基于改进的维奥拉积分方法提取心音信号包络

Heart sound signal envelope extraction based on modified Viola integral method

心音是评价人体心脏功能的重要生理信号之一,分析心音信号特征对心脏疾病的早期诊断和治疗具有重要意义。分别采用希尔伯特黄变换、归一化香农能量以及维奥拉积分方法提取心音信号的特征包络,并针对维奥拉积分法存在的不足,结合香农能量方法可抑制高、低强度信号的特性,提出了一种改进的心音信号特征包络提取方法,最后采用阈值法实现了对心音信号的分段。采用改进的维奥拉积分方法对55例心音信号提取特征包络并进行分段处理,经验证该方法能有效的提取心音信号特征包络,对正常信号分段正确率达到了100%,异常信号分段正确率平均达到了95.32%,适合于心音信号分段。

Heart sound is one of the most important physiological signals evaluating human heart function,analyzing heart sound signal characteristic is of great significance for the early diagnosis and treatment of heart disease.In this paper,the Hilbert-Huang transform, normalized Shannon energy and Viola integral method are used to extract the characteristic envelope of heart sound signal.Aiming at the deficiency of Viola integral method,considering the characteristic that Shannon energy method can restrain the high and low intensity signals,a modified heart sound signal envelope extraction method is proposed.Finally,the threshold method is adopted to achieve the heart sound signal segmentation.The modified Viola integral method was adopted to extract the characteristic envelopes and conduct segmentation processing of 55 cases of heart sound signals,and the experiment result shows that the proposed method can effectively extract the envelopes of normal and abnormal heart sound signals,the segmentation accuracy is 1 00% for normal heart sound signal and 95.32% for abnormal heart sound signal,which indicates that the modified Viola integral method is suitable for heart sound signal segmentation.