Human identification by walk recognition using relevance vector machine

Abstract

Biometric identification has gained a large popularity in recent years. Biometric identification of a person is giving more accurate results comparing to other security methods. In this paper we have proposed a method to recognize a person by the way he walks. This paper mainly focuses on moving subject (human) identification using Gaussian filter, background reduction, joint point plotting and feature mapping, vectorization. We have utilized the CASIA dataset. We have implemented an algorithm which is a mapping method where the generated silhouettes are mapped with point and eclipse on joints and other body parts. We have used Relevance vector machine which is a machine learning technique which use Bayesian interface to produce results for classification. With RVM we have classified and identified the person using feature extracted from given algorithm.