Document identification using shape trees

Article type: Research Article

Authors: Henker, Uwe^{a; *} | Petersohn, Uwe^b

Affiliations: [a] DOCexpert Computer GmbH, Kirschhäckerstraße 27, D-96052 Bamberg, Germany | [b] Department of Computer Science, Dresden University of Technology, D-01062 Dresden, Germany

Correspondence: [*] Corresponding author. E-mail: [email protected]

Abstract: We present a technique to identify documents using an abstraction of a scanned image. The types of documents to identify must be known to the system a priori. To this end, the necessary features are saved in a case base [8] as shape trees. This file also contains rules for possible further processing. In an extremely reduced image, it is possible to filter out the significant, distinguishing information from the image and recognize it using case-based reasoning (CBR) [8]. This method has been demonstrated and proven by an example of experiments using medical order forms. An average of 97% of the forms were correctly identified; none were identified incorrectly.

Keywords: Document identification, shape tree, document processing, case based reasoning, optical mark reading, similarity, segmentation

DOI: 10.3233/HIS-2009-0062

Journal: International Journal of Hybrid Intelligent Systems, vol. 6, no. 1, pp. 15-26, 2009