Anterior Chamber Angle Shape Analysis and Classification of Glaucoma in SS-OCT Images

Table 3

The average classification accuracy and the standard deviation of fuzzy kNN classifier with various feature groups (𝑚=1.2,  𝑘=5).

Features	Accuracy (%)		Error rate (%)		Sensitivity (%)		Specificity (%)		AUC
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
𝑇𝐼,𝐸𝐼,𝐻iris	94.69	0.71	5.24	0.71	99.9	0.24	77.71	3.39	0.88	0.02
mAOD, ATsl	98.21	0.82	1.79	0.82	1.00	0.00	92.14	0.38	0.96	0.01
FD	98.93	0.62	1.07	0.62	1.00	0.00	95.43	2.69	0.98	0.013
FD, mAOD, ATsl	99.11	0.76	0.89	0.76	99.9	0.24	96.11	0.24	0.98	0.012

𝑇_𝐼: triangularity, 𝐸_𝐼: ellipticity, 𝐻_iris: mean curvature of iris, mAOD: mean of continuous AOD, FD: fractal dimension, AT_sl: area of trapezoidal profile of iridocorneal angle, AUC: area under the curve, and SD: standard deviation.