642 Binarization and Thinning of Fingerprint Images by Pipelining Shadrokh Samavi, Farshad Kheiri, Nader Karimi Department of Electrical and Computer Engineering. Isfahan Univ. of Technology, Isfahan, Iran 84156 Abstract : Two critical steps in fingerprint recognition are binarization and thinning of the image. The need for real time processing motivates us to select local adaptive thresholding approach for the binarization step. We introduce a new hardware for this purpose based on pipeline architecture. We propose a formula for selecting an optimal block size for the thresholding purpose. We also present in this paper a new pipeline structure for implementing the thinning algorithm. Keywords: Fingerprint, Binarization, thinning, Pipeline Processing. 1 INTRODUCTION In today’s life style, security is one of the most important concerns. Technology could provide us with this purpose. Perhaps the first step is to identify a person. Identification cards or simple identification numbers are two examples of this system. By identifying a person more precisely, the system’s security can improve. One of the methods to identify a person is biometrics. It is defined as a science which studies human’s behaviors and physical characteristics, to identify him/her [1]. This identification is done by recognition of face, hand, voice, retina, iris and fingerprint. Although iris recognition is one of the most precise methods, it is not acceptable by all the people as a non-invasive method. It seems that fingerprint recognition can be the next choice. Fingerprint recognition was first employed by Scotland Yard in 1901 as an identification system for the first time [2]. They used Henry-Galton system. This system identifies five types of fingerprint which are shown in Figure 1. The categories that are shown in Figure 1 are arches, tented arches, left loops, right loops, and whorls. Figure.1: Five categories of fingerprints [2]. Later a more precise system, based on minutiae points was introduced. Minutiae points consist of two characteristics, ridge ending and ridge bifurcation. Extracting minutiae points and matching them in two fingerprints are two main tasks in the recognition process. To extract minutiae, a fingerprint image should have 40 to 100 minutiae [2]. To match two fingerprints, between 10 to 16 minutiae should match. The variation is related to different human races [2]. 3rd MVIP, Feb. 2005, Vol. 2 University of Tehran, Tehran, Iran