Recognition of South Indian Sign Languages for Still Images Using Convolutional Neural Network

Abstract

Recognition of gestures is a growing area of study. Being an integral part of hand gestures for non-verbal communication plays a crucial role in our everyday lives. The Hand Gesture recognition system provides us with a artistic, natural, comprehensible way to connect with the machine that human beings are more familiar with. This paper aims to present a hand gesture recognition based on input image segmentation and hand gesture classification based on the Convolutionary Neural Network by keeping in mind the resemblance of the human hand form with four fingers and one thumb. Here we have suggested a method for understanding the gestures of isolated images of  South Indian Sign languages (Kannada and Telugu) that is a stride towards supporting and refining people with hearing and speech disability. With a Convolutional Neural Network, we have carried out the classification of gestures. In the elements gestures recognition, and it's needed to differentiate the each spatial and worldly developments and this paper proposes a way for dynamic gesture recognition.