We introduce a new nonlinear filter for signal and image restoration, the h
ybrid order statistic (HOS) filter. Because it exploits both rank- and spat
ial-order information, the HOS realizes the advantages of nonlinear filters
in edge preservation and reduction of impulsive noise components while ret
aining the ability of the linear filter to suppress Gaussian noise. We show
that the HOS filter exhibits improved performance over both the linear Wie
ner and the nonlinear L filters in reducing mean-squared error in the prese
nce of contaminated Gaussian noise. In many cases it also performs favorabl
y compared with the Ll and rank-conditioned rank selection filters. (C) 200
1 Optical Society of America OCIS code: 100.2000.