Advanced YOLO Tracking: Football Teams, Ball & Possession (part2)

Turn a basic YOLO model into a broadcast-quality football tracker In Part 1,    • How to Train a Custom YOLO Object Detectio...   In this video, I cover: → Hardware Optimization: Forcing YOLO to use Apple Silicon (MPS) or Nvidia GPUs without crashing. → Smart Annotations: Replacing default boxes with clean triangles, referee rings, and a dedicated ball marker using OpenCV. → Zero-Label Team Detection: Using K-means clustering to automatically group players by jersey color. → Advanced Tracking: Writing the logic for a live, real-time ball possession bar. 🔗 Links: GitHub Repo: https://github.com/ajwise9/Match-node... Full written guide & code snippets: coming soon... Part 1 (Training the YOLO Model):    • How to Train a Custom YOLO Object Detectio...   Python → https://www.python.org/ OpenCV Documentation → https://docs.opencv.org/ 📌 Chapters: 0:00 — intro 0:35— Context & Problem Setup 1:03— Outlines 2:18 — Session 1: Hardware Optimization (Apple Silicon / GPU) 6:18 — Session 1.2: benchmark-restuls 6:53 — Session 2: Custom Smart Annotations 8:56 — Session 3: Auto-Detecting Teams (K-Means Logic) 14:34 — Session 3.2: Live Ball Possession 17:40 — Session 4: Calculating player speed 25:00 — The Final Codebase & Results 🚀 Next video (Part 3): we deploy the model locally and on the clould. Hit subscribe so you don't miss it!