Automated software has been developed that can shoot sporting events from different angles and create a video stream without the need for camera operators or editors. ESPN has already experimented with the software and is interested in using it, news reports said.
Called the Autonomous Production of Images based on Distributed and Intelligent Sensing (APIDIS), the system combines the video stream from several cameras, said Christophe De Vleeschouwer, at the Catholic University of Louvain (UCL) in Belgium in an interview with New Scientist.
Tracking a ball across various video streams is relatively easy, De Vleeschouwer said; however, viewers also want to see what the players are doing. APIDIS aims for a shot of the action that is a compromise between focusing on the ball and wider views of the pitch by tracking the ball and players simultaneously, calculating which camera captures the most detail.
APIDIS can be tailored to viewers’ demands, De Vleeschouwer said, by giving preference to shots containing particular players. For example, perhaps you’re Chelsea captain John Terry’s wife and you want to keep an eye on him for the entire Premiership match against Everton. That was apparently done in a test of the software.
APIDIS has been tested on several basketball matches, a game chosen because of its fast pace. The resulting footage was good enough to attract the interest of ESPN, De Vleeschouwer said.
To see a video of APIDIS in action, click here.
