Gait recognition based on the feature fusion

A gait recognition algorithm is proposed that fuses motion and static features of sequences of silhouette images - the wavelet moment and the widths capture the motion and static characteristic of gait. A subspace transformation, principal component analysis (PCA), is applied to process the spatial templates. It aims essentially at reducing data dimensionalities. Finally, nearest neighbor classifier is adopted to recognize subjects. Experimental results show that the method is efficient for human identification, and has a recognition rate of around 88% on the CASIA data set, furthermore, the performance is compared with other algorithms.