Image processing device and image processing method for determining character data from image data based on misrecognition character data and outputting character data in accompaniment to image data

Patent metadata

Field	Value
Publication number	US-9614990-B2
Application number	US-201514808158-A
Country	US
Kind code	B2
Filing date	Jul 24, 2015
Priority date	Jul 30, 2014
Publication date	Apr 4, 2017
Grant date	Apr 4, 2017

Abstract

A font recognition section recognizes a font of characters depicted by character image data included in image data upon first character data corresponding to the character image data being acquired through character recognition. A font character recognition section compares the character image data to font character image data depicting characters in the recognized font and acquires second character data corresponding to font character image data when the font character image data matches the character image data. A recognized character determination section determines whether or not the first character data matches the second character data, and when determining that the first character data does not match the second character data, adopts the second character data, instead of the first character data, as recognized character data for the character image data.

First claim

What is claimed is:

1. An image processing device comprising:
an image reading section that reads image data from a document;
a character recognition section that acquires first character data, the first character data corresponding to character image data included in the image data and indicating text data identified by a character code;
a misrecognition character determination section that determines whether or not misrecognition character data is present in the first character data, the misrecognition character data being pre-registered as easily misrecognized character data and indicating text data identified by a character code;
a font recognition section that recognizes a font of characters depicted by the character image data when the misrecognition character determination section determines that the misrecognition character data is present;
a font character recognition section that compares the character image data to font character image data depicting characters in the recognized font and acquires second character data when the font character image data matches the character image data, the second character data corresponding to the font character image data and indicating text data identified by a character code; and
a recognized character determination section that determines whether or not the first character data matches the second character data, wherein
the image reading section
outputs the first character data in accompaniment to the image data when the misrecognition character determination section determines that the misrecognition character data is not present,
outputs the second character data in accompaniment to the image data when the recognized character determination section determines that the first character data does not match the second character data, and
outputs the first character data in accompaniment to the image data when the recognized character determination section determines that the first character data matches the second character data.

2. The image processing device according to claim 1 , wherein
when the font recognition section recognizes two or more fonts for characters depicted by the character image data that compose a single word as a result of performing font recognition with respect to each character depicted by the character image data, the font recognition section recognizes a most frequently recognized font among the two or more fonts as the font of the characters depicted by the character image data.

3. An image processing method comprising
implementation by a processor of:
reading image data from a document;
acquiring first character data, the first character data corresponding to character image data included in the image data and indicating text data identified by a character code;
determining whether or not misrecognition character data is present in the first character data, the misrecognition character data being pre-registered as easily misrecognized character data and indicating text data identified by a character code;
recognizing a font of characters depicted by the character image data when determining that the misrecognition character data is present in the first character data;
comparing the character image data to font character image data depicting characters in the recognized font and acquiring second character data when the font character image data matches the character image data, the second character data corresponding to the font character image data and indicating text data identified by a character code;
determining whether or not the first character data matches the second character data;
outputting the first character data in accompaniment to the image data when determining that the misrecognition character data is not present in the first character data;
outputting the second character data in accompaniment to the image data when determining that the first character data does not match the second character data; and
outputting the first character data in accompaniment to the image data when determining that the first character data matches the second character data.

Assignees

• Kyocera Document Solutions Inc

Inventors

• Hayashi Hironori

Classifications

• G06V30/12 Primary

  Detection or correction of errors, e.g. by rescanning the pattern · CPC title

• H04N1/00798 Primary

  Circuits or arrangements for the control thereof, e.g. using a programmed control device or according to a measured quantity · CPC title

• G06V30/245

  Font recognition · CPC title

• G06V30/10

  Character recognition · CPC title

• G06K9/03

  Physics · mapped topic