Pengenalan Pola Huruf Hijaiyah dengan Metode Local Binary Pattern Menggunakan Algoritma Fuzzy K-Nearest Neighbor

Abstract

A letter is a form, stroke, or symbol writing system. Any information obtained from a sentence depends on the letters are written clearly. Finding written hijaiyah letters can be recognized by humans, but will be difficult if a computer tries to recognize them. The reason system is difficult is because of the large variety of different letters. This study aims to make it easier for someone to learn to recognize hijaiyah letters by using the Local Binary Pattern method for the feature extraction process. The results of feature extraction will take the maximum value of the histogram of each letter. And results feature extraction will be carried out classification process using the Fuzzy K-Nearest Neighbor algorithm until finally hijaiyah letters can be recognized. Based on experimental results that have been carried out, the highest level of accuracy is obtained when the amount of training data is 154 data and the number of data testing is 29 data, resulting in an accuracy rate of 96.55%.