LOGICAL STRUCTURE EXTRACTING DEVICE

摘要

PROBLEM TO BE SOLVED: To provide a logical structure extracting device which can extract a logical structure with high precision for documents in various styles without placing any load on a user and the whole system. SOLUTION: The user specifies the logical structure to be extracted through a logical structure specification part 9. An area division part 4 divides a document image into areas according to parameters for area division that a parameter setting part 3 sets. An area structure decision part 5 decides whether the result of the area division is the logical structure assumed by an area structure extraction part 7. When not, the parameters are changed by the parameter setting part 3 and the image is divided again into areas. When the area structure that the logical structure extraction part 7 assumes is obtained, a character recognition part 6 recognizes characters in all the areas and the logical structure extraction part 7 extracts the logical structure that the user specified. The extracted logical structure is sent to a structured document process part 8.