臺灣博碩士論文加值系統

本篇提出一個可自動化心電圖判讀系統，利用2個導程Lead II及V1來判別心電圖病變，其所面臨的問題包含有如何抓取心電圖的QRS波位置和如何從心電圖辨別心律不整的種類，由於所錄製的心電圖常受到身體晃動的低頻漂移雜訊，和肌電訊號耦合的高頻雜訊影響，因此心電圖訊號會先經前置處理來達到訊號穩定性，再利用支援向量機(Support Vector Machines，SVM)做QRS波位置的辨別，以此為心電圖的波形標記基準，來擷取辨識的波形。利用自我組織學習模糊類神經網路 (Self-constructing Neural Fuzzy Inference Network，SoNFIN) 分類器做波形的辨識，包含正常節律(Normal Sinus Rhythm)、早發性心室收縮(Premature Ventricular Contraction)、左束枝傳導阻滯(Left Bundle Branch Block)、右束枝傳導阻滯(Right Bundle Branch Block)等4種，研究中所採用的心電圖來自MIT-BIH Arrhythmia資料庫，共選用了29個病患的資料，在QRS波位置辨識正確率達99.8%，在心電圖的心律不整波形辨識率亦可達到98.8%。

In this thesis, an automated system was built for electrocardiogram (ECG) arrhythmias classification that use ECG’s Lead II and V1 signals as input. Two problems must be overcome how to pick up ECG’QRS position and how to classify different arrhythmias. The ECG signals often coupled the different noises, like as the body motion causing the based line shift, or the electromyogram causing the high frequency noise. Therefore, the ECG signals will pass through the pre-processing to achieve the signal stability, first. Support vector machine was used to determine the QRS wave position. Self-constructing neural fuzzy inference network (SoNFIN) was used to classify 4 different arrhythmias including Normal Sinus Rhythm, Premature Ventricular Contraction, Left Bundle Branch Block, Right Bundle Branch Block. The ECG signals taken from MIT-BIH arrhythmia database. There are 29 patients to test the performance of our proposed system. The results show that the accuracy could achieve 99.8% for determining the QRS wave position, and 98.8% for classifying the arrhythmias’ types.

博、碩士論文授權書................................................................................I
論文口試委員審定書(中文).....................................................................II
論文口試委員審定書(英文)....................................................................III
中文摘要.................................................................................................IV
Abstract......................................................................................................V
致謝..........................................................................................................VI
圖目錄......................................................................................................IX
第一章 緒論..............................................................................................1
1.1 研究背景.......................................................................................1
1.2 研究動機與目的...........................................................................9
第二章 分類器介紹................................................................................11
2.1 支援向量機(SVM).....................................................................11
2.1.1線性SVM...........................................................................11
2.1.2非線性SVM.......................................................................16
2.2自我組織學習模糊類神經網路分類器(SoNFIN).....................19
2.2.1 SoNFIN架構......................................................................19
2.2.2 TSK型模糊類神經網路之線上學習演算法....................22
第三章 研究方法....................................................................................28
3.1 心電圖資料介紹.........................................................................28
VIII
3.1.1 資料庫...............................................................................28
3.1.2 前置處理...........................................................................34
3.2支援向量機(SVM)......................................................................37
3.3自我組織學習模糊類神經網路分類器(SoNFIN).....................39
第四章 實驗結果....................................................................................44
4.1 數位濾波.....................................................................................44
4.2支援向量機(SVM)......................................................................45
4.2自我組織學習模糊類神經網路分類器(SoNFIN).....................51
第五章 討論與結論................................................................................57
參考文獻..................................................................................................64
圖1 (a)心臟構造(b)心電圖波形產生.......................................................2
圖2 波形各區段時間長度.......................................................................2
圖3 肢導程位置.......................................................................................4
圖4 胸導程位置.......................................................................................5
圖5 心電圖雜訊干擾...............................................................................6
圖6可分為2類示意圖..........................................................................11
圖7 不可分為2類示意圖.....................................................................15
