Algorithmical and hardware advances create many opportunities for image- and vision-based intelligent systems that are human-centric. Computing is ubiquitous in every household. Computers are becoming smaller, more portable, and embedded in many common appliances and devices. In addition, digital cameras are becoming pervasive in society. They are appearing in many varieties, and are embedded in many devices from cars to telephones. 

      This talk focuses on the problem of camera networks for security applications. We will present the Hyperionframework (deployed to several Mass Transit sites around the U.S) which involvesthe computation of an extensive set of video-analytics based on human and crowd activity monitoring, automatic camera placement, camera-to-camera tracking, semi-autonomous calibration, and video forensics analysis. An innovative user interface allows a single user to monitor thousands of cameras.  We augment the system capabilities by pairing cameras with robots in order to provide swift mobility in case that the data requires so. Finally, we try to create an engineering/scientific solution which is respectful of design, privacy, and societal issues.

Suggested reading:

• Jeremy Schiff, Marci Meingast, Deirdre K. Mulligan, Shankar Sastry, and Ken Goldberg (2007) Respectful Cameras: Detecting Visual Markers in Real-Time to Address Privacy Concerns, International Conference on Intelligent Robots and Systems (IROS), San Diego
  download 1.09mb pdf