This (currently in progress) project aims to provide an interactive learning experience for those interested in learning the ASL alphabet! The web app combines a simple, randomized, flashcard style learning. To detect proper handsigns, the app uses a machine learning model trained on a dataset created by David Lee. The model is an extension of resnet50 from the pytorch library. Foundations of the models code come from the github repository by MLWhiz. Additionally, hand tracking from Handtrack.js was used to aid in better detection and prediction of signs.
Install requirements with the following command:
pip install -r requirement.txt
From the outermost folder, run the server. I use the following command:
python -m server
With the server running, navigate to the asl-alphabet-tutor folder and you can launch the front end with:
npm start
A learning environment that functions according to a Modified-Leitner algorithm to help aid in learning