Research Of Coronary Artery Stenosis Intelligent Diagnosis Algorithm Based On The Analysis Of Diastolic Murmurs

Coronary heart disease is one of the most common heart diseases,mostly caused by coronary artery organic stenosis or obstruction which leads to myocardial dysfunction and organic disease.It is a serious harm to human health and the key to its prevention and treatment is early diagnosis and early treatment.When the early coronary artery stenosis occurring,there is generally no pain to patients or ECG detection does not appear abnormalities.So it is vital to study an algorithm that can detect early coronary stenosis.Heart sound reflects the situation of the heart and cardiovascular system mechanical movement,and reflects the physiological and pathological information of the various parts of the heart and their interaction.The diastolic murmur caused by coronary blood turbulence contains the pathological features of coronary heart disease.Therefore,analyzing and studying diastolic heart sound signal has a very important significance for the intelligent diagnosis of coronary heart disease.This paper mainly studies the intelligent diagnosis algorithm of coronary stenosis based on diastolic heart sound analysis.According to the analysis of the relationship between the time-frequency characteristics of diastolic heart sound signals and coronary stenosis,the diastolic heart sound feature extraction algorithm is extracted based on the marginal energy under the multi-frequency threshold and Hilbert Vibration Decomposition(HVD).Then,a support vector machine suitable for small samples is constructed,and the intelligent diagnosis of coronary artery occlusion is realized by using the heart sound signal acquired by the clinic.And it achieves a better diagnosis effect.The main contents of this paper are as follows:(1)The paper expounds the background of intelligent diagnosis of heart disease based on heart sound analysis,and analyzes the research achievements and the future development trend of heart sound diagnosis technology at home and abroad.(2)The paper describes the basic characteristics of the heart sound signal,including the generation principle of heart sound,basic characteristics,acquisition mode and the experimental environment and data source of this paper.The diagnosis mechanism of coronary heart disease is emphatically expounded,which is the theoretical basis and technical support of intelligent diagnosis of coronary heart disease.(3)The diagnosis algorithm of diastolic heart sound is proposed based on the marginal energy under multiple-frequency threshold.Firstly,empirical mode decomposition(EMD)is used to obtain the Intrinsic Mode Function(IMF)component which can represent the heart sound signal.Secondly,the basic principles and algorithm steps of Hilbert transform are studied.The difference between the Fourier spectrum and the marginal energy spectrum is verified by the simulation signal,and the energy ratio under the multi-frequency threshold is obtained by the marginal energy spectrum.Then,the principal component analysis is adopted to reduce the dimension and remove the redundant features.Finally,the diastolic heart sound signals of normal and coronary heart disease patients are analyzed.The results show that the energy ratio of diastolic heart sound signal of coronary heart disease patients is higher than the normal and has a significant difference compared with the normal under multi-frequency threshold.This indicates that the energy ratio can effectively reveal the information that coronary occlusion can produce high-frequency heart murmur.(4)The diagnosis algorithm of diastolic heart sound is proposed based on improved Hilbert vibration decomposition,including HVD,improved endpoint extension,low-pass filter parameter setting,feature extraction,and support vector machine.The boundary effect is generated since the unprocessed signal passing through Hilbert vibration decomposition.So firstly,an improved boundary extension scheme is proposed,and the effectiveness is verified by simulation signal.Secondly,the instantaneous frequency mean values of different combined harmonic components after HVD are used as the eigenvalue.Finally,the diastolic heart sound signals of normal and coronary heart disease patients are analyzed.The result shows that the mean value of instantaneous frequency of coronary heart disease patients is higher than the normal and has a significant difference compared with the normal.(5)The intelligent diagnosis algorithm for coronary heart disease is developed based on support vector machine(SVM).The diagnosis performance based on the marginal energy of multi-frequency threshold and the characteristics of improved Hilbert Vibration Decomposition is discussed,and a high classification accuracy is obtained.The feature-level fusion algorithm and its performance are studied.The results show that the feature-level fusion algorithm has higher classification accuracy and better real-time performance.The paper provides a new idea for sound intelligent diagnosis technology by studying the analysis of diastolic heart sound signals and the extraction of pathological features,and lays a solid theoretical foundation and effective technical support.