Study on Wavelet Energy Entropy and Its Application to Bioelectrical Signal De-Noising

Bioelectrical Signal is a weak signal which is buried in the noisy environment, how to recover the weak signal in a strong noisy environment is very important. The paper gets SEMG and EEG as examples, and proposes a de-noising method based on wavelet energy entropy to bioelectrical signal de-noising. First multi-scale wavelet decomposition of bioelectrical signal is performed by discrete wavelet transform; To calculate the intervals´ wavelet energy entropy at every decomposed scales, and the filtering value threshold of high-frequency coefficients can be determined by the wavelet energy entropy´s distribution character; Finally, the de-noised bioelectrical signal is obtained by the wavelet reconstructed from the remain low-frequency coefficients and the high-frequency coefficients which are processed. The method has the advantage of setting threshold value adaptively and easily to operate. The experiments show that the method based on wavelet energy entropy has good performance in de-noising, and reserving the detailed information.