Inspiration
We were inspired by a video called Dino Game using Hand Gestures. We wanted to scale this idea up to using an app like YouTube.
What it does
You can use YouTube keyboard shortcuts with a simple series of hand gestures. This would be especially useful if you hands were dirty and you didn't want to phyiscally touch the keyboard.
How we built it
We built this project in python utilized the OpenCV and CVZone libraries to detect hands from video input. Using the hand tracking data obtained using the above libraries, we track particular gestures such as putting up an index finger to go 10 seconds backwards.
Challenges we ran into
The different libraries we used were only compatible with an older version python(3.11), so we had to set up an virtual enviroment. Moreover, intitally we had an issue with tracking gestures because the frames update very quickly. To solve this issue, we had to track the previous state and only perform actions when the state changed.
Accomplishments that we're proud of
1) Setting up a virtual environmnet 2) Learning how to use OpenCV to detect hands 3) Gesture tracking Algorithm that keeps track of previous states to prevent lag
What we learned
We learned how to build on existing CV frameworks for novel applications. We are excited to apply this knowledge to new use cases as well.
What's next for GestureProject
1) Improve gesture detection to prevent false positives. 2) Add additional functionality such as scrolling and clicking
Log in or sign up for Devpost to join the conversation.