Class-specific kernel linear regression classification for face recognition under low-resolution and illumination variation conditions

EURASIP Journal on Advances in Signal Processing

Table 2 Performance (%) comparisons of different methods on AR in size of 8 × 6 pixels

Methods\training data	Left lighting	Right lighting	Full lighting	Average
PCA+Euclidean	41.67	34.17	8.75	28.20
PCA+Mahalanobis	40.00	32.08	22.50	31.53
KPCA-p	41.25	31.25	12.50	28.33
KPCA-g	42.42	34.58	16.67	31.22
LDA+Euclidean	74.42	44.00	50.25	56.22
LDA+Mahalanobis	52.92	34.17	54.17	46.09
KLDA-p	75.25	47.33	41.50	54.69
KLDA-g	73.50	51.75	47.17	57.47
LRC	75.33	54.08	42.00	57.14
RLRC	74.42	57.75	41.90	58.03
SRC	80.00	72.50	34.17	62.22
LPP	60.83	42.92	30.83	44.86
NPE	82.08	52.50	30.42	55.00
IPCR	41.25	31.25	8.33	26.94
URC	43.75	32.92	17.50	31.39
LDRC	75.00	48.75	55.42	59.72
LBP	61.25	55.42	46.67	54.45
KLRC-p	75.42	57.92	51.25	61.53
KLRC-g	73.75	59.17	53.75	62.22