AUTHORS:
Tomas Chrpa, Pavel Hrubes, Pavel Paclik, Jiri Prchal, Pavel Ptak, Lukas
Vydra,
PUBLISHED:
KEYWORDS:
ABSTRACT:
Face profile may serve as an important information source for human
emotion recognition. In this report, a novel approach for profile segmentation
is presented. Changes of the face profile under various emotions are measured
from real images and discussed in order to provide a knowledge base for
further feature selection process. Movements of basic face parts like lips
and chin are considered to be of the crucial importance for emotion recognition.
The whole algorithm of face profile segmentation was implemented under
Matlab and ISFER framework.