Research Article
A Smoothed Matrix Multivariate Elliptical Distribution-Based Projection Method for Feature Extraction
Figure 6
The classification performance of SMEDP for varying values of λ, μc, and ρ on ORL. (a) The average recognition rate of SMEDP with different values of λ and μc. (b) The average recognition rate curves of SMEDP with varying values of ρ. (c) The part of Figure 6(b) where the recognition rate ranges from 87% to 93%.
| (a) |
| (b) |
| (c) |