Open AccessRegion based Facial Expression Recognition using Gradient Directions
Author(s) -
Vinola C.*,
V. Lakshmi Devi,
K. Valarmathi,
V. Manjula,
S. Muthurajkumar
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.l3427.029420
Subject(s) - facial expression recognition , emotion recognition , facial expression , artificial intelligence , computer science , matching (statistics) , pattern recognition (psychology) , field (mathematics) , expression (computer science) , three dimensional face recognition , computer vision , facial recognition system , subject (documents) , speech recognition , mathematics , face detection , statistics , pure mathematics , programming language , library science
Facial Expression Recognition (FER) has gained significant importance in the research field of Affective Computing in different extents. As a part of the different dimensional thinking, aiming at improving the accuracy of the recognition system and reducing the computational load, region based FER is proposed in this paper. The system is an emotion identifying system among the basic emotions, through subject independent template matching based on gradient directions. The model designed is tested on the Enhanced Cohn-Kanade (CK+) dataset. Another important contribution of the work is using only eye (including eyebrows and the nose portion near eyes) and mouth regions in the emotion recognition. The emotion classification result is 94.3% (CK+ dataset) for 6-class FER.