基于ARM的室性波识别技术
摘要
室性心动过速(VT:Ventricular Tarchycardia)与心室纤颤(VF:Ventricular Fibrillation)是恶性心律,易引发心脏猝死现象。猝死发生前可无任何先兆,据美国AHA统计,在猝死病例中,约有高达30 %的死者没有或从来没有发现过心脏病。绝大部分心脏病猝死情况发生在医院外面,如果能提前发现和采用迅速的治疗可以提高这部分人存活的机会。因此对猝死和恶性室性心律失常的预测是心律失常学研究的一个重要课题。本课题也是以此为主要内容进行研究的。
心电信号(ECG)是一种微弱的非线性、非平稳的随机信号,常规的心电信号一般是mV级。由于在心电信号的采集过程中,采集端需要与人体直接接触,因而采集端采集到的心电信号中包含了大量的干扰信号,如工频干扰、基线漂移、肌电干扰、电极接触噪声、运动伪迹等。这些噪声给心电信号的分析、识别和诊断带来了严重的干扰,因此必须对采集的心电信号进行有效滤波。小波变换(WT)是一种时频局部化、多分辨分析方法,具有自动“变焦距”的功能,可用来提取和识别那些淹没在噪声中的微弱电生理信号,在获得信噪比增益的同时,能够保持对信号突变信息的良好分辨。因此WT是处理心电信号的一种可行有效的方法。本文采用coif4小波对采集的心电信号进行去噪。在经典的软、硬阈值去噪算法的基础上,对中间尺度的小波系数采用软、硬阈值折衷的去噪算法。仿真结果表明,coif4小波十分适合用于心电信号滤波,它能有效地去除心电信号中常见的肌电干扰、工频干扰、基线漂移等干扰和噪声,并且在去除肌电干扰的同时,有效地保留了有用的心电信号成份。
最近的研究表明心脏动力学是复杂、非线性的。因此,非线性动力学与非线性数学模型可以有效地来分析心电信号。本文引用了一种非线性算法——引入Hurst指标来识别NSR、VT与VF。该算法基于多尺度分析与非线性动力学,通过量化心电信号的运动状态,提取相应的动力学参数来识别NSR、VT与VF。利用此算法在研究心脏机能和状态时,不再是对心电信号时间序列进行特征提取,而是针对心电信号变化的动力学原因和运动学特征进行研究,目的是揭示心脏在发生病变及猝死前的情况下,寻找心电信号从正常的窦性节律到心动过速,到心室纤颤变化的动力学指标趋势,从而为临床心电医学的研究提供可靠的证据。Complex Measure算法被认为是经典的恶性心律检测算法,它是通过在时域上量化心电信号的动力学特征来识别VT与VF,大量的仿真结果表明:当WL (Window Length) =7s时,VT、VF的检测率可达到100%;本课题中采用的Hurst Index算法是在时频域上来量化心电信号的动力学特征.,当WL=5s时,VT、VF的检测率即可达到100%,并且,当用Hurst算法与Complex Measure算法分别处理相同长度的数据时,结果表明,前者比后者处理速度快很多。另外,Complex Measure算法处理的VT仅是单形态的,而Hurst Index算法处理的VT既包含单形态的,也包含多形态的。综合算法识别的准确性、实时性与普遍性,本文的Hurst Index算法在临床应用中比Complex Measure算法有更大的潜能。
本课题主要进行的是恶性室性心律失常识别与检测的算法研究。课题所使用的心电数据主要来自于MIT-BIH Arrhythmia Database (MITDB)、MIT-BIH Malignant Ventricular Ectopy Database (VFDB)与CU Ventricular Tachyarrhythmia Database (CUDB),这三个数据库均来源于美国麻省理工大学创建的生理信号处理数据库(PhysioBank)被世界公认为标准、权威的心电信号数据库。
嵌入式ARM微处理器技术是目前最热门的一项技术,它的应用几乎已经深入到各个领域,如:工业控制领域、消费类电子产品领域、通信系统领域、网络系统领域、无线系统领域等。ARM技术正在逐步渗入到我们生活的各个方面。ARM微处理器具有体积小、功耗低、成本低、性能高、专用性强、系统精简等特点。将ARM微处理器应用到便携式心电监护系统的设计上可以提高运算速度和存储能力、降低电路的复杂度、减小系统的体积、进一步提高心电监护质量。便携式心电监护仪是及时检测出病人发生的心律失常信号并实时发出报警的心电监护设备,它具有体积小、便于携带、操作简单等优点,对挽救院外病人的生命发挥着重要作用。目前,便携式心电监护仪受到越来越多消费者的青睐,并且,越来越多的医学工作人员进行便携式心电监护仪的研制与改进。本课题中选用Philips公司的LPC2368作为微处理器,并以此为核心对便携式心电监护仪的设计进行了探讨,使便携式心电监护仪具有集成度高、体积小、反应速度快、稳定及可靠性强等特点。这为本课题的算法应用搭建了硬件平台,也为以后的研究工作提供了参考方向。
Ventricular tachycardia (VT) and ventricular fibrillation (VF) are catastrophic and life-threatening ventricular arrhythmia, which are prone to cause a sudden death. There is no sign before it. According to statistics by AHA, up to 30 percent of people died of a sudden death have no or never have heart disease. Almost all of them died outside of hospital. If the disease can be detected ahead and then treatment were carried out immediately, it can improve the opportunity of subsistence rate. Therefore, it is an important topic of forecasting a sudden death and life-threatening ventricular arrhythmia, which is right the main content of this paper.
