Identification of image blur parameters by the method of generalized cross-validation

Generalized cross-validation (GCV) is introduced to address the blur identification problem. Motivated by the success of GCV in identifying optimal noise smoothing parameters for image restoration, the method is extended to the problem of identifying blur parameters as well. Experiments are presented which show that GCV is capable of yielding good identification results. Some potential advantages of GCV over the maximum-likelihood approach are discussed