Brain-Tumor-Classification

Description

This is a final year thesis that implements a neural network for classifying different types of brain tumors from MRI scans. The system can identify three types of tumors: glioma, meningioma, and pituitary tumors using convolutional neural networks (CNNs).

The project uses VGG16 as the primary model architecture with LeNet5 implemented for performance comparison. The system achieves approximately 95% accuracy with VGG16 and 85-90% accuracy with LeNet5.

Why?

This project addresses the possibilities of critical need for fast and accurate brain tumor detection in medical imaging.
The motivation includes:

• Medical Impact: Providing radiologists with an automated tool for more accurate diagnosis
• Patient Care: Helping improve prognosis for patients with brain tumors through early and precise detection
• Efficiency: Offering a rapid analysis method for MRI scans
• Learning Opportunity: Gaining hands-on experience with CNNs and applying cutting-edge research in medical image analysis

The proposed method involves pre-processing MRI scans to extract relevant features, training CNNs using datasets of MRI scans with and without brain tumors, and fine-tuning architecture and hyperparameters for optimal performance.

Quick Start

Prerequisites: Google Colab or similar Jupyter notebook environment

1. Open BrainTumorDetection.ipynb in Google Colab or any similar technologies
2. Update directory paths for your dataset location
3. Run the training code blocks
4. Expected results: ~95% accuracy (VGG16), ~85-90% accuracy (LeNet5)

⚠️ Note: AI models (~2.5 GB) are not included in the repository due to size constraints.

Usage

Training the Model

# Run the training code block in the notebook
# Make sure to update dataset directories before running

Testing the Model

# Update the testing directories to point to your test data
# Run the evaluation blocks to see model performance

Model Comparison

The notebook includes both VGG16 and LeNet5 implementations for performance comparison:

• VGG16: Primary model with ~95% accuracy
• LeNet5: Baseline comparison model with ~85-90% accuracy

Contributing

Contributions are welcome! This is an educational project with potential for real-world medical applications.

How to Contribute

1. Fork the repository
2. Create a feature branch (git checkout -b feature/improvement)
3. Make your changes and test thoroughly
4. Update documentation as needed
5. Submit a pull request

Areas for Improvement

• Model optimization and hyperparameter tuning
• Additional tumor type classification
• Data augmentation techniques
• Model deployment and web interface
• Performance optimization for larger datasets

Have fun exploring medical AI! 🧠🔬