用QRS检测的神经网络自适应匹配滤波

• 作 者: 汤孟兴
• 出 处: 国外医学.生物医学工程分册
• 发表日期: 1993年01期
• 关键词: 匹配滤波 QRS 白化滤波器 复合波 低频成分 人工神经网络 非稳态 噪声特性 心律不齐 卷第

这里,我们介绍一种用于QRS检测、基于人工神经网络(ANN)的自适应匹配滤波方法。用一个ANN自适应白化滤波器模拟ECG固有的非线性、非稳态的低频成分。包括绝大多数高频QRS复合波能量的剩余信号,经过一个线性匹配滤波器来检测QRS复合波的位置。我们提出一个算法,根据从ECG信号中检测到的QRS复合波,自适应地校正匹配滤波器的模板,以便对不同的对象来制定模板。这种ANN

Here, we introduce an adaptive matching filter based on Artificial Neural Network (ANN) for QRS detection. An ANN adaptive whitening filter was used to simulate the inherent non-linear, non-steady-state, low frequency components of the ECG. The remaining signal, which includes most of the high-frequency QRS complex energy, passes through a linear matched filter to detect the QRS complex position. We propose an algorithm that adaptively corrects the template of the matched filter based on the QRS complex detected from the ECG signal to create a template for different objects. This kind of ANN