Step 1: Merge the two databases.

Step 2: Reduce all images to 512x512 resolution

Step 3: Calculate wavelet coefficient for each image (vertical, horizontal and diagonal)

Step 4: Use the obtained coefficient for Anisotropic Diffusion coefficient

Step 5: Apply Histogram equalization to the images

Step 6: Randomly distribute the processed images as Training/Validation-Test (80%-20%)

Step 7: Apply classical U-Net and V-Net models

Step 8: Apply proposed U-Net and V-Net models (Changes in the architectural structure are detailed in the code section and in the article.)

Step 9: Apply Seg-Net Model with hyperparameters

Step 10: Evaluate all the obtained findings in terms of performance metrics (Dice, Jaccard Index..)

Note: All codes are implemented with tensorFlow library tools and Python programming language on Jupyter Notebook. Additionally, necessary explanations are added in the source code.