HUMAN IDENTIFICATION USING GAIT BIOMETRIC

M.SivaSangari1, M.Yuvaraju2, B.Mathivanan3
1

Student -ME Software Engineering(Part Time),Anna University Coimbatore, India 3 AP/CSE / SREC*, Coimbatore, India 1 ssangari.cbe@gmail.com 3 mathisri@rediffmail.com 2 L/CSE / Anna University Coimbatore, India 2 rajaucbe@gmail.com

Abstract: Gait is an emergent biometric aimed essentially to recognize people by the way they walk.
Gait’s advantages are that it requires no contact, like automatic face recognition, and that it is less likely to be obscured than other biometrics. Gait has allied subjects including medical studies, psychology, human body modeling and motion tracking. These lend support to view that gait has clear potential as a biometric. Essentially, we use computer vision techniques to derive a gait signature from a sequence of images. From the sequence of images being taken the Gait of the person is identified using some algorithms and is stored as a database. Thus stored database images are being checked when it’s being implemented.
Keywords: Gait Recognition, Biometric

1

INTRODUCTION

Nowadays since most of the systems are being computerized more and more authentication and authorization  systems are in need for human computer interaction. As a result, the concept of biometrics evolved. A robust biometric  authentication system has to be implemented so that it can have some of the features like easy authentication, less time  consumption, efficiency and security. Characteristics of human motion, such as walking, running or jumping vary from  person to person. GAIT refers to the style of walking of an individual. Differences in human motion enable people to  identify oneself. It’s a behavioral biometric and is subject to change. The accurate estimation of the gait features is  achieved by gathering the periodic nature of human walking as a database and computing the features over many steps.  The goal is to propose and demonstrate the feasibility with a prototype system.
2 LITERATURE REVIEW

Gait Biometric has an interesting history of the development of concept human authentication M.Murray (1967) has shown “Gait as a total pattern of movement” in the American Journal of Physical Medicine, 46(1). In 1967 Murray conducted [Murray67] a systematic study to characterize the coordinated movement patterns of the various parts of the body that constitute the walking act. Murase and Sakai developed [Murase96] a method to efficiently calculate the spatio-temporal correlation for model-free moving object recognition. A. Neri, S. Colonnese, G. Russo, and P. Talone. (1998) had given the basics of Automatic moving object and background separation and W. I Scholhorn, Nigg B.M, and J.Stephanshyn, and W. Liu, (2002) proposed the Identification of individual walking patterns using time discrete and time continuous data sets. Nicholas R. Howe and Alexandra Deschamps (2003), Northampton, proposed the Better Foreground Segmentation through Graph Cuts algorithms for many tracking and surveillance applications, background. A. Kale, N. Cuntoor, B. Yegnanarayana, A.N Rajagopalan, and R. Chellappa, Center for Automation Research, University of Maryland, and College Park adopted an appearance-based approach to the problem of gait recognition. The width of the outer contour of the binarized silhouette of a walking person is chosen as the basic image feature. Subtraction provides an effective means of segmenting objects moving in front of a static background.

3 GAIT BIOMETRIC SIGNIFICANCE IN AUTHENTICATION

When there is a lot of Biometric used to identify a person they all need the support of the person either by their direct or indirect contact. The identity verification (i.e.) authentication can be a combination of knowledge(which can be a password or PIN number), possession(which can be a key or smart card) and the biometrics (such as fingerprint, Iris pattern, Retina pattern, voice and facial) for which the user acceptance to coordinate with the system is required. But when compared to other biometric Gait does not need any direct or indirect contact of the person. Also this can be used in surveillance without the knowledge of the person. There is no sensor needed .Only thing needed is a surveillance camera that can capture the images from a notable long distance .This cost is low when compared to any other biometric.
4 GAIT FEATURES

