In video surveillance and web-based browsing, enormous amounts of data are received and stored. Processing this data for its information content relevant to a particular application can become quite cumbersome to the point that the data is rendered mostly useless.

University of Maryland researchers have developed new software to create an optimized video summary of footage based on the important criteria of covering many portions of the video and ensuring that these portions are as distinct as possible. The novel approach, "Video Précis," creates a video summary that identifies key aspects of footage over a wide range of scenarios within a user specified summary length. The summary provides a quick glimpse of the contents of a video and significantly improves a user’s browsing experience and bandwidth savings for video-hosting websites.

File Number: IS-2009-105 

Web site: http://otc.umd.edu