We consider the following problem a black and white image is observed in digitized form. Unfortunately the real image is not observed at some stage the image has been distorted with noise. Our objective is to remove as much of the noise as possible, to get approximately the original image back. In a more mathematical setting, let x be an m by n array, with entries 0 and 1 x is considered to be a realization of a random variable X. We do not observe the image x. Instead we observe y, a noisy version of x that is a realization of the random variable Y, where the distribution of Y depends on x. We want to estimate x on the basis of y.
This article is from 'Computing Science and Statistics: Proceedings of the Symposium on Interface Critical Applications of Scientific Computing (23rd): Biology, Engineering, Medicine, Speech Held in Seattle, Washington on 21-24 April 1991,' AD-A252 938, p583-586.