Recovery of Transient Signals in Noise by OptimalThresholding in Wavelet Domain

The recpvery of transient signals corrupted by additive white Gaussian noise means of the discrete wavelet transform was studied. An optimum threshold scheme in the discrete wavelet domain-presented, which involved a hypothesis testing for wavelet coef Fcients. This threshold was designed using likelihood ratio, Neyman-Pearson criterion and the minimum mean-squared error. The effectiveness of the proposed optimum threshold was demonstrated by computer simulations of signal recovery in poor signal一to-noise ratio scenarios.