臺灣博碩士論文加值系統

本研究目的在於建立一套心搏辨識輔助診斷系統，使能提供不正常心律之判斷基準，輔助醫師診斷疾病。傳統心電圖機僅能觀測特定時間之電氣活動，瞬息發生的心律不整不一定能被完整記錄。因此，開發即時辨識系統渴望成為上述問題之解決方法。
本系統可分為訊號擷取、R點偵測、特徵擷取、類神經網路分類及辨識結果呈現。訊號擷取使用廣泛被認定之MIT-BIH資料庫測試，搭配資料擷取卡將訊號傳送到個人電腦分析。R點偵測方面，使用帶通濾波移除對訊號前處理，再將訊號透過疊代法估測峰值及真實R點搜尋機制將QRS波定位，離線偵測法則之正確率為96.64%，若加以簡化之後，即時偵測法則的平均為94.07%，雖然即時偵測法的正確率略為下降，但偵測一個R點平均時間只需0.00038秒。特徵擷取使用高階統計以描述小波轉換後之特定次頻帶，並透過適當次頻帶選擇可以濾除雜訊。類神經網路分類器的權重值訓練方式使用三種不同比例的訓練樣本組合，目的在於提升樣本較少的疾病類型在分類時的靈敏度，調整後之最佳訓練樣本辨識率為98.48%。此外，我們使用三種測試方法來驗證類神經網路效能，並檢視分類結果之可信度。最具代表性之測試方法搭配最佳樣本組合得到98.33%之疾病辨識率。辨識結果的呈現分為離線與即時系統測試，離線系統為實現即時系統之前身，負責類神經網路權重值訓練，以用於即時系統。最後結果，於離線模擬測試下，整體辨識率平均達到98.63%，即時系統測試之整體辨識率平均為98.34%，證實系統用於即時辨識之效能與可行性。

In this thesis, we proposed an assistive diagnosis system for ECG beat recognition. The system aimed to establish a fundamental recognition system for beats, hoping to provide assistive diagnosis for the physician. Conventional methods for arrhythmia detection were usually based on standard resting ECG. However, it could only provide a snapshot of the patient’s cardiovalscular activity in time and was difficult to make reliable diagnosis for transient arrhythmias. Therefore, one of the possible solutions for the problem is to perform real-time diagnosis.
The system consisted of five parts, including data acquisition, QRS detection, feature extraction, classification, and result display. To acquire ECG signals, we employed on NI DAQ card, which converted analog signal to digital signal for further data processing. QRS detection was a crucial part of the entire system, because the result of feature extraction was strongly affected by the locations of the QRS complexes . A sequence of processing steps include band-pass filters , which rejected 60Hz interference and baseline wander, fiducial mark localization, iterative thresholds computation, and adjustment of real QRS complexes . Using the MIT-BIH database for verification, the algorithm achieved the R-peak detection rate of 96.64% for offline analysis and 94.07% for a simplified version which is designed for the requirement of short detection time. Despite a slight decrease in the detection rate, the recognition time for the real-time algorithm was only 0.00038sec per beat, which met the requirement of the real-time system. The discrete wavelet transform was then employed to decompose the ECG signal in order to appropriately select subband components to increase noise rejection and improve beat discrimination. Due to its extremely short processing time in classifier testing, the feed-forward back-propagation neural network was exploited to perform beat classification. To elevate the sensitivity of arrhythmias which were of small training size, different profiles designated as 1st, 2nd, and 3rd profile were considered when training the weight of the FFBNN. When evaluating the proposed method, we designed three trials to compare the influence of different sample selection strategies. Finally, we developed two platforms, namely offline platform and real-time GUI platform, to demonstrate the feasibility of the proposed system. Prior to the real-time GUI platform, offline platform was utilized to develop the initial prototype and adjust the weight of the neural network. This system achieved a recognition rate of 98.63% with the offline platform, and 98.34% for the real-time analysis. The outstanding performance demonstrates the effectiveness and efficiency for the system.

中文摘要 I
Abstract III
目錄 V
圖目錄 VIII
表目錄 X
第一章 緒論 1
1.1 研究動機 1
1.2 文獻回顧 2
1.3 論文架構 3
第二章 研究方法與背景 4
2.1 心電圖原理 4
2.1.1 心臟傳導系統 4
2.1.2 心電圖訊號量測 5
2.1.3 心臟疾病簡述 6
2.2 理論背景 8
2.2.1 小波轉換 8
2.2.2 高階統計 10
2.2.3 倒傳遞類神經網路 13
2.3 訊號擷取概述 15
2.3.1 訊號擷取系統 15
2.3.2類比輸入 16
第三章 實驗流程 18
3.1 類比訊號擷取 18
3.1.1 訊號擷取流程 19
3.2 數位訊號 23
3.2.1 數位訊號取得 23
3.2.2 資料庫 25
3.3 R點偵測 25
3.3.1 離線R點偵測 25
3.3.2 即時R點偵測 29
3.4 特徵擷取 30
3.4.1 小波轉換後的高階統計特徵 30
3.4.2 RR間隔相關特徵 31
3.4.3 特徵正規化 32
3.5 疾病分類 32
3.5.1 分類樣本選取 32
3.5.2 分類 38
3.6 系統模擬 40
第四章 實驗結果與討論 42
4.1 R點偵測 42
4.1.1 離線偵測 42
4.1.2即時偵測 43
4.1.3 偵測結果分析 44
4.2 疾病辨識 47
4.2.1 類神經網路效能評估 47
4.2.2 樣本選取對辨識率的影響 48
4.2.2.1 1st Profile與2nd Profile 51
4.2.2.2 2nd Profile與3rd Profile 52
4.2.2.3 測試樣本選取比較 53
4.2.2.4 R點調整對辨識率的影響 54
4.3 R點偵測與疾病辨識整合 57
4.3.1 離線系統測試 57
4.3.2 即時系統測試 58
第五章 結論與未來展望 61
5.1 結論 61
5.2 未來展望 62
參考文獻 63

[1] J. J. Oresko, Z. Jin, J. Cheng, S. Huang, Y.Sun, H. Duschl, and A.C. Cheng, “A Wearable Smartphone-Based Platform for Real-Time Cardiovascular Disease Detection Via Electrocardiogram Processing”, IEEE Transactions on Information Technology Biomedicine, vol. 14, pp. 734-740, 2010.
[2] W. F. Zhu, H. M. Zhao and X. P. Chen, “A new QRS Detector Based on Empirical Mode Decomposition”, IEEE 10th International Conference on Signal Processing, pp. 1757-1760, 2010.
[3] J. Pan and W.J. Tompkins, “A Real-time QRS Detection Algorithm”, IEEE Transactions on Biomedical Engineering, vol.32, pp.230-235, 1985.
[4] V.X. Afonso, W.J. Tompkins, T.Q. Nguyen, and S. Luo, “ECG Beat
Detection Using Filter Banks”, IEEE Transactions on Biomedical Engineering., vol.46, pp.192-201,1999.
[5] C.W. Li, C.X. Zheng, C.F. Tai, “Detection of ECG Characteristic Points
Using Wavelet Transforms”, IEEE Transactions on Biomedical Engineering, vol.42,
pp.21-28, 1995.
[6] Y.Z. Qiu, X.F. Ding, J. Feng, and Z.W. Mo, “QRS Complexes Detection
Based on Mexcian-hat Wavelet”, Journal of Biomedical Engineering, vol.23, pp.1347-1349, 2006.
[7] D. Benitez, P.A. Gaydecki, A. Zaidi, and A.P. Fitzpatrick, “The Use of
the Hilbert Transform in ECG Signal Analysis”, Computers in Biology and Medicine, vol.31, pp.399-406, 2001.
[8] Y.L Chen, “Algorithm research on QRS detection of ambulary ECG
automatic analysis”, Doctoral Dissertations, Zhejiang University, 2006.
