摘要
人耳识别技术是一种新的生物识别技术,它以人耳作为识别媒介来进行身份鉴别,但人耳识别的相关理论和方法还不太完善。首先介绍了独立成分分析方法(ICA)和基于核的主成分分析方法(KPCA)的基本原理,然后通过实验得到在分别采用ICA和KPCA方法时,在不同人耳库上的特征提取时间以及采用不同分类器时的人耳识别率。最后通过分析比较实验结果得到基于ICA方法的识别技术和基于KPCA方法的识别技术各自的优点和缺点。
Human ear recognition is a new technology of biologic recognition and it use human ear as the recognition media to identify different people, but the mature theory or method has not been found yet. First, this paper describes the basic principles of independent component analysis (ICA) and kernel principal component analysis (KPCA); then we use experiment to get the time of feature extraction using different human ear storage and the human ear recognition ratio using different classifiers when we use the method ICA or KPCA separately. In the end, it describes the advantages and defaults t through comparing and analyzing the experiment result which base on ICA recognition method and KPCA recognition method.
出处
《电脑开发与应用》
2009年第2期14-15,18,共3页
Computer Development & Applications