ESTIMATION OF THE OPTIMAL REGULARIZATION PARAMETER OF AN ITERATIVE WAVELET ALGORITHM FOR SIGNAL RECOVERY
Yu. E. Voskoboynikov
Novosibirsk State University of Architecture and Civil Engineering, ul. Leningradskaya 113, Novosibirsk, 630008 Russia
voscob@mail.ru
Keywords: signal and image recovery, nonlinear regularization algorithms, threshold wavelet filtering, estimation of optimal regularization parameter
Abstract
Two methods of statistical estimation of the optimal regularization parameter of a nonlinear algorithm for signal recovery are proposed, one of which is based on the optimality criterion and the other one on the residual principle. These methods have simple algorithmic implementation and do not need a priori information about the quantitative characteristics of the sought solution. The results of numerical experiments show the effectiveness of the method based on the optimality criterion.
|
