Space-variant two-dimensional filtering of noisy images

摘要: Publisher Summary The necessity of processing real images calls for algorithms where a rapid switching the filter characteristics is allowed. This chapter discusses simple and efficient method to meet this requirement. main feature analytical construction image model. Starting from smoothness, stochastic, inhomogeneity assumptions, nonstationary state-space representation obtained without onerous identification procedures or priori knowledge autocorrelation function. model space varying according presence edges. In way, edge-defocusing phenomenon greatly reduced. adaptive behavior by including information on edge locations into transitions in correspondence are not result heuristic but justified theoretical basis because they strictly related