Skip to content

README.md

Latest commit

 

History

History
102 lines (76 loc) · 1.96 KB

README.md

File metadata and controls

102 lines (76 loc) · 1.96 KB

HealthGen - Medical Image Disease Classifier

Deep Learning system for brain MRI classification and segmentation using CT Scan and X-Ray images to diagnose multiple diseases.

Features

  • Brain MRI Classification: ResNet18-based classifier for 4 brain conditions
  • Segmentation Masks: UNet XL model for generating attention masks
  • Real-time Prediction: Fast inference on medical images
  • Visualization: Side-by-side comparison of original, prediction, and mask

Supported Diseases

  • Glioma
  • Meningioma
  • No Tumor
  • Pituitary Tumor

Installation

git clone https://github.com/TnsaAi/HealthGen.git
cd HealthGen
pip install -r requirements.txt

Dataset Structure

Organize your dataset:

archive/Training/
├── glioma/
├── meningioma/
├── notumor/
└── pituitary/

Usage

1. Train Classifier

python train_medical_classifier.py

2. Test System

python test_system.py

3. Prepare Dataset

python prepare_dataset.py

4. Single Prediction

python predict.py path/to/image.jpg

Model Architecture

  • Classifier: ResNet18 (modified for grayscale input)
  • Segmentation: UNet XL with encoder-decoder architecture
  • Input Size: 128x128 grayscale images
  • Output: Disease classification + attention mask

Files

  • train_medical_classifier.py - Main training script
  • rl_mri_model.py - UNet model definition
  • prepare_dataset.py - Dataset preparation utilities
  • predict.py - Single image prediction
  • requirements.txt - Dependencies

Requirements

torch>=1.9.0
torchvision>=0.10.0
PIL>=8.0.0
matplotlib>=3.3.0
numpy>=1.21.0
tqdm>=4.62.0

Model Performance

  • Training: Mixed precision with data augmentation
  • Validation: Real-time accuracy monitoring
  • Inference: GPU/CPU compatible

License

MIT License

Contributing

  1. Fork the repository
  2. Create feature branch
  3. Commit changes
  4. Push to branch
  5. Create Pull Request