Inspiration
After witnessing the daily challenges faced by our disabled friend, including communication barriers, we felt compelled to embark on a project aimed at alleviating these struggles. Our goal is to develop a solution that empowers individuals with disabilities, like our friend, by providing them with tools to overcome communication obstacles and navigate the world more independently. By leveraging technology, we aim to create a platform that facilitates smoother interactions and fosters inclusivity in society.
What it does
Creating a system that dynamically detects American Sign Language (ASL) and converts it to text involves several steps and can be a complex task.
How we built it
Using opencv and mediapipe were able to detect hands and landmark them. Using the data from the detections we built an LSTM machine learning model to convert the signs into text. We then passed this text through flask to our react.js front end which then got displayed with tailwind.css.
Challenges we ran into
In our project, putting all the pieces together was tough. Two parts gave us the most trouble: first, getting the OpenCV camera to show up on the website was difficult. It took a lot of trying different things. Second, making sure all the data was sent and received in real-time was tricky. We had to figure out a lot of stuff and make it work smoothly. But, we kept going and worked together, and in the end, we got it all to work! It's cool because now our project can help people with disabilities talk easier.
Accomplishments that we're proud of
We are most proud that we came out of this hackathon with happy faces and a working product. Often times working in a team with a tight deadline is stressful, however our team came together as one unit to create our final working product.
What we learned
Aside from all we learned about web applications and fullstack development we learned what it meant to be a team. Our conditions were far from perfect but we trudged on knowing we had each others backs.
What's next for team Mise
Adding more data to our model to diversify our dataset.
Built With
- flask
- flask-cors
- keras
- mediapipe
- next.js
- numpy
- opencv
- react
- tailwind.css
- tensorflow
- typescript
Log in or sign up for Devpost to join the conversation.