SYNTHESIS OF THE ALGORITHM FOR NONLINEAR FILTRATION OF MULTIDIMENSIONAL DISCRETE-VALUED MARKOVIAN RANDOM PROCESSES
Igor Trubin · Telecommunications and Radio Engineering · 2010
Multidimensional Markovian processes were used to propose an efficient method for synthesizing algorithms designed for recovery and compression of digital half-tone images and video sequences distorted by additive Gaussian white noise. The synthesized algorithms have a clear physical significance and their implementation requires minimal computational resources.