The University of Maryland Department of Electrical and Computer Engineering

Search
 
» INFO FOR:   Prospective Students | Current Students | Alumni | Industry & Government | Faculty & Staff | Family | Media
 
 
 
 
 
 
 
 
 
  The A. James Clark School of Engineering

Join our group on LinkedIn
Follow us on Twitter
Follow Us on Facebook
Directory

ECE Google Apps Mail

ECE Web VPN

Help Desk

Technical Operations

University Libraries

ECE Site Feedback


Give to ECE: Great Expectations Campaign





ECE Spotlight on Research



Human Gait Analysis & Recognition
Prof. Rama Chellappa
Dr. Rama Chellappa
Dr. Rama Chellappa

Analysis and recognition of human gait and activities has applications in surveillance, elderly care and video indexing. We have developed a compact signature for characterizing human gait and a set of activities corresponding to humans carrying objects such as a back pack, handbag or briefcase. Our approach is based on analyzing surfaces in space and time, which form graph results in a twisted helical pattern resembling a “figure-8,” or Double Helical Signature (DHS). It is shown that the patterns sufficiently characterize human gait and activities. The advantages of DHS are: (1) it naturally codes appearance and articulations; (2) it reveals an inherent geometric symmetry (FriezeGroup); and (3) it is effective for representing gait and activities. The structure of human gait is illustrated below.

Human Gait graph

The mathematical background supporting our work is the geometric group theory. To simultaneously learn and extract the structure of DHS, an iterative local curve embedding algorithm is applied. We study the geometric constraints that are useful in matching DHS across different viewing directions and individuals.

We have integrated the gait DHS into a real-time surveillance system. It is capable of localizing pedestrians without requiring landmarks. Our system is robust to target size, moving direction, and shadows, as well as occlusions. Our system can match spatio-temporal signatures generated by human motion in different views and time instants. It is also effective for classifying activities such as carrying objects. By recognizing various symmetries, we provide a robust solution that does not depend on silhouettes and landmarks. Extensive experiments indicate that the approach is superior to many existing methods in terms of accuracy and reliability.

For further details please visit:
http://www.cfar.umd.edu/~rany/Research-DNA.htm
http://www.cfar.umd.edu./~rama/research.html

More Information:

UM Gait and Face Recognition Technology Could Move Video Surveillance to a New Level




return to spotlight on research

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

 

↑ Back to Top



© Copyright 2005-2013, University of Maryland
University of Maryland A. James Clark School of Engineering Department of Electrical and Computer Engineering