ECG (Electrocardiogram) signals are weak, nonlinear, non-stationary and random. The common amplitudes of them are just millivolt-high. Because of the direct contact of skin with ECG apparatus, there are a lot of noises in the collected signals, such as power-line interference, baseline-drift, muscle contraction (EMG), electrode contact noise, motion artifacts etc. Theses noises lead to a serious interference in analyzing and detecting ECG signals as well as in an effective treatment, thus a highly effective restraint of noises should be carried out firstly. Wavelet transform (WT) is a multi-resolution method based on time-frequency domain, which can be viewed as a camera adjusting its focus automatically. WT can extract and identify the tiny physiological signals from noises. It can not only improve the signal-noise ratio, but also achieve an excellent resolution when signals change suddenly. Therefore, WT is an effective tool in analyzing ECG signals. In this paper, coif4 wavelet is chosen to remove noises. Based on the typical soft-threshold and hard-threshold de-nosing algorithm, an algorithm of a tradeoff of the two thresholds is used. The simulating results show that coif4 wavelet is very suitable for ECG de-nosing. It can effectively get rid of the common noises, such as muscle contraction, power-line interference, baseline-drift and so on, and at the same time, it can effectively keep the useful ECG information.
Recent studies have shown that the cardiac dynamics are complex and nonlinear. Thus the nonlinear dynamics or nonlinear mathematical models are considered to be suitable tools for analyzing ECG signals. This paper introduces a nonlinear technique called Hurst index algorithm for VT and VF discrimination. In brief, the Hurst index is defined in the multi-scale domain as a feature to quantify the nonlinear dynamics behavior of the ECG signals for detecting the life-threatening ventricular arrhythmia. In this paper, when studying the cardiac mechanism and state, it is not based on the extraction of the ECG features from the time sequences, but is in accordance with the changing ECG dynamic mechanism and dynamic features to uncover the changing trends of dynamic index from healthy NSR to VT, and to VF when the heart is broken and is prone to a sudden cardiac death. Thus, the variational dynamic index can supply a reliable reference to clinical ECG study. One typical method to detect life-threatening ventricular arrhythmia is Complex Measure algorithm, which detects VT and VF through quantifying ECG dynamic features in time domain. After a lot of simulations, it can be concluded that when the WL (Window Length) is equal to 7s, the Complex Measure algorithm can achieve 100% detecting accuracy in discriminating NSR, VT, and VF from each other, while using the Hurst Index algorithm, the WL just needs 5s. Furthermore, upon applying Hurst index algorithm and Complexity measure algorithm to the same ECG episodes, the Hurst index has a better real-time capability than the Complexity measure. Besides, the extracted VT in Complexity measure is monomorphic, while in this study the extracted VT contains both monomorphic and polymorphic waves. Due to the real-time capability and prevalence as well as reliability, the Hurst index has a better potential for clinical application than the Complexity measure.
This study puts an emphasis on algorithm. The mainly used databases in this study are MIT-BIH Arrhythmia Database (MITDB)、MIT-BIH Malignant Ventricular Ectopy Database (VFDB) and CU Ventricular Tachyarrhythmia Database (CUDB), which are all from a physiological signal database called PhysioBank. This physiological signal database is established by American MIT University, and known as standard and authoritative by the world.
