A Novel Statistical Feature Analysis-Based Global and Local Method for Face Recognition

Table 6

The descriptive analysis of the five tests.

Methods	Yale B dataset		FEI dataset
	Accuracy	Standard deviation	Accuracy	Standard deviation
Proposed method/random forest	94.54	0.71	93.16	0.62
GLCM/random forest	42.29	0.87	39.97	0.68
BOW/random forest	70.62	0.90	82.21	0.73
FLBP/random forest	87.10	0.73	85.09	0.68
Proposed method/NN	93.61	0.44	95.27	0.45
GLCM/NN	39.21	0.95	46.32	0.95
BOW/NN	75.42	0.85	86.42	0.88
FLBP/NN	86.44	0.76	89.72	0.47