Abstract

In previous publications we presented an image-restoration algorithm for quantum-limited noncoherent imagery that is based on maximum-likelihood estimation [1]. This work was extended from the work of others in nuclear-medicine imaging [2,3]. A practical issue of using this method is in its speed of execution. This method is derived from the iterative expectation-maximization (EM) algorithm [4], and as such its convergence can be slow. We have observed this issue in our simulations as others have for positron-emission tomography [5].

© 1989 Optical Society of America