針對用於遠距醫療的量測裝置之心電圖特徵描繪演算法開發
No Thumbnail Available
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
隨著遠距醫療的發展，在過去的幾十年中，已有許多用於心電圖量測的可攜式裝置被研發出來。這些科技進步的成果使我們能夠在居家環境中進行心電圖量測，甚至讓操作變得更方便。然而，在這種模式下，我們無法保證訊號品質的良好。因為透過可攜式裝置量測而得的心電訊號，容易因為電極黏貼不穩造成嚴重的電極接觸雜訊干擾。此外，遠距醫療模式也使得資料可能蒐集自不同的國家，因而造成資料受到不同的電力線干擾。這些問題將導致許多心電圖特徵描繪演算法無法正常運作甚至失效。因此，本研究針對上述問題提出了一套新的特徵描繪演算法，其中包含：
1. 提出短時Goertzel演算法來偵測電力線頻率，並以零相位延遲的Butterworth陷波濾波器來處理電力線干擾。
2. 使用截止頻率為0.8Hz的零相位延遲Butterworth高通濾波器來處理心電訊號的基準線飄移。
3. 提出一個基於粗粒化流程，並配合微分、平方與移動平均的方法來處理電極接觸雜訊。
4. 改良Pan-Tompkins演算法並將其用於偵測心電訊號中的R波。
5. 改良Elgendi演算法並將其用來偵測心電訊號中的P、T波。
最後，本研究提出的這套特徵描繪演算法已被用於偵測CMATE®心電圖量測裝置所量測的資料中的心電訊號特徵。而實驗結果也證明了本研究提出的演算法的可行性與不錯的表現。
With the growth of telemedicine, kinds of portable devices for ECG measurement have been developed over the past few decades. It makes in-home ECG measurements possible and convenient. However, a good signal quality could not be always guaranteed. ECG signals measured by a portable device may be contaminated by serious electrode contact noise (ECN) due to misconnection of electrode. Besides, for ECG signals collecting from different countries, signals could be contaminated by power line interferences with different frequencies. These drawbacks make lots of conventional delineation algorithms fail to work. In this study, we propose a new delineation algorithm to overcome this difficulty. This algorithm consists of the following parts: 1. A short-time Goertzel algorithm was proposed to detect the frequency of the power line interference, and then a zero phase Butterworth notch filter is designed to attenuate the power line interference. 2. The baseline wandering is filtered by using a zero phase Butterworth high pass filter with cutoff frequency at 0.8 Hz. 3. A novel algorithm based on coarse grain, derivative, squaring and moving averaging is developed to handle the electrode contact noise. 4. A modified Pan-Tompkins’ algorithm is used to detect R peaks of ECG signals 5. Finally, P and T waves are detected by modified Elgendi’s algorithm. The proposed algorithm has been implemented to delineate ECG signals measured by CMATE® ECG devices successfully. Experimental results showed the feasibility and superiority of the proposed algorithm.
With the growth of telemedicine, kinds of portable devices for ECG measurement have been developed over the past few decades. It makes in-home ECG measurements possible and convenient. However, a good signal quality could not be always guaranteed. ECG signals measured by a portable device may be contaminated by serious electrode contact noise (ECN) due to misconnection of electrode. Besides, for ECG signals collecting from different countries, signals could be contaminated by power line interferences with different frequencies. These drawbacks make lots of conventional delineation algorithms fail to work. In this study, we propose a new delineation algorithm to overcome this difficulty. This algorithm consists of the following parts: 1. A short-time Goertzel algorithm was proposed to detect the frequency of the power line interference, and then a zero phase Butterworth notch filter is designed to attenuate the power line interference. 2. The baseline wandering is filtered by using a zero phase Butterworth high pass filter with cutoff frequency at 0.8 Hz. 3. A novel algorithm based on coarse grain, derivative, squaring and moving averaging is developed to handle the electrode contact noise. 4. A modified Pan-Tompkins’ algorithm is used to detect R peaks of ECG signals 5. Finally, P and T waves are detected by modified Elgendi’s algorithm. The proposed algorithm has been implemented to delineate ECG signals measured by CMATE® ECG devices successfully. Experimental results showed the feasibility and superiority of the proposed algorithm.
Description
Keywords
遠距醫療, 心電圖, 特徵描繪演算法, Telemedicine, Electrocardiogram, Delineation algorithm