Fig. 1From: Scale-aware dense residual retinal vessel segmentation network with multi-output weighted lossa shows the general flow of network computing loss, and b shows the general flow of network computing loss using deep supervisionBack to article page