以Fuzzy C-Means法辨識心電圖的心跳類別

Heartbeat Case Determination Using the Fuzzy C-Means (FCM) Method on ECG Signals

郭建得(Jen-Der Guo)

指導教授 ： 葉雲奇

健行科技大學/電資學院/電子工程系碩士班/碩士(2015年)

https://doi.org/10.6824/UCH.2015.00013

摘要

本論文提出一個簡單且有效的心電圖(ECG)心跳種類辨識法，它包含了：(1)ECG信號的前置處理器：目的是放大從病患身上所取到的ECG信號，並做各種雜訊的去除處理；(2)ECG信號的傳送：將處理後的ECG信號以Wi-Fi無線通訊技術傳送到接收器；(3)計算原始特徵點的特徵值：依據從Wi-Fi接收器收到之ECG信號中的QRS複合波、P波及T波位置，計算各原始特徵點的特徵值；(4)主要特徵點的選取：以主成份分析法(Principal Component Analysis ; PCA)選取主要特徵點，目的是減少心跳種類的辨識時間；(5)心跳種類的辨識：以模糊分群平均(Fuzzy C-Means)法辨識心臟病患的心跳種類，本論文能辨識五種較常發生的心跳種類，包含正常的心跳(NORM)及四種不正常的心跳。四種不正常的心跳分別為：左束分支阻斷(LBBB)、右束分支阻斷(RBBB)、心室過早收縮(VPC)、及心房過早收縮(APC)等。最後，本論文以MIT-BIH心律不整資料庫中相關檔案來評估所提出方法的效能，經實際的測試，辨識心跳類別NORM，LBBB，RBBB，VPC，及APC的Se分別可達98.28%，90.35%，86.97%，92.19%，及94.36%。總平均正確判斷率TCA為93.57%。

關鍵字

心電圖(ECG) ； Wi-Fi無線通訊技術 ； QRS複合波 ； 模糊分群平均法(Fuzzy C-Means Method) ； MIT-BIH心律不整資料庫。

並列摘要

This paper presents a simple and effective electrocardiogram (ECG) heartbeat species identification method, which includes: (1) ECG signal pre-processor: the aim is to enlarge the body taken from the patient to the ECG signal, and do all kinds of miscellaneous information removal process; (2) ECG signal transmission: the post-processing of the ECG signal to Wi-Fi wireless communication technology is transferred to the receiver; and (3) calculation of the original feature points feature value: according to the received Wi-Fi receiver ECG signal to the middle of the QRS complex, P T wave spread position, a characteristic feature of the original value of each point; selection (4) the main features of the point: the principal component analysis (Principal Component Analysis; PCA) to select the main feature points the aim is to reduce the time the heartbeat species identification; (5) the heartbeat species identification: fuzzy clustering average (Fuzzy C-Means) method to identify the type of cardiac patients heartbeat, the heartbeat of this paper can identify five species occur more frequently, contain normal heartbeat (NORM) and four kinds of abnormal heartbeat. Four kinds of irregular heartbeat were: a left bundle branch block (LBBB), right bundle branch block (RBBB), ventricular premature contractions (VPC), and atrial premature contraction (APC) and so on. Finally, this paper MIT-BIH arrhythmia database related files to assess the effectiveness of the proposed method, the actual testing, identification heartbeat category NORM, LBBB, RBBB, VPC, and APC's Se were up 98.28%, 90.35 %, 86.97%, 92.19%, and 94.36%. The total average rate of correct judgment TCA was 93.57%.

並列關鍵字

Electrocardiogram (ECG) ； Wi-Fi wireless technology, ； QRS complex wave ； Fuzzy C-Means Method (FCMM) ； MIT-BIH arrhythmia database.

參考文獻

[1]黃蓮奇譯，心臟循環系統，台灣光復書局，1996。

[2]楊杞柳譯，心電圖判讀之快速學習法，財團法人徐氏基金會，1991。

[3]楚萃瑤，在馬達的電流波形上使用Fisher’s線性鑑別分析法辨識馬達的品質類別。清雲科技大學電子工程研究所碩士論文，2014。

[4]G. Vijaya, V. Kumar, and H. K. Verma, ANN-based QRS-complex analysis of ECG, J. Med. Eng. Technol., vol. 22, no. 4, pp.160-167, 1998.

[5]S. Kadambe, R. Murray, and G. F. Boundreaux-Bartels, Wavelet transform-based QRS complex detector, IEEE Trans. Biomed. Eng., vol. 46, no. 6, pp.838-848, 1999.