An Analytical Study of Signs Used in Baby Sign Language Using Mobilenet Framework

Abstract

A communication via gestures is a method of non-verbal communication which uses body stances and motions pass on the message, thoughts, realities, feeling to the viewers, and in this method of self-expression each body part assumes a significant job is known as Sign Language. Gesture based communication is valuable to not only deaf and dumb community, but also beneficial for individuals suffering from Autism, downs Syndrome, Apraxia of Speech for correspondence. The Baby Sign Language uses gestures which clearly expressing the emotions and desires of toddler and serves as a communication medium between parents and toddlers. In this paper we present research work based on a study of the existing literature on various Sign Languages has been carried out and then prepared a data set of static images for 53 odd baby signs, which were classified using transfer learning technique based on Deep Learning by utilizing a MobileNet, a class of Convolutional Neural Network. This model is further fine-tuned which reduces errors and increases classification accuracy results. A classification accuracy of 85.8% has been achieved on the dataset prepared.