心肌梗死诊断中的SDEC可分性准则研究

Study of Separability Criterion Based on SDEC in Myocardial Infraction Diagnosis

多数现存的可分性准则在高维情况下往往存在实现上的困难。针对心肌梗死心电信号(ECG)特征提取过程中的基于标准差和欧氏中心距(SDEC)的可分性准则进行研究,研究数据取自PTB诊断数据库,包括:正常状态病人、早期心肌梗死、急性期心肌梗死。实验结果表明,基于SDEC的可分性准则应用于心肌梗死心电信号特征提取和分类能有效地克服现存可分性准则实现上的困难,并与心肌梗死实际演变过程中的可分性相一致。

Most of existing separability criteria is difficult to be implemented, especially in the hyperdimentional cases. In this investigation, the myocardial infraction features were extracted from electrocardiogram （ECG） for the classification, and the data was collected from PTB diagnostic database including normal condition, early infraction and acute infraction. The separability based on standard deviation and Euclidean center distance （SDEC） were analysed during the feature extraction. The experimental results show that the separability criterion based on the SDEC is able to overcome implementary difficulties effectively in myocardial infraction ECG feature extraction and classification, and consistent with the evolution process of myocardial infraction.