Automated renal compartment segmentation of the kidneys

Thesis project to automatically segment the kidney tissue to separate the medulla and cortex based on the T1-weighted MRI scans in human subjects

NB: This project is a part of ongoing Application of Functional Renal MRI to improve the assessment of chronic kidney disease (AFiRM) study. The dataset used in this project contains MRI images of kidneys taken from live voluntary subjects.

The report for this project can be found here

Running the Code

Prerequesites

• Make sure you have Python > 3.9 and TF > 2.8 installed
• Navigate to the root of this repository
• You can Download the Dataset from here
• Code for dataset analsyis and mask generation you can find here
• Code for model execution you can find here
• FMRIB Automated Segmentation Tool was used to evaluate the dataset in 3D and for generating automated ground truth.