cqu_5853+ATTACHMENT02+ATTACHMENT02.4.pdf (398.86 kB)
Download file

Selection and fusion of facial features for face recognition

Download (398.86 kB)
journal contribution
posted on 06.12.2017, 00:00 by Xiaolong FanXiaolong Fan, Brijesh VermaBrijesh Verma
This paper proposes and investigates a facial feature selection and fusion technique for improving the classification accuracy of face recognition systems. The proposed technique is novel in terms of feature selection and fusion processes. It incorporates neural networks and genetic algorithms for the selection and classification of facial features. The proposed technique is evaluated by using the separate facial region features and the combined features. The combined features outperform the separate facial region features in the experimental investigation. A comprehensive comparison with other existing face recognition techniques on FERET benchmark database is included in this paper. The proposed technique has produced 94% classification accuracy, which is a significant improvement and best classification accuracy among the published results in the literature.

History

Volume

36

Issue

3P2

Start Page

7157

End Page

7169

Number of Pages

13

ISSN

0957-4174

Location

UK

Publisher

Elsevier

Language

en-aus

Peer Reviewed

Yes

Open Access

No

External Author Affiliations

Faculty of Business and Informatics; Institute for Resource Industries and Sustainability (IRIS);

Era Eligible

Yes

Journal

Expert systems with applications.