The aim of the StarTT project STRADA is to develop a tool that collects, analyses and learns from sports data. We aim to capture available data streams, such as heart rate monitors, video streams, and lap times in real-time and use all these data streams to automatically detect highlights during training or competition, e.g., an attack of your favorite rider in track cycling, an overtake in short track speed skating, …
The framework generates short, personalized video clips that might attract the broader attention of sports fans and broadcasting companies. STRADA is our answer to keep sports fans engaged in a time where on-demand information has become mainstream, and viewers are often saturated with an abundance of long live race coverage. Moreover, these filtered video clips also have value for athletes and coaches to improve technical and tactical skills.
The main challenges are the synchronization of different data streams, the pattern recognition in fused data streams and the automatic event detection required to create the video clips.
The infrastructure that will be further developped was already tested in numerous events including track and skiing contexts. Feel free to watch the video to get an idea of how ‘‘a day at the office’’ looks like for the STRADA team.