Image analysis – Image segmentation – Separating document regions using preprinted guides or markings
Reexamination Certificate
1996-12-04
2001-08-14
Mehta, Bhavesh (Department: 2621)
Image analysis
Image segmentation
Separating document regions using preprinted guides or markings
C382S176000, C382S317000
Reexamination Certificate
active
06275608
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to an image processing apparatus of an OCR (optical character recognizing) apparatus, a copying machine, a facsimile, or the like and, more particularly, to image processing method and apparatus and memory medium which can properly extract characters from an input image including a table image.
The invention also relates to image processing method and apparatus and memory medium which can discriminate a format of an input table and properly extract a character image when the character image is extracted from an input table image.
The invention also relates to image processing method and apparatus and memory medium which can automatically discriminate attributes of a table image included in an input image without manually designating the attributes by the user.
2. Related Background Art
An OCR apparatus having an automatic area dividing function performs processes to automatically discriminate an image surrounded by rules as a table image and extract a cell surrounded by the rules of a table as one character area.
After characters were recognized, by inserting a tab code between character trains of a result of the recognition obtained from one character area and inserting a return code between lines, information of the cell extracted from the table image is reflected to document data after completion of the recognition of characters, thereby enabling data to be easily supplied to table calculating software or the like.
According to the foregoing conventional technique, however, when the table image serving as a target of the character recognition is a table without lateral rule, as shown in
FIG. 7
, since a plurality of items arranged in the vertical direction are discriminated as one character area, a document of a form different from that of the input table is obtained and the operation to correct the form to that of the input table is forced on the operator, so that a burden on the operator is increased.
SUMMARY OF THE INVENTION
According to the invention, as mentioned above, a line image is extracted from image information, attributes of the image information are discriminated in accordance with characteristics of the extracted line image, and the discriminated attributes are stored in correspondence to the image information, so that table attributes indicating whether a table image inputted is a table with lateral rules or a table without lateral rule and the like don't need to be manually designated by the operator, the table attributes can be automatically discriminated, and an operability can be improved.
According to the invention as described above, by setting the image information to the image information surrounded by the rules, a plurality of items included in one cell surrounded by the rules can be automatically discriminated.
According to the invention as described above, by setting the image information to the image information which was optically read out, the optical read-out image can be automatically processed without operating keys or the like.
According to the invention as described above, by recognizing the line image as a character and outputting a line delimiter code together with a result of the character recognition, the form of the table can be maintained even after a document was formed by the character recognition.
According to the invention as described above, by displaying a document in accordance with the outputted information, a result of the character recognition of the input image and the form of the table can be confirmed on a display picture plane.
According to the invention as described above, the input image information is divided into small image areas surrounded by rules, line images are extracted from the image information of the divided small image areas, and the line images included in the input image information are ordered in accordance with the positional relation of the small image areas where the line images are located and the positions in the small image areas, so that a form of a special table without lateral rule can be properly discriminated.
According to the invention as described above, by discriminating the attributes of the image information and deciding whether the extraction of the line images is executed or not in accordance with a result of the discrimination of the attribute discriminating means, a proper process is executed in only the proper case, so that the form of the table can be maintained even after a document was formed by the character recognition.
REFERENCES:
patent: 4998285 (1991-03-01), Suzuki et al.
patent: 5048107 (1991-09-01), Tachikawa
patent: 5075895 (1991-12-01), Bessho
patent: 5091964 (1992-02-01), Shimomura
patent: 5111514 (1992-05-01), Ohta
patent: 5119437 (1992-06-01), Kuwamura et al.
patent: 5452380 (1995-09-01), Uehara et al.
patent: 5668892 (1997-09-01), Itonori
patent: 5680478 (1997-10-01), Wang et al.
patent: 5729635 (1998-03-01), Fast et al.
patent: 5737442 (1998-04-01), Alam
patent: 5774580 (1998-06-01), Saitoh
patent: 6-96275 (1994-04-01), None
“Public Domain Optical Character Recognition”, by Michael D. Garris, et al., Document Recognition 2, San Jose, Feb. 6-7, 1995, vol.2422, pp. 2-14.
Canon Kabushiki Kaisha
Fitzpatrick ,Cella, Harper & Scinto
Mehta Bhavesh
LandOfFree
Image processing method and apparatus and memory medium does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing method and apparatus and memory medium, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing method and apparatus and memory medium will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2474907