心电异常自动检测的研究

Study of Ecg Anomaly Detection

心电图能够直观地反映心脏的活动状况,能够通过心电图特征诊断各种心脏有关疾病和症状。文章提出了一种自动检测心电异常的方法,使用小波变换法对心电图进行基线漂移的校正和噪声的过滤,对预处理后的心电信号使用极值法和阈值的比较实现特征点的提取,通过特征参数的判断查找正常心电模板,计算正常心电模板和待检测心电图的相关系数来进行模板匹配,通过相关系数判断两者的匹配度,进而分析心电图是否异常。文章使用MIT-BIH数据库进行试验结果分析,能够较准确地分析出异常波形。

ECG has become an important detection item of a clinical diagnosis.The diagnosis of all kinds of heart disease and symptoms can be diagnosed by ECG.This paper proposed a kind of ECG anomaly detection,using the wavelet transform method to estimate the baseline drift of the ECG and the noise filtering.Use the extreme value and the threshold to realize the feature points extracting and search the normal ECG template by the characteristic pretreatment.Compare normal ECG and ECG to make Template Matching.Correlation coefficient calculating by making Template Matching is used to determine whether the ECG is abnormal or not.The test result is analyzed by using the MIT-BIH database.The abnormal waveforms can be analyzed accurately in the paper.