Transforms for nifti images to use for data augmentation training models in pytorch on 3D MRI data.

Python code that transforms a series of dicom images into a nifti file

Utility scripts to build a population template with ANTs, register an atlas template plus label maps into template space, and propagate template-space masks back to individual subjects using inverse transforms. Designed for NIfTI workflows and label-safe NN interpolation.

High-performance 3D Brain Tumor Segmentation pipeline using MONAI and PyTorch. Processes multi-modal MRI (FLAIR, T1, T1ce, T2) via volumetric UNet to segment TC, WT, and ET regions. Features automated 3D transforms, Dice/HD95 metrics, and Kaggle-ready checkpoint recovery. Trained on the MSD Task 01 dataset using NIfTI files for clinical accuracy.