CereBro AI - Brain Tumor Detection System

A comprehensive web application for detecting and classifying brain tumors from MRI scans using deep learning.

System Architecture

The system follows a microservice architecture with three main components:

1. Frontend (React/TypeScript): User interface for uploading MRI scans, viewing results, and managing user data.
2. Backend API (Node.js/Express): Main API server handling authentication, user management, and analysis storage.
3. ML Service (Python/FastAPI): Specialized microservice for ML inference that processes MRI images and returns tumor detection results.

Architectural Diagram

+-------------------+      +--------------------+      +--------------------+
|                   |      |                    |      |                    |
|  React Frontend   |----->|  Node.js Backend   |----->|  Python ML Service |
|                   |      |                    |      |                    |
+-------------------+      +--------------------+      +--------------------+
        ^                          |
        |                          v
        |                  +-------------------+
        +------------------|    MongoDB        |
                           +-------------------+

Setup Instructions

Prerequisites

• Node.js (v14+) and npm
• Python 3.8+ and pip
• MongoDB database (local or Atlas)

ML Service Setup (Python/FastAPI)

1. Navigate to the server directory:

   cd server
   

2. Install Python dependencies:

   pip install -r requirements.txt
   

3. Configure environment variables by editing .env.fastapi file:

   • Set MODEL_PATH to the location of your model file
   • Set PORT to the desired port (default: 6000)

4. Start the ML Service:

   python main.py
   

Backend API Setup (Node.js/Express)

1. Navigate to the server directory:

   cd server
   

2. Install Node.js dependencies:

   npm install
   

3. Configure environment variables by editing .env file:

   • Set MONGODB_URI to your MongoDB connection string
   • Set JWT_SECRET and other security settings
   • Ensure ML_SERVICE_URL points to your running ML Service

4. Start the Node.js server:

   npm run dev
   

Frontend Setup (React/TypeScript)

1. Navigate to the frontend directory:

   cd cortex-view-app-main
   

2. Install dependencies:

   npm install
   

3. Start the development server:

   npm run dev
   

Model Information

The system uses a deep learning model trained on MRI brain scans to detect and classify four categories:

1. Meningioma: A tumor that grows in the membranes that cover the brain and spinal cord
2. Glioma: Tumors that grow in the brain or spinal cord, starting in the glial cells
3. Pituitary: Tumors that affect the pituitary gland
4. No Tumor: Normal brain tissue without abnormal cell growth