VGG16 in Batik Classification based on Random Forest

摘要

Batik is one of cultural heritage in Indonesia which has been recognized by UNESCO. It has a unique pattern related to such region in Indonesia. Since Indonesia consist of various culture, a system for preserving culture is highly needed. To accommodate this issue, in this study, we propose a deep learning method which is used to classify Batik types. Our method is based on VGG16 architecture with a Random Forest as the classifier. We test our method using public Batik dataset which is widely used in the previous study. We measure the precision, recall, F-score, and accuracy to evaluate our proposed method performance. The result shows that our proposed method out-performed the existing method which is based on color and texture features. The precision, recall, F-score, and accuracy reach ± 97%.