Brain MRI Tissue Segmentation

Developed a full pipeline for multi-class brain tissue segmentation on T1-weighted MRI volumes. Implemented unsupervised and classical vision techniques, including K-Means clustering, Multi-Otsu thresholding, and gradient-based Watershed segmentation. Evaluated performance on both 2D slices and full 3D volumes using Dice coefficient, IoU, precision, recall, and confusion-matrix analysis.

View Project