The embedded ARM microprocessor is the hottest technique at present. It has been widely used in every domain, such as industry control domain, consumed-electrical products domain, communication system domain, network system domain, wireless system domain and etc. It is gradually permeating everywhere in our daily life. The ARM technique has many advantages, for example, small volume, low power cost, low expense, high quality, excellent specificity, concise system and so on. Applying ARM microprocessor to the design of easy-taking ECG monitoring system can decrease the complexity of circuit, the volume of the system and improve the computing speed, storage capability as well as the monitoring quality. The easy-taking ECG monitoring system can detect the ventricular arrhythmia in time and give rise to an alarm at the same time. It possesses such traits as small volume, easy-taking, smart operations etc, and plays an important role in saving patients’life outside hospitals. Presently, more and more people become familiar with the easy-taking ECG monitoring system and more and more medical staffs carry on the study and improvement of easy-taking ECG monitoring system. This study suggests adopting Philips Company’s LPC2368 as microprocessor and a discussion on the system structure based on LPC2368 is given in order to supply a hard platform for the Hurst Index algorithm and a reference for future study.
心电信号(ECG)是一种微弱的非线性、非平稳的随机信号,常规的心电信号一般是mV级。由于在心电信号的采集过程中,采集端需要与人体直接接触,因而采集端采集到的心电信号中包含了大量的干扰信号,如工频干扰、基线漂移、肌电干扰、电极接触噪声、运动伪迹等。这些噪声给心电信号的分析、识别和诊断带来了严重的干扰,因此必须对采集的心电信号进行有效滤波。小波变换(WT)是一种时频局部化、多分辨分析方法,具有自动“变焦距”的功能,可用来提取和识别那些淹没在噪声中的微弱电生理信号,在获得信噪比增益的同时,能够保持对信号突变信息的良好分辨。因此WT是处理心电信号的一种可行有效的方法。本文采用coif4小波对采集的心电信号进行去噪。在经典的软、硬阈值去噪算法的基础上,对中间尺度的小波系数采用软、硬阈值折衷的去噪算法。仿真结果表明,coif4小波十分适合用于心电信号滤波,它能有效地去除心电信号中常见的肌电干扰、工频干扰、基线漂移等干扰和噪声,并且在去除肌电干扰的同时,有效地保留了有用的心电信号成份。
最近的研究表明心脏动力学是复杂、非线性的。因此,非线性动力学与非线性数学模型可以有效地来分析心电信号。本文引用了一种非线性算法——引入Hurst指标来识别NSR、VT与VF。该算法基于多尺度分析与非线性动力学,通过量化心电信号的运动状态,提取相应的动力学参数来识别NSR、VT与VF。利用此算法在研究心脏机能和状态时,不再是对心电信号时间序列进行特征提取,而是针对心电信号变化的动力学原因和运动学特征进行研究,目的是揭示心脏在发生病变及猝死前的情况下,寻找心电信号从正常的窦性节律到心动过速,到心室纤颤变化的动力学指标趋势,从而为临床心电医学的研究提供可靠的证据。Complex Measure算法被认为是经典的恶性心律检测算法,它是通过在时域上量化心电信号的动力学特征来识别VT与VF,大量的仿真结果表明:当WL (Window Length) =7s时,VT、VF的检测率可达到100%;本课题中采用的Hurst Index算法是在时频域上来量化心电信号的动力学特征.,当WL=5s时,VT、VF的检测率即可达到100%,并且,当用Hurst算法与Complex Measure算法分别处理相同长度的数据时,结果表明,前者比后者处理速度快很多。另外,Complex Measure算法处理的VT仅是单形态的,而Hurst Index算法处理的VT既包含单形态的,也包含多形态的。综合算法识别的准确性、实时性与普遍性,本文的Hurst Index算法在临床应用中比Complex Measure算法有更大的潜能。
本课题主要进行的是恶性室性心律失常识别与检测的算法研究。课题所使用的心电数据主要来自于MIT-BIH Arrhythmia Database (MITDB)、MIT-BIH Malignant Ventricular Ectopy Database (VFDB)与CU Ventricular Tachyarrhythmia Database (CUDB),这三个数据库均来源于美国麻省理工大学创建的生理信号处理数据库(PhysioBank)被世界公认为标准、权威的心电信号数据库。
嵌入式ARM微处理器技术是目前最热门的一项技术,它的应用几乎已经深入到各个领域,如:工业控制领域、消费类电子产品领域、通信系统领域、网络系统领域、无线系统领域等。ARM技术正在逐步渗入到我们生活的各个方面。ARM微处理器具有体积小、功耗低、成本低、性能高、专用性强、系统精简等特点。将ARM微处理器应用到便携式心电监护系统的设计上可以提高运算速度和存储能力、降低电路的复杂度、减小系统的体积、进一步提高心电监护质量。便携式心电监护仪是及时检测出病人发生的心律失常信号并实时发出报警的心电监护设备,它具有体积小、便于携带、操作简单等优点,对挽救院外病人的生命发挥着重要作用。目前,便携式心电监护仪受到越来越多消费者的青睐,并且,越来越多的医学工作人员进行便携式心电监护仪的研制与改进。本课题中选用Philips公司的LPC2368作为微处理器,并以此为核心对便携式心电监护仪的设计进行了探讨,使便携式心电监护仪具有集成度高、体积小、反应速度快、稳定及可靠性强等特点。这为本课题的算法应用搭建了硬件平台,也为以后的研究工作提供了参考方向。
Ventricular tachycardia (VT) and ventricular fibrillation (VF) are catastrophic and life-threatening ventricular arrhythmia, which are prone to cause a sudden death. There is no sign before it. According to statistics by AHA, up to 30 percent of people died of a sudden death have no or never have heart disease. Almost all of them died outside of hospital. If the disease can be detected ahead and then treatment were carried out immediately, it can improve the opportunity of subsistence rate. Therefore, it is an important topic of forecasting a sudden death and life-threatening ventricular arrhythmia, which is right the main content of this paper.
