the use of traditional Indonesian textile motifs is influenced by elements of nature, environment and culture that develop in the community. Materials, motifs and techniques for making traditional textiles are different from one region to another. The use of information technology in preservation and providing knowledge to the public can made in the form of software that can match the pattern of images or photos while performing traditional textile classifications. The input in this study is an image of a patterned texture. The software through the pattern recognition process will perform calculations and produce values ??that can be matched in the sample database that has previously been processed. The method used is the color feature extraction with the Local Color Histogram method, texture feature extraction with the Co-occurrence matrix method, and the extraction of shape features with the Canny Edge Detection method. Each feature extraction will produce a vector. The data used consisted of 10 textile groups namely endek bali, songket bali, tenun dayak, tenun ikat, sasirangan, kain besurek, ulos, sutera bugis, kain gringsing. Classification method by using K-Nearest Neighborhood (K-NN). The results of this study are the highest accuracy values ??for extraction of texture features and the lowest accuracy values ??for the combination of extraction of shape and color features.
Copyrights © 2018