Automatic quantification of superficial foveal avascular zone in optical coherence tomography angiography implemented with deep learning

Visual Computing for Industry, Biomedicine, and Art

Table 2 Segmentation performance of the proposed method in each parameter group

Parameter group	Dice similarity coefficient (mean ± SD)	Sensitivity (mean ± SD)	Specificity (mean ± SD)	R
G1	0.977 ± 0.001	0.972 ± 0.017	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G2	0.975 ± 0.011	0.967 ± 0.022	0.999 ± 0.001	0.997 (p = 0.000 < 0.5)
G3	0.977 ± 0.001	0.972 ± 0.017	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G4	0.976 ± 0.010	0.972 ± 0.018	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G5	0.976 ± 0.010	0.972 ± 0.018	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G6	0.976 ± 0.010	0.973 ± 0.018	0.999 ± 0.001	0.997 (p = 0.000 < 0.5)
G7	0.977 ± 0.009	0.974 ± 0.016	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G8	0.975 ± 0.011	0.971 ± 0.020	0.999 ± 0.001	0.998 (p = 0.000 < 0.5)
G9	0.972 ± 0.013	0.975 ± 0.018	0.999 ± 0.001	0.996 (p = 0.000 < 0.5)