ARAÚJO, A. C.; ARAÚJO, Arthur Cruz de.
Abstract:
This paper reports the activities developed by Arthur Cruz de Araújo, an Electrical
Engineering undergraduate student at Federal University of Campina Grande, during
his integrated 660 hours long internship at Vsoft Tecnologia. The company has been
specializing in the development of software related to digital image processing, focusing
on face and fingerprint biometrics. The classification of digital fingerprints can be defined
as the problem of assigning a given fingerprint to a class in a consistent and reliable
manner. The main focus during this internship was the implementation of a structural
algorithm for fingerprint classification based on inexact graph matching. The developed
activities correspond to the algorithm’s subsystems: it passes through the generation of
a fingerprint’s orientation image, then moves on to its segmentation into regular regions
and the posterior creation of a representative graph. Prototype graphs for the classes of
interest are created and, using them, the inexact graph matching is performed, generating
a measure of distance between the graphs which would allow classification to happen. The
activities were developed in C++ language, using the computer vision library OpenCV.