Person Re-identification on PRID 2011 (test)
96.6Rank-1B-BOT + Attn-CL loss
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| B-BOT + Attn-CL losshyperparameter optimized=true, pre-trained=MARS dataset2019.11 | 96.6 | 100 | — | 100 | — | |
| B-BOT + Attn-CL losshyperparameter optimized=true, pre-training dataset=MARS2019.11 | 96.6 | 100 | — | 100 | — | |
| SOTA (Zeng, Tian, and Wu, (2018))2019.11 | 96.1 | 99.5 | — | — | — | |
| SOTA (Zeng, Tian, and Wu, (2018))2019.11 | 96.1 | 99.5 | — | — | — | |
| GLTRSource=ICCV192021.04 | 95.5 | 100 | — | 100 | — | |
| MGHSource=n/a2021.04 | 94.8 | 99.3 | — | 100 | — | |
| AdaptiveGraphSource=arXiv192021.04 | 94.6 | 99.1 | — | 100 | — | |
| ADFDSource=CVPR192021.04 | 93.9 | 99.5 | — | 100 | — | |
| B-BOT + Attn-CL losshyperparameter optimized=true2019.11 | 93.3 | 98.9 | — | 100 | — | |
| B-BOT + Attn-CL losshyperparameter optimized=true2019.11 | 93.3 | 98.9 | — | 100 | — | |
| STANSource=CVPR182021.04 | 93.2 | — | — | — | — | |
| SnippetSource=CVPR182021.04 | 93 | 99.3 | — | 100 | — | |
| QANprotocol=10-fold cross validation, setting=fine-tuning, information_used=appearance2017.04 | 90.3 | 98.2 | 99.32 | 100 | — | |
| CNN+AvePoolprotocol=10-fold cross validation, setting=baseline, aggregation=average pooling2017.04 | 81.3 | 96.6 | 98.5 | 99.6 | — | |
| SeeForestSource=CVPR172021.04 | 79.4 | 94.4 | — | 99.3 | — | |
| CNN+XQDASource=ECCV162021.04 | 77.3 | 93.5 | — | 99.3 | — | |
| ASTPNSource=ICCV172021.04 | 77 | 95 | — | 99 | — | |
| MGNSupervision=Fully2019.04 | 74.6 | — | — | — | — | |
| Local CNNSupervision=Fully2019.04 | 74.2 | — | — | — | — | |
| W-MGNSupervision=Weakly2019.04 | 72.7 | — | — | — | — | |
| MGN w/o TriSupervision=Fully2019.04 | 72.5 | — | — | — | — | |
| BaselineSupervision=Fully2019.04 | 71.8 | — | — | — | — | |
| W-Local CNNSupervision=Weakly2019.04 | 71.6 | — | — | — | — | |
| W-MGN w/o W-TriSupervision=Weakly2019.04 | 70.7 | — | — | — | — | |
| CNN+RNNprotocol=10-fold cross validation, information_used=appearance and spatio-temporal2017.04 | 70 | 90 | 95 | 97 | — | |
| CNN+Min(cos)protocol=10-fold cross validation, setting=baseline, aggregation=minimal cosine distance2017.04 | 69.8 | 91.3 | 97.1 | 99.8 | — | |
| W-BaselineSupervision=Weakly2019.04 | 68 | — | — | — | — | |
| P2SNetSupervision=Fully2019.04 | 60.5 | — | — | — | — | |
| TDLprotocol=10-fold cross validation2017.04 | 56.7 | 80 | 87.6 | 93.6 | — | |
| eSDCprotocol=10-fold cross validation2017.04 | 48.3 | 74.9 | 87.3 | 94.4 | — | |
| LADFprotocol=10-fold cross validation2017.04 | 47.3 | 75.5 | 82.7 | 91.1 | — | |
| LFDAprotocol=10-fold cross validation2017.04 | 43.7 | 72.8 | 81.7 | 90.9 | — | |
| STFV3Dprotocol=10-fold cross validation2017.04 | 42.1 | 71.9 | 84.4 | 91.6 | — | |
| NFSTSupervision protocol=Fully supervised, Feature type=Fusion2016.03 | 40.9 | 64.7 | 73.2 | 81 | — | |
| DVRprotocol=10-fold cross validation2017.04 | 40 | 71.7 | 84.5 | 92.2 | — | |
| XQDASupervision=Fully2019.04 | 39 | — | — | — | — | |
| KISSMEprotocol=10-fold cross validation2017.04 | 34.4 | 61.7 | 72.1 | 81 | — | |
| TopRankprotocol=10-fold cross validation2017.04 | 31.7 | 62.2 | 75.3 | 89.4 | — | |
| NFSTSupervision protocol=Fully supervised, Feature type=LOMO2016.03 | 29.8 | 52.9 | 66 | 76.5 | — | |
| XQDASupervision protocol=Fully supervised, Feature type=LOMO2016.03 | 26.7 | 49.9 | 61.9 | 73.8 | — | |
| L2Supervision=Fully2019.04 | 25 | — | — | — | — | |
| KFLDASupervision protocol=Fully supervised, Feature type=LOMO2016.03 | 22.4 | 46.5 | 58.1 | 68.6 | — | |
| MFASupervision protocol=Fully supervised, Feature type=LOMO2016.03 | 22.3 | 45.6 | 57.2 | 68.2 | — | |
| KISSMESupervision=Fully2019.04 | 18.2 | — | — | — | — | |
| Metric EnsemblesSupervision protocol=Fully supervised2016.03 | 17.9 | 39 | 50 | 62 | — | |
| MAHALSupervision=Fully2019.04 | 16 | — | — | — | — | |
| RPLMSupervision protocol=Fully supervised2016.03 | 15 | 32 | 42 | 54 | — | |
| KCCASupervision protocol=Fully supervised, Feature type=LOMO2016.03 | 14.3 | 37.4 | 47.6 | 62.5 | — |