Sample group and misplaced atom dictionary learning for face recognition

Authors

MENG WANG
ZHENGPING HU
ZHE SUN
MEI ZHU
MEI SUN

DOI

10.3906/elk-1702-49

Abstract

Latest research results have demonstrated the effectiveness of both sparse (or collaborative) representation and dictionary learning for problem solving in face recognition and other signal classification. Considering the fact that an informative dictionary helps a lot in sparse coding, a novel model that consists of group dictionary learning and high-quality joint kernel collaborative representation was proposed in this paper, where rich information from original and virtual space was mined and constructed as a sample group space to improve classification accuracy. Meanwhile, joint kernel collaborative representation with an $\ell_{2}$-regularization-based classifier was used to capture more nonlinear structure and minimize the time cost. Experiments showed that the proposed method outperformed several similar state-of-the-art methods in terms of accuracy and computational complexity.

Keywords

Sample group, misplaced atom, dictionary learning, joint kernel collaborative representation, face recognition

First Page

4421

Last Page

4430

Recommended Citation

WANG, MENG; HU, ZHENGPING; SUN, ZHE; ZHU, MEI; and SUN, MEI (2017) "Sample group and misplaced atom dictionary learning for face recognition," Turkish Journal of Electrical Engineering and Computer Sciences: Vol. 25: No. 5, Article 74. https://doi.org/10.3906/elk-1702-49
Available at: https://journals.tubitak.gov.tr/elektrik/vol25/iss5/74