Back to results
US 2003/0156754 A1Published application
Method and system for extracting title from document image
AssigneeRicoh Company, Ltd.
InventorSHIGEKI OUCHI
Life of this patent
Estimating the term…
Abstract
A title is determined from a document image of an unspecified format based upon likelihood of containing a title in each of minimal circumscribing rectangles. The likelihood is determined based upon information obtained during character recognition of the image contained in the minimal circumscribing rectangles. The information includes a font type, a recognition assurance level, character code and so on.
Claims36
A method of determining a title from a document image, comprising:
- dividing the document image into minimal circumscribing rectangles which contain a character image;
- recognizing characters in said minimal circumscribing rectangles;
- and
- determining a title of the document image based upon a likelihood of each of said minimal circumscribing rectangles containing a title, said likelihood being determined based upon information obtained during said character recognition.
1 more element
A system for determining a title from a document image, comprising:
- a character row area determination unit for dividing the document image into minimal circumscribing rectangles which contain a character image;
- a character recognition unit connected to said character row area determination unit for recognizing characters in said minimal circumscribing rectangles;
- and
- a title evaluation point determination unit connected to said character recognition unit for determining a title of the document image based upon a likelihood of each of said minimal circumscribing rectangles containing a title, said likelihood being determined based upon information obtained during said character recognition.
1 more element
Dependencies are inferred from the claim text.