Image Restoration and Super-Resolution by Novel Applications of a Neural Network

Gordon Chin Stephen L. Mahan and William E. Blass

A new instrument paradigm is proposed based on the discovery of a method to determine a robust inverse point spread function for a scientific observing instrument modeled as a linear system. As a result of this discovery, it is possible to contemplate an instrumental extension which results in the recovery of a major portion of lost resolution due to the blurring effects of the PSF. Implementation of the instrumental extension and the resulting resolution enhancement is independent of prior knowledge of or access to the observed data. The method is applied to HST images as well as several one dimensional spectral data sets. Results of HST recoveries are compared to Richardson-Lucy

Pages 49-57

Full paper available in HTML and compressed PostScript