Recursive M-type filter for image processing

Proposes a class of nonlinear filters-M-type filters, which are based on the M-estimates of location parameters in statistical theory. Moving-window M-filters can be defined starting from M-estimates of location. Then, several modified structures are given. It is shown that recursive M-type filter can perform better than nonrecursive M-type filter, running mean and median filter in white noise suppression, while it can be designed to be comparable to the median filter in edge preservation in the presence of noise. (To suppress impulse noise, the proposed filter generates output with a modified reference output: to reduce calculation complexity, M-type filter with linear FIR substructure is introduced.) Finally, experimental results for one- and two-dimensional signals are presented