Title: Recognition of the Tamil sign alphabet using image processing technique with angular-based analysis of left hand to aid deaf-dumb people
Authors: P. Subha Rajam; G. Balakrishnan
Addresses: Department of Information Technology, J.J. College of Engineering and Technology, Tiruchirappalli 620 009, Tamilnadu, India. ' Indra Ganesan College of Engineering, Tiruchirappalli 620 012, Tamilnadu, India
Abstract: Human-computer interaction is an important research work for enabling convenient communication between deaf-dumb and normal people. This paper introduces a Tamil sign recognition system for one of south Indian languages. Tamil alphabets have 12 vowels, 18 consonants and one Ayutha Ezhuthu. This proposed method introduces a set of 31 sign combination of binary images. Each image representing five finger positions in the form of binary 'UP' and 'DOWN' type. The images represent the left hand palm side which is to be taken ten times at different distances as static 320 images. Fingertip position identification of 'UP' or 'DOWN' is identified by using efficient and fast algorithm image processing techniques. Then a binary type of fingertip position is converted into corresponding Tamil letters. The experiments revealed that the system is able to recognise Tamil sign language with 98.44% better accuracy.
Keywords: Tamil sign language; sign language recognition systems; image processing; pattern recognition; binary conversion; Tamil letters; human-computer interaction; HCI; deaf and dumb people.
International Journal of Information and Communication Technology, 2012 Vol.4 No.1, pp.76 - 88
Received: 08 May 2021
Accepted: 12 May 2021
Published online: 27 Feb 2012 *