Geometric Feature-Based Face Normalization for Facial Expression Recognition

In this paper, we propose a robust facial expression recognition approach using ASM (Active Shape Model) based face normalization and embedded hidden Markov model (EHMM). Since the face region generally varies as different emotion states, the face alignment procedure is a vital step for successful f...

Full description

Saved in:
Bibliographic Details
Published in2014 2nd International Conference on Artificial Intelligence, Modelling and Simulation pp. 172 - 175
Main Authors Dong-Ju Kim, Myoung-Kyu Sohn, Hyunduk Kim, Nuri Ryu
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.11.2014
Subjects
Discrete cosine transforms
EHMM
Emotion recognition
Face
Face recognition
Facial Expression Recognition
Facial features
Hidden Markov models
Online AccessGet full text

Cover

More Information
Summary:In this paper, we propose a robust facial expression recognition approach using ASM (Active Shape Model) based face normalization and embedded hidden Markov model (EHMM). Since the face region generally varies as different emotion states, the face alignment procedure is a vital step for successful facial expression recognition. Thus, we first propose ASM-based facial region acquisition method for performance improvement. In addition, we also introduce the EHMM-based recognition method using two-dimensional discrete cosine transform (2D-DCT) feature vector. Here, we apply large window size during feature extraction of 2D-DCT. The reason is that the facial feature of large window size will represent better facial expression characteristic than that of small window size. The performance evaluation of proposed method was performed with the CK facial expression database and the JAFFE database, and the proposed ASM-based method showed average performance improvements of 7.9% and 5.3% compared to eye-based method for CK database and JAFFE database, respectively.
DOI:10.1109/AIMS.2014.52