Deep Learning-Based 3D Segmentation & Classification of Brain Tumors

Cairo University — Graduation Project 2024

Our project focused on developing an intelligent system to classify and segment brain tumors from MRI scans, aiming to support doctors in diagnosis and surgical planning.

By combining deep learning, advanced imaging, and augmented reality, we created an end-to-end tool that brings AI into real-world healthcare.

MRI scans present complex patterns that make manual tumor detection difficult and time-consuming.

We trained CNN-based models (InceptionV3, SegNet) in TensorFlow to accurately classify and segment tumors.

We built a DICOM viewer app to visualize MRI data in 2D/3D, improving doctors’ workflow.

Finally, we integrated an AR tool using Unity + C, enabling interactive 3D visualization of medical models.

Key Highlights

Developed deep learning models for tumor segmentation and classification.
Created a desktop DICOM viewer for MRI visualization.
Integrated augmented reality (Unity + C) for interactive 3D exploration.
Award: Best Graduation Project in Biomedical Engineering, and top 18 across 185 projects at Cairo University.

Watch Demo Video

I’m grateful to my professors, teammates, and family for their guidance and support throughout this journey.

Watch Full Project on YouTube →