Comparative Study of Retinal Vessel Segmentation Based on Global Thresholding Techniques

Table 4

Performance of different segmentation methods on DRIVE database.

Method	Average accuracy	Average sensitivity	Average specificity
Human observer	0.9473	0.7761	0.9725
Chaudhuri et al. [6]	0.8773	0.3357	0.9794
Staal et al. [14]	0.9442	0.7345	0.9773
Niemeijer et al. [20]	0.9416	0.7145	0.9801
Zana and Klein [18]	0.9377	0.6971	0.9769
Jiang and Mojon [11]	0.9212	0.6399	0.9625
Marín et al. [19]	0.9452	N/A	N/A
Ricci and Perfetti [24]	0.9646	N/A	N/A
Martinez-Perez et al. [17]	0.9181	0.6389	0.9496
Soares et al. [13]	0.9466	N/A	N/A
Vlachos and Dermatas [26]	0.9285	0.7468	0.9551
Akram and Khan [42]	0.9469	N/A	N/A
Amin and Hong [37]	0.9191	0.6608	N/A
Mendonça and Campilho [28]	0.9463	0.7315	N/A
Tagore et al. [38]	0.9424	N/A	N/A
CLAHE and average filter (ATC) using Otsu	0.9494	0.6535	0.9780
CLAHE and Gaussian filter (ATC) using Otsu	0.9498	0.6729	0.9765
CLAHE with average and Gaussian filters (ATC) using Otsu	0.9427	0.6416	0.9717
CLAHE and adaptive filter (ATC) using Otsu	0.9368	0.6160	0.9677
CLAHE and average filter (ATC) using ISODATA	0.9516	0.6163	0.9780
CLAHE and Gaussian filter (ATC) using ISODATA	0.9500	0.6701	0.9770
CLAHE with average and Gaussian filters (ATC) using ISODATA	0.9510	0.6027	0.9846
CLAHE and adaptive filter (ATC) using ISODATA	0.9521	0.6435	0.9819
Phase congruence with IDM-based threshold (MO)	0.9430	0.7152	0.9650
Phase congruence with IDM-based threshold (MOMF)	0.9360	0.7425	0.9546
Phase congruence with IDM-based threshold (ATC)	0.9377	0.7391	0.9569