Detection of similarity in batik motifs using backpropagation method

One of the famous batik producing regions in Malaysia is Kelantan. Batik has a variety of motifs and colours, and each batik motifs has its own uniqueness but seems similar, which make them difficult to be identified. This research aims to detect the similarities of floral batik motifs by calculating the accuracy. At the same time, the characteristics of mathematical elements of floral motifs for batik and the relationship between mathematics and culture are identified. There are three types of motifs were selected, which are Bunga Cempaka, Daun Keladi, and Pucuk Rebung. The detection process used is the backpropagation method since this method is well known to be one of the most accurate in recognizing the pattern. The backpropagation algorithm is divided into three phases which are forward and backward propagation phases, changes in the weights, and biases phase. The result for Bunga Cempaka motifs shows the least accuracy, with 36% among the motifs. However, the backpropagation algorithm still can be used well to recognize floral batik motifs with 72% of average accuracy rate. The result also shows that these three motifs contain a reflection transformation through physical observation.