Static Hand Gesture Based Human Computer Interaction Using Convolution Neural Network

Abstract

Human computer interaction is a hot topic among researchers primarily due to its application in healthcare, intelligent homes and helping people with disabilities. Through this paper we aim to propose a solution to bridge the gap between the computer and human. The proposed solution uses static gestures selected from the American sign language (ASL) and uses artificial neural network(Alex net). We were able to achieve an accuracy of 95% with our dataset of more than 30,000 images split into train and test images in the ratio 20:1.