Jointly Learning the Discriminative Dictionary and Projection for Face Recognition

Table 5

The recognition accuracies (the top average recognition accuracy ± standard deviation %) of methods on the CMU PIE database under various feature dimensions.

Method	dim = 50	dim = 100	dim = 150	dim = 200	dim = 250
PCA + SRC	89.55 ± 0.72	91.44 ± 0.70	92.08 ± 0.60	92.26 ± 0.62	92.23 ± 0.59
DRSRC	89.04 ± 0.84	91.38 ± 0.57	91.89 ± 0.62	92.08 ± 0.59	92.11 ± 0.70
LCK-SVD	74.80 ± 1.84	88.12 ± 1.10	88.91 ± 0.99	89.31 ± 0.89	89.16 ± 1.03
FDDL	78.52 ± 1.42	89.47 ± 0.86	91.72 ± 0.58	92.54 ± 0.44	92.90 ± 0.61
LSD	76.98 ± 0.51	89.43 ± 0.69	90.97 ± 0.55	92.37 ± 0.71	93.12 ± 0.93
JDDRDL	83.75 ± 2.27	91.59 ± 0.65	92.94 ± 0.65	93.34 ± 0.70	93.30 ± 0.71
DSRC	84.42 ± 1.54	88.37 ± 1.28	91.24 ± 0.97	91.65 ± 0.63	92.44 ± 0.53
JNPDL	86.63 ± 1.19	91.23 ± 1.35	93.18 ± 0.69	92.64 ± 0.77	92.15 ± 0.64
JLDDP	90.09 ± 1.36	92.87 ± 0.99	94.07 ± 0.82	93.89 ± 0.87	93.23 ± 0.79