Research On Visualization Analysis And Recognition Of Heart Sound Signal Based On Variable Value Model

Cardiac auscultation is important in the diagnosis of heart disease.The initial diagnosis of congenital heart disease is mainly based on the clinician auscultation for preliminary diagnosis.The heart sound signal is a weak audio signal,which contains rich information of heart activity and heart sound signal.Research from the perspective of signal analysis has developed into a hotspot and a difficult point in the field of biomedical signal processing.Most scholars at home and abroad have analyzed the heart sound signal from the perspective of traditional signal processing.This paper tries to find another way.By studying the morphological characteristics of heart sound signals and applying the variable value logic method,the heart sound signal after enveloping is extracted and subjected to a series of variable value logic transformation,and finally a two-dimensional scatter plot is formed,so that different direct observations can be observed.The two-dimensional scatter plot of the disease,the different distribution characteristics of different pathological heart sound signals in the corresponding two-dimensional scatter plot,the contents include the following:1.Heart sound signal preprocessing.Conventional heart sound signal preprocessing usually includes noise reduction,envelope extraction,and the like for the heart sound signal.However,since the noise threshold is difficult to control and the pathological heart noise is easily removed,the preprocessing of this study selects the envelope signal with good quality heart signal without denoising.2.Envelope extraction method selection.The Hilbert transform method,the Shannon energy method and the Viola integral method are used to extract the normal heart sound signal and the pathological heart sound signal,and the extracted envelope is algorithmically labeled to extract various envelopes.The advantages and disadvantages of the method are analyzed,and the characteristics of the algorithm are combined.Finally,the Viola integral method is selected as the envelope extraction method of heart sound signal.3.Variable value model conversion of heart sound signals.The extracted heart sound signal envelope is subjected to a variable logic algorithm to convert the heart sound signal into 10 measures required for visual analysis.4.Visual analysis identification.Through the analysis of 10 measures,one or two of the measures are selected for visualization,and the one-dimensional scatter plot and the two-dimensional scatter plot of the heart sound signal are respectively made,and the specific characteristics of different heart sound signals are combined,and then the corresponding scatter is analyzed.The characteristics of the dot map.In this paper,a comprehensive study of the heart sound signals collected in the clinic is carried out,and the pathological heart sound signals and normal heart sound signals are converted into corresponding scatter plots,and the one-dimensional scatter plots are analyzed and identified.The experimental results show that the recognition rate of pathological heart sound signal(Sensitivity)reaches 81.3%,the recognition rate of normal heart sound signal reaches 76%(Specificity),accuracy rate can reach 80%,the system has a good recognition effect.