Inspiration
In a world where all the young generation can think about is "67", what better way to turn the trending brain rot hand gesture into an ACTUALLY meaningful activity? Juggling is a hobby where anyone can see themselves quickly progress, and with Juggle Buddy, users can improve their juggling skills at a rate of 67%.
What it does
This software tracks the ball trajectories of a user juggling 3 balls, and measures it against an “ideal” ball path. This ideal ball path depends on which pattern the user wants to train. There is a set of pre-trained patterns used as references. The software compares in real-time the ball trajectories of the user and returns a score based on how closely it follows the ideal trajectory.
How we built it
We used computer vision (CV) with OpenCV’s Hough Circle Transform framework.
Challenges we ran into
Our biggest bottlenecks were:
- Ideation (seven hours)
- Dependencies of Mediapipe not working
- Upgrading to Cursor pro with other Github (lowkey)
- Generating clean reference paths from the reference video
Accomplishments that we're proud of
Presenting this and having a vision.
What we learned
Be more prepared.
What's next for Juggle Buddy
Ideas:
- Overlay reference path after scaling to real-time user
- Compare real-time juggling data to reference path
- Compute scores and overlay on webcam footage
- App integration to gamify the experience
Log in or sign up for Devpost to join the conversation.