Singular Regions Detection in Fingerprint Images

Singular regions detection from a fingerprint is a laborious process which required the completion of various stages starting from a set of fingerprint images obtained through simulation with the help of the SFinGe program, which offered the possibility to generate fingerprints with different features that could be selected (form, type and position of singular points, noises, and so on). A prior stage of the proper algorithm for singular regions detection is to enhance the initial image through a series of morphological transformations of this, but which was not detailed in this paper. The algorithm starts with the directional field determination, continuing with the detection of singular points or singular regions. Introducing the possibility to alter a fingerprint by simulation through the central rotation, an algorithm was developed to determine if a fingerprint is natural or altered. This new method for singular regions detection was tested on natural and altered fingerprints. The importance of this paper is given by the fact that nowadays, many delinquents or immigrants alter their fingertips not to be recognized by the police. The experimental results, as will be shown, turned that the proposed method can detect accurately singular regions and based on this, the difference between natural and altered fingerprints can be made.