ASPCS
 
Back to Volume
Paper: Maximum Likelihood Estimator and Bayesian Reconstruction Algorithms with Likelihood Cross-Validation
Volume: 25, Astronomical Data Analysis Software and Systems I
Page: 210
Authors: Nunez, Jorge; Llacer, Jorge
Abstract: In this paper we describe the cross-validation method as applied to astronomical image reconstruction and present results of reconstructing images from the Hubble Space Telescope Faint Object Camera. Both Maximum Likelihood Estimator (MLE) and Bayesian reconstruction methods with Entropy prior benefit from the approach: in the MLE case, cross-validation provides a stopping rule which results in optimized images; in Bayesian reconstruction with Entropy prior, the cross- validation allows specifying de hyperparameter that controls the relative weight of Likelihood vs. Entropy in the reconstruction to obtain the best Maximum a Posteriori image.
eAccess Cost: $9.00
Back to Volume