Video Surveillance to New Level of Advancement

Written by Bradley Siddell

tracking walk University of Maryland electrical and computer engineering professor Rama Chellappa has created an artificial-intelligence-based real-time computer video monitoring system that can identify suspicious activities or individuals, which may remove some of the burden from security guards who must monitor many video screens simultaneously. person walking

Chellappa, a pioneer in pattern recognition and computer vision software, developed a digital signature for the human gait, called "human gait DNA."

Deviations in this normal gait pattern cause asymmetries that the system can recognize and analyze, but for now a concealed object not effecting one's gait would not be noticed. Gait is also used to identify specific individuals, as is face recognition software. To locate and observe actual pedestrians, Chellappa has used corrective algorithms to compensate for changing light, shadows, and viewing angles. Other recognition technologies are currently being worked on with support from the Department of Homeland Security, including an algorithm to estimate the heights of subjects in the field of view of a camera, and a program that can find unattended packages using a structured representation known as attribute grammars.

Source: AScribe Newswire