| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a document processor and the like for detecting coordinates of a frame, without creating format information for every document, for documents wherein the positions and sizes of character frames and field frames are different in each document, and arrangement relationships of the frames are different from one another in spite of the same document type. <P>SOLUTION: A document is divided into partial areas and a plurality of partial format information are created on an area basis. In recognizing the document, an input image is collated with partial formats for each partial area and an optimum partial format is selected. Format information for the whole document is created by combining the optimum partial formats in the respective partial areas. The coordinates of the frame are extracted from the format information dynamically created like that. According to this application, a quasi-typical document can accurately be recognized by using the partial format information. In addition, creation manhours of the format information can be reduced and the capacity of the format information can be reduced as compared with a conventional one. <P>COPYRIGHT: (C)2006,JPO&NCIPI |
