Optimal Heart Sound Segmentation Algorithm Based on K-Mean Clustering and Wavelet Transform

The accurate localization of S1 and S2 is essential for heart sound segmentation classification. However, current direct algorithms have poor noise immunity low accuracy. Therefore, this paper proposes a new optimal algorithm based on K-means clustering Haar wavelet transform. includes three parts. Firstly, method uses the Viola integral Shannon’s energy-based to extract function envelope energy. Secondly, time–frequency domain features acquired are extracted from different dimensions peak searched adaptively dynamic threshold. Finally, transform implemented localize sounds in time domain. After validation, recognition rate reached 98.02% that 96.76%. model outperforms other effective methods been implemented. has high robustness immunity. it can provide feature extraction analysis signals collected clinical settings.