Research Article
Ensemble Learning-Based Person Re-identification with Multiple Feature Representations
Table 1
Matching rates of different approaches on the VIPER dataset.
| Rank | | | 1 | 10 | 25 | 50 | 1 | 10 | 25 | 50 |
| Our approach | 0.147 | 0.513 | 0.701 | 0.867 | 0.285 | 0.763 | 0.915 | 0.987 | RS-KISS | 0.098 | 0.405 | 0.608 | 0.765 | 0.245 | 0.666 | 0.847 | 0.930 | RDC | 0.091 | 0.344 | 0.535 | 0.697 | 0.157 | 0.539 | 0.752 | 0.879 | LFDA | 0.101 | 0.388 | 0.593 | 0.766 | 0.202 | 0.632 | 0.826 | 0.928 | MFA | 0.100 | 0.391 | 0.596 | 0.769 | 0.201 | 0.655 | 0.843 | 0.938 | Adaboost | 0.042 | 0.020 | 0.350 | 0.503 | 0.082 | 0.366 | 0.582 | 0.909 | Bhat | 0.038 | 0.124 | 0.203 | 0.295 | 0.047 | 0.166 | 0.266 | 0.402 | PLS | 0.023 | 0.082 | 0.142 | 0.232 | 0.27 | 0.109 | 0.204 | 0.329 | Xing’s | 0.036 | 0.121 | 0.203 | 0.295 | 0.047 | 0.166 | 0.266 | 0.415 |
|
|