Algorithm 2. The VGG-19 model description

Input: thyroidal images of dimension (500 px, 500 px) from the train dataset.

Output: VGG model weights

1. for each image in the dataset

2. Resize image to (224 px, 224 px)

3. Normalize the image pixels values between [0, 1].

4. end

5. Load the VGG-19 model pre-trained on ImageNet dataset.

6. Remove the last layer of the model.

7. Make non-trainable all the layers of the model.

8. Add a Flatten layer on the model output to obtain a 1-D array of features.

9. Add a fully connected layer with 256 hidden units.

10. Apply a dropout for inactivate neurons in the previous layer.

11. Add a fully connected layer with 4 hidden units and a softmax activation function.

12. Optimize the model with Adam optimizer.

13. Train the model for 100 epochs.

14. Save the final model.