The term gait recognition is typically used to signify the identification of individuals in image sequences `by the way they walk'. There is an increased interest in gait as a biometric, mainly due to its non-intrusive as well as non-concealable nature. Considerable research efforts are being devoted in the computer vision community to characterize and extract gait dynamics automatically from video. The objective is to use gait as a filter (or indicator) to effectively enhance the overall recognition performance of a system that uses multiple modalities. In this proposal, we describe two different gait recognition methods; a non-parametric method that uses the self-similarity plot of a walking sequence as the input feature for classification; and a parametric method that estimates the
spatiotemporal parameters of gait (the cadence and stride length) and exploits their linear relationship as a cue for identification. Finally, because carried loads are gait-altering, we also present a motion-based method to detect whether a walking person carries an object (load).