BIM Sign Language Translator Using Machine Learning (TensorFlow)

Abstract

This project is developed to provide an application service to assist the deaf community where the deaf community in Malaysia has been facing difficulties conveying their messages to the non-deaf community. This project aims to design an offline and stand-alone BIM sign language translator, develop a BIM sign language translator system using TensorFlow machine learning, and conduct acceptance and functionality testing of the prototype to the deaf community non-fluent BIM user. The study uses Waterfall Model as a software development guide and uses TensorFlow, Python, OpenCV, and Qt as a core development library. The working application that translates sign language is then tested on the target users. The testing result shows that the target user can easily use the application to translate their sign language and shows positive indications while using the application. Even though there exist some limitations, such as the lack of a sign language dictionary that the application can translate. Still, it is hoped that this application will benefit the users.