Fusion of multiple gait cycles for human identification

In this paper, a gait recognition system fusing multiple gait cycles is presented for human identification. First, the cycle length is estimated by calculating the autocorrelation of the foreground sum signal. After gait cycle partitioning, we extract two kinds of gait feature, gait energy image (GEI) and motion silhouette image (MSI). To identify individual, the outputs of the nearest neighbor classifiers are fused at the abstract level based on majority voting. Our proposed system is tested on the CASIA gait dataset A and the SOTON gait database. Compared to previous works, our empirical results show extraordinary performance in terms of correct classification rate.