Researchers have devised software that churns through sports video and automatically grabs the highlights. A New Scientist article describes several of these projects–from snooker analysis at Trinity College to soccer game scanning at the University of Florence. Computer vision problems are notoriously difficult, but, according to the article, "as sports follow fixed rules, and take place in predictable locations, computers ought to be able to pick out the key pieces of play and string them together."
Link
Last year in Lab Notes, I wrote about a similar project at UC Berkeley. Computer scientists there are using soccer and ballet video footage to demonstrate machine vision software that recognizes humans and their activities. Link