The walking style of a person is otherwise known as the GAIT of the person. There are about 24 gait features which will uniquely identify a person in this world. Approaches to gait recognition can be broadly classified as being model-based and model-free. Model based approach assumes a priori models that match the two- dimensional (2-D) image sequences to the model data. Feature correspondence is automatically achieved once when the matching between the images and the model data is established. Here several ellipses are fitted to different parts of the binarized silhouette of the person and the parameters of these ellipses such as location of its centroid, eccentricity, etc. are used as a feature to represent the gait of a person. Recognition is achieved by template matching. Model-free methods establish correspondence between successive frames based upon the prediction or estimation of features related to position, velocity, shape, texture, and color. Here optical flow algorithm is used to derive a motion image sequence for a walk cycle. Principal Components Analysis Algorithm (PCA) is then applied to the binarized silhouette to derive what are called Eigen gaits. 4.1 Gait Cycle A “Gait Cycle” is the time of heal strike between the same leg. Cyclic gait analysis – Extracting key frames from a test sequence. These frames are compared to training frames by using nearest neighbor matching among correlation scores.

Gait Cycle -1

Gait Cycle -2

Fig.1: A Complete Gait Cycle
4.1 Width Vector An important issue in gait is the extraction of salient features that will effectively capture gait characteristics. In order to be robust to changes in clothing and illumination it is reasonable to consider the binarized silhouettes of the subject. We choose the width of the outer contour of the silhouette as the feature vector. The physical structures of the subject as well as the swing of the limbs and other details of the body are retained in the width vector. The width along a given row is computed as the difference in the locations of the right-most and the left-most boundary pixels in that row and a width vector is generated for each frame. From the temporal width plots, it is clear that the width vector is roughly periodic and gives the extent of movement of different parts of the body. The brighter a pixel, the larger is the value of the width vector in that position.

4.2 Stride Analysis The stride analysis variables are most commonly used to describe a gait pattern:

Step length is the distance between the point of initial contact of one foot and the point of initial contact of the opposite foot. In normal gait, right and left step lengths are similar. • Stride length is the distance between successive points of initial contact of the same foot. Right and left stride lengths are normally equal. • Cadence or walking rate is calculated in steps per minute. • Velocity, the product of cadence and step length, is expressed in units of distance per time. "Free speed" refers to the individual's comfortable walking speed. Since individuals walk at different speeds depending on the situation, normal velocity values are somewhat arbitrary. 4.3 Extracting the Silhouette from the images

The images are being captured from the video sequence and are being stored as the database. The next step is to extract the silhouette form those captured images. Silhouette is nothing but the outline of the images representing the body structure. 4.4 The Description of the System Design The ultimate aim of the proposed work is to authenticate a person based on his walking postures. Some of the unique features of a person like velocity, dynamics and anatomy features are considered to authenticate him. The system has to strictly follow certain constraints.

Surfac e

Fig.2: Extracting Human Body Points

Fig.3 Centroid & Distance Calculation from Background Subtracted Silhouette

• • • •

The constraints are given below: The video clips should be in AVI format. The person should be present from the starting frame of the video. The frame rate should be 15 frames per second. The camera used for taking the video should be fixed.

4.4.1. Extracting Gait Information To extract the gait of a person. 1. A person is made to pass through the web camera and his video is captured. 2. A pair of video clipping of the person walking is considered. From the images the on the person the silhouette is segmented from the video using the Background Registration Technique and Graph cut algorithms[5] 3. Once the database is created for a particular set of people, the second file is taken and processed to extract the gait features and these real time extracted gait features is compared with the gait information available in the database 4. When the same person walks in front of the camera his gait features are compared in real time and is compared with the template available. When there is a match in the template displayed then the person is authenticated. If no match is found automatically the authentication of the person is denied. When the video of the person is captured a mask is being provided on the captured image. Then on the mask of the person is used to identify the various gait features like height, maximum swinging width, chest width, shoulder width, and speed of walking of the person. Similarly a pair of video clips is taken for more people. Each video clip from the pair is processed for each person and the Gait information is extracted. After extracting the frames of silhouettes from the video sequences the velocity of Gait of the particular person is being calculated. The formula for calculating the velocity is

