Open AccessPengenalan Citra Huruf Hijaiah Menggunakan Metode Gray Level Co-Occurrence Matrices (Glcm) Dengan 4 Sudut Orientasi Dan Jaringan Syaraf Tiruan Backpropagation
Author(s) -
Muhlis Fathurrahman,
Ramaditia Dwiyansaputra
Publication year - 2021
Publication title -
jurnal teknologi informasi, komputer dan aplikasinya
Language(s) - English
Resource type - Journals
ISSN - 2657-0327
DOI - 10.29303/jtika.v3i1.140
Subject(s) - sentence , arabic , natural language processing , computer science , artificial intelligence , alphabet , backpropagation , gray level , speech recognition , artificial neural network , sri lanka , gray (unit) , pattern recognition (psychology) , linguistics , image (mathematics) , history , philosophy , medicine , ethnology , south asia , radiology
Arabic is one of the international languages according to the United Nations (UN) which was adopted by General Council resolution 3190 (XXVIII) as the official language and working language of the General Council and Main Committees on 18 December 1973. Arabic can be found in the holy book Al - Qur'an. For a Muslim, it is obligatory to learn and master Arabic in order to read and understand the contents of the Al-Qur'an. job applicant from Indonesia is also have to learn Arabic. The Hijaiiyah letter has the same role as the alphabet, which is to compose every word and sentence. Humans have a natural intelligence to be able to recognize each Hijaiiyah letter based on the special characteristics or patterns contained in each letter. However, natural intelligence has deficiencies such as inconsistencies in assessing the similarity of each handwritten Hijaiiyah letter from different people. Therefore this research will develop a system for identifying or recognizing Hijaiiyah handwritten patterns using the Gray Level Co-occurrence Matrices (GLCM) method with 4 orientation angles and Backpropagation Artificial Neural Network (ANN). Data was collected using the Autodesk Sketchbook application so that can reduce the noise. The purpose of this research is to know the level of accuracy and precision of the classification of the Hijaiiyah letter pattern. In this research, the amount of data used was 1500 images of Hijaiiyah letters. The highest accuracy is 45.1111% with a precision 45.1111%.