Image Restoration from Multiple Sources
Abstract
This paper proposes a new method of image restoration. The proposed method allows to combine information from several sources, taking the perceived credibility of each into account. It is applicable to both ordinal (e.g., gray level image) and non-ordinal (e.g., classified forest map) categorized images. The accuracy checks have shown the method to be robust with respect to the prior information and the accuracy of the sources. Two application examples are provided.