Adaptive Enhancement of Gray Level and True Color Images with Quantitative Measurement Using Entropy and Relative Entropy

Image adaptive contrast stretching enhancement is a popular and practical approach in automated image recognition systems. Its applications may consist of medical diagnosis, homeland security, telecommunication, weather forecasting and environment protection. Under conditions of an improper or disturbed illumination, this scheme adapts to intensity distributions to improve the image quality. This technique is suitable for both gray level images and trimulus true color images. In general, quality improvement via adaptive image enhancement can be observed directly from enhanced images. Considering the accuracy requirement in decision supports, quantitative measures are necessary. In this case, concepts from the information theory are proposed to evaluate the effects of adaptive image enhancement in this article. Particularly, discrete entropy, relative entropy and mutual information have been applied to indicate the impact of adaptive image contrast enhancement techniques for both gray level images and true color images. These useful results can be expanded to other image processing technologies for a variety of practical implementation problems.