Inspiration
This project was based on an idea that we had many months ago but were never able to begin working on! The original idea was to use hand detection to be able to control a drone but for the sake of this hackathon, we used the hand detection idea to make an ASL detection software.
What it does
There are two separate mode, one for the user to create their own data, and another for the user to use the dataset we built prior.
How we built it
In order to build this we used Python and many libraries which included, OpenCV, MediaPipe, Scikit-learn, Tkinter, and Pickle
## Challenges we ran into Because this was our first time with any of the technologies we used today, we had no prior knowledge and had to learn everything we used from scratch. We also had trouble with the libraries and figuring out what version of python we should use.
Accomplishments that we're proud of
We are proud of this project, there was a lot of trial and error and we scrapped a lot of code and ideas in the process.
What we learned
We learned how to integrate machine learning models with a live video input, landmark techniques for hand tracking, data visualization, and data processing.
What's next for ASL-ML Translator
The next step might be to use a framework to make the small program into a proper web app, perhaps flask or django. We could also transfer over the existing code into java? which would allow it to function a a web app. If we decide to stick to the language and technologies we currently have, we could clean up the gui and add more features such as a word dataset.
Built With
- mediapipe
- opencv
- pickle
- python
- scikit-learn
- tkinter