Object based video synopsis

Tens of thousands of surveillance cameras all over every corner of the world send large amounts of data 24 hours a day, but most of these videos are redundant data lacking of activity. It is time consuming to browse all of those raw data. This paper has researched object based video synopsis technology with the surveillance video of a neighborhood and an indoor surveillance video as research target, presenting all necessary activities in a short time by displaying objects concurrently, even if they originally occurred at different time. The result shows that we can create a synopsis with maximum activity while avoiding conflicts between objects as much as possible.