A nonlinear technique for image contrast enhancement and sharpening

Contrast represents the extent of variation in light intensity or gray value in a specified region of an image. In this paper, we present an approach to contrast sharpening based upon the calculation of a local measure and the use of this computed quantity as part of a nonlinear contrast enhancement method. The contrast enhancement method presented here uses the concept of a local mean edge gray value, as well as gray scale partitioning into discrete subintervals, as the basis for processing the image. This technique maps the intensity values in each of the subintervals in a continuous fashion, to intensities nearer to those of the upper and lower endpoints of each subinterval. In the case of an image corrupted by Gaussian noise, the image is first processed by a tandem of pyramidal lowpass filters and then by the contrast enhancement algorithm. The result is a very smooth, sharp image