A Wavelet-Based Learning Model Enhances Molecular Prognosis in Pancreatic Adenocarcinoma

Figure 3

The CNN model structure for feature extraction. The proposed deep model consists of 9 layers, namely, convolutional layer, max-pooling layer, convolutional layer, max-pooling layer, batch normalization layer, flatten layer, hidden layer, dropout layer, and output layer, with the size and link noted above.