A critical appreciation on algebraic image restoration

The authors consider the aspect of unconstrained algebraic restoration of a gray-level image, degraded due to defocusing, additive noise, etc., with additional nonnegative constraint on the individual pixel of gray-level image. The negative values in the gray level introduce absurdity in the entire representation. It is demonstrated that under certain conditions the constrained and unconstrained solutions for the restored image are the same. It is shown that using P. Wolfe's algorithm (1959) it is possible to handle efficiently the nonnegative constraint of the gray-level image. Necessary and sufficient conditions of the optimal constrained restoration were tested.<
>