小波变换在心电信号特征提取中的应用研究
|
摘要
本课题是小波变换在心电信号特征提取中的应用研究，它是小波变换在心电信号处理领域一个重要应用。心电图的QRS波群、P波、T波包含了人体心脏的丰富信息，它的识别的好坏对临床诊断具有重要意义。但是以前主要是通过临床医生经验进行判断，这不但使医生的工作繁重而且往往会因为个体的原因而产生误判。随着信息技术的发展，计算机在信号处理中得到了越来越多的应用。但心电信号包含的高频的工频干扰、肌电干扰以及低频的基线漂移干扰，这为特征点的提取以及疾病的识别带来难度。本课题就是基于这种背景把小波变换理论应用在心电信号的处理中，不但对信号的特征点进行检测，而且还探讨了对室早和室上早疾病的识别。
论文主要是在以下三个方面对心电信号进行处理：
1) QRS波的检测：运用小波多尺度分析的特性，提出了跨尺度的模极值检测方法，同时运用动态阈值算法并结合人体生理特点的综合检测策略进行复检。这不但有效地提高了QRS波群的检测准确率，还大大的节省了检测程序运行的时间。
2) P波和T波检测：充分利用小波变换抑制噪声的能力，结合过零点检测方法实行纵向检测。为提高精度，又在原信号上运用神经元网络(BP算法)，并结合人体生理特点进行横向检测。
3) 在室上早和室早的识别方面，识别疾病较好的方法是神经元网络，但是BP算法脱离不了梯度算法求局部精确解的本质。本文引入遗传算法求全局最优解的特点，再结合神经元网络的求局部精确解的优势，运用到室早和室上早疾病识别中。在这方面本文做了有益的探索。
论文在这三面做了理论的探讨，并根据所提的方法具体在编程中得以实现。但由于本人知识水平的有限，难免会有不足与错误之处，恳请专家给予批评指导。
The paper is Wavelet Transform-Based ECG Charactericstic Detector. It is the more important application in signal process based Wavelet Transform(WT). QRS waves and P, T waves in electrocardiogram include abundance information about person heart, which are detected well or not provide with important meaning for clinic diagnoses. The clinician diagnosing electrocardiogram with using theirs experience before , but it is not only taking heavy work for doctor but also bringing mistakes in diagnoses because of doctor individual reason. With the information technology developing, The computer are more and more used in signal process. But ECG signal includes 60Hz power line interferce, muscle noise, and motion noise. It take difficulty in ECG signal process. This paper apply WT in ECG signal process under this background, not only detecing Charactericstic point but also discussing symptom detector.
The paper mainly process ECG signal by three aspects:
1) QRS waves detector: Using WT's Multi-resolution Analysis(MRA) characteristic, bringing forward madule max value pair's detector spaning scales and dynamic threshold arithmetic combining with body physiology and synthesis strategy which takes duplicate detector. It not only take a more detector right rate of QRS waves but also shorten the period of computing.
2) P waves and T waves detector: Using WT's restraining interferce capability, combining with acrossing zero detector method, practising vertical detector. For improving precision, taking landscape orientation detector in S=24 scale using BP arithmetic and at the same time combining with body physiology.
3) The symptom detector: BP arithmetic is the preferable method of detecting
symptom. But BP arithmetic gets local precision value of grads arithmetic in essence. The project introduce Genetic Algorithms(GA)'s characteristic of getting whole optimization value at first , then combining with BP arithmetic's predominance of getting local precision value , to detect symptom. Taking a availability explore under this direction in this paper.
These arithmetic above were finished in computer program and take a preferably result. But because of finity of knowledge myself, The paper must take on deficiency or mistake, and hope experts bring forward valuable advice and comment.
