SISTEM KLASIFIKASI RELIEF CANDI BOROBUDUR MENGGUNAKAN LOCAL BINARY PATTERN DAN ALGORITMA NEAREST NEIGHBORS

Abstract

Reliefs are sculptures and carvings that are usually made on stone. At Borobudur Temple there are more than 2400 relief panels that are used to tell all the teachings of the Buddha. But it turns out that there are still many people both at home and abroad who still don't know about the story behind the reliefs. Therefore, through the design of the system and the manufacture of the Borobudur Temple Relief Classification System, it is hoped that it can help people who want to know the history of the reliefs that exist in Borobudur Temple. This system is built using the Local Binary Pattern Uniform and Local Binary Pattern Variance methods with point radius 8.1, 16.2, 24.3 and 32.4 as feature extraction methods and while the classification method uses Nearest Neighbors combined with Euclidean, Cosine and Manhattan as the distance calculation method. After testing, the highest accuracy results are 66.33% with the correct classification results as much as 199 data from 300 test data.