Vector operators based on robust order statistics have proved successful indigital multichannel imaging applications, particularly color image filteringand enhancement, in dealing with impulsive noise while preserving edges andfine image details. These operators often have very high computationalrequirements which limits their use in time-critical applications. This paperintroduces techniques to speed up vector filters using the minimaxapproximation theory. Extensive experiments on a large and diverse set of colorimages show that proposed approximations achieve an excellent balance amongease of implementation, accuracy, and computational speed.
展开▼
