Gait Recognition using Empirical Mode Decomposition

Gait recognition is identifying human beings by the manner in which they walk. It has been shown by several researchers that human beings have the ability to recognize other people by their gait. Machine recognition of gait is becoming increasingly important for surveillance, awarespaces etc. A number of methods have been proposed by different researchers in the recent past for this purpose. Most of these methods analyze gait as a linear and stationary signal. However, recent research shows that gait is nonlinear and nonstationary. Hence, linear analysis would be insufficient for analysis of gait. In this paper, we present our research involving nonlinear, nonstationary analysis of gait, by using a technique called empirical mode decomposition. A novel method is proposed for recognizing gait using the resulting intrinsic mode functions. Preliminary investigations reveal that this method is very effective for gait recognition. In this paper, we discuss the method, experiments, and results of the experiments in detail.