Adaptive Heuristic Histogram Equalisation Approach Combined With Gamma Transformation For Image Contrast Enhancement

Abstract

In image processing applications, histogram equali-sation (HE) is an extensively used contrast enhancement method. The algorithm is simple to implement. However, HE has a num-ber of flaws, such as large brightness change, unnatural effects, and over-enhancement, that make it inappropriate for many applications. To evade these issues, a new adaptive heuristic HE combined with gamma transformation approach is implemented. First, gamma transform is applied on input image histogram. The transformed histogram is then split into under-exposed and over-exposed sub-histograms. Probability distribution function (PDF) of the sub-histograms are calculated. Then adaptive parameters are calculated by taking the ratio of maximum to mean values of PDFs of sub-histograms. Thereafter, PDFs and cumulative distribution functions (CDF) are modified by imposing threshold limit to that adaptive parameters. Finally, second adaptive parameter of sub-histograms are calculated by using their respective modified CDFs and new CDFs are obtained by using these second adaptive parameters. Traditional HE is then applied on sub-histograms separately with the new CDFs to get the enhanced output image. The implemented method outperforms the traditional HE method in both subjective and objective tests for most of the images. The method improves image contrast while retaining the original properties of the input images, with no over-enhancement or undesirable effects in the output images.