Facial expression recognition algorithm based on equal probability symbolization entropy

Fa Zheng, Bin Hu*, Xiangwei Zheng

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Electroencephalogram (EEG) records brain activity using electrophysiological markers and is a comprehensive representation of the dynamic activity of human brain neurons. EEG can be used to study human facial expression recognition. In fact, entropy values of EEG can fully reflect changes in facial expressions. This paper improves the sample entropy and the permutation entropy by introducing equal probability symbolization and applies the equal probability symbolization entropy to facial expression recognition. The original permutation entropy, sample entropy and equal-probability symbolization entropy values are calculated for the three expressions of anger, fear and happiness. The results demonstrate that equal-probability symbolization entropy can distinguish human facial expressions clearly and accurately.

Original languageEnglish
Title of host publicationComputer Supported Cooperative Work and Social Computing - 13th CCF Conference, ChineseCSCW 2018, Revised Selected Papers
EditorsXiaolan Xie, Yuqing Sun, Tun Lu, Hongfei Fan, Liping Gao
PublisherSpringer Verlag
Pages469-477
Number of pages9
ISBN (Print)9789811330438
DOIs
Publication statusPublished - 2019
Externally publishedYes
Event13th CCF Conference on Computer Supported Cooperative Work and Social Computing, ChineseCSCW 2018 - Guilin, China
Duration: 18 Aug 201819 Aug 2018

Publication series

NameCommunications in Computer and Information Science
Volume917
ISSN (Print)1865-0929

Conference

Conference13th CCF Conference on Computer Supported Cooperative Work and Social Computing, ChineseCSCW 2018
Country/TerritoryChina
CityGuilin
Period18/08/1819/08/18

Keywords

  • EEG signal
  • Equal probability symbolization entropy
  • Facial expression recognition

Fingerprint

Dive into the research topics of 'Facial expression recognition algorithm based on equal probability symbolization entropy'. Together they form a unique fingerprint.

Cite this