Enhanced Hand Gesture Recognition System

Abstract

This paper implements a hand gesture recognition technique by using the SIFT based feature extraction.The matching point threshold has been calculated by using the neural network on the basis of the input gesture and the training.The paper implements the work by using the MATLAB and analyzes the results on the various sign of ASL.The system shows the accuracy of 93.3% i.e. improved by approx.&% as compared to the existing 85.9%.The system discussed in the paper is also robust as it shows the accurate results on the manipulated gestures.