Blind deconvolution: errors, errors everywhere

This paper focuses on a class of methods that accounts for uncertainty in the model as well as the data. Our example concerns spectroscopy - the attempt to reconstruct a true spectrum from an observed one. The problem we're considering is sometimes called blind deconvolution, because we're trying to unravel not only the spectrum, but the function that caused the blurring. These problems also arise in image deblurring.