METHOD OF EXTRACTION OF FIELD FROM SCANNING IMAGE OF DOCUMENT AND DATA-PROCESSING SYSTEM

摘要

PURPOSE: To automatically extract the field of an image from a document in an improved method by using viewpoint structure which characterizes the interior of the field and a peripheral section when the image is extracted for recognizing a character. CONSTITUTION: A format definition data set 400 is accessed through the use of a format ID and the whole fields 405 of the document which is scanning- inputted from it. In this process, coordinates where the respective fields exist are adjusted by using a value indicating the inclinationdegree of the specified document which is scanning-inputted and field viewpoint structure for the scanning-inputted document is compared with field viewpoint structure 408 in the data set. A system judges whether or not a viewpoint basic element exists in the data set concerning the inside the field and the section surrounding the field and it is compared with the viewpoint basic element of the corresponding section in the scanning-inputted document. Then, the whole coincident viewpoint basic elements are removed from the scanning-inputted image.