EURASIP Journal on Image and Video Processing

Table 1 Performance of proposed method on different datasets

From: Retinal vessel segmentation with constrained-based nonnegative matrix factorization and 3D modified attention U-Net

STARE dataset		Predicted condition
		Prediction positive	Prediction negative
True condition	Condition positive	TP=87800	FN=11064
	Condition negative	FP=16627	TN=817366
DRIVE dataset		Predicted condition
		Prediction positive	Prediction negative
True condition	Condition positive	TP=85926	FN=15948
	Condition negative	FP=17268	TN=789017

Back to article page