Frontal view gait based recognition using PCA

Abstract

Gait is a biometric identification technology, which can identify people from distance, without user cooperation. This paper proposes a simple and efficient automatic gait recognition method taking frontal view silhouette of walking person using principal component analysis. Here for each image, background subtraction method is applied to extract the moving silhouette. These silhouette images are used to extract features using principal component analysis algorithm. Here principal component analysis method is applied to reduce the dimensionality of feature vectors. These reduce features vector represent the most relevant information of walking person, which are able to distinguish one people from others. Our result shows that taking frontal view image for recognition, gives good results. In our work the recognition rate is 97.50%.