embellishIn addition in each stage of encoder the channels of all feature maps are reduced to 128 before inputting the FPE module In FPE module the first two blocks adopt 5×5 and 7×7 average pooling

In addition, to enhance the visual appeal, the encoder stage reduces the number of channels in all feature maps to 128 before passing them through the FPE module. Within the FPE module, the first two blocks utilize a 5×5 and 7×7 average pooling operation. Conversely, the last two blocks employ a 3×3 and 5×5 pooling kernel. Notably, padding operations ensure that all edge feature maps maintain the same size. Ultimately, the decoder produces an output that is bilinearly interpolated to match the size of the input image