ECG (Electrocardiogram) signals are weak, nonlinear, non-stationary and random. The common amplitudes of them are just millivolt-high. Because of the direct contact of skin with ECG apparatus, there are a lot of noises in the collected signals, such as power-line interference, baseline-drift, muscle contraction (EMG), electrode contact noise, motion artifacts etc. Theses noises lead to a serious interference in analyzing and detecting ECG signals as well as in an effective treatment, thus a highly effective restraint of noises should be carried out firstly. Wavelet transform (WT) is a multi-resolution method based on time-frequency domain, which can be viewed as a camera adjusting its focus automatically. WT can extract and identify the tiny physiological signals from noises. It can not only improve the signal-noise ratio, but also achieve an excellent resolution when signals change suddenly. Therefore, WT is an effective tool in analyzing ECG signals. In this paper, coif4 wavelet is chosen to remove noises. Based on the typical soft-threshold and hard-threshold de-nosing algorithm, an algorithm of a tradeoff of the two thresholds is used. The simulating results show that coif4 wavelet is very suitable for ECG de-nosing. It can effectively get rid of the common noises, such as muscle contraction, power-line interference, baseline-drift and so on, and at the same time, it can effectively keep the useful ECG information.
Recent studies have shown that the cardiac dynamics are complex and nonlinear. Thus the nonlinear dynamics or nonlinear mathematical models are considered to be suitable tools for analyzing ECG signals. This paper introduces a nonlinear technique called Hurst index algorithm for VT and VF discrimination. In brief, the Hurst index is defined in the multi-scale domain as a feature to quantify the nonlinear dynamics behavior of the ECG signals for detecting the life-threatening ventricular arrhythmia. In this paper, when studying the cardiac mechanism and state, it is not based on the extraction of the ECG features from the time sequences, but is in accordance with the changing ECG dynamic mechanism and dynamic features to uncover the changing trends of dynamic index from healthy NSR to VT, and to VF when the heart is broken and is prone to a sudden cardiac death. Thus, the variational dynamic index can supply a reliable reference to clinical ECG study. One typical method to detect life-threatening ventricular arrhythmia is Complex Measure algorithm, which detects VT and VF through quantifying ECG dynamic features in time domain. After a lot of simulations, it can be concluded that when the WL (Window Length) is equal to 7s, the Complex Measure algorithm can achieve 100% detecting accuracy in discriminating NSR, VT, and VF from each other, while using the Hurst Index algorithm, the WL just needs 5s. Furthermore, upon applying Hurst index algorithm and Complexity measure algorithm to the same ECG episodes, the Hurst index has a better real-time capability than the Complexity measure. Besides, the extracted VT in Complexity measure is monomorphic, while in this study the extracted VT contains both monomorphic and polymorphic waves. Due to the real-time capability and prevalence as well as reliability, the Hurst index has a better potential for clinical application than the Complexity measure.
This study puts an emphasis on algorithm. The mainly used databases in this study are MIT-BIH Arrhythmia Database (MITDB)、MIT-BIH Malignant Ventricular Ectopy Database (VFDB) and CU Ventricular Tachyarrhythmia Database (CUDB), which are all from a physiological signal database called PhysioBank. This physiological signal database is established by American MIT University, and known as standard and authoritative by the world.
