Heart sound total variation filtering method based on pathology self-adaption

Abstract

The invention relates to a heart sound total variation filtering method based on pathology self-adaption, and belongs to the technical field of medical calculation biological feature recognition. The method comprises the steps that firstly, additive Gaussian white noise is added to an input original heart sound signal, a noise adding signal meeting the requirement is obtained through screening, then the noise adding signal is subjected to improved empirical mode decomposition, an intrinsic mode function under the maximum similarity is obtained through screening, and the original heart sound signal is segmented; information entropies of different segments of an original heart sound signal are calculated by using a linear coefficient-based tracking evolution algorithm, the information entropies are used as weight values to construct l1 of the heart sound signal and limit l1 regularization total variation equations, and improved Split-Bregman algorithm is used to respectively solve, and finally the heart sound signal fused with the two filtering results is obtained by using a moving average method. Compared with the prior art, the heart sound signal denoising method has the advantages that the original heart sound signal can be denoised more accurately, pathological information can be reserved, and a basis is provided for digital analysis of a phonocardiogram.

Description

technical field [0001] The invention relates to a heart sound full variation filtering method based on pathological self-adaptation, in particular to a heart sound filtering method which integrates different pathological features of heart sound and full variation filtering, and belongs to the technical field of medical computing biological feature recognition. Background technique [0002] For a long time, cardiovascular disease has seriously threatened human health and life because of its high morbidity and high mortality. According to the World Health Organization (WHO), 36 million people die from non-communicable diseases such as cardiovascular diseases, diabetes, respiratory diseases and malignant tumors every year in the world, accounting for 2 / 3 of the total global deaths. By 2020, The number will climb to 44 million. [0003] "China Cardiovascular Disease Report (2018)" shows that in 2016, the mortality rate of cardiovascular disease in China still ranked first in th...

AI Technical Summary

Problems solved by technology

[0006] However, the traditional frequency-domain noise reduction methods (Butterworth filter, Chebyshev filter, etc.) have the following problems: (1) Pathological noises are not considered during denoising, such as hyperactivity and beating in mitral valve stenosis When the amplitude of the S3 segment is high, it will be misdiagnosed as the noise of the measurement and filtered out; (2) when the heart sound is collected by mobile , due to the close distance between the heart and lungs, the mitral valve area is easily affected by the left lung, and the heart sounds and lung sounds are aliased. Since the frequencies of the two are similar, the denoising algorithm based on frequency domain analysis cannot effectively remove the lung sounds

[0007] And at present, because the accuracy and reliability of the intelligent analysis of phonocardiogram cannot be compared with the auscultation analysis of cardiologists, so its proportion in actual clinical application is not large

Images