Histogram Equalization based on Binary Optimization Model with Subjective and Objective Consistency

Abstract

Histogram equalization (HE) is the most famous method for image enhancement due to its low computation complexity and wide application scope. However, existing HE-based methods often lead to over-enhancement, under-enhancement and unnatural visual perception. To overcome these defects, an adaptive HE algorithm is proposed in this paper. The core idea is to adjust the histogram with the optimal gamma correction parameter. Firstly, a sequence of images with gradually changing quality is obtained by traversing the gamma parameters, and two measurement values are calculated for each image in the sequence. Then a binary optimization model is proposed to search for the optimal parameter. Importantly, in order to compensate for the difference between the objective model and the subjective perception, a novel correction factor is designed to adjust the optimal parameter from the model. Finally, secondary gamma correction is performed by inverting the final parameter to preserve image details and prevent under enhancement. Experimental results show that the proposed algorithm outperforms those state-of-the-art HE-based algorithms.