Semantic image search system of madurese cultural batik with impression, color and shape features

Batik is one of Indonesian cultural products in the form of traditional fabrics which has many motifs. One of famous batik is madurese batik. Here, Madura is one of Indonesia islands which has a distinctive motif in terms of colors and motif characteristic that distinguishes it from other regions in Indonesia, but the Madurese communities themselves have not understood the meaning of Madurese batik motif. Therefore, to identify Madurese batik motif from the meaning and origin; it was designed a program that could help the younger generation to know the meaning of Madurese batik motif. Also, it provided traction and preserve to Madurese batik. The program was used to identify the meaning of Madurese batik motif by using some references on Madurese batik and from the experts that understand well on Madurese batik. We used the 3D feature extraction method-CVQ to extract color feature, and Hu Moment Invariant for feature extraction forms. The Madurese batik image searching used the image of madurese batik motif and based on its meaning. In this study, to analyze the performance, we conducted an experiment using image features like query that displayed the results in the form of impressions, and multi-image. While the query displayed the results in the form of images that have the same impression to the closest query. The experiments used color features as the query image which showed the performance of precision was 46.72%, when using the image form as query, it showed performance of precision about 80%. The last, the experiments using the similarity measurements showed multi impression as query performance was 48.3% precision as comparison to the color metric.