基于多模板匹配的室性早搏判别算法  

PVC Recognition Algorithm Based on Multiple Template Matching

作　　者：杨波[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,深圳518055

出　　处：《计算机工程》2010年第16期291-292,F0003,共3页Computer Engineering

基　　金：湖南省科技重大专项基金资助项目"新型临床医疗检测产品研制与产业化"(2007FJ1004)

摘　　要：提出一种基于可变宽度模板的多模板匹配的室性早搏判别算法。该方法通过样本数据创建初始模板库,根据不同的RR间期确定模板宽度,使用基于模板队列匹配和模板库匹配的二级匹配策略,利用相关系数和RR间期比相结合的算法检测室早,模板库可自动扩充。经MIT-BIH心律失常数据库测试,该算法达到了98.66%的灵敏度(正确率),平均错误率为1.11%。An algorithm based on multiple variable-width template matching for Premature Ventricular Contraction（PVC） recognition is presented.A template database is created by means of training samples,in which the width of each template is based on its RR interval.A two-stage strategy is proposed,which is based on template queue matching and template database matching,combining correlation coefficient with ratio of RR intervals.And the new template can be added to the database automatically.The performance of this algorithm is evaluated on the MIT-BIH arrhythmia database and it shows that the sensitivity for PVC reaches 98.66%,and AER is 1.11%.

关 键 词：多模板匹配 可变宽度模板 RR间期 室性早搏 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]