基于变分模态分解的肺音去噪算法  

Lung sound denoising algorithm based on variational mode decomposition

作　　者：孙文慧 张乙鹏 林冬梅[3] 陈扶明 SUN Wenhui;ZHANG Yipeng;LIN Dongmei;CHEN Fuming(School of Information Engineering,Gansu University of Chinese Medicine,Lanzhou 730000,China;Medical Security Center,the 940th Hospital of Joint Logistics Support Force of Chinese People's Liberation Army,Lanzhou 730050,China;College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]甘肃中医药大学信息工程学院,甘肃兰州730000 [2]中国人民解放军联勤保障部队第940医院医疗保障中心,甘肃兰州730050 [3]兰州理工大学电气与信息工程学院,甘肃兰州730050

出　　处：《中国医学物理学杂志》2024年第4期479-485,共7页Chinese Journal of Medical Physics

基　　金：国家自然科学基金(61901515,62361038);甘肃省自然科学基金(22JR5RA002)。

摘　　要：目的:为有效提高肺音信号质量,提出一种基于变分模态分解的肺音去噪方法。方法:首先利用经验模态分解对带噪肺音信号进行分解,根据本征模态函数特征确定最佳分解层数,然后根据分解层数对原始带噪肺音进行变分模态分解处理,接着根据皮尔逊系数选取有用模态,最后采用阈值方法对各模态函数去噪,重构后得到没有噪声干扰的肺音信号。结果:通过与维纳滤波和FIR滤波进行对比,本文方法的语音质量感知评价、短时间客观可读性和源信号失真比均更优。结论:本文方法能有效对肺音信号进行去噪处理。Objective To propose a lung sound denoising method based on variational mode decomposition(VMD) for effectively improving the quality of lung sound signals.Methods Empirical mode decomposition was utilized to decompose the noisy lung sound signal,and the optimal decomposition level was determined based on the intrinsic mode function features.Subsequently,the original noisy lung sound was processed with VMD according to the decomposition level,and the useful modes were then selected based on Pearson correlation coefficient.Finally,a threshold method was applied to the denoising of each mode function,and the lung sound signal without noise interference was obtained after reconstruction.Results Compared with Wiener filtering and finite impulse response filtering,the proposed method exhibited superior performance in perceptual evaluation of speech quality,short-time objective intelligibility,and signal-to-distortion ratio.Conclusion The proposed method can effectively remove the noise from lung sound signals.

关 键 词：肺音去噪 变分模态分解 经验模态分解 

分 类 号：R318[医药卫生—生物医学工程] TP912.35[医药卫生—基础医学]