Klasifikasi Sidik Jari Menggunakan Support Vector Machine Dengan Feature Berbasis Minutiae Menggunakan Metode Region

Abstract

Manual fingerprint classification proceeds by carefully inspecting the geometric characteristics of major ridge curves in a fingerprint image. This research proposed an automatic approach of identifying the geometric characteristics of ridges based on minutiae position and angle. Position and angle of minutiae are analyzed using region method. Region was used to discretize the number of minutiae. Positions of minutiae in a region give information about its relation to another region. Angles of minutiae indicate the ridge flow direction. Support vector machine, a binary classifier, is used to classify the fingerprint based on those characteristics. In this research, the classes are left loop, right loop, whorl, arch and tented arch. This research used 2695 fingerprint images from NIST. Classification performance is measured using 3-fold cross validation method. This research achieved 79.7 % accuracy. As we only used local features in this research, further research needs to be conducted especially to investigate global features.