基于coif-5小波的心音自适应阈值降噪方法

为改善小波阈值及阈值函数在心音信号降噪处理中性能,通过使用coif5小波做5层小波分解,抑制分量极少的层次的系数,保留有用分量层次的系数,然后引入相关统计参量进行阈值自适应估计,并按提出的阈值函数对有用分量层次系数阈值化处理,最后通过处理后的系数重构出降噪信号；所提方法与硬阈值、软阈值、线性中间阈值函数降噪性能进行了对比,并采用了三类评价指标做出了有效评价。实验结果表明：提出方法降噪后的SNR更高、RMSE更小、听诊MOS更高。可见,提出方法降噪性能更优。

A Wavelet Adaptive Threshold Method for Heart Sound De-noising Based on Coif-5

In order to improve the performance of wavelet threshold and threshold function in noise reduction processing of heart sound signals, there use the coif5 wavelet to do 5 layers decomposition, restrain the coefficient of a low level of component and retain the useful component level. Then, statistical parameters are introduced to estimate the threshold adaptively, and threshold of useful components is thresholding according to the threshold function. Finally, reconstructing the signal after the noise reduction. The proposed method is compared with hard threshold, soft threshold and linear intermediate threshold function, and use three evaluation indexes to evaluate. The experimental results show that the proposed method has a higher SNR, a smaller RMSE and a higher MOS. Therefore, the proposed method has better performance.