Sequential Binary Image Estimation Algorithms

Abstract

The general problem that we are trying to solve is sequential image-signal interpretation with the applicability of our work to the improvement of hearing aids. As part of this bigger problem, we would like to find a good way to sequentially estimate binary images from noisy gray images. The sequentially requirement is the main difficulty of our work; it arises from the real-time aspect of the application. In this work, we describe two methods for estimating the prior. They are both based on the Metropolis Algorithm with Gibbs priors and an annealing schedule. These two prior estimation methods lead to two sequential binary image estimation algorithms.