The embedded ARM microprocessor is the hottest technique at present. It has been widely used in every domain, such as industry control domain, consumed-electrical products domain, communication system domain, network system domain, wireless system domain and etc. It is gradually permeating everywhere in our daily life. The ARM technique has many advantages, for example, small volume, low power cost, low expense, high quality, excellent specificity, concise system and so on. Applying ARM microprocessor to the design of easy-taking ECG monitoring system can decrease the complexity of circuit, the volume of the system and improve the computing speed, storage capability as well as the monitoring quality. The easy-taking ECG monitoring system can detect the ventricular arrhythmia in time and give rise to an alarm at the same time. It possesses such traits as small volume, easy-taking, smart operations etc, and plays an important role in saving patients’life outside hospitals. Presently, more and more people become familiar with the easy-taking ECG monitoring system and more and more medical staffs carry on the study and improvement of easy-taking ECG monitoring system. This study suggests adopting Philips Company’s LPC2368 as microprocessor and a discussion on the system structure based on LPC2368 is given in order to supply a hard platform for the Hurst Index algorithm and a reference for future study.
引文
[1] 郭继鸿, 张萍. 动态心电图学[M]. 北京: 人民卫生出版社, 2003.
[2] 李翠微, 郑崇勋. ECG 自动分析技术的发展[J]. 国外医学生物医学工程分册, 1994, 17(1): 15-22.
[3] 宁新宝. 生物医学电子学[M]. 长沙: 湖南科学技术出版社, 1985.
[4] 刘广芝. 心电图学概论[M]. 贵州:贵州科技出版社, 2003: 234-235.
[5] 刘广芝. 心电图学概论[M]. 贵州: 贵州科技出版社, 2003: 257-257.
[6] Langer A, Heilman MS, Mower MM. Considerations in the development of the automatic implantable defibrillator[J]. Medical Instrumentation, 1976, 10(3): 163-167.
[7] Ripley KL, Bump TE, Arzbaecher RC. Evaluation of techniques for recognition of ventricular arrhythmias by implanted devices[J]. IEEE Transactions on Biomedical Engineering, 1989, 36(6): 618-624.
[8] Chen S, Thakor NV, Mover MM. Ventricular fibrillation detection by a regression test on the autocorrelation function[J]. Med Biol Eng Comput, 1987, 25(3): 241-249.
[9] Thakor NV, Natarajan A, Tomselli G. Multiway sequential hypothesis testing for tachyarrhythmia discrimination[J]. IEEE Transactions on Biomedical Engineering, 1994, 41(5): 480-487.
[10] Chen SW, Clarkson PW, Fan Q. A robust detection algorithm for cardiac arrhythmia classification[J]. IEEE Transactions on Biomedical Engineering, 1996, 43: 1120-1125.
[11] Lin D, Jenkins JM, DiCarlo LA, MacDonald RS. Arrhythmia diagnosis using morphology and timing from atrial and ventricular leads[J]. Computers in Cardiology, 1988: 159-162. (September)
[12] Throne RD, Jenkins JM, DiCarlo LA. A comparison of four new time-domain techniques for discriminating monomorphic ventricular tachycardia from sinus rhythm using ventricular waveform morphology[J]. IEEE Transactions on Biomedical Engineering,1991, 38(5): 561-570.
[13] Clayton RH, Murray A, Campbell RW. Comparison of four techniques for recognition of ventricular fibrillation from the surface ECG[J]. Med Biol Eng Comput, 1993, 31: 111-117.
[14] Kuo S, Dillman R. Computer detection of ventricular fibrillation[J]. Comput Cardiol,1978: 347-349.
[15] Barro S, Ruiz R, Cabello D, Mira J. Algorithmic sequential decision-making in the frequency domain for life threatening ventricular arrhythmias and imitative artifacts: a diagnosticsystem[J]. J Biomed Eng, 1989, 11(4): 320-328.
[16] Afonso VX, Tompkoins WJ. Detecting ventricular fibrillation: selecting the appropriate time-frequency analysis tool for the application[J]. IEEE Engineering in Medicine and Biology, 1995: 152-159.
[17] Paul A, Jamie W, Gareth C, etc. Evaluating arrhythmias in ECG signals using wavelet transforms[J]. IEEE Engineering in Medicine and Biology, 2000: 104-109.
