©2010 International Journal of Computer Applications (0975 – 8887) Volume 1 – No. 2 30 Web Enabled Based Face Recognition Using Partitioned Iterated Function System ABSTRACT Many conventional methods of face recognition depend solely on appearance and model based. However there is inbuilt degree of self-similarity in the image of faces, which can be efficiently utilized through representation exploiting self-transformations, known as Iterated Function System (IFS). Interestingly, virtually all images of natural or man-made objects, show region wise self similarity although they may not be globally self similar. Such objects can be represented by Partitioned Iterated Function System (PIFS) very compactly. Hence, we propose is to carry out the face recognition using Partitioned Iterated Function System. This approach has been tested upon the 106 images of 27 persons using FERET database. The results obtained under the variance, rotation and scaling are outperforming. In this approach we carried out face recognition based on PIFS representation and matching carried out in the PIFS code domain, which is more efficient than correlation in the image domain. The recognition method is efficient in terms of time complexity as the PIFS code of reference faces are built off-line and recognition of query object involves only comparison of its PIFS code with those in the database online. Keywords Alpha, Luminance, Fractals 1. INTRODUCTION In recent years, face recognition has received more and more attention due to its benefit of being a passive, nonintrusive system to verify personal identity in a natural and friendly way. It has many potential application areas ranging from access control, mug shots searching, security monitoring, and surveillance systems. Face recognition is among the most challenging tasks in pattern recognition research due to its scientific challenges and potential applications. There have been a lot of methods proposed for overcoming the difficulty of face recognition. Methods of face recognition can be divided into two approaches namely, feature geometry based and subspace analysis techniques. In feature geometry based approach, recognition is based on the relationship between human facial features such as eye(s), mouth, nose and face boundary. Subspace analysis approach attempts to capture and define the face as a whole. The face is treated as a two-dimensional pattern of intensity variation. The original image representation is highly redundant, and the dimensionality of this representation could be greatly reduced when only the face pattern is of interest. Apart from these two approaches of face recognition In this paper, we consider the indexing problem for a class of images where it is possible to state fairly accurately the notion of a background and a foreground. Our experiments revolve around Given a library of reference images: I1, I2,...,In, and a query image Q, we want to preprocess the reference images to produce indices such that we can find the „closest‟ image Ii to Q.Important subset of this class, namely, photographs of humans (such as those used in corporate identity cards, or those clicked by an automatic teller machine camera). Unlike images generated under structured lighting conditions (such as those of nuts and bolts in factory plants), faces with facial and tonsural hair growth have a predominant texture. Traditional segmentation based techniques do not work well in such cases, and many interesting [2, 12] approaches fail. Fractals are important mathematical entities that have the ability to represent natural unstructured entities such as face, hair, and trees against the background in a photograph. Fractal descriptors are also compact, and therefore, have been used for compression. Indeed, the fractal subdivision method of chopping an image may be viewed as an automatic segmentation algorithm. The biggest impediment in using fractal descriptors for indexing is the one-to- many relationship between an image and fractal descriptors. Many descriptors can converge (using the fractal paradigm, made precise in Section 3) to the same image. In this paper, we study the use of fractal indices for general image indexing, and exemplify it with faces as the domain. Note that no assumption is made of “zeroing background” unlike approaches such as the venerable eigenfaces [11]. The rest of this paper is organized as follows. In the next section, we summarize previous work in this area. In Section 3 we provide the theoretical background for this work. In Section 4 our implementation is discussed along with sample results. Final remarks are made in Section 5 2. RELATED WORK Although many researchers for the purpose of image compression have utilized the Iterated Function System (IFS) structure of fractal object representation, there have been very few attempts directed towards object indexing or recognition. In [4] the authors have presented a somewhat restricted recognition scheme applicable to the specific domain of L-System fractals and tested their technique on binary synthetic plant images generated by the L-System. In [8] a recognition method is suggested which (i) works on binary images, and (ii) which is based on applying the reference Dr.S.G.Bhirud Asst. Prof. Computer Engg ept VJTI Mumbai Amol D.Potgantwar Lecturer Computer Engg Dept SITRC Nashik