Many of today’s smart phones and cameras have enough memory to record hours of video, so they are increasingly used to capture interesting moments. But often, that interesting action is preceded by long minutes of uneventful footage.
Not a problem, say computer scientists at Carnegie Mellon University. They developed software that automatically identifies the action segments of video shot by smartphone and other cameras, including Google Glass.
The algorithm called LiveLight edits out the boring and repetitive parts from the recording and plays back only the shots that involve some action.
The developers say LiveLight could help officials quickly review recordings from highway traffic or wildlife cameras, and even make short action clips that can be immediately uploaded to social media sites, such as Facebook or Twitter.
The software is now being prepared to be offered to the market.