Block based intensity-pair distribution for image contrast enhancement

Abstract

In this paper, we present an image contrast enhancement method based on block-wise intensity pair distribution. Instead of using the intensity-pair distribution of the whole image, this proposed algorithm takes the intensity-pair distribution block-wise and maps the intensity of the center pixel according to an expansion function. To control the contrast stretch and over-enhancement, a linear magnitude mapping function is used instead of a non-linear one. This linear mapping preserves the relative contrast enhancement ratio between the gray levels. The local information from blocks easily facilitates contrast enhancement, brings out subtle edge information, and removes noises from the image. The proposed algorithm can be best suited for detail analysis of image features. © 2006 IEEE.