Image contrast enhancement and quantitative measuring of information flow

Contrast enhancement is an effective approach for image processing and pattern recognition under conditions of improper illumination. It has a wide variety of applications, such as on object identification, fingerprint verification and face detection. At the same time, unpleasant results might occur when certain types of noises are amplified at the same time. Thus, adaptive image enhancement can be conducted to avoid this drawback, which is used to adapt to the intensity distribution within an image. To evaluate the actual effects of image enhancement, some quantity measures should be taken into account instead of on a basis of intuition exclusively. In this study, a set of quantitative measures is proposed to evaluate the information flow between original and enhanced images. Concepts of the gray level energy, discrete entropy and relative entropy are employed to measure the goodness of the adaptive image enhancement techniques. The images being selected are the scenery picture, architecture picture, static object picture and living creature picture.