Mathematical Problems in Engineering / 2016 / Article / Tab 2 / Research Article
Sparsity Preserving Discriminant Projections with Applications to Face Recognition Table 2 The best recognition rate and the corresponding standard deviation of the seven algorithms under the different size of the training set on ORL (
is the training sample size).
Methods SPDP 0.8283 ( 0.023) 0.8939 ( 0.019) 0.9266 ( 0.022) 0.9523 ( 0.015) 0.9686 ( 0.021) 0.9802 ( 0.015) 0.9839 ( 0.017) LPDP 0.8109 ( 0.032) 0.9006 ( 0.017) 0.9305 ( 0.015) 0.9579 ( 0.010) 0.9772 ( 0.009) 0.9863 ( 0.013) 0.9906 ( 0.015) PCA 0.6709 ( 0.026) 0.7576 ( 0.035) 0.8204 ( 0.036) 0.8626 ( 0.023) 0.8866 ( 0.027) 0.9045 ( 0.033) 0.9116 ( 0.026) LDA 0.7241 ( 0.021) 0.8276 ( 0.022) 0.8958 ( 0.031) 0.9231 ( 0.027) 0.9360 ( 0.033) 0.9465 ( 0.041) 0.9563 ( 0.044) LPP 0.7833 ( 0.023) 0.8657 ( 0.019) 0.9060 ( 0.016) 0.9289 ( 0.022) 0.9432 ( 0.027) 0.9527 ( 0.025) 0.9546 ( 0.026) NPE 0.7869 ( 0.015) 0.8689 ( 0.017) 0.9047 ( 0.022) 0.9331 ( 0.023) 0.9469 ( 0.021) 0.9565 ( 0.028) 0.9584 ( 0.023) SPP 0.7324 ( 0.028) 0.8055 ( 0.022) 0.8442 ( 0.025) 0.8704 ( 0.031) 0.8935 ( 0.026) 0.9162 ( 0.035) 0.9397 ( 0.034)