Skip to main content

Table 6 Results comparison with existing methods on complete STARE database.

From: A framework for retinal vasculature segmentation based on matched filters

No Type Methods Year Sen Spe ACC AUC
1   2nd human observer 0.8951 0.9384 0.9348
2 Unsupervised methodology Hoover et al. [15] 2000 0.6747 0.9565 0.9264
3 Jiang et al. [18] 2003 0.9009
4 Mendonca et al. [29] 2006 0.6996 0.9730 0.9440
5 Lam et al. [30] 2008 0.9474 0.9392
6 Al-Diri et al. [13] 2009 0.7521 0.9681
7 Lam et al. [12] 2010 0.9567 0.9739
8 Fraz et al. [28] 2011 0.7311 0.9680 0.9442
9 You et al. [39] 2011 0.7260 0.9756 0.9497
10 Zhao et al. [32] 2014 0.7187 0.9767 0.9509
11 Proposed method 2015 0.7413 0.9825 0.9569 0.9682
12 Supervised methodology Staal et al. [33] 2004 0.9516 0.9614
13 Soares et al. [35] 2006 0.7207 0.9747 0.9479 0.9671
14 Ricci et al. [11] 2007 0.9584 0.9602
15 Marin et al. [37] 2011 0.6944 0.9819 0.9526 0.9769
16 Fraz et al. [41] 2012 0.7548 0.9763 0.9534 0.9768
17 Vega et al. [38] 2015 0.7019 0.9671 0.9483
  1. –, means the value is not answered in the reference paper
  2. Italic numbers denote the best cases among unsupervised methods and the comparatively better results among supervised ones in terms each evaluation criterion