PlasmoCount is an ML tool for the automated detection and staging of malaria parasites in Giemsa-stained thin blood smears. PlasmoCount uses a collection of ResNets for the detection of individual red blood cells, classification of infection, and continuous staging of the asexual development of the parasite. You can access PlasmoCount's prototype on www.plasmocount.org (password for access available on request) or run it locally following the steps below.

PlasmoCount is built on a Flask framework with ReactJS for front-end development. You will need the following three packages installed: Python, Yarn, and Node.js. Before you start, make sure you install all Python and Node dependencies. You will also need to create an uploads folder and specify its path in the Flask config file. Finally, download the models using git lfs pull.

• src/: contains all the dynamic components in React.
• api/: contains the Flask project.
• api/app.py: contains all API endpoints.
• api/models/: stores all the ML models.

Start the production build of the React app, by running yarn build. Then run the Flask server with yarn start-api. Navigate to port 5000 to starting using PlasmoCount.

If you would like to work on the user interface, you can start the React development server with yarn start. You can then navigate to port 3000 to track your file changes in the development environment.

Please note that in this local version, all form inputs except for the uploaded files are disabled.

For questions or more information, please contact the Baum lab.