How can document analysis help in capturing five million pages?

This paper describes how document analysis techniques like OCR, layout analysis, model based recognition and interpretation can be fruitfully applied in the field of high-volume, high-accuracy document capturing with very hard time constraints. We describe the way we set up a workflow that enables reliable capturing of real-estate registration documents. Techniques from document analysis are used to speed up the archiving process and to raise its quality. In particular an automatic determination of the positions for input of new text in partially filled text columns is described. This enables to bridge the gap between the non-coded archived documents and the coded information which is used to update the documents later