Vision based Human Gait Recognition System: Observations, Pragmatic Conditions and Datasets

Background: Gait uniquely distinguishes the persons by using their physiological and psychological state. For this reason, human gait recognition at a distance recently gained wider interest from research community. In recent days, the concept of fusion based biometric systems has attracted many researchers and it is proved that gait can be valuable solution when all other biometric sources failed. However gait is still open research area. Method: The review explains the observations with the help of available data and scientific arguments of the experts in their documented works are also taken for our study. Findings: This paper provides a comprehensive survey of current developments on gait recognition approaches and emphasizes on three major phases involved in gait recognition system, namely representation, dimensionality reduction and classification. We also discuss in detail the pragmatic conditions of gait and efficient utilization of standard datasets in order to use gait recognition systems in masses. Also we highlighted the prominent guidelines to internal and external gait challenges. Further we provide the motivations for fusion of physiological and psychological biometric sources to accomplish practical scenarios. Application/Improvement: The study concludes that all previous attempts are restricted to only few external gait variations and evaluated on limited data set. However, there is no single attempt which addresses most of the external gait variations. Hence there is a scope for further exploration and evaluation.