Kenneth Sauers
Jarret Torres
Mykola Maslych
Use this app to take a picture of a skin disorder and get a preliminary diagnosis.
Built in 36 hours during SwampHacks 2020 at University of Florida.
The images show Start Screen, Evaluation Screen, Results Screen (result is the scientific name for a common mole)
Built on Native Android, the app takes a picture and saves it to local storage. When user clicks to send the image, app converts it to a byte array, packs it into a JSON Object and sends a POST request to the server. The server responds with a string of the name of a potential skin disorder. Depending on the response, app's result screen presents a disclainer and shows additional information about the diagnosis.
Our Flask server is hosted on GCP's App Engine. It receives a POST request, decodes the byte array into an image and feeds it into a trained CNN.
We trained a CNN on a public dataset to categorize a picture into one of the seven skin disorders.