Skip to content

Vaxier8/ASLModel

Repository files navigation

STEELHACKS ASLModel

Winner of Best Overall for SteelHacks 2023

Hosted Here!

Members

Brayden Nguyen, Logan Warren, Xavier Bear, Chris Landingin

Description

Our ASL Letter Recognition web application is designed to recognize American Sign Language (ASL) letters in real-time from live video input. The application works by taking in live video footage from the device's webcam of the user's hands, processing it using a neural network that we have trained to 75% accuracy, and then writing the corresponding ASL letter on the screen.

To train our model, we used over 25,000 pictures of ASL letters. We created tensors using TensorFlow, which is an open-source machine learning library developed by Google, and then used these tensors to input into our model. By training our model on this data, it was able to learn the patterns and features that are specific to each letter in the ASL alphabet.

Once our model has been trained, we integrated it into our web application using TensorFlow.js. TensorFlow.js is a JavaScript library that allows us to run machine learning models directly in the browser, without the need for any server-side processing. This means that our ASL Letter Recognition application can run entirely on the client-side, making it fast and responsive.

Overall, our ASL Letter Recognition web application is a powerful tool for helping people communicate more effectively with those who use ASL. It is a proof of concept for deeper and more complex ASL to be translated almost instantly. By leveraging the power of machine learning and deep learning algorithms, we are able to create an application that is highly accurate and responsive, making it easy for anyone to use.

MIT License

Allows:

  • Commercial use
  • Modification
  • Distribution
  • Private use Limitations:
  • Liability
  • Warranty

About

This is neural network assisted live ASL alphabet predicting tool that uses tenorflow mediapipe and opencv

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors