Hand Gesture Recognition Algorithm to Identify American Sign Language

Abstract

Sign language is the most expressive and natural way which is used by hearing and visually impaired people for communication. These people use their hands for communicating with each other, which is not understandable to the normal person. And overall leads the isolation of these peoples from rest of the word. Also, different country has different languages for communication between normal as well as people who have severe hearing impairments. Therefore, the objective of this paper is filling the gap between normal and deaf people through design an algorithm to recognize the different characters of specifically American Sign Language by gesture recognition of the human hand. This proposed algorithm used the information based on the contours and convexity shortcomings to identify the American Sign Language alphabets and digits. The key aim of the proposal is to identify the 17 characters and 6 digits of the American Sign Language without support of any specialized hardware, or any other complex machine learning algorithms. Simulation results validate that, the proposal recognize the American Sign Language alphabets and digits with accuracy more than 85 percent with simple and less complex algorithm within a specified time limits.

 Keywords: Sign Language, American Sign Language, Gesture Recognition, OpenCv, Contours, Convexity.