A method for almost exact histogram matching for two digitized images

Abstract A method for the almost exact matching of the histograms between two digitized images, with noise level comparable to the separation between two consecutive discrete gray levels, is presented. The exact matching between the histograms of the two images can be easily obtained if the original histograms are given as continuous functions of the gray level in the images. But using the existing methods, it is not possible to do so in general for digitized images where the gray level for each pixel is one of the allowed discrete values. If the noise level in the images is comparable to the separation between the discrete gray levels, then the exact value for the mean gray level for each pixel can be any value in a small interval about the specified discrete gray level, such that the specification of the discrete gray level for that pixel does not change. Using the above fact, an almost exact histogram matching can be obtained in the discrete case.