T波交替计算机检测研究

【作者】 王娟；

【导师】 刘正春； 黄忠朝；

【作者基本信息】 中南大学 ， 生物医学工程， 2010， 硕士

【摘要】 心脏疾病-一种具有突发性和偶然性的疾病,一直以来都对人类生命健康造成重大威胁。由各种心脏疾病引起的心脏性猝死(sudden cardiac death, SCD)对人类健康的威胁尤其大。T波电交替(T-wave alternans, TWA)作为预测心脏性猝死的独立的、具有统计意义的指标,在临床上的应用价值已经得到了充分的肯定。TWA是一种非常微弱的电生理现象,极易受到心率、噪声以及异位心搏等因素的影响,因此检测起来十分困难。本论文在总结分析TWA研究进展和现状的基础上,针对TWA的检测开展研究,主要进行了以下几个方面的工作：(1)心电信号预处理。用一种简单的去低频信号方法对心电信号进行预处理,去除了基线漂移。(2)心电信号特征点定位。首先用小波变换方法检测R波,R波定位后,用一种简单的几何方法准确定位T波所在区间,在此区间上用小波变换方法检测T波,完成了特征点定位。(3)TWA检测。在完成了T波矩阵构建的基础上,将奇异值分解和增强的谱分析方法相结合检测TWA,对检测结果进行了分类分析,验证算法的有效性。与以往的方法相比,本文的T波检测方法有以下优点：T波检测区间的定位能够避免心率和P波对T波检测的影响。在进行T波顶点检测时,在阈值判据的基础上加入了斜率判据,进一步提高T波定位的精确性。T波电交替检测中,奇异值分解能够提取出“干净”的心电信号,去除噪声对TWA检测的影响,增强的谱分析方法适合非平稳信号的检测,两种方法结合使用,极大地提高了TWA检测的准确性。更多还原

【Abstract】 Cardiopathy, as a type of sudden and chance diseases, has always beenone of the main threats to the health of human being. Sudden cardiacdeath caused by kinds of cardiac disease is an especially large threat.T-wave alternans has been fully recognized as an independent andstatistically significant index for clinical forecasting SCD. TWA is a veryfeeble electrophysiologic phenomenon, which is easily interfered byheart-rate, noise and ectopic beat and accordingly detected difficultly.Onthe basis of summarizing and analyzing the progress and status quo ofTWA research, this dissertation focused on the the detection of TWA. Themajor contributions include the following aspects: 1. Preprocess of electrocardiosignal. A simple method was introducedto remove baseline drift for the preprocessing of electrocardiosignal. 2. Characteristic points location of electrocardiosignal. R wave wasdetected by wavelet transform method. Then T-wave detection intervalwas determined with a simple method and the T-wave apex was detectedin this interval through wavelet transform. 3. Detection of TWA. TWA detection was realized by fusing singularvalue decomposition and enhanced spectral method after T-wave matrixwas constructed. The detection results were assorted and analyzed tovalidate the algorithm. Compared with previous methods, the present T-wave detectionmethod shows advantages as follows. T-wave detection interval wasdelineated to avoid interfernce from heart-rate and P-wave. Slope wasintroduced as a new criterion on the basis of threshold criterion to give amore accurate T-wave apex detection. For TWA detection, the singularvalue decomposition is able to remove noise and give cleanelectrocardiosignal. Enhanced spectral method is suitable for thedetection of non-stationary signal. Fusing these two methods improvedenormously the accuracy of TWA detection.■更多还原