最优改进对数谱幅度估计与小波结合的心音降噪

Combination of optimally-modified log-spectral amplitude estimator and wavelet for heart sound denoising

小波阈值降噪为心音降噪的一种常用方法。本文提出了使用最优改进对数幅度谱估计与小波阈值降噪相结合的方法对心音降噪。在正常心音和一些常见疾病的心音中加入不同强度的白噪声和粉红噪声,构造不同信噪比的心音信号,并将本文所提出的方法和仅用小波阈值降噪方法的去噪效果进行了定量的对比。结果表明本文方法降噪效果总体优于仅使用小波阈值降噪达到的效果。

Wavelet thresholding is a frequently-used method for heart sounds denoising.Herein optimally-modified logspectral amplitude estimator is combined with wavelet thresholding to remove heart sound noises.After that the white noises and pink noises of different intensities are artificially added into normal heart sounds and the heart sounds of some common diseases to construct heart sounds of different signal-to-noise rates,traditional wavelet thresholding and the proposed method are used for heart sounds denoising.The quantitative comparison results show that the denoising effect of the proposed method is better than that of traditional wavelet thresholding.