Skip to main content
Fig. 3 | BMC Medical Informatics and Decision Making

Fig. 3

From: MSAL-Net: improve accurate segmentation of nuclei in histopathology images by multiscale attention learning network

Fig. 3

a Demonstrates the residual blocks, where each block consists of \(3 \times 3\) convolution and b shows the network architecture, where the downsampling layers (In-Block, Residual-Blocks) receive the input of RGB image, then DDC is applied in the middle part, and upsampling layers (Decoder-Block) deliver the final output. The green arrows represent the skip connections between the contraction to expansion path. Furthermore, each convolutional block uses a conv, batchnorm, and relu layer, whereas each deconvolutional block uses a deconv, batchnorm and relu layer

Back to article page