Automated detection of glaucoma using structural and non structural features

SpringerPlus

Table 3 Comparison Of Isolated Features And hybrid feature classification

Features	Without PCA			With PCA
Features	Sensitivity	Specificity	Acc. %	PCA	Sensitivity	Specificity	Acc. %
Wavelets	0.9	0.75	86	142	0.9	0.78	87
LBP and color moments	0.87	0.9	88	15	0.88	0.9	89
Wavelets, color moments, Correlograms and Haralick	0.84	0.64	79	84	0.87	0.75	84
Wavelets and multi-wavelets	0.86	0.75	83	304	0.93	0.78	89
Multi-wavelets	0.8	0.64	76	18	0.86	0.82	85
Wavelets, color moments, Correlograms	0.84	0.57	77	3	0.87	0.68	82
Wavelets and color moments	0.88	0.75	84	77	0.9	0.78	87
Wavelets, color moments, Haralick feat.	0.84	0.75	82	192	0.88	0.82	86
Wavelets and Haralick feat.	0.82	0.75	78	185	0.83	0.86	84
LBP	0.86	0.64	80	9	0.84	0.7	81
LBP and multi-wavelets	0.86	0.75	83	18	0.86	0.82	85
LBP, wavelets, multi-wavelets	0.86	0.67	81	309	0.9	0.78	89
Haralick feat.	0.75	0.78	76	52	0.85	0.75	82
Color moments and correlograms	0.8	0.35	69	11	0.8	0.57	74
LBP, wavelets, color moments multi-wavelets	0.86	0.75	83	308	0.9	0.78	87