Nonlinear wavelet transform thresholding methods were widely used in the field of signal de-noising. This kind of method are nearly optimal in many signal space. But the Wiener filter is optimal for any signal in the sense of MSE. Therefore, the empirical Wiener filter is proposed to process signals in this paper and the results show better performance than thresholding ^°f ?.'n the end, translation invariant wavelet is introduced to suppress pseudo-Gibbs phenomena effectively.
展开▼