A New Distance Metric Learning Algorithm for Hand Posture Recognition

DOI

10.2991/icmii-15.2015.43How to use a DOI?

Keywords

Distance metric learning. Nearest classify. Hand posture recognition.

Abstract

A new distance metric learning algorithm for hand posture recognition is proposed in this paper. By adopting the Fourier descriptor of the hand shape as the sample feature, the algorithm exerts new constrains on the distance of inter-class pairs, and learns a Mahalanobis distance metric matrix by minimizing the distance of intra-class pairs and maximizing the distance of inter-class pairs simultaneously. Then k-nearest neighbor classifier is used to test the introduced method. The experimental results on multi-class hand posture recognition by k-nearest neighbor classifier show that the presented algorithm has less error rate than other distance metric learning algorithms.

Copyright

© 2015, 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  - Qiongli Liu
AU  - Dajun Xu
AU  - Zhiguo Li
AU  - Peng Zhou
AU  - Jingjing Zhou
AU  - Yongxia Xu
PY  - 2015/10
DA  - 2015/10
TI  - A New Distance Metric Learning Algorithm for Hand Posture Recognition
BT  - Proceedings of the 3rd International Conference on Mechatronics and Industrial Informatics
PB  - Atlantis Press
SP  - 232
EP  - 236
SN  - 2352-538X
UR  - https://doi.org/10.2991/icmii-15.2015.43
DO  - 10.2991/icmii-15.2015.43
ID  - Liu2015/10
ER  -