Indian Sign Language Alphabets Recognition from Static Images Using Correlation-Coefficient Algorithm with Neuro-Fuzzy Approach

Abstract

Sign Language is a language used by mute people to communicate with other person by expressing their thoughts in the form of signs. Signs are expressed by hand and face movement. This paper describes proposed methodology for design and implementation of Alphabets recognition system of Indian sign language where static images of hand gesture are used as an input. Methodology includes various image processing techniques to remove unnecessary noise and to make images smooth. This Paper describe Correlation-coefficient algorithm which has been used for feature extraction and Neuro-fuzzy algorithm which has been applied as recognition algorithm. Proposed technique has been implemented on database of 100 images for testing. Implementation has been done on Matlab. Average accuracy of proposed methodology is 92.30%.

Keywords: Indian Sign Language (ISL), American Sign Language (ASL), Canadian Sign Language (CSL), NF (Neuro-Fuzzy), ANN (Artificial Neural Network)

JEL Classification: Y60