Deep Learning dalam Mengindetifikasi Jenis Bangunan Heritage dengan Algoritma Convolutional Neural Network

Abstract

A heritage building is a building that has a distinctive style or tradition from a culture whose activities are carried out continuously until now and are used as a characteristic of that culture. The problems that occur in the community are the lack of knowledge to recognize the types of heritage buildings and the lack of digital documentation. Another problem that occurs in identifying heritage buildings is that there are similarities between heritage buildings and new buildings that imitate the architectural style of heritage buildings from ornaments. This can raise doubts in the information related to the original history of heritage buildings for the public or visitors. This study aims to apply the Convolutional Neural Network (CNN) to identify the types of heritage buildings. The benefits of this research can be found in the characteristics of a building based on ornaments so that it can be used to obtain information about the types of heritage buildings in Indonesia. A dataset of 7184 images of ornaments from heritage buildings were used which were taken directly at the Yogyakarta location, namely; Mataram Grand Mosque, Taqwa Wonokromo Mosque, Kalang House, Joglo KH Ahmad Dahlan and Ketandan. It is necessary to identify the heritage building because the object of the building can become extinct at any time, so to maintain it, documentation is needed as an effort to preserve culture and for education. Based on the evaluation of the performance of the tests carried out using the confusion matrix method from 391 ornamental images, the results obtained are 98% accuracy