Deep Learning for Batik's Pattern Classification

Batik has been recognize by UNESCO as one of cultural heritage in Indonesia. The preservation of batik is needed since it has rich pattern and can be seen as a reflection of region in Indonesia. One of solution is building an automatic system to help people recogize the type of batik. In our previous research, we applied a transfer learning using VGG16 followed by a Random Forest classifier to recognize Batik's pattern.

Fig. 1. Our diagram of our proposed method in [1].

The feature is firstly extracted using a pretrained VGG16. Then, we put this feature as an input on Random Forest. We compared the Random Forest with Support Vector Machine. The result showed that our method performed better in term of Precision, Recall, F-score, and Accuracy.

This is still an open issue because the dataset we used may be not covered all Batik's pattern. We can create an app which can classify all Batik's pattern in Indonesia.

Ref:

[1] D. M. S. Arsa and A. A. N. H. Susila, "VGG16 in Batik Classification based on Random Forest," 2019 International Conference on Information Management and Technology (ICIMTech), Jakarta/Bali, Indonesia, 2019, pp. 295-299, doi: 10.1109/ICIMTech.2019.8843844.