P. Foggia, C. Sansone, and M. Vento (Eds.): ICIAP 2009, LNCS 5716, pp. 739–747, 2009. © Springer-Verlag Berlin Heidelberg 2009 Reconnecting Broken Ridges in Fingerprint Images Nadia Brancati, Maria Frucci, and Gabriella Sanniti di Baja Institute of Cybernetics “E. Caianiello”, CNR, Pozzuoli (Naples), Italy n.brancati@cib.na.cnr.it, m.frucci@cib.na.cnr.it, g.sannitidibaja@cib.na.cnr.it Abstract. In this paper, we present a new method for reconnecting broken ridges in fingerprint images. The method is based on the use of a discrete direc- tional mask and on the standard deviation of the gray-levels to determine ridge direction. The obtained direction map is smoothed by counting the occurrences of the directions in a sufficiently large window. The fingerprint image is, then, binarized and thinned. Linking paths to connect broken ridges are generated by using a morphological transformation to guide the process. 1 Introduction Fingerprints are widely used in biometric techniques for automatic personal identifi- cation, though a number of other techniques exist, which involve other biometric features such as face, iris, ear and voice. In fact, fingerprints of any individual are unique (even in the case of identical twins), remain the same over lifetime, and are easy to collect. Fingerprints were initially introduced for criminal investigation, and their verifica- tion used to be performed manually by experts. Nowadays fingerprints are still used to identify suspects and victims of crimes, but are also involved in an increasing num- ber of applications, such as physical access control, employee identification, and information systems security. Moreover, the tedious manual matching work has been replaced by automatic fingerprint identification systems, which can work with data- bases including even several millions records of fingerprint images. Most of the fin- gerprint matching algorithms follow the forensic procedure of matching particular features in a fingerprint image, called minutiae. The minutiae are local discontinuities of ridgelines in a fingerprint image. Though minutiae could be classified in several classes, it is standard practice to use a classification in two minutiae only, namely termination and bifurcation. Fingerprints can be captured by different devices, each of which may produce a corrupted image. For automatic fingerprint identification, the quality of fingerprint images is of great importance, since low quality images severely affect the detection of the minutiae and, accordingly, a correct identification. For this reason, fingerprint images generally undergo a number of different processes, aimed at enhancement, computation of the ridge direction, thinning, and feature extraction. Enhancement largely contributes to the robustness of a system for fingerprint veri- fication/recognition and is a topic that has received much attention. In this paper, we focus on the problem of reconnecting broken ridges. We suggest a method based on