We describe an approach to quantitatively evaluating human-assisted failure-recovery tools and processes in the environment of modern Internet- and enterprise-class server systems...
Aaron B. Brown, Leonard Chung, William Kakes, Calv...
Human-area segmentation is a major issue in video surveillance. Many existing methods estimate individual human areas from the foreground area obtained by background subtraction, ...
We propose a method for human activity recognition in videos, based on shape analysis. We define local shape descriptors for interest points on the detected contour of the human a...
In many scenarios a scene is filmed by multiple video cameras located at different viewing positions. The difficulty in watching multiple views simultaneously raises an immediate...
We describe a “bag-of-rectangles” method for representing and recognizing human actions in videos. In this method, each human pose in an action sequence is represented by orien...