基于模糊推理的心拍分类

Beat Classification Based on Fuzzy Reason

心拍分类对于临床心律失常自动化检测非常重要。临床上对心拍分类的诊断标准存在一定的不确定性,模糊推理可以较好地表达心拍分类过程中的不确定性,而隶属度函数的设计是模糊推理系统的关键问题。本研究提取较为精确的QRS复合波间期和RR间期特征组成模糊输入量;通过对MIT-BIH心律失常心电数据库的所有正常拍和室性早搏模糊输入量进行统计分析,提出了一种设计隶属度函数的具体思路,并实现了一个用于心拍分类的模糊推理系统。通过对MIT-BIH心律失常心电数据库测试,该系统心拍分类结果较好,具有临床应用价值。

ECG （electrocardiogram） beat classification is very important for clinical automated detection of arrhythmia. Fuzzy reason could reflect the uncertain diagnosis rules of beat classification. Member function is the key to fuzzy reason system. In this paper, fuzzy input features were consisted of extracted accurately width of QRS complex waves and RR interval. A detail design of member function was presented based on fuzzy input statistics for all of normal and premature ventricular contraction beats of MIT-BIH arrhythmia database. Further, a fuzzy reason system was realized to classify ECG beats and was proved by experiment carrying out using all of ECG records in MITBIH arrhythmia database that better classification accuracy and clinical practical application are reached.