Classification of stamps and handmade batik based on pattern recognition

Abstract

Batik is a traditional cloth made by painting motifs with Indonesian culture nuance on plain cloth. It has always been used by the people, from their birth until marriage, even someone who died using batik cloth as a cover. Batik cloth made with philosophy and symbolic meaning. Every region in Indonesia has a characteristic according to the circumstances and character of the man himself. There are three ways to make batik, such as using stamps, handmade, printing. Every type of batik has different types of patterns, and there is a repetition of a pattern and a combination of several patterns. Based on the type of batik, handmade batik has the highest price, and printing batik has the lowest prices. For batik lovers, handmade batik would want to have, but not everyone knows how to characterize the stamps batik and often mistaken in the detection of the type of batik. Thus, only a few people can distinguish the differences between the three types of batik. It is difficult for ordinary people to recognize a batik pattern. The purpose of this research is to create a prototype system that can detect the patterns of batik motifs created by handmade or stamps. Cosine similarity is used in this research to recognize the pattern and classified handmade or stamps batik. The results of this research is 95% in detecting handmade and stamp batik, then 91% in detecting sum of pattern in stamps batik.