ASL Sign Practice
Inspiration
Typing speed tests like Monkey-Type are a great way to improve typing fluency and pass the time, and we wondered if we could create a similar experience for learning American Sign Language (ASL). Our goal is to provide a fun and interactive way for users to practice ASL fingerspelling.
What We Learned
While researching models to use for this project, we deepened our understanding of computer vision, real-time machine learning inference. We also gained hands-on experience working on user experience design and integrating a machine model live inference data with a web application.
How We Built It
The website is a minimalistic web app that prompts users with words to sign. A CNN model processes webcam input to classify hand signs corresponding to ASL letters. The frontend was built with HTML, CSS, and JavaScript, while model inference is handled by an existing open source asl model created by Akram.
Challenges
The Challenges we faced include identifying a model and incorporating it into our frontend design.
Moving Forward
Our project aims to make learning the ASL alphabet fun and easy to cater to a large audience. Even though as of now our website only teaches the alphabet and not full words, we hope that even this level of awareness can help bridge communication gaps and make a meaningful difference in everyday interactions.
Log in or sign up for Devpost to join the conversation.