论文主要是在以下三个方面对心电信号进行处理：
1) QRS波的检测：运用小波多尺度分析的特性，提出了跨尺度的模极值检测方法，同时运用动态阈值算法并结合人体生理特点的综合检测策略进行复检。这不但有效地提高了QRS波群的检测准确率，还大大的节省了检测程序运行的时间。
2) P波和T波检测：充分利用小波变换抑制噪声的能力，结合过零点检测方法实行纵向检测。为提高精度，又在原信号上运用神经元网络(BP算法)，并结合人体生理特点进行横向检测。
3) 在室上早和室早的识别方面，识别疾病较好的方法是神经元网络，但是BP算法脱离不了梯度算法求局部精确解的本质。本文引入遗传算法求全局最优解的特点，再结合神经元网络的求局部精确解的优势，运用到室早和室上早疾病识别中。在这方面本文做了有益的探索。
论文在这三面做了理论的探讨，并根据所提的方法具体在编程中得以实现。但由于本人知识水平的有限，难免会有不足与错误之处，恳请专家给予批评指导。
The paper is Wavelet Transform-Based ECG Charactericstic Detector. It is the more important application in signal process based Wavelet Transform(WT). QRS waves and P, T waves in electrocardiogram include abundance information about person heart, which are detected well or not provide with important meaning for clinic diagnoses. The clinician diagnosing electrocardiogram with using theirs experience before , but it is not only taking heavy work for doctor but also bringing mistakes in diagnoses because of doctor individual reason. With the information technology developing, The computer are more and more used in signal process. But ECG signal includes 60Hz power line interferce, muscle noise, and motion noise. It take difficulty in ECG signal process. This paper apply WT in ECG signal process under this background, not only detecing Charactericstic point but also discussing symptom detector.
The paper mainly process ECG signal by three aspects:
1) QRS waves detector: Using WT's Multi-resolution Analysis(MRA) characteristic, bringing forward madule max value pair's detector spaning scales and dynamic threshold arithmetic combining with body physiology and synthesis strategy which takes duplicate detector. It not only take a more detector right rate of QRS waves but also shorten the period of computing.
2) P waves and T waves detector: Using WT's restraining interferce capability, combining with acrossing zero detector method, practising vertical detector. For improving precision, taking landscape orientation detector in S=24 scale using BP arithmetic and at the same time combining with body physiology.
3) The symptom detector: BP arithmetic is the preferable method of detecting
symptom. But BP arithmetic gets local precision value of grads arithmetic in essence. The project introduce Genetic Algorithms(GA)'s characteristic of getting whole optimization value at first , then combining with BP arithmetic's predominance of getting local precision value , to detect symptom. Taking a availability explore under this direction in this paper.
These arithmetic above were finished in computer program and take a preferably result. But because of finity of knowledge myself, The paper must take on deficiency or mistake, and hope experts bring forward valuable advice and comment.
引文
[1] 张开滋 刘海洋 吴杰 心电信息学 北京技术文献出版社 1998
[2] 顾文荣等《心电图快速学习法》人民卫生出版社 1992
[3] 李翠微等 ECG自动分析技术的发展 国外医学生物医学工程分册1994
