Skip to main content
. 2018 Feb 7;8:2575. doi: 10.1038/s41598-018-19426-7

Figure 9.

Figure 9

Image transformation model of CNN. “Conv” is the abbreviation of convolution kernel. “DeConv” is the abbreviation of deconvolution kernel. The numbers before the “Conv” or “DeConv” are the numbers of kernels. “FC” is the abbreviation of fully connected layer. All the convolutional and deconvolutional layers use 3 × 3 kernels. All the layers use Relu as the activation functions43.