Indonesian Traditional Shadow Puppet Classification using Convolutional Neural Network

Abstract

The reduced interest and knowledge of the Indonesian people about traditional culture has threatened the wayang kulit culture with extinction. Young people have forgotten and do not know the wayang kulit culture. Therefore, computer vision technology is expected to help introduce wayang kulit to young congregations. The technology that utilizes the Convolutional Neural Network (CNN) method is widely used by researchers and allows it to provide high recognition performance. Therefore, this study applies the CNN method with the MobileNet architecture to help recognize shadow puppets, especially character name recognition. The results of the MobileNet architecture on CNN provide a higher recognition performance than using the VGG-16 architecture, which gives an increase of 1.67% become 86,67%.