@skalskip92 asked that I should animate a character next so I did. This is an implementation of Tracking4All for monocular motion capture.
The project is built with Python, Mediapipe, YOLO, C# and Unity
Works in real time as well
4/ Three solve modes: Beginner, Intermediate (2-look OLL/PLL) and Advanced (Kociemba). Code is MIT licensed. How do you handle glare in colour detection? https://t.co/cUXtWlO9Uu
1/ I built a Rubik's cube solver that scans the cube through a phone camera, then walks you through the solve on an animated 3D cube with spoken instructions. No deep learning, just classic image processing.
3/ Glare was the hardest problem. A washed-out sticker kept drifting toward white. What fixed it: making hue count more than brightness, and assigning all 54 stickers at once so every colour ends up with exactly 9.
I’m Khalifa, a CV/ML/Web developer based in Ilorin 🇳🇬. I’ve been posting my open-source computer vision projects, and the support has been amazing!
I’m looking to connect with people passionate about 3D reconstruction. Huge thanks to my mentors for the guidance. @skalskip92
I used object detection and key points detection to make an AI kick up counter. I open sourced the project and the code can be found on my GitHub. Questions and contributions are welcome.
#Ai#ComputerVision#MachineLearning#Football#Soccer
Used Computer Vision+ regression to show the use of the quadratic equation in "real life" to my students:
Track an object (ROI, YOLOv8, RF-DETR)
Collect (x,y) coords
Fit trajectory with using quadratic equation
Result: the ball’s predicted path.
Math meets motion