Methods to Increase the Contrast of the Image with Preserving the Visual Quality

Abstract

Contrast enhancement is a technique for increasing the contrast of an image to obtain better image quality. As many existing contrast enhancement algorithms typically add too much contrast to an image, maintaining visual quality should be considered as a part of enhancing image contrast. This paper focuses on a contrast enhancement method that is based on histogram transformations to improve contrast and uses image quality assessment to automatically select the optimal target histogram. Improvements in contrast and preservation of visual quality are taken into account in the target histogram, so this method avoids the problem of excessive increase in contrast. In the proposed method, the optimal target histogram is the weighted sum of the original histogram, homogeneous histogram and Gaussian histogram. Structural and statistical metrics of “naturalness of the image” are used to determine the weights of the corresponding histograms. Contrast images are obtained by matching the optimal target histogram. Experiments show that the proposed method gives better results compared to other existing algorithms for increasing contrast based on the transformation of histograms.