Home // International Journal On Advances in Networks and Services, volume 4, numbers 1 and 2, 2011 // View article


Combining Biometrics Derived from Different Classes of Nonlinear Analyses of Fronto-Normal Gait Signals

Authors:
Tracey Lee
Saeid Sanei
Mohammed M. Belkhatir

Keywords: nonlinear; chaos; Hilbert Huang Transform;EMD

Abstract:
With the advent of low cost high powered computing, cameras need not just be used to record multimedia data. Cameras become sensors as we process waveforms of gait signals from the video content of humans walking towards these cameras. This sensory data allows cameras to be incorporated into networks that monitor humans and their movements. This work introduces a novel analysis of gait for human recognition which uses and can be used for surveillance. Current approaches in human gait analyses employ linear signal decomposition techniques to obtain features such as frequency and phase. In contrast, we establish the nonlinear nature of fronto-normal (FN) gait. This motivates for the use of nonlinear analyses on FN gait as a biometric and opens up new avenues for research in gait recognition. Using these nonlinear analyses to derive features, we show that by themselves they may not provide sufficient discriminating ability. But by a novel combination of two different nonlinear measures, one exploiting chaosity and another representing regularity, this can be used to identify a person using gait. We apply this in a multi-biometric experiment to demonstrate its effectiveness.

Pages: 232 to 243

Copyright: Copyright (c) to authors, 2011. Used with permission.

Publication date: September 15, 2011

Published in: journal

ISSN: 1942-2644