A novel Krawtchouk moment zonal feature descriptor for user-independent static hand gesture recognition
Subhamoy Chatterjee, Piyush Bhandari, Maheshkumar H. Kolekar · 2016
A real-time static hand gesture recognition system involves user-independent interpretation of hand shapes by a computer. This work focuses on three main issues in developing a user-independent static hand gesture recognition system. They are 1) Introducing some well-known static zoning methods to improve user-independent gesture recognition 2) Introducing a novel Krawtchouk moment zonal feature for improved gesture recognition 3) Conducting two experiments; two-fold and tenfold, to test the user-independent gesture recognition performance of our proposed system. A comparative study of gesture recognition using global Krawtchouk moment, uniform and nonuniform zoning based Krawtchouk moment and our proposed novel Krawtchouk moment zonal feature has been reported in this work. Our proposed feature extraction method has shown gesture recognition accuracy of 96.40% and 98.67% for two-fold and ten-fold experiments respectively, which are better than the other used methods.