International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075, Volume-3 Issue-1, June 2013 217 Published By: Blue Eyes Intelligence Engineering & Sciences Publication Retrieval Number: A0913063113/2013©BEIESP Abstract— Recognition of any individual is a task to identify people. Human recognition methods such as face, fingerprints, and iris generally require user’s cooperation, physical contact or close proximity. These methods are not able to recognize an individual at a distance therefore recognition using gait is relatively new biometric technique without these disadvantages. Human identification using Gait is method to identify an individual by the way he walk or manner of moving on foot. Gait offers ability of distance recognition or at low resolution. In this paper, firstly binary silhouette of a walking person is detected from each frame. Secondly, feature from each frame is extracted using image processing operation. Here center of mass, step size length, and cycle length are talking as key feature. At last BPNN technique is used for training and testing purpose. Here all experiments are done on gait database and input video. Index Terms— Backpropagation neural network (BPNN), gait recognition, silhouette images, background subtraction, features extraction. I. INTRODUCTION In most of the metropolitan, identity cards and passports are used for authentication and verification of human beings. But now days, biometric identifications are most suitable for human recognition. Biometric means unique features of a person. Biometric identification aims to recognition of an individual from their physiological and behavioral characteristics. Different biometrics measures or vectors are used to identify an individual, physiological characterstics like fingerprints, palm geometry, DNA, iris, face recognition- they all are related to the body of person and other features such as voice, gait and they are related to the behavior of person. Gait is effective way of human recognition. Gait is unobtrusive and distance recognition. It overcomes all the disadvantages of physiological characterstics like- it needs user’s cooperation, also these physiological characterstics needs only high resolution images. Example: only few authorized doctors are allowed to go into operation theater, in this scenario gait analysis technique is used as, gait sequences of those authorized doctors are stored in hospitals’ database, therefore whenever an unauthorized person tries to enter into room, then his gait sequences will not match with stored sequences and a system will generates an alarm to alert the authorities of department for any action. Revised Manuscript Received on June 06, 2013. Oshin Sharma Department of Computer Science, Chitkara University, Baddi, India. Sushil Kumar Bansal, Department of Computer Science, Chitkara University, Baddi, India. Fig 1: Gait Recognition Scenario (Mark, 2002) Background subtraction, feature extraction and Recognition are three main parts of gait recognition system. Background subtraction is the first step of gait recognition system. In this process foreground objects in a particular scene are extracted and binary silhouette images will be obtained. Next is feature extraction process. In this step input data will be transformed into a reduced set of features. In this paper we are using model based approach of feature extraction. Final step of gait system is recognition. Here both the input and trained sequences in database are compared with each other. II. REVIEW OF RECENT RESEARCH Many researchers had given their contribution in model based approach of gait recognition. In this model based gait recognition system both the motion of lower leg rotation and motion of tigh describes walking and running [1]. In our paper we are also proposing a model based approach as this approach can handles self occlusion, noise, scaling and rotation. Earliest model based approach of gait recognition system able to obtain gait signatures, when human walking as a pendulum and representing the tigh motion with combination of velocity Hough transform and Fourier representation [2]. Model based approaches can handle self occlusion, noise, scaling and rotation. [5] They used static body parameters without analyzing gait dynamics for gait recognition. Model-based approaches aim to explicitly model human body or motion, and they usually perform model matching in each frame of a walking sequence so that the parameters such as trajectories are measured on the model. Model based approaches are further divided into two main classes. First class, state space method and second is spatiotemporal method [6]. Latest model based gait recognition provides good recognition rates, when both motion model of thigh and lower leg rotation describes walking and running[3]. In model based approach moving person is divided in different Gait Recogniton System for Human Identification using BPNN Classifier Oshin Sharma, Sushil Kumar Bansal