Real time Arabic Sign Language recognition

Abstract

Automated sign language recognition is an auxiliary tool for communication between deaf and vocal people. Additionally, it is one of the main methods of human-computer interface. A sign language is a sequence of gestures, where a gesture is a sequence of postures. There are two approaches to sign language recognition: instrumented glove-based systems and vision-based systems. The Arabic Sign language has not received enough attention as other sign languages such as the American, British, Australian, Chinese and Japanese. This thesis represents a pioneering work on the automation of the Arabic Sign Language recognition using the CyberGlove as an interface device. The principal components analysis is utilized as the feature extraction algorithm to implement a real time system for the automated translation of single handed signs of the Arabic Sign Language Dictionary to a spoken language.