Multi-exposure optimized contrast and brightness balance color image enhancement

Abstract

Color image enhancement has been very challenging for researchers in imaging systems. In this paper, optimal weighted multi-exposure histogram equalization model is proposed for image contrast enhancement. The proposed method divides the input image into two sub-images, specifically low-exposed and high-exposed regions. Then, exposure information obtained from these sub-images is used to divide the input histogram into sub-histograms, and the level of excessive enhancement is controlled by clipping and the optimal weighting process before applying histogram equalization. Next, each sub-histogram is equalized individually and integrated to preserve the naturalness and fine details of the input image. Finally, dual gamma correction is adopted to improve the contrast in dark areas. In the proposed model, a metaheuristic stud krill herd optimization algorithm is employed to determine the optimal constraints to optimize the degree of enhancement based on the objective function defined as discrete entropy or mean square error. Experimental results prove that the proposed technique produces a better visual quality image with an attractive color tone and improved natural appearance without introducing artifacts.