Efficient contrast enhancement with truncated adaptive gamma correction

Abstract

Gamma correction is an efficient tool typically used to adjust the contrast and brightness of digital images. Recently, an adaptive gamma correction method is proposed by relating gamma parameter with the cumulative distribution function (CDF) of gray level histograms. It gains satisfied contrast enhancement (CE) effects on dimmed images, but is invalid on bright regions which are universal in real scenarios, such as excessive exposure, white objects and highlight regions. In order to attenuate such a deficiency, we propose an improved adaptive gamma correction algorithm in this paper. Instead of directly applying CDF curves as the gamma parameter, we truncate them by thresholding with a reasonable limit. Accordingly, contrast of input dimmed images could be improved without losing details in bright regions. Both qualitative and quantitative experimental results demonstrate the promising performance of our proposed CE method in enhancing the contrast and brightness of dimmed images.