Klasifikasi Bahasa Isyarat Indonesia (BISINDO) Menggunakan Convolutional Neural Network dengan Arsitektur VGG16

Abstract

In Indonesia there are two sign languages, namely Sistem Isyarat Bahasa Indonesia (SIBI) and Bahasa Isyarat Indonesia (BISINDO), where SIBI is the official sign language taught in special schools, but most deaf people in Indonesia use BISINDO to communicate with each other. To overcome confusion due to language differences between the deaf and hard of hearing and the general public, we use a Convolutional Neural Network (CNN) architecture to translate BISINDO. In this research we got the results of the model performance reaching accuracy values ​​(96,07%). With these results, this model can be used to develop a computer vision system that allows BISINDO translation directly using a camera.