Velocity= Distance / No. of Frames

Distance Velocity=distance / No. of Frames

Fig 4.Velocity Calculation from images Walking base is the sum of the perpendicular distances from the points of initial contact of the right and left feet to the line of forward progression. Foot angle or toe out describes an angle between the line of progression and a line drawn between the midpoints of the calcareous and the second metatarsal head.
5 CONCLUSION

There are a lot of ways by which authentication could be implemented in which the popular one is biometric authentication. A parametric approach for human identification using height and stride parameters for walking gaits achieves its accuracy by exploiting periodic nature of human walking and computing the gait features

over many steps. Thus Gait Biometrics could be best alternative where the technology is robust, efficient, user friendly, and has no healthy hazards. Though gait biometrics is complex to implement and require more sophisticated hardware. This work shows height and stride are the useful discriminates for person identification. The best approach may be combinational features of biometrics such as face recognition, etc REFERENCES [1]. Zongyi Liu and Sudeep Sarkar,”Improved Gait Recognition by Gait Dynamics normalization” IEEE Transactions on Pattern
Analysis and Machine Intelligence, Vol. 28, No. 6, June 2006. [2]. Murat EK_INC_I,” Human Identification Using Gait”, Turk J Elec Engin, VOL.14, NO.2 2006, cT¨UB_ITAK. [3]. Naresh Cuntoor, Amit Kale and Rama Chellappa,Combining Multiple Evidences for Gait Recognition. [4]. Rong Zhang, Christian Vogler, Dimitris Metaxas, “Human Gait Recognition” [5]. S. Sarkar, P.J. Phillips, Z. Liu, I. Robledo-Vega, P. Grother, and K.W. Bowyer, “The Human ID Gait Challenge Problem: Data Sets, Performance, and Analysis,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 27, no. 2, pp. 162-177, Feb. 2005 [6]. Z. Liu and S. Sarkar, Simplest Representation yet for Gait Recognition: Averaged Silhouette, Proc. Int’l Conf. Pattern Recognition, vol. 4, pp. 211-214, 2004. [7]. D. Tolliver and R. Collins, Gait Shape Estimation for Identification, Proc. Int’l Conf. Audio and Video-Based Biometric Person Authentication, pp. 734-742, 2003. [8]. L. Wang, T. Tan, H. Ning, and W. Hu, “Silhouette Analysis-Based Gait Recognition for Human Identification,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 12, pp. 1505- 1518, Dec. 2003 [9]. Chiraz BenAbdelkader, Ross Cutler and Larry Davis (2003) ‘Person Identification Using Automatic Height and Stride Estimation’, University of Maryland, College Park, College Park, MD 20742, Microsoft Research. [10]. R. Collins, R. Gross, and J. Shi, “Silhouette-Based Human Identification from Body Shape and Gait,” Proc. Int’l Conf. Automatic Face and Gesture Recognition, pp. 366-371, 2002. [11]. Philippe C. Cattin (2002) ‘Biometric Authentication System Using Human Gait’, Swiss Federal Institute of Technology Zurich. [12]. L. Lee and W. E. L. Grimson, “Gait analysis for recognition and classification,” in Proc. IEEE Conf. Face and Gesture Recognition, 2002, pp.155–161. [13]. J. Little and J. Boyd, “Recognizing people by their gait: The shape of motion,” Videre, vol. 1, no. 2, pp. 1–32, 1998. [14]. Neri, S. Colonnese, G. Russo, and P. Talone, “Automatic moving object and background separation,” Signal Processing, vol. 66, pp. 219–232, Apr. 1998 [15]. J. Cutting and L. Kozlowski, “Recognizing friends by their walk: gait perception without familiarity cues,” Bull. Psycho. Soc., vol. 9, pp. 353–356, 1977. [16]. M. P. Murray, “Gait as a total pattern of movement,” Amer. J. Phys. Med., vol. 46, pp. 290–332, June 1967. [17]. M. P. Murray, A. B. Drought, and R.C. Kory, “Walking patterns of normal men,” J. Bone and Joint Surgery, vol. 46-A, no. 2, pp.335–360, 1964.