Retinal Vessel Segmentation by Deep Residual Learning with Wide Activation

Figure 11

Magnified patches of different network structures on DRIVE: (a) original image; (b) magnified RGB patch; (c) ground truth; (d) Network_1; (e) Network_2; (f) Network_3; (g) Network_4; (h) WA-Net.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)