Rotation invariant finger vein recognition

Authors:
Shaohua Pang;Yilong Yin;Gongping Yang;Yanan Li
Affiliations:
School of Computer Science and Technology, Shandong University, Jinan, China;School of Computer Science and Technology, Shandong University, Jinan, China;School of Computer Science and Technology, Shandong University, Jinan, China;School of Computer Science and Technology, Shandong University, Jinan, China
Venue:
CCBR'12 Proceedings of the 7th Chinese conference on Biometric Recognition
Year:
2012

Citing 6
Cited 0

Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Feature extraction of finger-vein patterns based on repeated line tracking and its application to personal identification

Machine Vision and Applications
Handbook of Multibiometrics (International Series on Biometrics)

Handbook of Multibiometrics (International Series on Biometrics)
Driver identification using finger-vein patterns with Radon transform and neural network

Expert Systems with Applications: An International Journal
A Palmprint Recognition Algorithm Using Phase-Only Correlation

IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
Finger vein recognition with manifold learning

Journal of Network and Computer Applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

Finger vein patterns have recently been recognized as an effective biometric identifier and many related work can achieve satisfied results. However, these methods usually suppose the database is non-rotated or slightly rotated, which are strict for preprocessing stages, especially for capture. As we all know, user-friendly capture tends to cause the rotation problem, which degrades the recognition performance due to the unregulated images or feature loss. In this paper, we propose a new finger vein recognition method to solve the common rotation problem of finger vein images. Two experiments are designed to evaluate the recognition performance in both verification mode and identification mode and to demonstrate the advantages and robustness of our method in different rotated databases compared with pattern binary based method like LBP. Experimental results show that the proposed method can not only achieve a lower EER (1.71%) than LBP (2.67%), but also can overcome the difficulties of rotation in rotated databases. The EERs of proposed method in three different rotated databases are 2.15%, 2.98% and 2.75% respectively, which shows that our method has better robustness in rotation than LBP.