Face Recognition Based on the Key Points of High-dimensional Feature and Triplet Loss

DOI

10.2991/icaset-17.2017.16How to use a DOI?

Keywords

Face alignment, High-dimensional feature, Multiscale, Triplet Loss

Abstract

Face recognition has been a hot issue in the flied of computer vision, and face recognition is increasingly applied in the actual life. However, some low dimensional features such as Gabor, LBP, SIFT couldn't achieve a good performance of face feature presentation. So an algorithm which based on the key points of high-dimensional feature is proposed. The extracted feature is transformed by Triplet Loss. The proposed algorithm firstly implement face alignment, and then extract multiscale feature. When high-dimensional features are presented, it need to be transformed by triplet loss matrix. The paper use LBP as a basic feature. Experiments results on two public three databases (LFW, PubFig) show that the propose method achieves promising results in face recognition and proves that our proposed method preforms well than the state-of-the-art single feature such as Gabor, LBP, SIFT.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhiming Li
PY  - 2017/05
DA  - 2017/05
TI  - Face Recognition Based on the Key Points of High-dimensional Feature and Triplet Loss
BT  - Proceedings of the 2017 7th International Conference on Applied Science, Engineering and Technology (ICASET 2017)
PB  - Atlantis Press
SP  - 85
EP  - 90
SN  - 2352-5401
UR  - https://doi.org/10.2991/icaset-17.2017.16
DO  - 10.2991/icaset-17.2017.16
ID  - Li2017/05
ER  -