A New Technique of Image Enhancement by Equalizing the Bivariate Probability Density of Brightness

Abstract

Histogram equalization is the most well-known intensity transformation technique and is widely used to enhance images owing to high efficiency, simplicity, and low computational cost. However, the traditional approach to image histogram equalization has a number of disadvantages, which significantly limit the possibilities of its use for images enhancement in automatic mode. In this paper, we propose a new technique of image intensity transformation based on equalizing the bivariate probability density of brightness. This approach provides effective image enhancement without the appearance of unwanted artifacts. A generalized description for intensity transformation based on equalizing the joint bivariate brightness distribution is presented. The relationship between the proposed generalized description and the traditional definition of histogram equalization is shown. The proposed technique provides an increase in the efficiency of image enhancement compared to the traditional technique of histogram equalization.