FFU-Net: Feature Fusion U-Net for Lesion Segmentation of Diabetic Retinopathy

Table 4

Comparative segmentation results of the proposed model against the state of the art on EX and SE.

		EX			SE
Methods	SEN	IOU	Dice	SEN	IOU	DICE
Dai et al.	0.8074	0.7843	0.8791	0.7006	0.7071	0.8284
Zhang et al.	0.8418	0.8137	0.8973	0.7523	0.7505	0.8575
Van Grinsven et al.	0.8031	0.7749	0.8732	0.6988	0.692	0.818
M-Net	0.8327	0.8083	0.894	0.7297	0.7156	0.8343
FC-DenseNet	0.8414	0.8099	0.8949	0.7554	0.7623	0.8651
Sambyal et al.	0.8421	0.8183	0.9001	0.7563	0.763	0.8656
FFU-Net	0.8755	0.8414	0.9138	0.7933	0.7876	0.8812
U-Net	0.7819	0.7602	0.8638	0.6713	0.6707	0.8029