Chest X-ray Analysis using Deep Learning

This project is the final streamlit app developed for Omdena Myanmar Chapter

Project Goal

The project's primary objective is to create an AI-driven healthcare solution for chest X-ray analysis, disease detection, and COVID-19 diagnosis, utilizing deep learning algorithms. To begin, an extensive literature review will be conducted, informing the project's goals. A diverse dataset of chest X-ray images, encompassing COVID-19 cases and other chest diseases, will be collected and preprocessed for model training. Exploratory data analysis will be performed to reveal insights and potential biases.

Choosing the most appropriate deep learning algorithm or architecture will be a key step, considering factors such as complexity, interpretability, and computational requirements. The selected model will be implemented, trained, and optimized to achieve precise disease detection. Additionally, a user-friendly web or mobile app will be developed to provide real-time results for uploaded chest X-ray images.

The final solution will be deployed in a secure, scalable hosting environment. Comprehensive project documentation and a final presentation will be delivered to showcase results and the solution's potential impact on healthcare. Knowledge sharing and open discussions about AI-driven healthcare solutions will be encouraged.

Requirements

To run this locally, you need to have the following dependencies installed:

• Python 3.8 or above
• Streamlit

pip install -r requirements.txt

Run Web Application Locally

To start the system, run the following command in your terminal:

streamlit run main.py

Once the application is running, you can access it through your web browser. Under the models section the user will be asked to upload the chest x-ray image. After providing the required inputs, click the "Predict" button to obtain the predicted diagnosis result.

Demo Video

Here's a video of me demonstrating this web application, link

Acknowledgments

I would like to acknowledge Omdena's Myanmar Chapter for providing the opportunity to work on this project. I also extend my gratitude to all the contributors who participated in developing this system.