Stationary Wavelet Transform Based Detection of Aortic Stenosis Using Seismocardiogram Signal

Aortic stenosis (AS) is one of the most common and severe valvular heart diseases, which can cause heart failure. Early detection and treatment are the most effective ways to prevent AS. This study proposes a stationary wavelet transform (SWT) based machine learning framework to detect AS using seismocardiogram (SCG) signal. First, the SCG signal is preprocessed and segmented into cardiac cycles. Then, SWT is deployed to decompose each cardiac cycle into subbands. Further, each subband is used to extract the novel statistical features, which include log-energy entropy, relative wavelet energy, multi-scale kurtosis, and median absolute deviation. Finally, these features are fed into the random forest classifier for automated detection of AS. The effectiveness of the proposed method is evaluated using data from two publicly available databases. Our method achieves an overall accuracy of 99.4%, sensitivity of 98.9%, and specificity of 99.9%. The proposed method provides comparable performance with the current state-of-the-art techniques. The impressive results of the proposed framework make it useful to detect AS in primary healthcare units.