A high capacity video surveillance may not be of much use if not accompanied by the necessary work of image analysis. Without this work, one should run with the risk that suspicious acts are overlooked, or they may raise false alarms.
The cost of knowing what is happening is very high. “The problem now is that we can not put more cameras, the problem is to observe what they show is very expensive and error prone,” said Paul Cohen, head of the department of computer science at the University of Arizona.
To help mitigate some of these limitations, Cohen and his collaborators are working to build an intuitive, not only to capture video, but also to detect suspicious human behaviors. (more…)









