Efficient Fingerprint Feature Extraction: Algorithm and Performance Evaluation L.R. Palmer, M.S. Al-Tarawneh, S.S. Dlay and W.L. Woo Newcastle University, United Kingdom Email: {lloyd.palmer, mokhled.al-tarawneh, w.l.woo, s.s.dlay}@ncl.ac.uk Abstract— This paper summarises a simple procedure for extracting fingerprint minutiae from a digital greyscale image. Many approaches were studied and the best selected and implemented. A sequential binarisation approach was chosen implementing Otsu’s method for binarisation, a parallel thinning algorithm and then a crossing number approach for the minutiae extraction steps. The algorithm was implemented in a Matlab environment using images from the FVC2004 database. In order to show the performance of this algorithm results are presented and compared to another similar algorithm found in literature. Introduction In an increasingly digital world, reliable personal authentication has become an important human computer interface activity. Most existing security measures rely on knowledge-based approaches like passwords or token- based approaches such as swipe cards and passports to control access to physical and virtual spaces. Though ubiquitous, such methods are not very secure; they cannot differentiate between an authorised user and a person having access to the tokens or knowledge. Biometric approaches are gaining citizen and government acceptance and provide means of reliable personal authentication that address the above problems. Fingerprints have several advantages over other biometrics, such as the following, high universality, high distinctiveness, high permanence, easy collectable, high performance and wide acceptability. In order to implement a successful Automatic Fingerprint Identification System (AFIS), it is necessary to understand the topology of a fingerprint. The fingerprint surface exhibits a quasiperiodic structure of ridges and valleys that serve as a friction surface for when we grip objects. This structure possesses very rich structural information when examined as an image. The fingerprint images can be represented by both global as well as local features. The global features include the ridge orientation, ridge spacing and singular points such as core and delta. The singular points are very useful from the classification perspective. For robustness purposes our paper concentrates on these local features. There are typically between 60 and 100 minutiae points on the fingerprint. Minutiae are local features characterised by ridge discontinuities. There are about 18 distinct types of minutiae features that include ridge endings, bifurcations, spurs and islands (Fig. 1). However, ridge endings and bifurcations are the most commonly used features in fingerprint recognition systems. A ridge ending occurs when the ridge flow abruptly terminates and a ridge bifurcation is marked by a fork in the ridge flow. Automatic fingerprint identification depends upon the comparison of these local ridge characteristics and their relationships to make a personal identification. A sequential binarisation approach was chosen over a direct greyscale approach as it’s simple in principle and efficient from a design and processing point of view, whereas the direct greyscale approach is generally complex and non-adaptive. A sequential binarisation approach generally consists of the following three stages: Binarisation, Thinning and Minutiae extraction.