心律不整心電圖訊號之 QRS 複合波群波偵測演算法

Abstract

目前心電圖(ECG)在判斷心臟病心律不整是非常有用的工具，心電圖自動 檢測演算法是目前的學術研究的主流方向，由於 QRS 複合波群是左右心室的快速 去極化的過程，而左右心室的肌肉組織比心房發達，使 QRS 複合波群的振幅比 P 波大，因此心電圖的自動檢測最首要的是找出 QRS 複合波群，得到 QRS 複合波群 後利用 R-R 區間來分析心率變異率(HRV)，透過心率變異率來檢測出心律不整的狀況，並確診是否有心臟相關的疾病。 本論文提出一種有別於 Pan-Tompkins 的 QRS 複合波群檢測演算法，且能夠可 靠性高的實時運行於嵌入式裝置。首先對心電圖訊號進行前處理後，利用訊號的 梯度與閾值，檢測出 QRS 複合波群的特徵，並依據 QRS 複合波群形態學來判別特徵。 最後在本論文中開發一套使用者介面檢測工具，整合幾種基於一階微分方法 的 QRS 複合波群檢測演算法，方便進行演算法的評估與測試，最後由 MIT-BIH 心 律不整資料庫作為驗證此演算法的標準，並且驗證本論文提出的演算法，符合美 國心臟協會的標準規範，達到 93.21%的靈敏度以及 89.28%的精確度。

At present, electrocardiogram (ECG) is a very useful tool in identifying arrhythmia. ECG automatic detection algorithm has now become the mainstream in academic research. The QRS complex corresponds to the rapid depolarization of left and right ventricles. The left and right ventricles are more muscular than the atrium, thus, the amplitude of QRS complex is larger than the P-wave. Therefore, the primary goal of ECG Automatic Detection is to identify the QRS complex, and then the heart rate variability (HRV) analysis can be performed using the R-R interval in QRS complex. The HRV can be used to detect arrhythmia and also to help diagnose whether there is heart-related disease or not. In this paper, we propose a highly reliable real-time embedded device which is different from the Pan-Tompkins algorithm for QRS detection. The ECG signal is pre-processed by the gradient and threshold to detect the features of QRS complex. The QRS complex morphology is then used to further determine feature. Finally in this paper, a set of user interface detection tools is developed. Several QRS complex detection algorithm based on the first derivative methods are integrated to facilitate the evaluation and testing of the algorithm. The MIT-BIH arrhythmia database was used as the standard in verifying this algorithm, the sensitivity of 93.21%, and the accuracy of 89.28% was achieved. This proved that the algorithm proposed in this paper conforms to the standards of American Heart Association.