Feature visualization of 2D and 3D MRI brain scans (Early stopping class used from repository Bjarten/early-stopping-pytorch !)
- 2D_transfarmations_visualization - an example of how 2D images were generate using learned weights from 5-layer CNN model and the effects of different transformations on generated images
- SBF_filter - Switching bilateral filter coded based on the paper by Lin, Chih Hsing et al., 2010 (DOI: 10.1109/TIP.2010.2047906), as a noise detection and removal method instead of total variation denoising and gaussian blurring
- Transformation_2D_3D.py contains transformation functions, both for a separate use with 2D images and functions that can be added to the torchvision.transforms.Compose (for both 2D and 3D images)