Image Recovery Using Test Trials of Processing Systems

Mathematical methods for restoration of multidimensional signals and images distorted and noisy during their processing by linear stationary systems using test trials of these systems are considered. The use of test signals and images allows, first, to perform non-blind recovery without identifying distortion functions, which may even be from the class of generalized ones, and, second, to use regularization technique to recover real non-deterministic signals and images with noise and uncertainties. The peculiarities of practical use of test trial methods are considered. Numerical experiments on 1D signal and 2D image restoration at different noise levels are presented.