Description: Introduced a whitening matched filter based on neural network of the QRS wave detection. We use neural networks to handle the whitening matched filter low frequency ECG signal to simulate the nonlinear and non-steady state characteristics. Processed ECG signal contains most of the high frequency components, let through a linear matched filter to detect the QRS wave and position. For large noise the ECG signal after the matched filter plus differential filter, such as taking the square and the moving average processing, improve the detection accuracy. Using this method we have M IT?B IH noise ECG database of 105 large data processing, testing rate was 9912 correct. In contrast, detection with digital band-pass filter, the correct rate of 9718 .
To Search:
File list (Check if you may need any files):
基于神经网络的波型检测方法.pdf
Download