Sport Video Labeling | Rugby Analysis
Project: Rugby Video Annotation for Action Detection We are working on a large-scale rugby video labeling project focused on temporal action detection using the VGG Image Annotator (VIA). The goal is to accurately tag specific rugby actions (e.g., passes, tackles, lineouts) across full game videos for sports analytics and machine learning training. Project Highlights 365.5 full games annotated over 6 months 36,550 minutes of footage (~609 hours) 109.65 million frames processed ~950 actions/game (min 720, max 1200) 347,225 total actions tagged 1,827+ annotation hours delivered 4–5 hours spent per game on average Tools & Quality Labeled using VIA tool Annotators trained with detailed rugby guidelines Internal reviews and ongoing client feedback ensured accuracy Ongoing Work This project is still in progress. We’ve also completed multiple projects with the same client, including a large-scale soccer annotation task.