The program applies object and pose detection to videos.
Counting soccer ball juggles is easy when you can do a couple of them, but what if you're really good at it and can juggle all day long? Then this program by Vriza Wahyu Saputra is for you. It can tell you how many times you've bounced the ball using object and pose detection as well as object tracking.
First, you'll need to download some Python dependencies and run the program on a video of you juggling. It uses YOLOv8, a state-of-the-art computer vision model, to detect key points and the soccer ball. Then, the tool will create a dynamic line that moves with your knee position and divides the body into upper and lower parts.
When the ball appears above the line, the program triggers the counter, which is disabled when the ball is below the line.
This is a simplified description of the process. If you'd like to see the code that makes it all work, check out the creator's LinkedIn post here.
The project was created with Roboflow's Supervision, which can help you with computer vision tools. Saputra's work together with Supervision and similar software can advance object tracking research and let the community come up with even more exciting ideas in the future.
Here are some tools that can help you aid the industry if you're up for it:
Find Saputra's repository here and join our 80 Level Talent platform and our Telegram channel, follow us on Instagram, Twitter, and LinkedIn, where we share breakdowns, the latest news, awesome artworks, and more.