A Fuzzy Logic Based Expert System for Quality Assurance of Document Image Collections

Huge document image collections in digital libraries are prone to reduced quality and require automatic quality assurance. This paper presents an approach for bringing together information automatically aggregated from a quality assurance tool and expert knowledge related to digital preservation. The main contribution of this work is the definition of fuzzy expert rules and the application of fuzzy logic in order to support digital preservation experts in decision making. Page duplicate detection in document image collections is demonstrated in detail. Another contribution is a multi level analysis approach that comprises not only image processing, but also collection metadata aggregation e.g. file names, file size, creation date, possible inconsistency detection. This expert system supports planning for long term preservation and ensures quality of the digitized content. Our goal is to create a reliable inference engine and human maintainable conclusions from the output of an image processing tool that detects duplicates based on methods of computer vision. Another goal is to give a system at hand that supports digital document handling for teaching and education. We employ artificial intelligence technologies (i.e. fuzzy logic, expert rules) to emulate reasoning about the knowledge base similar to a human expert. A statistical analysis of the automatically extracted information from the image comparison tool and the qualitative analysis of the aggregated knowledge are presented in the evaluation part of the paper.