基于多尺度熵的常见心脏疾病特征研究

发布时间：2019-07-04 21:07

【摘要】：近年来,中国的心血管疾病死亡率一直居所有疾病之首,患者人数在未来数年内仍将继续增加。心血管疾病在生活和经济上都给患者带来了巨大的负担。心电图(ECG)可以直观准确地反映心脏的电活动特性和表现心脏的工作状态,是目前临床医生判断大部分心血管疾病的常用参考手段。但是,伴随着心血管疾病患者人数的日益增加以及患者心电图监护数据的增加,完全让临床医生人工根据心电图来判断心血管疾病发生将给医生带来巨大的工作负担,且很容易发生误判和漏判。因此,心电自动分析技术应用于临床心血管疾病的判断逐渐成为当前心电信号处理研究领域的热点。多尺度熵(MSE)因其具有物理意义明确、分析更具有系统性等优点在生物医学信号处理领域正得到越来越多的应用。本文基于多尺度熵对充血性心衰(CHF)和房颤(AF)两种常见心脏疾病进行了特征研究,并提出了一种基于多尺度熵的充血性心衰判别算法和一种基于多尺度熵的房颤判别算法。本文主要研究内容如下:(1)基于多尺度熵进行了充血性心衰的特征研究,并比较了充血性心衰患者和正常人心率变异性之间的差异,发现健康人的多尺度熵的平均值大于充血性心衰患者,这说明充血性心衰患者心电信号的复杂度低于健康人。最后本文基于多尺度熵并结合连续相邻两个RR间期之间差值的均方根提出了一种新的充血性心衰判别算法,我们用MIT-BIH心电数据库中的心电数据验证了本文算法的性能。实验结果表明本文算法的判断准确率达到了91.67%,说明本文算法具有一定的临床应用前景。(2)基于多尺度熵进行了房颤的特征研究,并比较了房颤患者和正常人心率变异性之间的差异,发现健康人的多尺度熵的平均值大于房颤患者,这说明充血性心衰患者心电信号的复杂度低于健康人。接着本文基于多尺度熵和信号功率谱低频段能量与高频段能量的比值两个参数,设计了一种新的房颤判别算法。最后,我们用MIT-BIH心电数据库中的心电数据验证了本文算法的性能。实验结果表明本文算法的准确率、敏感度和阳性预测率分别为93.06%、91.67%和94.29%。
[Abstract]:In recent years, the death rate of cardiovascular disease in China has been the first of all diseases, and the number of patients will continue to increase over the next few years. Cardiovascular disease has a great burden on patients both in life and in the economy. Electrocardiogram (ECG) can directly and accurately reflect the electrical activity characteristics of the heart and the working state of the heart, and is a common reference for the current clinician to judge most of the cardiovascular diseases. However, with the increase of the number of patients with cardiovascular disease and the increase of the patient's ECG monitoring data, the clinician is completely allowed to manually judge the occurrence of the cardiovascular disease according to the electrocardiogram, which will cause great work burden to the doctor, and can be easily misjudged and missed. Therefore, the application of the automatic electrocardio-analysis technique to the clinical cardiovascular disease is becoming a hot spot in the current research field of ECG signal processing. The multi-scale entropy (MSE) is getting more and more applications in the field of biomedical signal processing because it has the advantages of clear physical meaning, more systematic analysis and the like. In this paper, the characteristics of two common heart diseases of congestive heart failure (CHF) and atrial fibrillation (AF) are studied based on the multi-scale entropy, and a multi-scale entropy-based judgment algorithm for congestive heart failure and an AF discrimination algorithm based on multi-scale entropy are presented. The main contents of this paper are as follows: (1) The characteristic study of congestive heart failure is carried out based on the multi-scale entropy, and the difference between the heart rate variability of the patients with congestive heart failure and the normal person is compared, and the average value of the multi-scale entropy of the healthy person is found to be larger than that of the patients with congestive heart failure. This indicates that the complexity of the cardiac electrical signal in patients with congestive heart failure is lower than that of a healthy person. In the end, based on the multi-scale entropy and combined with the root mean square of the difference between two consecutive RR intervals, a new algorithm for the determination of congestive heart failure is presented, and the performance of this algorithm is verified by the ECG data in the MIT-BIH ECG database. The experimental results show that the accuracy of the algorithm is 91.67%, which shows that the algorithm has a certain clinical application prospect. (2) Based on the multi-scale entropy, the characteristics of the atrial fibrillation were studied, and the difference between the heart rate variability of the patients with atrial fibrillation and the normal person was compared, and the mean value of the multi-scale entropy of the healthy person was found to be larger than that of the patients with atrial fibrillation, indicating that the complexity of the cardiac electrical signal in the patients with congestive heart failure was lower than that of the healthy person. Then, based on the multi-scale entropy and the ratio of the low-band energy of the signal power spectrum and the energy of the high-frequency band, a new algorithm for the determination of AF is designed. Finally, we use the ECG data in the MIT-BIH ECG database to verify the performance of the algorithm. The results show that the accuracy, sensitivity and positive predictive rate of the algorithm are 93.06%, 91.67% and 94.29%, respectively.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R540.4

【参考文献】