[18] Youngkyoo J, Willis JT. Detecting and Classifying Life-threatening ECG by ventricular Arrhythmias using Wavelet Decomposition[J]. Proceedings of the 25th Annual International Conference of the IEEE EMBS, Cancun, Mexico September 17-21, 2003.
[19] Dirk H, Bernd P, Hanspeter H, Ulrich Z. Nonlinear coordination of cardiovascular autonomic control[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (6): 17-21.
[20] Seidel H, Herzel H. Investigating the dynamics of atrioventricular delay[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (6): 22-25.
[21] Fojt O, Holcik J. Applying nonlinear dynamics to ECG signal processing[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (2): 96-101.
[22] Cohe ME, Hudson DL, Deedwania PC. Applying continuous chaotic modeling to cardiac signal analysis[J]. IEEE Engineering in Medicine and Biology, 1996, 15 (5): 97-102.
[23] Zhang HX, Zhu YS, Xu YH. Complexity information based analysis of pathological ECG rhythm for ventricular tachycardia and ventricular fibrillation[J]. International Journal of Bifurcation and Chaos, 2002, 12 (10): 2293-2303.
[24] Zhang XS, Zhu YS, Thakor NV. Detecting ventricular tachycardia and fibrillation by complex measure[J]. IEEE Transactions on Biomedical Engineering, 1999, 46 (5): 548-555.
[25] 李 安, 李 东. 小波分析法在心电信号数字滤波中的应用[J]. 计算机仿真, 2001, 18(6): 70-73.
[26] 王永国. 略谈心电图的干扰与抗干扰[J]. Journal of Prctical Electroc Ardiogy J S, 1996, 4(4): 104-107.
[27] 曹玉珍, 陈成, 胡勇等. 基于独立成份分析的肌电信号消噪方法研究[J]. 信号处理,2003, 19(4): 369-372.
[28] 杨福生. 小波变换的工程分析与应用[M]. 第一版, 北京: 科学出版社, 1999: 32-33.
[29] MALLAT S. A theory of multiresolution signal decomposition: The wavelet representation[J]. IEEE Trans on PAMI, 1989, 11(7): 674-693.
[30] 冷军发, 荆双喜, 李 臻. 基于小波多分辨分析的阈值去噪[J]. 焦作工学院学报(自然科版), 2002, 21(6): 443-446.
[31] 王亚, 吕新华, 王海峰. 一种改进的小波阈值降噪方法及 Matlab 实现[J]. 微计算机信息, 2006, 22(2-3): 259-261.
[32] Yan S, Kap LC, Shankar MK. Life-threatening ventricular arrhythmia recognition by nonlinear descriptor[J]. Biomedical Engineering Online, 2005, 4:6.
[33] Flandrin P. Wavelet analysis and synthesis of fractal Brownian motion[J]. IEEE Transactions on Information Theory, 1992, 38 (2): 910-917.
[34] Mandelbrot BB, VanNess JW. Fractal Brownian motion, fractional noises and applications[J]. SIAM review, 1968, 10 (4): 422-437.
[35] Mallat S. Zero-crossing of a wavelet transform[J]. IEEE Transactions on Information Theory, 1991, 37 (4): 1019-1033.
[36] Jean-Marc Bardet. Statistical study of the wavelet analysis of fractional Brownian motion[J]. IEEE Transactions on Information Theory, 2002, 48(4): 991-999.
[37] Chen G, Ren ZL, Sun HZ. Curve fitting in least-square method and its realization with Matlab[J]. O. I. Automation, 2005, 24 (3): 107-108.
[38] Wang K, Mao ZJ. Actualizing least-squares curve-fitting based on Matlab[J]. Journal of Beijing Broadcasting Institute (Science and Technology) 2005, 12 (2): 52-56.
[39]赖于树, 梁丁, 熊建等. ARM 微处理器与应用开发[M]. 北京: 电子工业出版社, 2007: 26-27.
[40] 李驹光, 聂雪媛, 江泽明等. ARM 应用系统开发详解[M]. 北京: 清华大学出版社, 2003: 2-3.
[41]赖于树, 梁丁, 熊建等. ARM 微处理器与应用开发[M]. 北京: 电子工业出版社, 2007: 33-34.
[42] 李驹光, 聂雪媛, 江泽明等. ARM 应用系统开发详解[M]. 北京: 清华大学出版社, 2003: 3-3.
[43] 杨刚, 肖宇彪, 陈江等. 32 位嵌入式系统与 SoC 设计导论[M]. 北京: 电子工业出版社, 2006.
[44] http://zlgmcu.com/home.asp.
