A Novel Local Human Visual Perceptual Texture Description with Key Feature Selection for Texture Classification

Table 2

The running time of feature extraction, feature selection, classification, and total processing of classifying textures in Brodatz database with different texture descriptors and different feature selection methods.

Method	Feature extraction(s)	Feature selection(s)	Classification(s)	Total(s)
LTM	2.12	-	65.63	67.75
LTM+PCA-MSPEM	2.12	2.74	54.65	59.51
LTM+PFA	2.12	0.10	21.63	23.85
LTM+PCA-APX	2.12	0.12	20.48	22.72
LTM+PCA-ORTH	2.12	0.18	21.10	23.40
TM	3.26	-	65.63	68.89
TM+PCA-MSPEM	3.26	2.74	54.65	60.65
TM+PFA	3.26	0.10	21.23	24.59
TM+PCA-APX	3.26	0.12	20.48	23.68
TM+PCA-ORTH	3.26	0.18	21.10	24.54
LBP	0.15	-	82.27	82.42
LBP+PCA-MSPEM	0.15	5.49	68.51	74.15
LBP+PFA	0.15	0.30	27.11	27.56
LBP+PCA-APX	0.15	0.32	25.67	26.14
LBP+PCA-ORTH	0.15	0.38	26.45	26.98
SIFT	1.19	-	73.95	75.14
SIFT+PCA-MSPEM	1.19	3.81	61.58	66.58
SIFT+PFA	1.19	0.14	24.37	25.70
SIFT+PCA-APX	1.19	0.17	23.08	24.44
SIFT+PCA-ORTH	1.19	0.25	23.77	25.21
Gabor	0.32	-	75.60	75.92
Gabor+PCA-MSPEM	0.32	4.22	62.94	67.48
Gabor+PFA	0.32	0.12	24.88	25.32
Gabor+PCA-APX	0.32	0.14	23.54	24.00
Gabor+PCA-ORTH	0.32	0.24	24.28	24.84