Image Enhancement using Bi-Histogram Equalization with Adaptive Sigmoid Functions

Among the contrast-enhancement methods, histogram equalization is the most popular. However, its major drawback is that it over-enhances the image and shifts its mean brightness and, consequently, it creates an unnatural look. In this paper, we propose a method that overcomes this problem by splitting the image histogram into two sub-histograms, using the mean as a threshold, and replacing their cumulative distribution functions with two smooth sigmoids with their origins placed on the median of the sub-histograms. Our method has been tested on gray scale images taken from the USC-SIPI database. Experimental results have shown that the proposed method outperforms other state-of-the-art methods in terms of contrast-enhancement and brightness-preservation.