Summary

Research in analyzing physiological signal utilizing digital signal processing has been growing rapidly for the past few years. It is to the interest of finding biological pattern or characteristic which relate to the condition of human body either to search for diseases or improving the quality of life. This book focuses on methods and approaches in analyzing heart sound to find heart disease. Heart sound analysis can provide information regarding the heart condition whether it is normal or abnormal. Heart sound is time-varying signal where it exhibits some degree of non-stationary. Due to this characteristic, three techniques are proposed to analyze them. These techniques analyze the heart sound from very different aspect to reveal the most unique features. The techniques are Time-Frequency Distribution (TFD), Mel-Frequency Cepstrum Coefficient (MFCC) and Instantaneous Energy and Frequency Estimate (IEFE). The proposed approach is tested against sixty one real patients with normal and abnormal type of heart sound and murmurs. Artificial Neural Network is then used to train and classify the heart sound into normal and abnormal class