Heart sound connected network feature extraction method based on frequency correlation

Abstract

The invention discloses a heart sound connected network feature extraction method based on frequency correlation. According to the method, multi-channel heart sound pickup part points are used as network nodes, frequency correlation among the nodes is calculated, and a modular heart sound communication network is constructed on a large-scale level according to auscultation area division, so that feature extraction of the communication network is realized. Heart sound signals at different positions in the auscultation area module can be collected at the same time, and multi-channel heart sounddata with more physiological significance and physical significance than single-channel data is obtained; frequency correlation analysis and correlation matrix establishment are carried out; accordingto the method, isolated and dispersed data information is effectively and organically integrated together, so that the heart sound characteristics are extracted and represented, the heart sound communication network characteristic matrix is input into the support vector machine, the effect of heart sound classification and recognition can be achieved, and a foundation is laid for subsequent utilization and application of the heart sound characteristics.

Description

Technical field [0001] The invention relates to a heart sound feature analysis method, in particular to a heart sound connected network feature extraction method based on frequency correlation. Background technique [0002] The prior art can only perform auscultation on one location at a time, and cannot extract and analyze the correlation of heart sounds at different locations at the same time. Currently, there are single-channel electronic stethoscopes and 4-channel electronic stethoscopes, but single-channel electronic stethoscopes cannot solve the above problems. Due to the small number of channels in the four-channel stethoscope, and due to individual differences, the distance of the auscultation area is different. Fewer channels cannot correspond to the corresponding auscultation area for signal collection and correlation analysis. Therefore, it is necessary to perform detailed feature analysis on multi-channel heart sound signals, and then perform classification and recog...

AI Technical Summary

Problems solved by technology

[0002] The existing technology can only auscultate one position at a time, and cannot extract and analyze the correlation of heart sounds at different positions at the same time

Currently there are single-channel electronic stethoscopes and 4-channel electronic stethoscopes, but single-channel electronic stethoscopes cannot solve the above problems

Due to the small number of channels of the four-channel stethoscope, and due to individual differences, the distance of the auscultation area is different, and the fewer channels cannot be well corresponded to the corresponding auscultation area for signal acquisition and correlation analysis.

Images