Nonlinear filters with spatially connected neighborhoods

A new class of nonlinear filters using rank-order operations over spatially connected neighborhoods and fast algorithms for the design of these filters are presented. Many rank-order processing techniques may be implemented by using filters such as noise suppression, enhancing, and restoring images. The performance of the proposed filters for suppression impulsive noise, additive noise, and mixed additive and impulsive noise in test images is compared to conventional rank-order algorithms. The comparisons are made using a mean square error, a mean absolute error, and a subjective human visual error criteria. (C) 2001 Society of Photo-Optical Instrumentation Engineers.