Published Paper Details:

Sign Language is one among the media of communication for deaf people. One should learn signs to interact with them. Learning usually takes place in peer groups. There exist very few study materials for sign learning. Because of this, the method of learning signing learning may be a difficult task. Fingerspelled sign learning is the initial stage of sign learning and moreover, are used when no corresponding sign exists or signer is not aware of it. Most of the prevailing tools for sign learning use external sensors which are costly. This paper discusses SignQuiz, which can be an economical web-based fingerspelled sign learning application for Indian signing (ISL) utilizing automatic signing recognition technique. It works in two modes, learning and testing. In learning mode, signs are listed and one can learn the signs by clicking on the required ones. In testing mode, the user is tested for the learned signs. SignQuiz helps to find out signs with none external help. This is the first attempt in ISL for learning finger spelled signs using a deep neural network. The results indicate that SignQuiz is best than the printed medium for fingerspelled sign learning