Algorithm 3. The VGG-CNN ensemble description.

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

Output: prediction probabilities for each diagnosis class (autoimmune, micro-nodular, nodular, normal).

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 trained 5-CNN model.

6. Load the trained VGG-19 model.

7. Predict the images with CNN resulting a list of probabilities (P11, P12, P13, P14)

8. Predict the images with VGG resulting a list of probabilities (P21, P22, P23, P24)

9. Average the two lists of predictions of the two models.

10. for each class in the set of diagnosis

11.Output prediction probabilities for the diagnosis class.

12. end