Python image processing pipeline to support the diagnosis of Pectus Excavatum (PE). Allows automatic computation of existing clinical indexes (Haller Index) used in PE diagnosis and treatment with unprecedented advantages of avoiding user bias and saving time. Then with the obtained results from patient pre and post-surgery data, we trained a ML model, making it possible to predict post-surgical indexes.

• Description: Generation of synthetic images due to the lack of real data, allowing the calculation of the Haller index and the creation of a database.

• Description: Implementation of a neural network to predict surgical outcomes.
• Technologies: TensorFlow, neural networks.
• Purpose: Train the model with the synthetic image database and validate its accuracy.

• Description: Implementation of streamlit interface.
• Features: Allows uploading images, automatically calculates the Haller index, and provides predictions about surgical outcomes.

• Python 3.7 or higher
• TensorFlow 2.0 or higher
• Image processing libraries (OpenCV, PIL)
• Framework for the user interface (streamlit)

1. Clone the repository:

   git clone https://github.com/user/pectus-project.git
   cd pectus-project

2. Create a virtual environment and install dependencies:

   python -m venv env
   source env/bin/activate  # On Windows: env\Scripts\activate
   pip install -r requirements.txt

3. Run the user interface:

   streamlit run deploy/app.py

1. Start the application.
2. Upload a patient's chest image.
3. Automatically calculate the Haller index.
4. Get predictions about surgical outcomes.
5. Use the results to make informed treatment decisions.

Contributions are welcome.

For any questions or inquiries, please contact David at [d.huamanor@alum.up.edu.pe].

Thank you for your interest in PExpert! Together, we can make a significant difference in the lives of those affected by this condition.

David Huaman