Comments on "Two dimensional Bayesian estimate of images"

In a recent paper [1], a certain two-dimensional recursive estimation procedure was claimed to be optimal. Here, we show that it is not and point out a mistake in the proof given.