Tennis analysis using deep learning and machine learning.
Tennis is a sport played all over the world. Watching a tennis game, you could enrich the view automatically with numbers of serves through the middle or in the corners, how deep the ball is being played, preferences for left or right, depending on where the player is positioned. Such statistics can be provided by tools like Hawk-Eye, IBM Slamtracker and so on.
Hawk-Eye is a complex system comprising up to ten high-speed cameras, which are able to track the ball with high accuracy, obtain its real-world position and display a reconstruction of any bounces.
IBM Slamtracker, an application presenting real-time scores and statistics (between 15 and 25 parameters for each point) to augment the experience of fans. This system is extremely accurate but also highly sophisticated, comprising between 8 to 10 high-speed cameras (up to 1 000 fps) and an extremely powerful computer. The fact that this technology is equipment-intensive, costly and requires expertise to install it on a court restricts its availability to the high profile venues of major tournaments.
Our goal is to create a video analysis tool which includes ball tracking, court tracking, bounce detection and players tracking, based on only one single camera with 25–30 fps.
0 Comments