基于BAB算法的ECG身份识别解析特征选择方法  

Analytic feature selection of ECG based BAB algorithm for human identification

作　　者：杨向林[1] 严洪[1] 任兆瑞[1] 许志[1] 张煜[1] 姚宇华[1] 

机构地区：[1]中国航天员科研训练中心,北京100094

出　　处：《仪器仪表学报》2010年第10期2394-2400,共7页Chinese Journal of Scientific Instrument

基　　金：中国航天医学工程预先研究基金(SJ200903)资助项目

摘　　要：ECG作为一种新的活体生物特征用于身份识别有着广阔的应用前景。针对目前不同学者用于ECG身份识别的解析特征种类多、差异大问题。提出了ECG身份识别的特征选择问题,并提出了基于分支定界法的ECG身份识别解析特征选择方法,将所提出的解析特征与Gahi最新提出的解析特征送入神经网络进行比较。实验表明该算法所提特征稳定性高、特异性强,优于Gahi算法所提特征,可有效用于ECG身份识别。As a new live biometric feature for human identification, ECG has extensive application prospects. Aiming at the fact that the analytic features used in ECG human identification currently proposed by different scholars vary significantly. The issue of feature selection for ECG human identification is discussed in this paper. A method based on branch and bound approach for analytic feature selection of ECG is proposed. The method was used for human identification, and the obtained features are compared with the features presented by Gahi through neural network. Experiment demonstrates that the features obtained using the proposed method are comparatively stable and distinguishing, and superior to the features obtained using Gahi＇ algorithm. The proposed method can be effectively used for ECG human identification.

关 键 词：解析特征 心电图 分支定界法 特征选择 身份识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]