[9] S.S. Mehta, N.S. Lingayat, “SVM-based algorithm for recognition of
QRS complexes in electrocardiogram”, IRBM, vol.29, pp.310-317, 2008.
[10] T. Ince, S. Kiranyaz, and M. Gabbouj, “A Generic and Robust System for Automated Patient-Specific Classification of ECG Signals”, IEEE Transactions on Biomedical Engineering, vol.56, pp.1415-1426, 2009.
[11] X. Alfonso and T. Q. Nguyen, “ECG beat detection using filter banks” ,IEEE Transactions on Biomedical Engineering, vol. 46, pp. 192–202, 1999.
[12]J. L. Talmon, Pattern Recognition of the ECG. Berlin, Germany: Akademisch Proefscrift, 1983.
[13] Y. H. Hu, W. J. Tompkins, J. L. Urrusti, and V. X. Afonso, “Applications
of artificial neural networks for ECG signal detection and classification” ,
J. Electrocardiol., vol. 26, pp. 66–73, 1994.
[14] Y. Hu, S. Palreddy, and W. J. Tompkins, “A patient-adaptable ECG beat
classifier using a mixture of experts approach” , IEEE Transactions on Biomedical Engineering, vol. 44, pp. 891–900, 1997.
[15] D. A. Coast, R. M. Stern, G. G. Cano, and S. A. Briller, “An approach to
cardiac arrhythmia analysis using hidden Markov models” , IEEE Transactions on Biomedical Engineering, vol. 37, pp. 826–836, 1990.
[16] 梅約醫學中心，“心臟小百科”，天下生活出版
[17] http://academic.ntsu.edu.tw/ezcatfiles/b004/img/img/227/k8.pdf
[18] http://csm.jmu.edu/biology/danie2jc/heart.htm
[19] 劉芳岑，“以次頻帶分解法進行雙導程心電圖搏動辨識之研究”，中正大學電機所99年
[20] 中山醫學大學附設醫院急診醫學部， “十二導程心電圖操作講義”
[21] S.N. Yu and Y.H. Chen, “Noise-tolerant electrocardiogram beat classification
based on higher order statistics of subband components,” Artificial Intelligence in
Medicine, vol. 46, pp. 165-178, 2009.
[22] 陳書鋒，“基於心電圖及皮膚電活動辨識情緒的研究”，中正大學電機所99年
[23] Mallat, S. G., “A Theory for Multiresolution Signal Decomposition: The
Wavelet Representation”, Pattern Analysis and Machine Intelligence, IEEE
Transactions on, Vol. 11, pp. 674-693,1989.
[24] Shensa, M. J., “The Discrete Wavelet Transform: Wedding the à trous and
Mallat Algorithms”, IEEE Transactions on Signal Processing, pp. 2464-2482, Vol. 40, 1992.
[25] 電腦虛擬儀器圖形編程LabVIEW 實驗教材¸北京中科泛華測控技術有限公司
[26] I. Jekova, G. Bortolan, and I. Christov, “Assessment and comparison of different methods for heartbeat classification,” Medical Engineering & Physics, vol. 30, pp. 248-259，2008.
[27] 數位訊號處理，宏友圖書出版，2006
[28] Physiobank Archive Index, MIT-BIH Arrhythmia Database:
http://www.physionet.org/physiobank/database/
[29] S. N. Yu and Y. H. Chen, “Electrocardiogram beat classification based on wavelet transformation and probabilistic neural network, ” Pattern Recognition Letters, vol. 28, pp. 1142-1150, 2007.
[29] S.Osowski, L.T. Hoai, and T.Markiewicz, “Support vector machine-based expert system for reliable heartbeat recognition”, IEEE Transactions on Biomedical Engineering, vol.51, pp.582-589, 2004.
[30] M. Cvikl, A. Zemvab, “FPGA-oriented HW/SW implementation of ECG beat detection and classification algorithm”, Digital Signal Processing, Elsevier, pp.238-248, 2010.