Finger Spelled Signs in Sign Language Recognition Using Deep Convolutional Neural Network

Authors

  • Reshmi Rajendran Student, Department of Electronics and Communication Engineering, Sree Buddha College of Engineering, Pathanamthitta, India
  • Sangeeta Tulasi Ramachandran Professor, Department of Electronics and Communication Engineering, Sree Buddha College of Engineering, Pathanamthitta, India

Keywords:

accuracy, costlier, external sensors, self-learning, sign language

Abstract

Sign languages are a form of communication by deaf people. A person who knows sign language can easily communicate with them. It is not possible to learn sign language without the help of an expert and through continuous practice. Many versions of sign languages exist in our world. For a normal person to communicate with the deaf, they need to learn sign language that requires interest and guidance. Without continuous practice, learning signs seems difficult. This need triggers many inventions as well as innovations in the sign language learning area. Technology-based tools exist in different forms which depend on external sensors. But most of them are costlier and unaffordable. This paper discusses a GUI-based system that facilitates self-learning of Static American Sign Language (ASL). Web-based Graphical User Interface (GUI) allows operations easier and intuitive. Deep Convolutional Neural Network is used to classify finger-spelled signs that are captured via webcam. The accuracy of the system is also much better than many existing systems.

Downloads

Download data is not yet available.

Downloads

Published

24-06-2021

How to Cite

[1]
R. Rajendran and S. T. Ramachandran, “Finger Spelled Signs in Sign Language Recognition Using Deep Convolutional Neural Network”, IJRESM, vol. 4, no. 6, pp. 249–253, Jun. 2021.

Issue

Section

Articles