ABSTRACT
In this chapter, we propose an alternative solution to the problem of adaptive regularization by adopting a new global cost measure. It was previously seen that the newly formulated ETC measure is capable of distinguishing between the textures and edges in an image. It is further observed in this chapter that the distribution function of this measure value in a typical image assumes a characteristic shape, which is comparatively invariant across a large class of images, and can thus be considered a signature for the images. This is in contrast with the distribution function of other quantities such as the gray level values which varies widely from image to image, as can be confirmed by observing the gray level histograms of different images.
