Pengenalan Sidik Jari Menggunakan Algoritma Pencocokan Adaptif Berdasarkan Penjajaran Minutiae

Abstract

In fingerprint matching contain several method that can used for fingerprint recognition system.One of them is minutiae based matching. This method utilising number and location of the minutiae that different on every body so that it could be distinguishing between one person and the other. However, in this method contain two main problem that is image deformation and inexcact minutiae localisation. In this final project is made a fingerprint recognition program based on minutiae matching. Program is made in three main step, that are image enhancement, minutiae extraction, and minutiae matching. Image enhancement consist of segmentation, normalisation, orientation estimation, ridge frequency estimation, filtering, binerization, and thinning process. Minutiae extraction consist of minutiae detection, eliminating false minutiae, and unification of ending and bifurcation. Minutiae matching consist of alignment step and matching step. In minutiae matching process a boundary with two types that is fixed size and adaptive boundary are used as the reference to determines a pair of minutiae is match or not. Determining of two fingerprint images is match or not based on percentage of number minutiae match from those images.Testing is done using 165 fingerprint images from eleven respondents. Testing result shows average of percentage of recognize program that using fixed size boundary is 67,88% with the maximum value is 100% and minimum is 40%, whereas the program that using adaptive boundary has average of percentage of recognize as big as 92,12% with maximum value is 100% and minimum value is 86,67%. Thereby, usage of adaptive boundary make the system more robust to deformation and inexact localisation problem. Keywords: minutiae, boundary, alignment, minutiae matching