Conversion of Lip Gestures into Text using Machine Learning Model

Abstract

Generally, 15% of the total population are having disabilities of speech and are paralyzed. It is difficult for them to convey their feelings or any other information. So, this paper can create an interface to communicate with others. Also, dumb people communicate using lip movements and hand gestures but it is hard to understand. So, the vocal or lip movements are converted into text by the sensors called as Silent Speech Interface and it has given rise to the possibility of speech processing even in the absence of voice. This paper presents an algorithm that records the lip movement by webcam and the recorded signals are analyzed by extracting features using the Local Binary Pattern (LBP) operator and detected using a Haar Cascaded classifier. This development is done by OpenCV software which can be further implemented by hardware design.