Development And Application Of Heart Sound Signal Integrated Database Management System

Heart sound signal generated due to the beating of the heart periodically and impacted by manyfactors like that the blood strikes the heart wall. As one of the most important human signals, itcontains a lot of relevant information about the status of cardiac pathology, reflecting the heart andcardiovascular structure, physiology and pathology information. Therefore, the heart sound signalsacquisition, de-noising, location, time-frequency domain characteristics and heart sound signalrecognition research have materially significance.This paper adopts the different time-frequency analysis methods to analyse the characteristicparameter extraction of heart sound signals. And using Delphi and Matlab mixed programmingmethods to design a pre-diagnosis and management system of heart sounds, which with wellinteractive performance. The main work of the study is as follows:1. Pretreatment and the localization of heart sound signals: we need to make pretreatment to thecollection heart sound signals, remain the main component of heart sound signals, then useenvelope extraction and interpolation methods to analyse the heart sounds signal’s locationaccording to the time-domain characteristics of heart sound signals.2. We conducted a three-pronged approach for analyzes the heart sound signals. E.g: in timedomain, frequency domain and time-frequency domain analyze its characteristic parameters. Andusing wavelet transform theory, structure theory and S transform, constructed Gaussian heart soundwindow function with high acquaintance of heart sounds, and use it to improve the S transform. Theexperiment found that the improved S transform can analyze more details of the local characteristicsof heart sound signals, improve the resolution for weakly heart sound signal and have betterfocusing of energy.3. Using the method of MFCC feature value extract heart sound signal characteristics to obtainmultidimensional characteristic parameters. Considering the effectiveness and practicality ofparameters, we use PCA method to reduce the dimension of heart sounds characteristic parametersand ensure the validity no reducing. Finally, all of the characteristic parameters are stored in theheart sound database.4. Using Matlab and Delphi mixed programming means. In the Delphi environment,programmed a management systems, which has well human-computer interaction for heart sounds.According to heart sound signal processing and system module design rules, we planned rationally in system design. Due to the characteristics of heart sound signal value that is stored in the database,when the database is rich enough in heart sound signal types, more accurate identification of heartsound, the unknown heart sound signal can be predict through known heart sound signal’scharacteristics in the database, which can be identify the heart sound signals more effectively,analyze its possible heart sound condition, analyse the patient possible pathological characteristicsand severity, reduce economic losses, and provide references for further treatment.