Brain Tumor Segmentation using U-Net on BraTS Dataset

This repository contains an implementation of brain tumor segmentation using the U-Net architecture on the BraTS (Brain Tumor Segmentation) dataset. The code is organized into modular Python scripts for data loading, model definition, training, testing, and utilities.

Table of Contents

1. Overview
2. Repository Structure

---

Overview

The goal of this project is to segment brain tumors from MRI scans using the U-Net architecture. The BraTS dataset provides multi-modal MRI scans (T1, T1ce, T2, FLAIR) and corresponding ground truth labels for tumor regions. The U-Net model is trained to predict three tumor sub-regions:

• Whole Tumor (WT)
• Tumor Core (TC)
• Enhancing Tumor (ET)

---

Repository Structure

brats_unet/

• ├── data_loading.py # Script for loading and preprocessing BraTS data
• ├── model.py # U-Net model implementation
• ├── train.py # Script for training the U-Net model
• ├── test.py # Script for testing the trained model
• ├── utils.py # Utility functions (e.g., loss calculation, visualization)
• ├── config.py # Configuration file (hyperparameters, paths, etc.)
• ├── dataset.yaml # YAML configuration file