[4] 杨振野等 动态心电图关键技术：存在问题及解决途径 国外医学生物医学工程分册
[5] 范晓东等 国外动态ECG监护技术的研究现状 国外医学生物医学工程分册1990
[6] Hamilton PS and Tompkins WJ.Quantitative Invertigation of QRS Detection Rules Using the MIT/BIH Arrhythmia Database. IEEE Trans BME. 1986
[7] Pan J and Tompkins WJ. Real time QRS Detection Algorithm,IEEE Trans BME 1985
[8] Mallat Stephen. Multiresolution Approximation and Wavelet Orthonormal Basis of L~2.IEEE Trans AMS. 1989
[9] Mallat Stephen. A Theory for Multiresolution Signal Decomposition:Wavelet Representation. IEEE Trans PAMI. 1989
[10] 杨福生 小波变换的工程分析与应用 科学出版社1999
[11] 彭玉华等 小波变换与工程应用 科学出版社1999
[12] 王建中 小波变换及其在物理和工程中的应用 数学进展1992
[13] Mallat Stephen. Singularity Detection and Processing with wavelets.IEEE Trans Information Theory. 1992
[14] Mordechai Frish and Hagit Messer. The use of the wavelet transform in the detection of an unknown transient signal.IEEE Trans on Information Theory 1992
[15] Senhadju L.Bellanger JJ.Carrault G et al.Wavelet analysis of ECG signal. Proc Int Cont IEEE EMBS,Philadelphia 1990
[16] 王博亮等 心电信号中QRS波的实时检测算法 航天医学与医学工程1995
[17] 杨宗凯 小波去噪及其中在信号检测中的应用 华中理工大学学报1997
[18] 王斌等 小波去噪对模极大值的处理 华中理工大学学报 1997
[19] Cunwei L,et al. Detection Of Characteristic Points Using Wavelets Transforms.Ieee Tran On Biomed Eng, 1995
[20] Michael Unser, Akram Aldroubi,A Review of Wavelets in Biomedical Applications.Proceedings of the IEEE, 1996
[21] 朱贻盛等 用于奇异信号处理的小波滤波器 1995生物医学联合学术年会论文集
[22] 心电信号的分析方法及应用研究 中国科大博士学位论文 生物电子学专业1997
[23] N.V. Thakor et al,Estimation of QRS complex power spectra for design of a QRS filter, IEEE Trans.On BME, 1984
[24] 吕柏权等 基于BP神经网络与小波的控制研究 系统仿真学报2001
[25] 陈国良等 人工神经网络理论研究进展 电子学报 1996
[26] 陈哲等 一种墓于BP算法学习的小波神经网络 青岛海洋大学学报2001
[27] E.Pietka "Feature extraction in computerized approach to the ECG analysis" Pattern Recog 1991
[28] 卢才义 实用心电监没快速诊断学 北京科学出版社 1999
[29] 吴波等 基于小波变换的波形特征抽取与识别 华中理工大学学报 1993
[30] S.Mallat,Zero-crossing of a wavelet transform,IEEE Trans 1991
[31] B.Logan, Information in the zero-crossing of band pass signals, Bell Syst. Tech 1977
[32] M.L.Ahlstrom,W,J,Tompkins, "Digital filters for real-time ECG signal processing using microprocessors",IEEE Trans. BME 1985
[33] O.pahlm,"Software QRS detection in ambulatory monitoring-a review" Med&Biol Eng&Comput 1984
[34] Chen Zhe,Feng Tian jin. The application of wavelet neural net work for time series prediction and system modeling based on System.Man and Cybernetics(IEEE SMC)1999
[35] Zhang Jun, Walter G, Miao Y, et al.Wavelet neural net works for function learning IEEE Trans on SP 1995
[36] 徐佩霞 孙功宪 小波分析与实用实例 中国科技大学出版社 1996
[37] 刘希顺等 基于小波变换的心电图中QRS波综合波检测算法 国防科大学报1997
[38] 程俊 小波变换用于信号突变的检测 通信学报 1995
[39] 谭浩强 C程序设计 清华大学出版社 1997
[40] Richard C.Leinecker, Tom Archer Visual C++ 6 Bible 电子工业出版社
[41] 杨行峻 郑君里 人工神经网络与盲信号处理 清华大学出版社
[42] 陈国良 王煦法等 遗传算法及其应用 人民邮电出版社
[43] 虞和济 陈长征等 基于神经网络的智能诊断 冶金工业出版社
[2] 顾文荣等《心电图快速学习法》人民卫生出版社 1992
