Sign Language Detection using Convolutional Neural Network (CNN)

Abstract

Communication is important to express feelings of oneself. Effective communication helps in personal and professional growth. Communicating with a person having some disabilities, such as speech and hearing impairment, is always a major challenge. Deaf And Dumb people cannot communicate with person with no disability because of communication barrier between them and the other person not knowing the sign language. As Per India’s Census 2011, At all India levels, disabled people constitute 2.21% of the total population. In India, about 19% of disabled people have hearing disabilities and 7% in Speech impairment [1]. Sign language gestures are not always enough for communication of people with hearing disability or people with impairment of speech. The gestures/signs made by the people having disabilities often get mixed or difficult to understand for someone who does not understand the language. Thus, we have implemented two models to convert sign gestures to text using Convolutional Neural Network (CNN) in Python and AlexNet in MATLAB.