Avaliação de desempenho de wavelet shrinkage pela esparsidade dos coeficientes

The wavelet shrinkage approach has been considered a method of low computational cost and reasonable performance in signal denoising. However, this method is not entirely adaptive, because of its dependence on parameters such as the wavelet function and scale. This paper aims to evaluate the relation between the performance of the method and the sparsity of the wavelet coefficients. It was observed that the signal SNR increase, in db, relies greatly on the aforementioned parameters, and that the sparsity can be a tool to adapt the parameters to each sample set.