Fig. 9From: Retinal vessel segmentation with constrained-based nonnegative matrix factorization and 3D modified attention U-NetSegmentation results of the proposed methodology, from up to down is original retinal images, ground truth, and the corresponding predications, respectivelyBack to article page