心電訊號特徵擷取與心律不整判斷系統之研究

The ECG features detection and arrhythmia classification system

顏紹雍(Shao-Yung Yen)

指導教授 ： 江行全

元智大學/工程學院/工業工程與管理學系/碩士(2007年)

https://doi.org/10.6838/YZU.2007.00210

摘要

醫師與醫護人員通常使用心電圖(Electrocardiogram, ECG)來判讀心臟異常訊息，此工作之執行經常耗費大量的人力資源與時間。為了能協助專業醫療人員，即時判斷病患是否患有心律不整之可能性。本研究發展一演算系統來擷取心電訊號中的QRS複合波型，結合臨床病理的心律不整判斷準則，完成一個能夠將儀器所蒐集得之心電訊號資料作為輸入，將病患是否有心律不整之異常作為輸出之監測診斷系統。對於一般較規律且雜訊較少的心電圖而言，以斜率理論為基礎的演算方法有不錯的偵測效果，但對於雜訊影響較大的心電圖，仍然有不少誤判的情形。因此本研究針對以斜率為基礎之演算法提出修正，增加向前及向後搜尋資料並刪除振幅過小波型之規則以改進其缺失。以MIT-BIH心律不整資料庫做為訓練資料以設定參數，同時使用實務上之安心卡資料做驗證及測試。將演算法偵測所獲得之結果透過文獻所提出之心律不整判斷規則，進行心律不整異常之判讀。由於受到特徵擷取誤判的影響，後續分類結果所產生之誤判為將正常患者誤判成有心律不整之情形。此類誤判在MIT-BIH資料庫的27筆資料中有1筆，分類正確率為96.3%；在安心卡的52筆資料中有5筆，分類正確率為90.38%。

關鍵字

心電圖 (ECG),QRS複合波型,MIT-BIH資料庫,心律不整分類

並列摘要

Automated diagnostic system has become an established component of medical technology. The main concept of the medical technology is an inductive engine that learns the decision characteristics of the diseases and then can be used to diagnose future patients with uncertain diseases states. Electrocardiogram (ECG) is an important tool in diagnosing the condition of the heart. It provides valuable information about the functional aspects of the heart and cardiovascular system. This research develops an algorithm to extract the QRS complex wave and combines the clinical judgment criterion of the cardiac arrhythmia to model a classification system. The system input data is ECG signal and output is the classification of cardiac arrhythmia. The method is forward-backward algorithm that is revised So and Chan method and augments forward and backward searching rules and also deletes lower R amplitude to improve the detection performance. The ECG signals are taken form MIT-BIH arrhythmia database, which are used to classify 4 different arrhythmias for training. There are normal, premature ventricular contraction, ventricular flutter/fibrillation and 2o heart block (Tsipouras, 2005). Regard MIT-BIH arrhythmia database as training data in order to set up the parameters, verifying and testing by real medical ECG data. According to QRS complex extraction and clinical judgment criterion of the cardiac arrhythmia, the proposed approach provided 96.3% accuracy of classification on MIT-BIH database and 90.38% accuracy of classification on empirical medical ECG data.

並列關鍵字

Electrocardiogram (ECG), QRS complex, forward-backward algorithm, MIT-BIH database, cardiac arrhythmia classification

參考文獻

[28] 陳建德， “以斜率為基礎建立偵測心電圖R波之研究” 私立元智大學工業

[1] Afonso, V.X., Tompkins, W.J., Nguyen, T.Q., and Luo, S. (1999). “ECG beat detection using filter banks,” IEEE Transactions on Biomedical Engineering, vol. 46, pp. 192-202

[2] Biag., M.H., Rasool, A., and Bhatti, M.I. (2001). “Classification of electrocardiogram using SOM, LVQ and beat detection methods in localization of cardiac arrhythmias,” Procedings of the 23rd annul EMBS international conference, pp. 25-28

[3] Chazal, P.D., O’Dwyer, M., and Richard B.R. (2004). “Automatic classification of heartbeats using ECG morphology and heartbeat interval features,” IEEE Transactions on Biomedical Engineering, vol. 51, pp. 1196-1206

[4] Christov, I., G`erman, G.H., Vessela K., Irena, J., Atanas, G., and Karen, E. (2005). “Comparative study of morphological and time-frequency ECG descriptors for heartbeat classification” Medical Engineering and Physics, vol.28, pp. 876-887