Fig. 4

a Dblock module: schematic diagram of multiscale information acquisition based on dilated convolution; b Decoder module: schematic diagram of the upsampling process based on deconvolution; c e1, e2, e3, and e4 obtain an output of 1 × 256 × 256 through shape transformation and fusion with the output feature map of the decoder to obtain the final segmentation prediction result based on pixels; d gate module: schematic of feature map fusion based on spatial weights