Arab Sign language Recognition with Convolutional Neural Networks

Abstract

The implementation of an automatic recognition system for Arab sign language (ArSL) has a major social and humanitarian impact. With the growth of the deaf-dump community, such a system will help in integrating those people and enjoy a normal life. Like other languages, Arab sign language has many details and diverse characteristics that need a powerful tool to treat it. In this work, we propose a new system based on the convolutional neural networks, fed with a real dataset, this system will recognize automatically numbers and letters of Arab sign language. To validate our system, we have done a comparative study that shows the effectiveness and robustness of our proposed method compared to traditional approaches based on k-nearest neighbors (KNN) and support vector machines (SVM).