Gait recognition has been receiving growing interest within the computer vision community. In this paper, an automat
ic gait recognition system is proposed. It is based on obtaining a gait signature image from each gait silhouette sequence.
This gait signature image represents the shape and the movement of a person from its gait cycle and is used as
the feature for identification. The dimension of the feature space is then reduced using Radon transform which is applied
to this gait signature image. Experimental results of the CASIA database show the efficiency of the proposed approach