基于Levenberg-Marquardt算法的心电高斯模型研究
Levenberg-Marquardt Algorithm-Based Gaussian Model Research for ECG
摘要: 针对心电信号中各构成成分(P波、QRS波、T波)波形特征，本研究提出了一种基于伯格–马夸尔特算法创建心电高斯模型以实现心电波形特征研究进而对心电特征进行定性评价研究。首先，基于小波分解对心电信号进行预处理，以消除基线漂移以及背景噪声；其次，以心电信号各R峰间距为参考，按照1:2的分割比例对心电信号进行周期分割；最后，基于Levenberg-Marquardt (LM)算法创建心电高斯混合模型以刻画各成分波形特征。本算法通过对MIT-BIH心律失常数据库中48条心电数据进行分析，利用其拟合优度均值为0.9668对本算法可行性进行了验证。
Abstract: Aimed at the waveform characteristics of main components (P wave, QRS complex and T wave) to an electrocardiogram (ECG), this study proposes a novel method based on Levenberg-Marquardt algorithm (LMA) to build the Gaussian model for researching ECG signal, and then to character the features of each component. Firstly, the wavelet decomposition-based preprocessing method is employed to cancel baseline shift and background noise existing in ECG signal. Secondly, ECG signal is segmented into periodic signal based on the distance between tow sequence R peaks as well its ratio of 1:2. Finally, LMA-based Gaussian Mixture model is built to character each component waveform. The performance of the proposed method has been evaluated by using the 48 ECG records of MIT-BIH arrhythmia database, and the higher average of R-square value 0.9668, used to measure the goodness of fit, has been achieved.
文章引用：吴杰, 黄婷婷, 张弼强, 刘保进, 宋伟. 基于Levenberg-Marquardt算法的心电高斯模型研究[J]. 计算机科学与应用, 2019, 9(10): 1946-1954. https://doi.org/10.12677/CSA.2019.910218

参考文献

[1] 胡盛寿, 等. “中国心血管病报告2018”概要[J]. 中国循环杂志, 2019, 34(3): 209-220.
[2] Dima, S.M., Panagiotou, C., Mazomenos, E.B., et al. (2013) On the Detection of Myocadial Scar Based on ECG/VCG Analysis. IEEE Transac-tions on Biomedical Engineering, 60, 3399-3409. [Google Scholar] [CrossRef] 
[3] Kutlu, Y. and Kuntalp, D. (2012) Feature Extraction for ECG Heartbeats Using Higher Order Statistics of WPD Coefficients. Computer Methods and Programs in Biomedicine, 105, 257-267. [Google Scholar] [CrossRef] [PubMed]
[4] Yang, H., Kan, C., Liu, G. and Chen, Y. (2013) Spatio-temporal Differentiation of Myocardial Infarctions. IEEE Transactions on Automation Science and Engineering, 10, 938-947. [Google Scholar] [CrossRef] 
[5] El Bouny, L., Khalil, M. and Adib, A. (2018) ECG Signal Filtering Based on CEEMDAN with Hybrid Interval Thresholding and Higher Order Statistics to Select Relevant Modes. Multimedia Tools and Applications, 78, 13067-13089. [Google Scholar] [CrossRef] 
[6] De Lannoy, G., François, D., Delbeke, J. and Verleysen, M. (2012) Weighted Conditional Random Fields for Supervised Interpatient Heartbeat Classification. IEEE Transactions on Biomedical Engineering, 59, 241-247. [Google Scholar] [CrossRef] 
[7] 陈志博, 李智, 李健, 等. 基于RR间期和多特征值的房颤自动检测分类[J]. 生物医学工程学杂志, 2018, 35(4): 550-556.
[8] Mondéjar-Guerra, V., Novo, J., Rouco, J., Penedo, M.G. and Ortega, M. (2019) Heartbeat Classification Fusing Temporal and Morphological Information of ECGs via Ensemble of Classifiers. Biomedical Signal Processing and Control, 47, 41-48. [Google Scholar] [CrossRef] 
[9] Zhou, Q.Y., Liu, X. and Duan, H.L. (2006) ECG Beat Classification Using Mirrored Gauss Model. 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference, Shanghai, 17-18 January 2006, 5587-5590. [Google Scholar] [CrossRef] 
[10] 李文涛. 心电信号噪声处理的研究[J]. 数字技术与应用, 2016(8): 85-86.
[11] Banerjee, S., Gupta, R. and Mitra, M. (2012) Delineation of ECG Characteristic Features Using Multiresolution Wavelet Analysis Method. Measurement: Journal of the International Measurement Confedera-tion, 45, 474-487. [Google Scholar] [CrossRef] 
[12] 李颖, 林洪生. 基于相对误差的曲线最小二乘拟合[J]. 沈阳师范大学学报, 2012, 30(3): 338-342.
[13] Wright, S.J., Goldberg, A.V., Johnson, D.S., et al. (2000) Trust-Region Methods. Elsevier, Amsterdam.