| | Input: Processed Trained CXR Images (PTCXRIMG) |
| | Output: Training Model (Net) |
| (1) | Start Procedure |
| (2) | Model = DCNCCLayers.CreateModel() |
| (3) | Model.add (Input layer) |
| (4) | Model.add (New Convolution block1) |
| (5) | Model.add (Normalization layer1_1) |
| (6) | Model.add (ReLU layer2_1) |
| (7) | Model.add (Pooling layer3_1) |
| (8) | Model.add (Dropout rate layer4_1) |
| (9) | Model.add (New Convolution block2) |
| (10) | Model.add (Normalization layer1_2) |
| (11) | Model.add (ReLU layer2_2) |
| (12) | Model.add (Pooling layer3_2) |
| (13) | Model.add (Dropout rate layer4_2) |
| (14) | Model.add (Flatten layer5_2) |
| (15) | Model.add (New Convolution block3) |
| (16) | Model.add (Normalization layer1_3) |
| (17) | Model.add (ReLU layer2_3) |
| (18) | Model.add (Pooling layer3_3) |
| (19) | Model.add (Dropout rate layer4_3) |
| (20) | Model.add (Flatten layer5_3) |
| (21) | Model.add (FullyConnected layer) |
| (22) | Model.add (Sigmoid layer) |
| (23) | Model.add (Classification layer) |
| (24) | Opt = trainingOptions ( |
| (25) | Initial_Learning_Rate = 0.0001, |
| (26) | Initial_Drop_Rate = 0.5, |
| (27) | Batch_Size = 32, |
| (28) | Max_Epochs = 50) |
| (29) | Net = TrainNetwork (PTCXRIMG, Model, Opt) |
| (30) | Return Net |
| (31) | End Procedure |
