🧠 NeuroDetect | Brain Tumor Classification & Analysis System | Random Forest Algorithm

An Advanced Computer Vision System for Automated Brain Tumor Classification using Random Forest Ensemble Learning.

View Demo • Installation • How It Works • Tech Stack

---

📚 Overview

NeuroDetect is a powerful medical imaging analysis system designed to classify brain MRI scans into four categories: Normal, Glioma, Meningioma, and Pituitary Tumor.

It features a fully interactive Web Dashboard built with Flask, providing real-time diagnostics, performance analytics, and printable medical reports — bringing AI from code to clinical usability.

🌟 Key Features

• 🖥 Web-Based Dashboard
• ⚡ Real-Time Diagnosis
• 📊 Confusion Matrix & Class Distribution Charts
• 📈 Probability Breakdown for All Classes
• 🧾 Printable Medical Reports
• 🛡 Robust Validation & Error Handling

---

📸 Demo Preview

Interactive Dashboard	Diagnostic Result & Report

---

🧠 Classification Classes

Class	Description
0. Normal	Healthy brain with no detected tumor.
1. Glioma Tumor	Tumor occurring in the brain/spinal cord.
2. Meningioma Tumor	Tumor forming on protective membranes of brain/spinal cord.
3. Pituitary Tumor	Abnormal growth in the pituitary gland.

---

🛠 Technology Stack

Backend & AI

• Python 3.x
• Scikit-Learn (Random Forest Classifier)
• OpenCV (Image preprocessing)
• NumPy & Pandas

Web Framework

• Flask
• Jinja2

Frontend & Visualization

• HTML5 / CSS3
• Matplotlib & Seaborn
• JavaScript (AJAX)

---

⚙️ How It Works

1️⃣ Data Preprocessing

• Convert MRI → Grayscale
• Resize to 128×128
• Flatten to 1D vector
• Normalize for model compatibility

2️⃣ Model Training

• Random Forest Classifier (Ensemble Learning)
• Automatically trains on first app launch
• 80/20 train-test split
• High accuracy & reduced overfitting

3️⃣ Real-Time Prediction (Flask API)

• User uploads an image
• Processed via BytesIO (not saved on disk)
• Model outputs:
  • Predicted class
  • Confidence probability of each class
• UI dynamically updates with results

---

💻 Installation & Setup

1. Clone the Repository

git clone https://github.com/AmrBedir/NeuroDetect_RandomForest.git
cd NeuroDetect

2. Install Dependencies

pip install -r requirements.txt

If missing requirements.txt, install manually:

pip install flask scikit-learn opencv-python matplotlib seaborn pandas pillow

3. Dataset Structure

Ensure your dataset is placed as follows (After unzipping each file):

/brain_tumor_dataset
    /glioma_tumor
    /meningioma_tumor
    /normal
    /pituitary_tumor

4. Run the Application

python app.py

5. Access the Dashboard

Navigate to:

http://127.0.0.1:5000/

The model will auto-train on first run. Wait for the "System Ready" message.

---

📊 Model Performance

Metric	Score
Accuracy	~79.35%
Precision	80.11%
Recall	79.35%
F1-Score	79.26%

(Exact metrics may vary depending on dataset + random seed.)

---

🔮 Future Improvements

• 🚀 Switch to CNN Models (VGG16, ResNet50, EfficientNet)
• 🧬 Add 3D MRI support (NIfTI)
• ☁️ Cloud Deployment (AWS / Render / Heroku)
• 📱 REST API for Mobile Integration
• 🗂 Automated Reporting + Multi-User Accounts

---

🤝 Contributors

Developed by NeuroDetect Team — FCIS, Mansoura University

Amr Bedir GitHub • LinkedIn

Kareem Mohamed Gomaa GitHub • LinkedIn

Madonna Ebrahim GitHub • LinkedIn

Christine Amir GitHub • LinkedIn

Karoline Rushdy GitHub • LinkedIn

---

🎓 Supervised By

• Dr. Mohamed ElMogy
• Eng. Yasmeen Al-Saqqar

---

⚠️ Disclaimer

This project is for educational & research purposes only.
It is NOT intended for medical diagnosis or treatment.
Always consult certified medical professionals.

---

📝 License

Licensed under the MIT License.
See the LICENSE file for full details.

---

Made with ❤️ fueled by ☕