Research Article
Color Distribution Pattern Metric for Person Reidentification
Table 3
PUR and CMC scores on IPR dataset with
test individuals.
| r | Reference or model: 3 images for each individual | Reference: 2 images for each individual | Reference: 1 image for each individual | SARC3D | l | p | r | s | l | p | r | s | l | p | r | s |
| 1 | 76 | 87.5 | 89.5 | 88 | 89 | 84.3 | 86.8 | 86.2 | 86.3 | 76.8 | 79.7 | 79 | 79.8 | 2 | 85 | 95 | 95 | 94.5 | 95 | 92.7 | 94.3 | 94 | 93.8 | 88.8 | 90.8 | 89.8 | 90.8 | 5 | 93 | 99.5 | 99.5 | 99.5 | 99.5 | 98.2 | 98.3 | 98.2 | 98.3 | 97.5 | 98.2 | 97.7 | 98.3 | 10 | | 99.5 | 99.5 | 99.5 | 99.5 | 98.8 | 99.2 | 98.8 | 99.3 | 98.7 | 98.8 | 98 | 98.7 | PUR | ā | 84.0 | 85.9 | 84.1 | 85.4 | 79.6 | 81.8 | 81.3 | 81.7 | 72.2 | 75 | 74.1 | 75.5 |
|
|
accuracy of rank 10 in SARC3D is read from Figure in [18]. PUR score is calculated using figures read from Figure in [18].
|