Magic-Fingers

The letter A joint nodes

Inspiration

We wanted to create a new tool to help bridge the gap between people who know ASL and people who don't by creating an interpreter using wrnchAI.

What it does

The program is able to take in the data for the hand from a json file provided by the wrnch cloud API and will guess the letter that a person is signing.

How we built it

We made our own data sets by taking screenshots from YouTube video tutorials and using the wrnch cloud API. From there, we extracted the data for the hands and trained a keras NN to identify the letter signed.

Challenges we ran into

We initially wanted to make a real-time communication tool by using the edge sdk from wrnch

Accomplishments that we're proud of

Being able to understand the wrnch API and teach the program to recognize ASL alphabet letters.

What we learned

We learned how to use the wrnch API as well as how to train and use the keras NN

What's next for Magic-Fingers

The next step is to work on the real-time interpretation using the edge SDK from wrnch. We'd like for our program to be able to recognize and distinguish between actual words in the future.