Automatic Retinal Vessel Segmentation Based on an Improved U-Net Approach

Table 1

Quantitative results of the proposed method for the DRIVE fundus database.

Image	ACC	TPR	TNR
01	0.9646	0.8386	0.9779
02	0.9676	0.8001	0.9879
03	0.9577	0.6764	0.9900
04	0.9677	0.7663	0.9891
05	0.9621	0.6464	0.9947
06	0.9601	0.6831	0.9911
07	0.9630	0.6973	0.9938
08	0.9578	0.6093	0.9943
09	0.9654	0.6630	0.9938
10	0.9680	0.7683	0.9877
11	0.9650	0.7594	0.9882
12	0.9676	0.7592	0.9901
13	0.9619	0.7703	0.9838
14	0.9690	0.7661	0.9886
15	0.9715	0.7547	0.9897
16	0.9668	0.7636	0.9880
17	0.9633	0.6837	0.9909
18	0.9673	0.8114	0.9816
19	0.9729	0.8923	0.9811
20	0.9692	0.8035	0.9831