Klasifikasi Citra Motif Batik di Indonesia Menggunakan Ekstraksi Ciri Completed Local Binary Pattern dan Template Matching

Abstract

Batik has become an identity of Indonesian culture in the era of globalization since UNESCO designated it as an intangible cultural heritage for humanity on October 2, 2009. Indonesian batik has become a highly valued intellectual property recognized both nationally and internationally. Not only is it worn for traditional or religious activities, but batik has also become a trend in various circles from young to old in fashion behavior. Many batik motifs in Indonesia come from various regions, each with its own cultural philosophy. To facilitate the recognition of batik motifs and to preserve the original arts and culture of Indonesia, a classification technique for existing batik motifs is needed. This research uses the completed local binary pattern method for feature extraction of batik motifs. These features will then become reference images/models used to recognize similar objects in the original image using the template matching method. Before the classification stage, batik motif images are processed through the stages of resizing, grayscaling, and image edge detection to detect patterns in the batik motifs. After the testing conducted, it can be concluded that the findings of this research are quite good in classifying batik motif images with an accuracy rate of 92%.