Malaysian sign language & alphabets recognition using deep learning algorithm and image processing technique

Abdul Hakim, Muhammad Fauzan (2023) Malaysian sign language & alphabets recognition using deep learning algorithm and image processing technique. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Malaysian sign language & alphabets recognition using deep learning algorithm and image processing technique.pdf - Submitted Version Download (459kB)
	Text (Full Text) Malaysian sign language & alphabets recognition using deep learning algorithm and image processing technique.pdf - Submitted Version Restricted to Repository staff only Download (3MB)

Abstract

People who suffer from hearing difficulties use sign language as a way of communication and sign language translators quickly become inadequate to serve the entire deaf community, especially in Malaysia. To address the problem, this project aims to develop a Malaysian Sign Language recognition algorithm and translate it into text form. To achieve the objective, a dataset of Malaysian Sign Language and alphabets are constructed. Furthermore, image processing techniques to extract specific landmarks were used. The developed algorithm is trained using CNN architecture and PyCharm software is used to perform real-time gesture translation into text form. The algorithm shows a promising result with an accuracy of 96.16%. In addition, the result of precision, recall, and F1-Score for every predicted class is as high as 100%.

Actions (login required)

View Item