Document Management as a Database Problem
Rudolf Bayer · Very Large Data Bases · 1995
1. Document Acquisition There is a broad spectrum of techniques how to acquire documents in such a way, that they are in computerreadable form and can be stored in a document base. This spectrum ranges from fully automatic at low cost via semiautomatic using tools like scanners and optical character recognition (OCR) to manual acquisition according to elaborate rules and regulations. The purpose of high quality document acquisition is to capture the structure and the semantic content of a document not as far as possible but as far as affordable. Presently the state of the art of semiautomatic acquisition of paper documents is scanning followed by OCR. This yields a facsimile image and the text content, but no structure and no real semantics. In many cases the text produced by OCR is low quality and must be corrected to be useful for effective retrieval. Various projects are under way to capture structure and semantics automatically [1,12], but they have not reached sufficient maturity to be used in production environments like libraries or businesses.