Inspiration
TypeRacer inspires us; it's a website that lets people practice typing. We believe that learning a new skill through the power of gamification
What it does
Much like TypeRacer, SignRacer pits users against each other to compete against each other and sign the same sentence the fastest.
How we built it
Using a linear regression model for image classification, we were able to have a model interpret sign language.
Challenges we ran into
While developing, we ran across the issue of our model overfitting due to not having enough training data, and the training data that we did have was very biased towards lighter skin color due to that being the only images we started training with. Once we found more publicly available data, we were able to include a broader scope of users.
Accomplishments that we're proud of
We're proud of the fact that not only does it work, but it is actually quite a fun platform to practice fingerspelling in ASL.
What we learned
We learned ASL fingerspelling while developing this platform; this was also our first time working on something that involved computer vision. I would say that for a first, this was a pretty good attempt.
What's next for Sign Racer
PVE game mode for players to not only be able to compete against each other, but also play co-op against the computer.
Built With
- computervision
- datascience
- game
- interactive-data
- learningtool
- linearregression
- multiplayer
