Myanmar Sign Language Classification using Deep Learning

Abstract

Nowadays, the development progress based on Human Computer Interaction (HCI) becomes more and more wider. In the field of Information Technology, HCI based Sign Language Classification System can be used to develop a day to day life of deaf people. There are very little research works in Myanmar sign language recognition and it is still needed to develop. This paper studies component detection using skin color detection and component localization based on Region of Interest (ROI) and sign classification using Convolutional Neural Network (CNN). The main focus of this work is to create a vision-based Myanmar Sign Language Recognition System. Myanmar Sign Language is captured by a webcam as an image and the system is implemented by MATLAB.