Previous
abstract 
Next
abstract
When creating an enhanced image or estimating an object's parameters, it is important to establish the error associated with the processing. The Cramer-Rao bound is a tool to calculate the error. It gives the minimum variance that can be achieved by any estimation technique; when the technique is maximum likelihood, the estimator asymptotically approaches the bound, as the number of observations becomes large. We calculate the bound for HST WF/PC imagery and show how it can be used as an adjoint to any enhancement or estimation algorithm. The bound is sensitive to the accuracy of the point-spread function (PSF). We show results for an improved estimation of the PSF.
