Person Re-identification on CUHK03 Manual
91.5Rank-1Part-Aligned Bilinear Representations
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Part-Aligned Bilinear Representations2018.04 | 91.5 | 99 | 99.5 | 99.9 | |
| Su et al. (Pose)Feature=Pose2018.04 | 88.7 | 98.6 | 99.2 | 99.7 | |
| Zhao et al. (Pose)Feature=Pose2018.04 | 88.5 | 97.8 | 98.6 | 99.2 | |
| Part-aligned representationMethod category=deeply learnt feature representation (DF)2017.07 | 85.4 | 97.6 | 99.4 | 99.9 | |
| Sun et al.2018.04 | 85.4 | 97.6 | 99.4 | 99.9 | |
| Zhao et al.2018.04 | 85.4 | — | — | — | |
| DCSLMethod category=deep learning with matching subnetwork (DMN)2017.07 | 80.2 | 97.7 | 99.2 | 99.8 | |
| Bai et al.2018.04 | 80.2 | 97.7 | 99.2 | 99.8 | |
| WARCAMethod category=metric learning (M)2017.07 | 78.4 | 94.6 | 97.5 | 99.1 | |
| Varior et al.2018.04 | 76.6 | 94.6 | 98 | — | |
| Ustinova et al. (Bilinear)Architecture=Bilinear2018.04 | 69.7 | 93.4 | 98.9 | 99.4 | |
| GOGMethod category=feature extraction (F)2017.07 | 67.3 | 91 | 96 | — | |
| PersonNetMethod category=deep learning with matching subnetwork (DMN)2017.07 | 64.8 | 89.4 | 94.9 | 98.2 | |
| EnsemblesMethod category=metric learning (M)2017.07 | 62.1 | 89.1 | 94.3 | 97.8 | |
| Deep MetricMethod category=deeply learnt feature representation (DF)2017.07 | 61.3 | 88.5 | 96 | 99 | |
| Shi et al.2018.04 | 61.3 | 88.5 | 96 | 99 | |
| DNSMethod category=metric learning (M)2017.07 | 58.9 | 85.6 | 92.5 | 96.3 | |
| SSSVMMethod category=metric learning (M)2017.07 | 57 | 84.8 | 92.5 | 96.4 | |
| IDLAMethod category=deep learning with matching subnetwork (DMN)2017.07 | 54.7 | 86.4 | 93.9 | 98.1 | |
| LOMOMethod category=feature extraction (F)2017.07 | 52.2 | 82.2 | 92.1 | 96.3 | |
| KISSMEMethod category=metric learning (M)2017.07 | 47.9 | 69.3 | 78.9 | 87 | |
| DeepReIDMethod category=deep learning with matching subnetwork (DMN)2017.07 | 20.7 | 51.3 | 68.7 | 83.1 | |
| BoWMethod category=feature extraction (F)2017.07 | 18.9 | 36.2 | 46.8 | — | |
| NFSTMethod variation=Fusion, Supervision setting=Fully supervised2016.03 | 0.6255 | 0.9005 | 0.948 | 0.981 | |
| Metric EnsemblesSupervision setting=Fully supervised2016.03 | 0.621 | 0.891 | 0.943 | 0.978 | |
| NFSTMethod variation=LOMO, Supervision setting=Fully supervised2016.03 | 0.589 | 0.856 | 0.9245 | 0.963 | |
| Improved DeepSupervision setting=Fully supervised2016.03 | 0.5474 | 0.865 | 0.9388 | 0.981 | |
| XQDAMethod variation=LOMO, Supervision setting=Fully supervised2016.03 | 0.522 | 0.8223 | 0.9214 | 0.9625 | |
| DeepReIDSupervision setting=Fully supervised2016.03 | 0.2065 | 0.515 | 0.665 | 0.8 |