HAND GESTURE RECOGNITION USING ORIENTATION HISTOGRAM AND BACKPROPAGATION NEURAL NETWORK

Abstract

This paper presents to recognize American Sign Language (ASL) hand gestures, based on a pattern recognition technique by using orientation histograms and backpropagation neural network (BPNN). The static ASL digitized images of English alphabets are used in this hand gesture recognition system. In this system orientation histogram is used as feature vectors because of its robustness in lighting changes conditions of images and also the position of the hand within the image should not affect the feature vector. The advantage of using BPNN is that it can perform a particular function by adjusting the values of the connections between elements, so input feature vector leads to specific target output. This system consists of Image Processing, Training and Testing phases in BPNN. The output of the system will be displayed the corresponding alphabet letter with corresponding input feature vectors. This system will be beneficial between the deaf and hearing communities problems.