|
GRAND seminar Static and Motion Based Approaches to Biometric Gait Recognition Abstract Gait recognition is the process of recognizing an individual by the characteristic way that they walk. Many other biometrics, for example, face and iris, require high resolution images of the subject from a particular angle in order to give adequate recognition rates. The promise of gait recognition is that this limitation is removed, a subject can be recognized in low resolution from any viewing angle. This does not come without it's difficulties for example, accurate detection and tracking of the subject to recognize. The purpose of this talk is to provide an introduction to gait recognition and some of the challenges inherent to this problem. We will talk about model-based and model-free approaches to gait recognition. Model-based approaches locates and tracks the movements of individual parts of the body and performs recognition using this information alone. Model-free approaches use the silhouette of the subject (static or dynamic) to recognize the subject. This talk will discuss trade-offs of these approaches and present state-of-the-art solutions to gait recognition. Biography Ed Lawson received a M.S. in Computer Science from George Mason University in 2003. He is currently enrolled in the PhD program at George Mason University. His research interests are human motion understanding, biometrics, and video surveillance.
|
