A New and Effective Automatic Gait Recognition Method via Perceptual Shape Descriptors

Yanmei Chai,Jingping Jia,Rongchun Zhao

DOI: https://doi.org/10.3969/j.issn.1000-2758.2005.06.028

2005-01-01

Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University

Abstract:Existing methods of gait recognition[2∼6] suffer from some shortcomings (some among them are quite serious), which arc discussed at the beginning of the full paper. We aim to suppress these shortcomings as much as possible with a new and effective method. We propose a new automatic gait recognition method via PSDs (Perceptual Shape Descriptors), which can effectively preserve the instantaneous walker's silhouette shape consistent with human vision. Firstly, we detect the walker in each image sequence. Then we use inner boundary tracking algorithm to extract the walker's binary contour and express it by the PSD. Next, PCA (Principal Component Analysis) is used to reduce the dimensions of training and classification data. Finally, the 2D Minkowski similarity measure and two different simple classification methods (NN and KNN) are used to recognize different subjects. Using the leave-one-out cross validation method, we obtain the correct classification rates on the UCSD database as 85.71% and 76.2% for NN and KNN respectively. While these results are not quite satisfactory, they do show preliminarily that we have suppressed the shortcomings of existing methods to a considerable extent, thus indicating that our new method is quite effective.