圖8 非線性函數示意圖.........................................................................16
圖9 函數轉換示意圖.............................................................................17
圖10 SVM架構圖....................................................................................17
圖11 TSK型模糊類神經網路的架構圖.................................................20
圖12學習演算法示意圖........................................................................23
圖13 法則產生流程圖...........................................................................23
圖14 線上學習演算法之虛擬碼...........................................................24
圖15 線上學習演算法之流程圖...........................................................27
圖16 方法架構圖...................................................................................28
圖17 正常節律心電圖(a)Lead II(b)V1..................................................31
圖18 早發性心室收縮心電圖(a)Lead II(b)V1......................................32
圖19 左束枝傳導阻滯(a)Lead II(b)V1..................................................33
圖20 右束枝傳導阻滯(a)Lead II(b)V1..................................................34
圖21 頻率響應圖(a)dB(b)相位..............................................................35
圖22 可適性濾波器的示意圖...............................................................35
圖23 抓取波形示意圖...........................................................................37
圖24 SVM標示示意圖............................................................................38
圖25合併標記流程圖............................................................................39
圖26 RR區間消除標記流程圖...............................................................39
圖27 正規化示意圖...............................................................................41
圖28 SoNFIN標記示意圖.......................................................................42
圖29心電圖濾波(a)原始信號(b)低通濾波(c)高通濾波.......................45
圖30 SVM分類結果................................................................................47
圖32 No.221成功分類............................................................................58
圖33 No.214成功分類............................................................................58
圖34 No.231成功分類............................................................................59
圖35 Sonfin誤判圖..................................................................................60
圖36 消除SVM多餘標記.......................................................................62
表1各種干擾雜訊....................................................................................7
表2 kernel函數........................................................................................19
表3 原始波形總數.................................................................................30
表4 SVM標示錯誤總數..........................................................................48
表5 SVM標示錯誤個數總計..................................................................49
表6 SVM標示錯誤統計..........................................................................49
表7 SVM後置處理減少誤標數目..........................................................49
表8 SVM標記錯誤與候後置處理總表..................................................50
表9 SoNFIN訓練檔每個種類個數.........................................................51
表10 SoNFIN訓練檔每一個種類統計...................................................51
表11 SoNFIN測試檔案種類個數...........................................................52
表12 SoNFIN測試檔案種類統計準確率...............................................52
表13 SoNFIN門檻值來減少數目個數...................................................52
表14 QRS波統計個數.............................................................................53
表15 QRS波準確率.................................................................................53
表16 SVM與SoNFIN錯誤個數總表......................................................54
表17 SoNFIN QRS波統計個數..............................................................55
表18 SoNFIN QRS波準確率..................................................................55
表19 SVM分類個數................................................................................56

[1]心臟電氣生理學檢查, http://www.cgmh.org.tw/
[2]心電圖原理簡介, http://designer.mech.yzu.edu.tw/
[3]M. Okasa, “A digital filter for the QRS complex detection,” IEEE Transactions on Biomedical Engineering, 1979, Vol. BME-26, No. 12, pp.700-703.
[4]J. Pan and W. J. Tompkins, “A real-time QRS detection algorithm,” IEEE Transactions on Biomedical Engineering, 1985, Vol. BME-32, No. 3, pp.230-236.
[5]G.M. Friesen, T.C. Jannett, M.A. Jadallah, S. Yates, S.R. Quint and H. Troynagle, “A comparison of the noise sensitivity of nine QRS detection algorithms,” IEEE Transactions on Biomedical Engineering, 1990, Vol.37, No. 1, pp.85-98.
[6]P.E. Trahanias, “An approach to QRS complex detection using mathematical morphology,” IEEE Transactions on Biomedical Engineering, 1993, Vol. 40, Issue 2, pp.201-205.
