Color image contrast enhancement by co-occurrence histogram equalization and dark channel prior

A novel color image contrast enhancement approach is proposed that equalizes color components dependently using cooccurrence histogram. And then the enhanced color components are the trade-off between co-occurrence histogram equalization and dark channel prior. Histogram equalization (HE) and its derivatives are the most common used methods. They don't consider the spatial gray-level correlations so that the equalizations are applied for all gray-levels appeared in an image. However, in the fact, it will be more effective to equalize a gray-level with its spatial-adjacent gray-levels. Co-occurrence histogram equalization (COHE) for image contrast enhancement is proposed to overcome this problem. It is applied here to the RGB color space. But dependent COHE perhaps result in hue changes and artifacts. In this paper dark channel prior is used to solve this problem. Experimental results demonstrate that the proposed method can enhance the RGB images effectively.