pediatric-cancer

Brain MRI Image Classification

T1	T2	Flair

Example_Images - Example images to show how MRI scans look like
Presentation - Material associated with final Presentation
Public_Networks - The public ImageNet based Image Classification models tested, and sample images that showcase which kind of features the network is learning.
- inception_features - features which Inception V3 learns
- inception_resnet_features - features which InceptionResNet V2 learns
- resnet_features - features which ResNet50 learns

BRAIN_TUMOR - Folder to store the images from Grand Challenge Dataset (download and extract from Grand-Challenge)
- imagesTr - Training images in nii.gz format
- imagesTs - Testing images in nii.gz format
- labelsTr - Training labels in nii.gz format (for segmentation only)
- dataset.json - File descriptors
Stanford - Folder to store Stanford images
- DIPG - Diffuse Intrinsic Pontine Glioma images like DIPG-patient_*.dcm
- EP - Etoposide and Cisplastin images like EP-StanEP*.dcm
- MB - Medulloblastoma images like MB-StanMB*.dcm

Folder to store iPython notebooks

Generated by resnet_50.py, train_resnet_50.py and train_classifier.py. Contains the HDF5 files of AutoEncoder and Classification models.

Generated by resnet_50.py, train_resnet_50.py and train_classifier.py. Contains the Tensorboard logs of AutoEncoder and Classification models.

Generated by AutoEncoder-Normalized-Preprocessing.py. Contains the normalized Grand Challenge images.

Generated by AutoEncoder-PerPlane-Preprocessing.py. Contains the per-plane normalized Grand Challenge images.

Generated by Classification-Preprocessing.py. Contains the normalized Stanford Images.