised2023.05 | 82.1 | 91 | |
| PLIPBackbone=RN50, Pretrain=SYNTH, Protocol=fully supervised2023.05 | 81.7 | 91.1 | |
| PromptSGMethod Category=General, Re-ranking=without2025.12 | 81.6 | 91 | |
| AdaSPBackbone=ResNet-50, Input size=256×1282023.03 | 81.5 | 90.6 | |
| TransReID-SSL + DenoiseRepBackbone=ViT-small2024.06 | 81.22 | 88.72 | |
| MGN(R101)Backbone=ResNet-1012020.12 | 80.9 | 89.8 | |
| MPV2PMethod Category=Attribute-based, Re-ranking=without2025.12 | 80.9 | 91.9 | |
| AAFormerMethod Category=General, Re-ranking=without2025.12 | 80.9 | 90.1 | |
| CerberusMethod Category=Attribute-based, Re-ranking=without2025.12 | 80.7 | 91.1 | |
| TransReIDBackbone=ViT-B, Pretrain=IMG, Protocol=fully supervised2023.05 | 80.6 | 89.6 | |
| TransReID-SSLBackbone=ViT-small2024.06 | 80.4 | 87.8 | |
| TransReID + DenoiseRepBackbone=ViT-base2024.06 | 80.12 | 89.33 | |
| ISPBackbone=ResNet-502020.12 | 80 | 89.6 | |
| ISPBackbone=HRNet-W32, Input size=256×1282023.03 | 80 | 89.6 | |
| ISPBackbone=RN50, Pretrain=IMG, Protocol=fully supervised2023.05 | 80 | 89.6 | |
| ISPBackbone=HRNet-W322024.06 | 80 | 89.6 | |
| SCALBackbone=ResNet-502020.12 | 79.6 | 89 | |
| TransReIDBackbone=ViT-base2024.06 | 79.6 | 89 | |
| MGN*Backbone=ResNet-50, Framework=fast-reid2020.12 | 79.4 | 89 | |
| MGNBackbone=RN50, Pretrain=IMG, Protocol=fully supervised2023.05 | 79.4 | 89 | |
| ADRMethod Category=Attribute-based, Re-ranking=without2025.12 | 79 | 89.3 | |
| ABDNetBackbone=ResNet-502020.12 | 78.6 | 89 | |
| GCPBackbone=ResNet-502020.12 | 78.6 | 87.9 | |
| ABD-NetBackbone=ResNet-50, Input size=384×1282023.03 | 78.6 | 89 | |
| ABDNetBackbone=RN50, Pretrain=IMG, Protocol=fully supervised2023.05 | 78.6 | 89 | |
| GCPBackbone=RN50, Pretrain=IMG, Protocol=fully supervised2023.05 | 78.6 | 87.9 | |
| ABD-NetBackbone=ResNet-502024.06 | 78.6 | 89 | |
| MGNBackbone=ResNet-502020.12 | 78.4 | 88.7 | |
| MGNBackbone=ResNet-50, Input size=384×1282023.03 | 78.4 | 88.7 | |
| MGNBackbone=ResNet-502024.06 | 78.4 | 88.7 | |
| SONABackbone=ResNet-502020.12 | 78.3 | 89.4 | |
| Ours(R50)+BDBBackbone=ResNet-50, Pre-training=LUPerson, Base Architecture=BDB2020.12 | 77.4 | 88.7 | |
| BAT-netBackbone=GoogLeNet, Input size=256×1282023.03 | 77.3 | 87.7 | |
| BAT-netBackbone=GoogLeNet2024.06 | 77.3 | 87.7 | |
| MHNBackbone=ResNet-502020.12 | 77.2 | 89.1 | |
| CDNetBackbone=CDNet, Input size=256×1282023.03 | 76.8 | 88.6 | |
| CDNetBackbone=CDNet2024.06 | 76.8 | 88.6 | |
| BOTBackbone=ResNet-502020.12 | 76.4 | 86.4 | |
| BDBBackbone=ResNet-502020.12 | 76 | 89 | |
| SANBackbone=ResNet-502020.12 | 75.5 | 87.9 | |
| VAL-PATLearning=Unsupervised, Backbone=ResNet-502023.04 | 74.9 | 86.1 | |
| DGNetBackbone=ResNet-502020.12 | 74.8 | 86.6 | |
| GASMBackbone=ResNet-502020.12 | 74.4 | 88.3 | |
| DSABackbone=ResNet-502020.12 | 74.3 | 86.2 | |
| OSNetBackbone=OSNet2020.12 | 73.5 | 88.6 | |
| OSNetBackbone=OSNet, Input size=256×1282023.03 | 73.5 | 88.6 | |
| OSNetBackbone=OSNet2024.06 | 73.5 | 88.6 | |
| IANet2020.12 | 73.4 | 87.1 | |
| P2NetBackbone=ResNet-502020.12 | 73.1 | 86.5 | |
| C-ContrastLearning=Unsupervised2023.04 | 72.6 | 84.9 | |
| PCBBackbone=ResNet-502020.12 | 69.2 | 83.3 | |
| RLCCLearning=Unsupervised2023.04 | 69.2 | 83.2 | |
| SpCLLearning=Unsupervised2023.04 | 65.3 | 81.2 | |
| SSKDEvaluation Protocol=Protocol-2, Training Set=FastHuman, Transfer=R101-ibn to R502021.08 | 63.73 | 79.05 | |
| FastReID-R101-ibnEvaluation Protocol=Protocol-2, Training Set=FastHuman, Role=Teacher, Backbone=ResNet-101-ibn2021.08 | 63.19 | 78.64 | |
| GCLLearning=Unsupervised2023.04 | 62.8 | 82.9 | |
| Supervised LearningEvaluation Protocol=Protocol-2, Training Set=FastHuman, Backbone=ResNet-502021.08 | 62.53 | 79.44 | |
| Knowledge DistillationEvaluation Protocol=Protocol-2, Training Set=FastHuman, Transfer=R101-ibn to R502021.08 | 62.19 | 78.05 | |
| MMTLearning=Unsupervised2023.04 | 60.3 | 75.6 | |
| FastReID-R50Evaluation Protocol=Protocol-2, Training Set=FastHuman, Role=Student, Backbone=ResNet-502021.08 | 59.62 | 75.99 | |
| MixMatchEvaluation Protocol=Protocol-2, Training Set=FastHuman2021.08 | 59.49 | 77.02 | |
| MeanTeacherEvaluation Protocol=Protocol-2, Training Set=FastHuman2021.08 | 57.41 | 74.82 | |
| FixMatchEvaluation Protocol=Protocol-2, Training Set=FastHuman2021.08 | 56.84 | 74.55 | |
| HCTLearning=Unsupervised2023.04 | 50.7 | 69.6 | |
| MMCLLearning=Unsupervised2023.04 | 40.2 | 65.2 |