[7]Y.C. Yeh and W.J. Wang, “QRS complexes detection for ECG signal: the difference operation method,” Computer Methods and Programs in Biomedicine, 2008, Vol. 91, Issue 3, pp. 245-254.
[8]S.S. Mehta and N.S. Lingayat, “Identification of QRS complexes in 12-lead electrocardiogram,” Expert Systems with Applications, 2009, Vol.36, Issue 1, pp.820-828.
[9]S.S. Mehta, D.A. Shete, N.S. Lingayat and V.S. Chouhan, “K-means algorithm for the detection and delineation of QRS-complexes in electrocardiogram,” IRBM, 2010, Vol. 31, Issue 1, pp. 48-54.
[10]M.G. Tsipouras, D.I. Fotiadis and D. Sideris, “An arrhythmia classification system based on the RR-interval signal,” Artificial Intelligence in Medicine, 2005, Vol.33, Issue 3, pp.237-250.
[11]I. Guler and E.D. Ubeyli, “Multiclass support vector machines for EEG-signals classification,” IEEE Transactions on Information Technology in Biomedicine, 2007, Vol.11, No.2, pp.117-126.
[12]B.M. Asl, S.K. Setarehdan and M. Mohebbi, “Support vector machine-based arrhythmia classification using reduced features of heart rate variability signal,” Artificial intelligence in medicine, 2008, Vol.44, Issue 1, pp. 51-64.
[13]M. Moavenian and H. Khorrami, “A qualitative comparison of artificial neural networks and support vector machines in ECG arrhythmias classification,” Expert Systems with Applications, 2010, Vol. 37, Issue 4, pp. 3088-3093.
[14]C.J.C Burges. “A tutorial on support vector machines for pattern recognition,” Data Mining and Knowledge Discovery, 1998, Vol.2, pp.955-974.
[15]N. Cristianini and J. Shawf-Taylor, “An introduction to support vector machines and other kernel-based learning methods,” Cambridge University press, 2000.
[16]B. Scholkopf, C.J.C. Burges and A.J. Smola, “Introduction to support vector learning advances in kernel methods-support vector learning,” Cambridge, MA, MIT press, 1999, pp.1-15.
[17]V. Vapnik, “Statistical learning theory,” Wiley, 1998.
[18]J.N. Wang, “A study of multiclass support vector machines,” Master Thesis, Department and Graduate Institute of Information and Management, 2003.
[19]R. Fletcher, “Practical methods of optimization,” John Wiley and Sons, Inc., 2nd edition, 1987.
[20]C. Cortes and V. Vapnik, “Support vector network,” Machine Learning, 1995, Vol.20, pp.273-297.
[21]B.E. Boser, I.M. Guyon, and V. Vapnik, “A training algorithm for optimal margin classifiers,” Fifth Annual Workshop on Computational Learning Theory ACM, 1992, pp. 144-152.
[22]R. Courant and D. Hiber, “Methods of mathematical physics,” Wildy-Interscience, 1953.
[23]P.O. Shih, “SoNFIN for physical activities classification with accelerometer,” Master Thesis, Department and Graduate Institute of Biomedical Engineering and Material Science, 2009.
[24] MIT-BIH Arrhythmia Database,http://www.physionet.org/
[25]達楷生醫科技股份有限公司, "基本心電圖圖例簡介," 2006, http://www.biosensetek.com/
[26]B. Widrow, J. R. Glover, J. M. Mccool, J. Kaunttz, C.S. Williams, R. H. Hean, J.R. Zeidler, E. Dong and R.C. Goodlin, “Adaptive noise cancelling: principles and applications,” Proceedings of The IEEE, 1975, Vol. 63, No. 12,pp.1692-1716.
[27]S.K. Mitra, “Digital signal processing: a computer-based approach,” 3E, McGraw-Hill, 2006.
[28]LIBSVM-A Library for Support Vector Machines, http://www.csie.ntu.edu.tw/~cjlin/libsvm/