CV at the Dutch Tennis Federation: Utilizing YOLO to create insights for coaches
summary
A one camera pipeline using YOLOv8 plus homography and U‑transform/UNet court detection produces 10 Hz player coordinates and rule based rally starts so a 5 hour match yields roughly 20–30 minutes of playing footage for annotation. Audio peaks and color clustering support serve detection and player ID. It works well on high quality broadcast angles, struggles on low frame rate or Wimbledon white kits, and delivers a Tableau dashboard and Github code for coach adoption.
AI-written summary · about
chapters (9)
- 0:00Introduction and context
- 1:19Challenges in tennis analysis
- 2:51Project objectives
- 6:02System pipeline architecture
- 12:33YOLO and object detection
- 17:00Rally and player identification
- 18:52Data output and insights
- 20:24Stakeholder engagement
- 22:14Q&A and closing
related talks
27mAnalysis of sports scenes using computer vision models
10mLightning Talk: Ball Tracking and Detection in Soccer Videos - Comparison of...
35mFootball video analysis using Deep Learning
20mAI-Driven Real-Time Ball Tracking for Live Sports Streaming
24mBeyond the Frame: Using Computer Vision to Unlock the Potential of Video Data
42mDeep Learning for Automatic Basketball Video Production
35mLarge-Scale Video Intelligence
29mAI in Sports - Basketball Rule Violation Detection