SERR-U-Net: Squeeze-and-Excitation Residual and Recurrent Block-Based U-Net for Automatic Vessel Segmentation in Retinal Image

Table 6

AUC results of our proposed method on the 20 DRIVE images.

DRIVE	AUC	DRIVE	AUC	DRIVE	AUC	DRIVE	AUC
21_test.tif	0.9874	26_test.tif	0.9651	31_test.tif	0.977	36_test.tif	0.9862
22_test.tif	0.9824	27_test.tif	0.9796	32_test.tif	0.9618	37_test.tif	0.9831
23_test.tif	0.9624	28_test.tif	0.9758	33_test.tif	0.9772	38_test.tif	0.9842
24_test.tif	0.9756	29_test.tif	0.9789	34_test.tif	0.9824	39_test.tif	0.9838
25_test.tif	0.9768	30_test.tif	0.9772	35_test.tif	0.9780	40_test.tif	0.9841