Automatic Face Recognition System using P-tree and K-Nearest Neighbor Classifier Mohammad Kabir Hossain, Abu Ahmed Sayeem Reaz, Rajibul Alam, and Dr.William Perrizo* Department of Computer Science and Engineering, North South University, Dhaka 1217 *Department of Computer Science, North Dakota State University, Fargo, ND 58105, USA Emails: mkhossain@northsouth.edu , sayeem_reaz@yahoo.com , rajibulalam@hotmail.com , william.perrizo@ndsu.nodak.edu Abstract: Face recognition has recently received remarkable attention in both authentication and identification systems due to high acceptability and collectability, regardless its lower circumvention and uniqueness than other biometric verification technologies. The basic approach with face recognition commences with feature set construction from the relevant facial traits of the users, termed enrollment [1]. When a user is to be authenticated (i.e. the user's identity is to be verified), his/her facial sample is captured and a feature set is created. This feature set is then compared with the enrollment feature set. But feature set search mechanism is time consuming and sometimes exhaustive. In this paper, a very efficient and time saving search mechanism is proposed that exploits the advantages of Peano Count Tree and K- Nearest Neighbor Search techniques. Keywords: Verification, Identification, Feature set, Image processing, P-tree, bSQ format, Quadrant ID, Euclidean similarity function, Minkwoski similarity function, Manhattan distance, K-nearest neighbor classifier. 1. INTRODUCTION While research towards automatic face recognition began in the late 1960's, progress has been slow. Recently there has been renewed interest in the problem due in part to its numerous security applications ranging from identification of suspects in police databases to identity verification at automatic teller machines [1]. Though automatic face recognition systems exhibit higher FAR (False Acceptance Rate), the probability that a sample falsely matches the presented face identification record or feature sets, and FRR (False Rejection Rate), the probability that a sample of the right person is falsely rejected, than other successful biometric systems like fingerprint or iris recognition, it is attractive because of its wide-spread acceptability, universality and easier acquisition. Recognition is a step by step process and quite time- consuming in case it has to deal with a large problem domain. In applications like surveillance system, airport and banking security, database search time is significantly huge. Thus, development of real-time application is a challenging task. In this paper our prime concern has been reduction in database search time during face matching phase. In order to achieve that objective we have made use of the techniques of peano count tree and k-nearest neighbor algorithm. The rest of the paper is organized as follows: first a review of the technology behind such an identification system and then the proposed system. 2. RELATED CONCEPTS Before we go through the identifier we need to know the approaches and problem domains involved in such systems. 2.1 Approaches Automatic face recognition divides into roughly two lines of inquiry: feature based approaches which rely on a feature set small in comparison to the number of pixels, and direct image methods which involve no intermediate feature extraction stage. The later method is substantially susceptible to lighting condition. Whereas in principle, feature-based schemes can be made invariant to scale, rotation and/or illumination variations and thus we are interested in them. 2.2 Problem Domains Face recognition system serves two problem domains- verification and identification. When a user is to be authenticated (i.e. the user's identity is to be verified), samples will be captured from the device and again a feature set is created. This feature set is then compared with the enrollment feature set. If the resulting similarity value is above a predefined threshold, the user is considered to be authenticated. In contrast to the verification use case, with identification the (claimed) identity of the user is not known in advance, but shall be determined based on sample images of the user's face and a set (population) of feature sets with known identities. The identification system takes some samples of the user's face, generates a feature set from them and compares this feature set with each element of this population. The elements yielding the highest comparison values above a certain confidence threshold are candidates for the identity wanted. In this paper, we proposed a system that can serve both purposes. 3. PREVIEWS OF FACE IDENTIFIER There are three main parts of the system. They are, Image processing, Recognition algorithm, and Database searching mechanism.