Automatic edge-based localization of facial features from images with complex facial expressions

Automatic localization of facial features is an essential step for many systems of face recognition, facial expression classification and intelligent vision-based human-computer interfaces. In this paper, we present automatic edge-based method of locating regions of prominent facial features from up-right facial images. The proposed localization scheme was tested on several public databases of complex facial expressions. The method demonstrated high localization rates when localization accuracy was evaluated by both a conventional point error measure and a new rectangular error measure that takes into account the location of the feature in the image and the true feature size. (C) 2010 Elsevier B.V. All rights reserved.