Utilizing Facial Landmarks in Speech Research: A Review

Abstract

Facial landmarks are specific points on the face, such as the contours of the eyes and lips. Movements of these landmarks are closely related to the articulatory movements and acoustic features of speech, providing useful information for speech research. Recent advances in image processing technology have made it possible to easily identify the positions of these facial landmarks on facial images using public domain machine learning libraries. This paper reviews speech research using the facial landmarks and describes the process of extracting facial landmark coordinates from facial images with tools that can be used on web browsers.