Klasifikasi Citra Sidik Jari Menggunakan Convolutional Neural Network Berdasarkan Tipe Pattern pada Sistem Henry

Abstract

Fingerprint is one of the biometric forms used for identifying a person. Fingerprint has unique ridge arrangement on each human finger and do not change with age. The ridges on fingerprint create distinct and different patterns. The Henry system is a fingerprint identification system that utilizes the patterns on the surface of a finger to verify a person's identity. As a classification method, Henry's classification system is the most widely used for fingerprint classification. There are five types of fingerprint patterns: Arch, Left Loop, Right Loop, Tented Arch, and Whrol. Fingerprint classification is an important part of individual identification system. However, identifying fingerprint manually is difficult to do because the pattern is complicated and this is depends on individual capabilities and also inefficient in terms of time. This research aims to classify fingerprint patterns using a Convolutional Neural Network. The Convolutional Neural Network method is used in this research because it has shown the most significant results in image recognition. The image processing techniques applied in this research include contrast enhancement and thresholding. After testing the data, this method is capable to classify fingerprint images into five fingerprint patterns with 85% accuracy from test dataset containing 75 fingerprint images.