An Effective Retinal Blood Vessel Segmentation by Using Automatic Random Walks Based on Centerline Extraction

Table 1

Comparative performance of proposed method with other existing methods on the STARE database.

Method	Year	Sensitivity	Specificity	Accuracy
2nd human observer	Not available	0.8952	0.9385	0.9349
Hoover et al. [8]	2000	0.6734	0.9568	0.9267
Jiang and Mojon [33]	2003	Not available	Not available	0.9009
Martinez-Perez et al. [31]	2007	0.7506	0.9569	0.9410
Al-Rawi et al. [29]	2007	Not available	Not available	0.9090
Palomera-Pérez et al. [32]	2010	0.769	0.944	0.926
Budai et al. [30]	2013	0.5800	0.9820	0.9370
Chakraborti et al. [2]	2014	0.6786	0.9586	0.9379
Soomro et al. [20]	2017	0.7480	0.9220	0.9480
Abdallah et al. [17]	2018	0.6801	0.9711	0.9388
Leopold et al. [21]	2019	0.6433	0.9472	0.9045
Proposed method	2019	0.7581	0.9550	0.9401