Ji-Hyun Yoo, Seong-Yong Ohm, Min-Gyo Chung, Journal of Internet Computing and Services, Vol. 13, No. 3, pp. 61-74, Jun. 2012
10.7472/jksii.2012.13.3.61, Full Text:
Keywords: Image Contrast Enhancement, Mean Brightness Preservation, Histogram Specification, Histogram Equalization, Maximum Entropy, Gaussian Distribution

Abstract

This paper proposes a histogram specification based image enhancement method, which uses the brightness mean and variance of an image to maximize the entropy of the image. In our histogram specification step, the Gaussian distribution is used to fit the input histogram as well as produce the target histogram. Specifically, the input histogram is fitted with the Gaussian distribution whose mean and variance are equal to the brightness mean(${\mu}$) and variance(${\sigma}2$) of the input image, respectively; and the target Gaussian distribution also has the mean of the value ${\mu}$, but takes as the variance the value which is determined such that the output image has the maximum entropy. Experimental results show that compared to the existing methods, the proposed method preserves the mean brightness well and generates more natural looking images.

Statistics
Cite this article