3D feature array involved registration algorithm for multi-pose hand vein authentication

Yong Qi, Ya Zhou, Chang Zhou, Xinran Hu, Xiaoming Hu

科研成果: 书/报告/会议事项章节会议稿件同行评审

1 引用 (Scopus)

摘要

Traditional hand vein recognition technology is usually based on 2D infrared images in which the hand vein patterns are unavoidable distorted with hand posture changes. Point cloud matching vein recognition based on Kernel Correlation method provides a new perspective to hand vein recognition but also suffers from posture change. In this paper, a 3D point cloud registration algorithm is proposed to improve the vein authentication under multi-pose. An improved Normal Distributions Transform algorithm is introduced and the coarse alignment based on 3D feature array is designed, which is capable of eliminating the influence of global coordinate shift between the point clouds. Experiments show that the 3D feature array involved registration algorithm can effectively improve the recognition rate. After the point cloud registration, the system can keep more than 90% recognition rate with the corresponding error rate is as low as 4% when the hand posture changes in the range ± 20 degrees.

源语言英语
主期刊名2016 International Conference on Biometrics, ICB 2016
编辑Fernando Alonso-Fernandez, Arun Ross, Raymond Veldhuis, Julian Fierrez, Stan Z. Li, Josef Bigun
出版商Institute of Electrical and Electronics Engineers Inc.
ISBN(电子版)9781509018697
DOI
出版状态已出版 - 23 8月 2016
活动9th IAPR International Conference on Biometrics, ICB 2016 - Halmstad, 瑞典
期限: 13 6月 201616 6月 2016

出版系列

姓名2016 International Conference on Biometrics, ICB 2016

会议

会议9th IAPR International Conference on Biometrics, ICB 2016
国家/地区瑞典
Halmstad
时期13/06/1616/06/16

指纹

探究 '3D feature array involved registration algorithm for multi-pose hand vein authentication' 的科研主题。它们共同构成独一无二的指纹。

引用此