基于支持向量机的心律失常识别方法  

Recognition method of arrhythmias based on SVM

作　　者：杨煦[1] 张跃[1] 

机构地区：[1]清华大学深圳研究生院嵌入式系统与技术实验室,广东深圳518055

出　　处：《计算机工程与设计》2007年第18期4442-4445,共4页Computer Engineering and Design

摘　　要：心律失常判别是心电数据处理的重要研究方向。提出一种基于支持向量机的心律失常判别方法来实现心律失常识别。根据临床诊断标准、结合计算机图像处理方法,提取心律参数特征和QRS复波形状特征,然后使用支持向量机对其进行分类,用MIT-BIH数据库对识别结果进行判断,并根据统计结果对该方法做出评价。A method based on SVM for classification ofarrhythmias is presented. According to clinical rules for arrhythmia recognition,R-R and QRS intervals are used to describe rhythmic changes while shape matching method based on Zernike moments is used to describedifferent shapes of QRS complex. By the end, the method is verified with MIT-BIH database. Judging from the result, this method works well on various kinds of atrial premature beats and ventricular premature beats.

关 键 词：支持向量机 ZEMIKE矩 模板形状匹配 心拍建模 心律失常 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]