Putting together sports packages is not easy, especially during a fast moving sport like tennis as with some of the longer and more intense rallies. To make the task easier for producers, this year’s event will see the adoption of artificial intelligence for the first time. This comes via courtesy of IBM.
Wimbledon and IBM have spent recent months fine-tuning an image-based artificial intelligence system to automatically edit the top action, and present to producers the most suitable action clips to broadcast. Each day during the tournament, particularly during the early rounds, generates hundreds of hours of tennis play. There are eighteen tournament courts at the championship, with an average of three matches contested on each court each day.
The artificial intelligence, powered by IBM’s Watson Visual Recognition API , will use crowd noise, player actions and match data to edit down the many hours of digital video. The digital video clips are then provided to the All England Lawn Tennis Club (which runs Wimbledon) and sections will then be uploaded on its official website, plus onto apps and other social media channels. The platform will not be used for the major broadcast coverage.
Speaking with the BBC about the technology, Sam Seddon, who is the Wimbledon client and program executive at IBM, said: “Recognizing player emotion is based on analysis of the broadcast video, and on looking for gesticulation – so for example fist pumping, arms aloft, hand shaking at the end of a match, and strong demonstrations of emotion such as shouting.”
The development process began in 2017 when IBM began analyzing tennis champions across six broad categories including: passion, performance under pressure, serve effectiveness, stamina, how well the player either adapted their normal playing style to an opponent or was able to force an opponent to conform to their tactics, and the ability to return serves.
