mri - MRI Images nonmri - CT Scan Images
Split - 80% Train, 20% Test Split from 80% Train, 70% Train 25% Validation
Correct Amount: 60% 20% 20%
Changes:
datasetBinaryMRI - All the images within the original folder was move up one hierarchy. Healthy and Tumor images were mixed into one folder.
The original names were MRI and CT They were renamed to mri and nomri Each contained 4000 images to predict their labels respectively.
datasetBinaryTumor - Original contained all the 8000 images. However CT scans were not required. They were deleted.
The Healthy and Tumor folders within the mri folder were renamed to notumor and tumor.
The images were relabelled into new classes The images were resized to 224x224 The images were normalized The images WERE NOT SCALED BECAUSE DATA AUGMENTATION WAS DONE
noise removal?