A knowledge based architecture for the virtual restoration of ancient photos. (February 2018)

• Record Type: Journal Article
• Authors: Ardizzone, E.; Dindo, H.; Mazzola, G.
• Is Part Of: Pattern recognition. Volume 74(2018:Feb.)
• Page Start: 326
• Page End: 339
• Publication Date: 2018-02
• Subjects: Image restoration -- Historical photos -- Digitization -- Ontology -- Knowledge base; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2017.09.031
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: A solution is proposed to help non-expert users in restoring their damaged photos. We represented the restoration domain with an ontology. Classes and properties of the ontology are included into a knowledge base. The knowledge base guide the users through the restoration process. A prototypal tool and a web application version have been implemented. Abstract: Historical images are essential documents of the recent past. Nevertheless, time and bad preservation corrupt their physical supports. Digitization can be the solution to extend their "lives", and digital techniques can be used to recover lost information. This task is often difficult and time-consuming, if commercial restoration tools are used for the purpose. A new solution is proposed to help non-expert users in restoring their damaged photos. First, we defined a dual taxonomy for the defects in printed and digitized photos. We represented our restoration domain with an ontology and we created some rules to suggest actions to perform in case of some specific events. Classes and properties of the ontology are included into a knowledge base, that grows dynamically with its use. A prototypal tool and a web application version have been implemented as an interface to the database, and to support non-expert users in the restoration process.