Facial Expression Recognition Based on Multi-Feature Fusion and HOSVD

To the problem of low recognition rate for human-independent facial expression, this paper proposed a facial expression recognition algorithm based on geometric and texture fusion features and HOSVD(High-Order Singular Value Decomposition). The algorithm transforms the facial expression recognition...

Full description

Saved in:
Bibliographic Details
Published in2019 IEEE 3rd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) pp. 638 - 643
Main Authors He, Ying, He, Xiaoju
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.03.2019
Subjects
expression feature extraction
expression recognition
Eyebrows
Face
Face recognition
Feature extraction
high-order singular value decomposition
Mouth
Nose
tensor analysis
Online AccessGet full text

Cover

More Information
Summary:To the problem of low recognition rate for human-independent facial expression, this paper proposed a facial expression recognition algorithm based on geometric and texture fusion features and HOSVD(High-Order Singular Value Decomposition). The algorithm transforms the facial expression recognition problem into the tensor domain, and extracts human-independent expression features using HOSVD. Then the interference caused by individual face differences on expression recognition is effectively excluded. The algorithm was tested on the Japanese Female Facial Expression database, and the results showed that the method achieved better recognition rate in human-independent experiments.
DOI:10.1109/ITNEC.2019.8729003