基于双自适应提升算法的心音信号去噪研究

Heart sound noise reduction based on dual self-adaptive lifting algorithm

心音去噪是心音信号临床使用的前提。提出将双自适应提升算法用于心音去噪处理。该算法采用自适应更新和自适应预测构造小波函数,通过将传统的硬阈值和软阈值函数相结合,构造了一个改进的阈值函数进行心音信号去噪处理。对临床采集的80例心音信号进行了去噪实验,结果表明:该算法表现出良好的去噪效果,并增强了信号的局部特征。同普通小波去噪方法相比,其信噪比提高了46.5%,均方根误差减小了64.0%,而且运行速度快,可有效地用于临床心音信号的去噪处理。

Heart sound noise reduction is the premise of heart sound signals for clinical usage. A dual self-adaptive lifting algorithm for heart sound noise reduction was proposed. With the algorithm, a self-adaptive update and a self-adaptive prediction were adopted to construct a wavelet function, and the traditional hard threshold function was combined with the soft threshold one to construct an improved threshold function for heart sound signals' denoisng processing. 80 heart sound signals from clinical collection were used to perform denoising tests. The result showed that the proposed method has good denoising effects, and it enhances the local characteristics of the signals; compared with the ordinary wavelet denoising method, after using the new method, the signal-to-noise ratio increases 46. 5%, the root-mean-square error reduces 64. 0%; the new algorithm has a fast operation speed, it can effectively be used to denoise clinical heart sound signals.