Deep Learning Brain Tumor MRI Segmentation using U-Net

Project Overview & Abstract

Medical AI project utilizing a custom U-Net architecture implemented in PyTorch. Achieves a 92.8% Dice Similarity Coefficient across BraTS multimodal MRI datasets. Highlights necrotic, enhancing, and edema tissue contours with an interactive slice-by-slice diagnostic viewer.

Technologies & Tools Used

Deliverable Package File Contents