Retinal Vessel Segmentation by Deep Residual Learning with Wide Activation

Figure 7

Patches sampled for model training: (a) the preprocessed patches; (b) the corresponding ground truth.

(a)

(b)