Multiloss Function Based Deep Convolutional Neural Network for Segmentation of Retinal Vasculature into Arterioles and Venules

Table 4

Performance comparison with different AV-classification methods.

S. No	Algorithm	Pixel Level Accuracy	Segment Level Accuracy	Used Dataset
1	Grisan[21]	-	87.58%	Private
2	Kondarmann [24]	95.32%	-	Private
3	Niemeijer [25]	93.5%	-	Private
4	Vazquez [23]	87.68%	-	VICAR
5	Fraz [26]	83.27%	-	Private
6	Relan [27]	-	89.4%	DRIVE
7	Xu [3]	92.3%	-	DRIVE
8	Dashtbuzorg [28]	87.4%	-	DRIVE
9	Estrada [29]	91.97%	93.5%	DRIVE
10	Welikala [30]	91.97%	91.27%	DRIVE
11	Sufian and Fraz [11]	93.5%	-	Private
12	Proposed Methodology	94.9%	96.07%	DRIVE
		96.5%	98.13%	AVRDB
		96.5%	97.03%	AV-classification