IDENTIFIKASI KAIN TAPIS LAMPUNG MENGGUNAKAN EKSTRAKSI FITUR EDGE DETECTION (CANNY) DAN KLASIFIKASI PROBABILITY NEURAL NETWORK (PNN)

Abstract

Tapis fabric is a Lampung tribe women's clothing in the form of a sarong made of woven cotton threads with a sugi motif or decoration, silver thread or gold thread with an embroidery system. Lampung tapis is the result of woven cotton threads with motifs, silver threads or gold threads and becomes the typical clothing of the Lampung tribe. Tapis fabric can be distinguished by shape and pattern, each type has its own special characteristics. This study aims to identify the tapis fabric using Edge Detection feature extraction and Probability Neural Network (PNN) classification. Experiments were carried out using 525 images data, 450 images became training data, while the other 75 images became test data consisting of 3 types of filters Bintang Perak, Gunung Beradu, and Sasab. The results of the experiment are quite good. The smoothing value applied to the PNN has an effect on the accuracy.