Tumor Segmentation Network

Description

This individual coursework involved us completing tasks within a Jupyter Notebook, with the final main task being using pytorch to create a U-Net to be trained using brain tumor images for segmentation purposes.

Details

Skills: Computer Vision Segmentation Pytorch Python
Scope: University
Date: March 2023

Responsibilities

  • Implemented a 3 layer U-Net and trained it using Grayscale brain scans and tumor segmentation images.
  • Achieved a final test loss of 0.042, with the model being able to predict and identify the tumorous regions of the brain with high accuracy.