Heart sound recognition based on wavelet transform and SampEn
基于小波变换和样本熵的心音识别研究*

This paper proposed the algorithm which combined the nolinear dynamics with the time-frequency analysis to study heart sounds. Firstly, in order to effectively restrain noise, employed wavelet transform to preprocess the sample data, and then could obtained SampEn of the new signals. Finally, adopting SVM to achieve the optimal classification. Experimental results prove the reliability and veracity of the method.