Skip to main content
. 2021 Dec;11(12):4847–4858. doi: 10.21037/qims-21-168

Figure 3.

Figure 3

A classic example of FCN-based target segmentation. The model takes the whole image as input through encoder, extracts feature maps through a series of deconvolution layers in the decoder, and eventually generates the probability maps for multiple target classes, where each pixel is assigned to the class with the highest probability.