Identificação de impressões digitais baseada na extração de minúcias

Abstract

This work is dedicated to the study of dactyloscopy (the identification process of the human based on his fingerprints) applied to the AFIS (automatic fingerprint verification systems). The main objective of the study consists in the development of the personal identifications methods, used a minutiae-based system. We call minutiae the small details extracted in a fingerprint image, which is used to characterize it and match it against other template to obtain correlation level between them. It was made a bibliographic search covering the main methods that were used by the state of art to solve the problem,and a simple solution using the minutiae was chosen. Theses methods were all realized in MatLab codes, with a Graphical Environment for a visual accompaniment and a better understanding of the procedures. Standard database tests proved the validation of the method. It was considered successful, since satisfactory results were achieved using just a few pieces of information extracted from each image.