DOCUMENT IMAGE PROCESSING UNIT AND DOCUMENT IMAGE PROCESSING METHOD

Japanese Patent JP3768743

B2

PROBLEM TO BE SOLVED: To allow a processing unit for a document image to link character strings included in a drawing and a line drawing of a document with high precision to other parts of the document even when it is difficult to accurately extract character information of the drawing and the line drawing or the like from the document.
SOLUTION: When a 1st region and a 2nd region being objects of linking are designated on a document image (ST101), (ST104), character recognition is applied to the 1st region (ST102), and recognized words are registered together with positional information (ST103). Furthermore, character recognition is applied to the 2nd region (ST105), and a recognized character string is collated with the registered word in the 1st region (ST106). Then in the case that the collation between the recognized character string in the 2nd region and the registered word in the 1st region indicates agreement, link information between the positional information of the character string in the 2nd region and the positional information of the registered word that are in agreement as a result of the collation is generated (ST107).

Hiroaki Kubota
Mitsuyoshi Okazaki

JP26521299A

April 19, 2006

September 20, 1999

Click for automatic bibliography generation  

Toshiba Corporation

G06K9/00; G06T1/00; H04N1/00; (IPC1-7): H04N1/00; G06T1/00; G06K9/00

JP10063674A

Takehiko Suzue
Sadao Muramatsu
Atsushi Tsuboi
Ryo Hashimoto
Satoshi Kono
Makoto Nakamura
Shoji Kawai

Previous Patent: PAPER FEEDING ROLLER MECHANISM FOR PRINTING DEVICE

Next Patent: SEMICONDUCTOR DEVICE AND MANUFACTURING METHOD THEREOF