An Improved Facial Expression Recognition Algorithm

Authors

  • Yilihamu Yaermaimaiti School of Electrical Engineering, Xinjiang University, Urumqi 830047, China

Abstract

The recognition of facial expression images is susceptible to non-uniform illumination factors, whichmay reduce the recognition rate. In view of this, an
improved facial expression recognition algorithm is proposed. Firstly, the pattern-oriented edge magnitudes (POEM) histogram of the corresponding
facial expression image is obtained through calculating the characteristic quantity of the facial expression image by the POEM. The histogram is
created as the POEM texture histogram of the central characteristic point and the texture characteristic information of the facial expression feature
points are obtained Secondly, the improved incremental non-negative matrix factorization (IINMF) algorithm is used to train the category information
of face image samples to extract the face image representation vector. Canonical correlation analysis (CCA) is then used to combine the characteristic
information of the POEM texture histogram and the eigenvector of the facial expression image extracted by IINMF to obtain the syncretic eigenvector
of the facial expression image. Finally, the nearest neighbor classifier is used to classify and obtain the final identification result. The experimental
results show that the proposed algorithm has a high recognition rate for facial expression recognition under non-uniform illumination and has excellent
robustness and real-time results

Keywords: Facial expression; non-uniform illumination; POEM; IINMF; CCA.

Cite As

Y. Yaermaimaiti, “An Improved Facial Expression Recognition Algorithmâ€, Engineering Intelligent Systems, vol. 28 no. 2,
pp. 125-130, 2020.




Downloads

Published

2020-06-01

Issue

Vol. 28 No. 2 (2020): International Journal of Engineering Intelligent Systems

Section

General Submission

License

The submission of a paper implies that, if accepted for publication, it will not be published elsewhere in the same form, in any language, without the prior consent of the publisher. Before publication, authors are requested to assign copyright to CRL Publishing Ltd. This allows CRL to sanction photocopying, and to authorize the reprinting of issues or volumes according to demand. Authors' traditional rights will not be jeopardized by assigning Copyright in this way, as they retain the right to reuse the material following publication, and to veto third-party publication.