One form of art passed down by the ancestors of the Indonesian nation is batik, batik in every region in Indonesia has a variety of colors and motifs. The diversity of colors and motifs of batik makes it difficult for many Indonesians to know the type of batik they are wearing. Every batik has a pattern, every pattern has a texture. Texture and color are the distinguishing elements between one batik and another, both are forms of feature extraction that can be used to group batiks that have similar patterns. In this study, a combination of Gray Level Co-Occurrence Matrix, Local Binary Pattern, and HSV Color Moment features was used to obtain texture and color characteristics from batik images, while K-Nearest Neighbor was used to classify batik images. Test results on scenarios using different feature combinations, a combination of features Gray Level Co-Occurrence Matrix, Local Binary Pattern, and HSV Color Moment using 200 batik image datasets consisting of 10 batik classes, obtain the highest accuracy value of 0.29 on the neighbor value K=5, on the other hand, in the test scenario using a different number of classes, the highest accuracy value is obtained when using 5 classes, each class consisting of 10 batik images, the accuracy value is 0.68 at the neighbor value K = 4.
Copyrights © 2023