Connected Component Based Efficient Gait Recognition System

Abstract

Abstract

Gait recognition is one of the major security concerns for biometric authentication. In this work, an efficient gait recognition system using connected component based segmentation algorithm is proposed. This proposed system works on two different sub-phases, named as extraction phase and recognition phase. In the extraction phase, connected component based segmentation algorithm is applied on different gait images, taken from different sources and with the help of threshold value, decided with multiple experiments performed on gait image segments, the number of segments are reduced. While in recognition phase, artificial neural network is used. This proposed system helps to enhance the overhead problem of standard PCA algorithm. Experiments are carried out on different datasets, collected from CASIA dataset, which are in the form of silhouette images. Experimental results show that proposed system efficiency, in terms of recognition rate as well as time, is better than original PCA algorithm while reducing number of segments.

Keywords: Connected component, PCA algorithm, PCA, gait recognition, segmentation, threshold

Cite this Article

Yogesh Sharma, Manuj Mishra, Rajendra Singh Kushwaha. Connected Component Based Efficient Gait Recognition System. Journal of Image Processing & Pattern Recognition Progress. 2017; 4(2): 13–16p.