[45] 曹茂永, 王霞, 孙农亮. 仪用放大器 AD620 及其应用[J]. 电测与仪表, 2000, 37(418): 49-52.
[46] 侯立亚, 李艳峥, 吴水才等. 基于 ARM 的心电采集系统设计[J]. 医疗卫生装备, 2006, 27(1): 1-2.
[47] http://www.alldatasheet.com/view.jsp?Searchword=CD4051. ? 2000 Fairchild Semiconductor Corporation DS005662, Revised August 2000.
[48] http://www.samsung.com/Products/Semiconductor/Flash/TechnicalInfo/datasheets.htm.
[2] 李翠微, 郑崇勋. ECG 自动分析技术的发展[J]. 国外医学生物医学工程分册, 1994, 17(1): 15-22.
[3] 宁新宝. 生物医学电子学[M]. 长沙: 湖南科学技术出版社, 1985.
[4] 刘广芝. 心电图学概论[M]. 贵州:贵州科技出版社, 2003: 234-235.
[5] 刘广芝. 心电图学概论[M]. 贵州: 贵州科技出版社, 2003: 257-257.
[6] Langer A, Heilman MS, Mower MM. Considerations in the development of the automatic implantable defibrillator[J]. Medical Instrumentation, 1976, 10(3): 163-167.
[7] Ripley KL, Bump TE, Arzbaecher RC. Evaluation of techniques for recognition of ventricular arrhythmias by implanted devices[J]. IEEE Transactions on Biomedical Engineering, 1989, 36(6): 618-624.
[8] Chen S, Thakor NV, Mover MM. Ventricular fibrillation detection by a regression test on the autocorrelation function[J]. Med Biol Eng Comput, 1987, 25(3): 241-249.
[9] Thakor NV, Natarajan A, Tomselli G. Multiway sequential hypothesis testing for tachyarrhythmia discrimination[J]. IEEE Transactions on Biomedical Engineering, 1994, 41(5): 480-487.
[10] Chen SW, Clarkson PW, Fan Q. A robust detection algorithm for cardiac arrhythmia classification[J]. IEEE Transactions on Biomedical Engineering, 1996, 43: 1120-1125.
[11] Lin D, Jenkins JM, DiCarlo LA, MacDonald RS. Arrhythmia diagnosis using morphology and timing from atrial and ventricular leads[J]. Computers in Cardiology, 1988: 159-162. (September)
[12] Throne RD, Jenkins JM, DiCarlo LA. A comparison of four new time-domain techniques for discriminating monomorphic ventricular tachycardia from sinus rhythm using ventricular waveform morphology[J]. IEEE Transactions on Biomedical Engineering,1991, 38(5): 561-570.
[13] Clayton RH, Murray A, Campbell RW. Comparison of four techniques for recognition of ventricular fibrillation from the surface ECG[J]. Med Biol Eng Comput, 1993, 31: 111-117.
[14] Kuo S, Dillman R. Computer detection of ventricular fibrillation[J]. Comput Cardiol,1978: 347-349.
[15] Barro S, Ruiz R, Cabello D, Mira J. Algorithmic sequential decision-making in the frequency domain for life threatening ventricular arrhythmias and imitative artifacts: a diagnosticsystem[J]. J Biomed Eng, 1989, 11(4): 320-328.
[16] Afonso VX, Tompkoins WJ. Detecting ventricular fibrillation: selecting the appropriate time-frequency analysis tool for the application[J]. IEEE Engineering in Medicine and Biology, 1995: 152-159.
[17] Paul A, Jamie W, Gareth C, etc. Evaluating arrhythmias in ECG signals using wavelet transforms[J]. IEEE Engineering in Medicine and Biology, 2000: 104-109.
[18] Youngkyoo J, Willis JT. Detecting and Classifying Life-threatening ECG by ventricular Arrhythmias using Wavelet Decomposition[J]. Proceedings of the 25th Annual International Conference of the IEEE EMBS, Cancun, Mexico September 17-21, 2003.
[19] Dirk H, Bernd P, Hanspeter H, Ulrich Z. Nonlinear coordination of cardiovascular autonomic control[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (6): 17-21.
[20] Seidel H, Herzel H. Investigating the dynamics of atrioventricular delay[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (6): 22-25.
[21] Fojt O, Holcik J. Applying nonlinear dynamics to ECG signal processing[J]. IEEE Engineering in Medicine and Biology, 1998, 17 (2): 96-101.
