Intelligent sign language recognition using enhanced fourier descriptor: A case of Hausa sign language

Abstract

Hausa sign language (HSL) is the main communication medium among deaf-mute Hausas in northern Nigeria. HSL is so unique that a deaf-mute individual from other part of the country can rarely understand it. HSL includes static and dynamic hand gesture recognitions. In this paper we present an intelligent recognition of static, manual and nonmanual HSL using an enhanced Fourier descriptor. A Red Green Blue (RGB) digital camera was used for image acquisition and Fourier descriptor was used for features extraction. The features extracted chosen manually and fed into artificial neural network (ANN) which was used for classification. Thereafter particle swarm optimization algorithm (PSO) was used to optimize the features based on their fitness in order to obtain high recognition accuracy. The optimized features selected gave a higher recognition accuracy of 90.5% compared to the manually selected features that gave 74.8% accuracy. High average recognition accuracy was achieved; hence, intelligent recognition of HSL was successful.