This application is designed for research and educational purposes, using AI models to detect fractures in pediatric supracondylar humerus X-rays. It employs a Twin Convolutional Neural Network to enhance and analyze X-ray images.

This application is for research and educational purposes only. The AI models utilized herein may produce inaccurate or unreliable results. Always consult a medical professional for clinical diagnosis and treatment.

• Upload X-ray images in JPG, PNG, or JPEG formats.
• Enhance uploaded images using adaptive histogram equalization, sharpening, and contrast stretching.
• Automatically crop the region of interest in the X-ray image.
• Generate predictions for fractures with confidence scores.
• Visualize Class Activation Maps (CAM) to highlight areas of interest in the X-ray.

1. Clone the repository:

   git clone https://github.com/Weston0793/SCHF.git
   cd SCHF

2. Create and activate a virtual environment (optional but recommended):

   python3 -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

4. Download the pre-trained models and place them in the models folder. (Refer to the repository or contact the authors for the models.)

1. Run the Streamlit app:

   streamlit run webapp.py

2. Open your web browser and navigate to the provided URL (usually https://localhost:8501).

3. Use the interface to upload an X-ray image, and view the enhanced image, cropped image, predictions, and Class Activation Map (CAM).

Alternatively, you can check out the hosted version of the application at SCHF Diagnostics.

• Aba Lőrincz^1,2,3,*
• András Kedves²
• Hermann Nudelman^1,3
• András Garami¹
• Gergő Józsa^1,3
• Zsolt Kisander²

1. Department of Thermophysiology, Institute for Translational Medicine, Medical School, University of Pécs, 12 Szigeti Street, H7624 Pécs, Hungary; [email protected] (AL)
2. Department of Automation, Faculty of Engineering and Information Technology, University of Pécs, 2 Boszorkány Street, H7624 Pécs, Hungary
3. Division of Surgery, Traumatology, Urology, and Otorhinolaryngology, Department of Paediatrics, Clinical Complex, University of Pécs, 7 József Attila Street, H7623 Pécs, Hungary

The source code for this project is available on GitHub: GitHub Repository

This project is licensed under the GPL-3.0 License. See the LICENSE file for more details.