Human Gait Recognition System

Abstract

Biometric identification like fingerprints, retina, palm and voice recognition needs subject’s permission and physical attention, but Human Gait recognition works on the gait of walking subjects to identify people without them knowing or without their permission. The purpose of this report is to summarize the research and related work done in the past on human gait recognition system and to solve algorithms which helps in identify walking subject from distance without any permission and interference of the walking subject. This can be achieved by making a database in which we record a video of the subject from a camera and then converting the video into frames of still images and later applying feature extraction techniques to get the silhouettes and training the silhouettes with principal component analysis. The input images are then matched with the one that already exist in the database used to identify walking subject.