DISTORTION-INVARIANT FILTER FOR NONOVERLAPPING NOISE

Anew heuristic filter based on the optimum filter for disjoint noise d eveloped by Javidi and Wang [J. Opt. Sec. Am. A 11, 2604 (1995)] is pr esented. In this new filter a number of optimum filters built from sin gle training images are combined Linearly by use of the synthetic disc riminant function (SDF) approach into a distortion-invariant filter fo r disjoint noise. Like the traditional SDF approach, this summation te chnique makes it possible to control the height of the correlation pea k easily, for example, if a uniform filter response is needed. The fil ter is compared with the distortion-invariant version of the optimum f ilter on images with low contrast and high levels of nonoverlapping cl utter. The new filter shows good results, demonstrating that it is, wi th very simple heuristic methods, possible to improve the performance of distortion-invariant filters for nonoverlapping noise. (C) 1998 Opt ical Society of America.