基于小波变换和似然无偏估计的运动心电信号伪差消除法

Exercise ECG Signal De-noising Using Unbiased Risk Estimate and Wavelet Transform

介绍了一种基于小波变换并结合似然无偏估计来消除运动心电信号中基线漂移和肌电噪声的新方法 ,且提出了评价心电消噪算法有效性的两个指标。该方法利用小波变换多分辨率分析的特性 ,将原始运动心电信号进行多尺度分解及单支重构 ,根据运动心电信号的自身特征 ,结合似然无偏估计针对不同的心电细节成分进行阈值消噪处理。研究结果表明 ,该方法能有效消除运动心电信号中的干扰成分 ,为进一步研究运动心电信号的特征识别分析提供了新途径。

In this paper a filtering method for EECG （Exercise ECG） signal is proposed which is based on wavelet transform （WT） and Stein＇s unbiased risk estimate （SURE）. This algorithm was used to decompose original EECG signals into detail signals on different frequency bands by using WT and get different thresholds with SURE. According to EECG signal features and by using the above thresholds, the method amended several detail signals so that the main interferences in EECG signal can be removed efficiently. The authors also put forward two indexes to estimate the validity of such algorithms. Our experimental results demonstrate that this is an efficient de-noising method for EECG.