心音訊號分析演算法之研究

Abstract

心跳是重要生命跡象，而由心音訊號可辨別出許多疾病徵兆，因此快而準確的心電圖信號分析對於病情診斷相當有幫助，心電圖信號處理包含以下幾個部分：特徵萃取、特徵選取、心律變異分析，針對不同的生物醫學信號，如心電圖、腦波圖、肌電圖，有不同的演算法設計，一方面提取出信號特徵，另一方面有效的消除信號雜訊，由於生物醫學信號通常是非穩定信號、時變率高，較不適合使用傅立葉轉換分析，使用時域分析方法會較合適[5]。聽診是診斷心臟疾病重要的檢查方法之一，心音、心律等的變化往往是心臟疾病最早出現的體徵，如主動脈狹窄、心動過速、心律不整…等，都是可以透過聽診直接判斷的病症，所以心音訊號之分析對於心血管疾病的診斷也具有重要的意義，這就是為什麼心音訊號分析會成為國內外的熱門研究題目了。本論文的題目「心音訊號分析演算法之研究」，目的是藉由自行開發之心音訊號感測系統量取一般真人受測者及收集 SAMⅡ假人心肺音標準模擬系統的原始心音數據為基礎，進行時域分析、特徵擷取、病症決策，進而撰寫出一套完整的演算法，進行時域分析且比較以整理出各別的生理特徵及特性並進一步用於正常與異常心音之初步診斷，並對於此套心音量測系統進行整合以實現心音感測醫療器材之雛型化開發[5]。

Heartbeat is an important sign of life, and many signs of disease can be discerned by heart signals. Therefore, fast and accurate physiological signal analysis for the diagnosis of the disease is helpful. Physiological signal processing contains the following parts: feature extraction, feature selection, analysis of heart rate variability. For different biomedical signals, such as ECG, EEG, EMG, there are different algorithm design. Can the extract the signal characteristics, and effective elimination of signal noise. Since biomedical signals are usually unstable signals. Therefore, it is not suitable to use Fourier transform analysis. Time domain analysis method will be more appropriate. Auscultation is one of the important ways to diagnose heart disease. Heart sounds, abnormal heart rhythm, often the earliest signs of heart disease. Such as aortic stenosis, tachycardia, irregular heartbeat ... and so on. Are diagnosed by auscultation disorders. Therefore, the analysis of heart sound signal is also of great significance for the diagnosis of cardiovascular diseases. This is why the heart sound signal analysis will become a hot topic. The topic of this thesis, “The Study of Heart Sound Signals Analysis Algorithm". The purpose is to measure the average patient by developing a heart sound signal sensing system. And collecting heart sound raw data from a standard simulated system of SAM II. Research time domain analysis, feature extraction, disease decision-making. Use time-domain analysis to write algorithms. And compare the different physiological characteristics. Initial diagnosis of normal and abnormal heart sounds Achieve heart sound sensing medical equipment prototype development. And heart sound measurement system integration.