Research Article
A Smoothed Matrix Multivariate Elliptical Distribution-Based Projection Method for Feature Extraction
Table 1
Average recognition rate (%) and standard deviation of different methods on ORL.
| Methods | l = 5 | l = 6 | l = 7 | l = 8 |
| PCA | 0.7466 ± 0.03254 | 0.7886 ± 0.02666 | 0.8245 ± 0.02511 | 0.8593 ± 0.03828 | LPP | 0.7455 ± 0.02937 | 0.7876 ± 0.03026 | 0.8233 ± 0.02761 | 0.8563 ± 0.03456 | SPP | 0.7265 ± 0.03429 | 0.7806 ± 0.03386 | 0.8315 ± 0.03525 | 0.8643 ± 0.03805 | CRP | 0.8230 ± 0.02229 | 0.8658 ± 0.02426 | 0.8950 ± 0.02630 | 0.9140 ± 0.02797 | NMRP | 0.8185 ± 0.02423 | 0.8595 ± 0.02735 | 0.8975 ± 0.02860 | 0.9170 ± 0.02597 | SMEDP | 0.8427 ± 0.02548 | 0.8776 ± 0.02273 | 0.9073 ± 0.02671 | 0.9253 ± 0.02521 |
|
|