Optimization of grey-scale morphological statistical estimators
Edward R. Dougherty · 1990
General gray-scale morphological filters are interpreted as statistical estimators by employing the Matheron representation for increasing translation-invariant signal-to-signal mappings. Mean-square optimization is discussed, including the minimal search problem and the manner in which classical moving statistical filters fit into the more general morphological paradigm. Key to the method is the automatic generation of structuring elements based upon image statistics.>