Advances in artificial intelligence (AI) and machine learning have opened up unprecedented analytics capabilities in a variety of team and individual sports, including baseball, basketball and tennis.
In Playly the application of AI techniques to soccer has opened up many new opportunities to solve existing scientific, methodological, and practical problems.
In Playly, we process video of a soccer match using several tools: we identify players and the ball, identify the actions of a particular player, estimate the player's speed and physical parameters (fatigue and engagement), and track the player's location on the pitch.
Based on the information obtained for each player, we create a multi-vector index of his participation in a particular soccer match.
Gathering the statistics extracted from the video for each player, we combine these statistics with all the existing data (age, «event stream providers» data) and create a model of the player, the current team and even footbal club.
We provide a search for the right player for the team both based on the «similarity» of players and based on «white spots», «suitability» of a particular player for a particular team.