[3] 李翠微等 ECG自动分析技术的发展 国外医学生物医学工程分册1994
[4] 杨振野等 动态心电图关键技术：存在问题及解决途径 国外医学生物医学工程分册
[5] 范晓东等 国外动态ECG监护技术的研究现状 国外医学生物医学工程分册1990
[6] Hamilton PS and Tompkins WJ.Quantitative Invertigation of QRS Detection Rules Using the MIT/BIH Arrhythmia Database. IEEE Trans BME. 1986
[7] Pan J and Tompkins WJ. Real time QRS Detection Algorithm,IEEE Trans BME 1985
[8] Mallat Stephen. Multiresolution Approximation and Wavelet Orthonormal Basis of L~2.IEEE Trans AMS. 1989
[9] Mallat Stephen. A Theory for Multiresolution Signal Decomposition:Wavelet Representation. IEEE Trans PAMI. 1989
[10] 杨福生 小波变换的工程分析与应用 科学出版社1999
[11] 彭玉华等 小波变换与工程应用 科学出版社1999
[12] 王建中 小波变换及其在物理和工程中的应用 数学进展1992
[13] Mallat Stephen. Singularity Detection and Processing with wavelets.IEEE Trans Information Theory. 1992
[14] Mordechai Frish and Hagit Messer. The use of the wavelet transform in the detection of an unknown transient signal.IEEE Trans on Information Theory 1992
[15] Senhadju L.Bellanger JJ.Carrault G et al.Wavelet analysis of ECG signal. Proc Int Cont IEEE EMBS,Philadelphia 1990
[16] 王博亮等 心电信号中QRS波的实时检测算法 航天医学与医学工程1995
[17] 杨宗凯 小波去噪及其中在信号检测中的应用 华中理工大学学报1997
[18] 王斌等 小波去噪对模极大值的处理 华中理工大学学报 1997
[19] Cunwei L,et al. Detection Of Characteristic Points Using Wavelets Transforms.Ieee Tran On Biomed Eng, 1995
[20] Michael Unser, Akram Aldroubi,A Review of Wavelets in Biomedical Applications.Proceedings of the IEEE, 1996
[21] 朱贻盛等 用于奇异信号处理的小波滤波器 1995生物医学联合学术年会论文集
[22] 心电信号的分析方法及应用研究 中国科大博士学位论文 生物电子学专业1997
[23] N.V. Thakor et al,Estimation of QRS complex power spectra for design of a QRS filter, IEEE Trans.On BME, 1984
[24] 吕柏权等 基于BP神经网络与小波的控制研究 系统仿真学报2001
[25] 陈国良等 人工神经网络理论研究进展 电子学报 1996
[26] 陈哲等 一种墓于BP算法学习的小波神经网络 青岛海洋大学学报2001
[27] E.Pietka "Feature extraction in computerized approach to the ECG analysis" Pattern Recog 1991
[28] 卢才义 实用心电监没快速诊断学 北京科学出版社 1999
[29] 吴波等 基于小波变换的波形特征抽取与识别 华中理工大学学报 1993
[30] S.Mallat,Zero-crossing of a wavelet transform,IEEE Trans 1991
[31] B.Logan, Information in the zero-crossing of band pass signals, Bell Syst. Tech 1977
[32] M.L.Ahlstrom,W,J,Tompkins, "Digital filters for real-time ECG signal processing using microprocessors",IEEE Trans. BME 1985
[33] O.pahlm,"Software QRS detection in ambulatory monitoring-a review" Med&Biol Eng&Comput 1984
[34] Chen Zhe,Feng Tian jin. The application of wavelet neural net work for time series prediction and system modeling based on System.Man and Cybernetics(IEEE SMC)1999
[35] Zhang Jun, Walter G, Miao Y, et al.Wavelet neural net works for function learning IEEE Trans on SP 1995
[36] 徐佩霞 孙功宪 小波分析与实用实例 中国科技大学出版社 1996
[37] 刘希顺等 基于小波变换的心电图中QRS波综合波检测算法 国防科大学报1997
[38] 程俊 小波变换用于信号突变的检测 通信学报 1995
[39] 谭浩强 C程序设计 清华大学出版社 1997
[40] Richard C.Leinecker, Tom Archer Visual C++ 6 Bible 电子工业出版社
[41] 杨行峻 郑君里 人工神经网络与盲信号处理 清华大学出版社
[42] 陈国良 王煦法等 遗传算法及其应用 人民邮电出版社
[43] 虞和济 陈长征等 基于神经网络的智能诊断 冶金工业出版社