A competitive sample selection method for palmprint recognition

Authors:
Jiajun Wen;Jinrong Cui;Zhihui Lai;Jianxun Mi
Affiliations:
Bio-Computing Research Center, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, China,Key Laboratory of Network Oriented Intelligent Computation, Shenzhen, China;Bio-Computing Research Center, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, China,Key Laboratory of Network Oriented Intelligent Computation, Shenzhen, China;Bio-Computing Research Center, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, China,Key Laboratory of Network Oriented Intelligent Computation, Shenzhen, China;Bio-Computing Research Center, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, China,Key Laboratory of Network Oriented Intelligent Computation, Shenzhen, China
Venue:
IScIDE'12 Proceedings of the third Sino-foreign-interchange conference on Intelligent Science and Intelligent Data Engineering
Year:
2012

Citing 7
Cited 0

Matrix computations (3rd ed.)

Matrix computations (3rd ed.)
Palmprint recognition using eigenpalms features

Pattern Recognition Letters
Online Palmprint Identification

IEEE Transactions on Pattern Analysis and Machine Intelligence
Competitive Coding Scheme for Palmprint Verification

ICPR '04 Proceedings of the Pattern Recognition, 17th International Conference on (ICPR'04) Volume 1 - Volume 01
Research of Palmprint Recognition Based on 2DPCA

ISNN 2009 Proceedings of the 6th International Symposium on Neural Networks: Advances in Neural Networks - Part II
Palmprint Recognition Based on 2DPCA-Moment Invariant

ICIG '09 Proceedings of the 2009 Fifth International Conference on Image and Graphics
Supervised sparse representation method with a heuristic strategy and face recognition experiments

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

In the field of palmprint recognition, traditional classification methods such as PCA and LDA only exploit training samples to build the model of classifiers while ignoring the feedback of test sample during the process of recognition. In this paper, both types of samples will be taken into account to make an optimal representation of the test sample. We first exploit nearest neighbor principle to construct the initial optimal training sample set for representation. This optimal set will be updated by adding a training sample which works well with current optimal set to obtain the least representation error to the test sample. The sample selection step does not stop until we acquire the optimal set with sufficient number of training samples for classification. Comparative experiments have been conducted on PolyU multispectral palmprint database which validates the proposed method. Moreover, a study on parameter points out that a small number of selected training samples is good for recognition in blue and green channels while much more selected training samples are suitable for improving recognition in red channel.