An image enhancement technique and its evaluation through bimodality analysis

It is well known that the most effective method for comparing images at present is subjective human evaluation. One factor complicating the design and evaluation of a given image treatment is the lack of an objective measure of picture quality. In the present paper a contrast enhancement and noise filtering technique is developed and evaluated through bimodality measure analysis. The proposed algorithm makes the different classes of an image statistically separable without sensibly modifying the contours. This method can be used as an aid in gray-level thresholding.