PENGENALAN POLA MOTIF KAIN TENUN GRINGSING MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK DENGAN MODEL ARSITEKTUR ALEXNET

Abstract

Gringsing is one of the traditional fabrics that are characteristic of the Tenganan Pegringsingan Village. Gringsing is quite unique because in its manufacture it uses double-ikat techniques, where this technique can only be found in three places in the world, such as Japan, India, and Tenganan Pegringsingan Village. In 2016 Gringsing was certified by the Ministry of Law and Human Rights of the Republic of Indonesia as a Geographical Indication. This study aims to building a deep learning model to recognize Gringsing motifs and know the performance of the model, so that people can more easily recognize Gringsing motives without having special abilities. The model was built using the Convolutional Neural Network (CNN) method with the AlexNet architectural model. Tests are conducted to determine the performance of the model such as training time, accuracy, precision, recall, and f-measure value. Based on the test results the model built was able to complete 100 epoch training with a time of 19,33 hours, and has an accuracy value of 76%, 74.1% of precision, 72.3% of recall, and 0.73 of F-measure.