Robust ear based authentication using Local Principal Independent Components

作者：

Highlights：

• Information set which consider both membership value and pixel intensity together is introduced in this paper.

• New type of features are derived on the bases of information sets and these are used to modify PCA.

• The concept of IPC classifier is very simple and can be implemented with very small computation.

• Experiments carried out on the ear based authentication under both constrained and unconstrained environments are very promising.

• Time complexity of LPIC is less as compared to PCA.

摘要

•Information set which consider both membership value and pixel intensity together is introduced in this paper.•New type of features are derived on the bases of information sets and these are used to modify PCA.•The concept of IPC classifier is very simple and can be implemented with very small computation.•Experiments carried out on the ear based authentication under both constrained and unconstrained environments are very promising.•Time complexity of LPIC is less as compared to PCA.

论文关键词：PCA,LPIC,Ear biometric,Effective information (EI),Energy feature (EF),Sigmoid feature (SF),Multi Quadratic feature (MQD),Inner product classifier (IPC),Euclidean classifier (EC)

论文评审过程：Available online 11 May 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.05.020