Next: About this document
Up: Image Restoration and Super-Resolution
Previous: Image and Spectrum
As a result of our explorations, we can set forth several points in conclusion.
- The convolution network has proven to be a robust method for
determination of an effective solution to a class of Fredholm integral
equations of the first kind.
- The convolution network was originally developed to search for an
efficient, data independent method of image reconstruction of Hubble Space
Telescope imagery. This has proven to be highly satisfactory and work is
- The convolution network has proven successful in resolution
enhancement of many imaging systems.
- The convolution network can successfully be applied for the
processing of sinc apodized spectra to obtain a positive semi-definite
representation of the spectra. This provides a means for obtaining a
baseline for the sinc apodized spectra.