[22] Cohe ME, Hudson DL, Deedwania PC. Applying continuous chaotic modeling to cardiac signal analysis[J]. IEEE Engineering in Medicine and Biology, 1996, 15 (5): 97-102.
[23] Zhang HX, Zhu YS, Xu YH. Complexity information based analysis of pathological ECG rhythm for ventricular tachycardia and ventricular fibrillation[J]. International Journal of Bifurcation and Chaos, 2002, 12 (10): 2293-2303.
[24] Zhang XS, Zhu YS, Thakor NV. Detecting ventricular tachycardia and fibrillation by complex measure[J]. IEEE Transactions on Biomedical Engineering, 1999, 46 (5): 548-555.
[25] 李 安, 李 东. 小波分析法在心电信号数字滤波中的应用[J]. 计算机仿真, 2001, 18(6): 70-73.
[26] 王永国. 略谈心电图的干扰与抗干扰[J]. Journal of Prctical Electroc Ardiogy J S, 1996, 4(4): 104-107.
[27] 曹玉珍, 陈成, 胡勇等. 基于独立成份分析的肌电信号消噪方法研究[J]. 信号处理,2003, 19(4): 369-372.
[28] 杨福生. 小波变换的工程分析与应用[M]. 第一版, 北京: 科学出版社, 1999: 32-33.
[29] MALLAT S. A theory of multiresolution signal decomposition: The wavelet representation[J]. IEEE Trans on PAMI, 1989, 11(7): 674-693.
[30] 冷军发, 荆双喜, 李 臻. 基于小波多分辨分析的阈值去噪[J]. 焦作工学院学报(自然科版), 2002, 21(6): 443-446.
[31] 王亚, 吕新华, 王海峰. 一种改进的小波阈值降噪方法及 Matlab 实现[J]. 微计算机信息, 2006, 22(2-3): 259-261.
[32] Yan S, Kap LC, Shankar MK. Life-threatening ventricular arrhythmia recognition by nonlinear descriptor[J]. Biomedical Engineering Online, 2005, 4:6.
[33] Flandrin P. Wavelet analysis and synthesis of fractal Brownian motion[J]. IEEE Transactions on Information Theory, 1992, 38 (2): 910-917.
[34] Mandelbrot BB, VanNess JW. Fractal Brownian motion, fractional noises and applications[J]. SIAM review, 1968, 10 (4): 422-437.
[35] Mallat S. Zero-crossing of a wavelet transform[J]. IEEE Transactions on Information Theory, 1991, 37 (4): 1019-1033.
[36] Jean-Marc Bardet. Statistical study of the wavelet analysis of fractional Brownian motion[J]. IEEE Transactions on Information Theory, 2002, 48(4): 991-999.
[37] Chen G, Ren ZL, Sun HZ. Curve fitting in least-square method and its realization with Matlab[J]. O. I. Automation, 2005, 24 (3): 107-108.
[38] Wang K, Mao ZJ. Actualizing least-squares curve-fitting based on Matlab[J]. Journal of Beijing Broadcasting Institute (Science and Technology) 2005, 12 (2): 52-56.
[39]赖于树, 梁丁, 熊建等. ARM 微处理器与应用开发[M]. 北京: 电子工业出版社, 2007: 26-27.
[40] 李驹光, 聂雪媛, 江泽明等. ARM 应用系统开发详解[M]. 北京: 清华大学出版社, 2003: 2-3.
[41]赖于树, 梁丁, 熊建等. ARM 微处理器与应用开发[M]. 北京: 电子工业出版社, 2007: 33-34.
[42] 李驹光, 聂雪媛, 江泽明等. ARM 应用系统开发详解[M]. 北京: 清华大学出版社, 2003: 3-3.
[43] 杨刚, 肖宇彪, 陈江等. 32 位嵌入式系统与 SoC 设计导论[M]. 北京: 电子工业出版社, 2006.
[44] http://zlgmcu.com/home.asp.
[45] 曹茂永, 王霞, 孙农亮. 仪用放大器 AD620 及其应用[J]. 电测与仪表, 2000, 37(418): 49-52.
[46] 侯立亚, 李艳峥, 吴水才等. 基于 ARM 的心电采集系统设计[J]. 医疗卫生装备, 2006, 27(1): 1-2.
[47] http://www.alldatasheet.com/view.jsp?Searchword=CD4051. ? 2000 Fairchild Semiconductor Corporation DS005662, Revised August 2000.
[48] http://www.samsung.com/Products/Semiconductor/Flash/TechnicalInfo/datasheets.htm.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |