A New Approach to Image Enhancement Based on Modified Histogram Equalization

Abstract

The problem of improving the efficiency of contrast enhancement for complex low-contrast images in an automatic mode with an acceptable level of computational costs is considered. A new approach to image enhancement by parameter-free intensity transformation based on an estimate of the bivariate distribution of brightness is proposed. Analysis of the bivariate distribution of brightness enables to significantly improve the efficiency of image contrast enhancement. Based on this approach, a new generalized description for the procedure of intensity transformation based on the analysis of the marginal cumulative distribution function of image brightness is proposed. The relationship between the proposed generalized description of the intensity transformation and the traditional definition of histogram equalization is shown. Based on proposed description, a new technique of modified histogram equalization is proposed. The research confirms the effectiveness of the proposed technique. The proposed technique of modified histogram equalization provides an increase in the efficiency of contrast enhancement compared to the well-known traditional technique of histogram equalization.