CHARACTER RECOGNITION APPARATUS, CHARACTER RECOGNITION METHOD, AND COMPUTER-READABLE MEDIUM

• US 20140093161A1
• Filed: 05/17/2013
• Published: 04/03/2014
• Est. Priority Date: 10/01/2012
• Status: Active Grant

• Current Assignee: Fujifilm Business Innovation Corp. (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Inventors: ODA, Hideto
• Granted Patent: US 9,135,525 B2
• Time in Patent Office: Days
• US Class Current: 382/159
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06F 18/253 of extracted features
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/19147 Obtaining sets of training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06V 30/1916 Validation; Performance evaluation

Abstract

A character recognition apparatus includes an evaluation-value output unit, a generation unit, a learning unit, and a determination unit. The evaluation-value output unit outputs evaluation values for each of different character recognition programs. Each evaluation value indicates a degree to which an inputted character pattern corresponds to each of character codes to be recognized using the character recognition program. The generation unit generates feature information for the character pattern. The feature information includes, as elements, the evaluation values. The learning unit learns classifications for feature information on a character-code-by-character-code basis based on feature information generated for a character pattern for which a character code is specified in advance. The determination unit determines a character code of an unknown character pattern whose character code is unknown, based on which classification among the learned classifications includes feature information generated for the unknown character pattern.

Claims

1. A character recognition apparatus comprising:

an evaluation-value output unit that outputs a plurality of evaluation values for each of a plurality of character recognition programs which are different from each other, each of the plurality of evaluation values indicating a degree to which a character pattern which has been input corresponds to each of a plurality of character codes to be recognized using the character recognition program;

a generation unit that generates feature information for the character pattern which has been input, the feature information including, as elements, the plurality of evaluation values being output by the evaluation-value output unit;

a learning unit that learns classifications for feature information on a character-code-by-character-code basis on the basis of feature information generated by the generation unit for a character pattern for which a character code is specified in advance; and

a determination unit that determines a character code of an unknown character pattern on the basis of which classification among the learned classifications for feature information on a character-code-by-character-code basis includes feature information generated by the generation unit for the unknown character pattern, the unknown character pattern being a character pattern whose character code is unknown.

2. The character recognition apparatus according to claim 1,

wherein the plurality of character recognition programs include a first character recognition program and a second character recognition program, and

wherein at least some of a plurality of character codes to be recognized by the first character recognition program match at least some of a plurality of character codes to be recognized by the second character recognition program.

3. The character recognition apparatus according to claim 1, further comprising:

a specifying unit that, for each of a plurality of character patterns for which character codes are specified in advance, compares a character code determined by the determination unit for the character pattern with the character code specified for the character pattern and that specifies a character code whose accuracy rate is equal to or less than a threshold on the basis of a result obtained through the comparison; and

a unit that adds a character recognition program which is to recognize the character code specified by the specifying unit to the evaluation-value output unit.

4. The character recognition apparatus according to claim 2, further comprising:

a specifying unit that, for each of a plurality of character patterns for which character codes are specified in advance, compares a character code determined by the determination unit for the character pattern with the character code specified for the character pattern and that specifies a character code whose accuracy rate is equal to or less than a threshold on the basis of a result obtained through the comparison; and

a unit that adds a character recognition program which is to recognize the character code specified by the specifying unit to the evaluation-value output unit.

5. A non-transitory computer readable medium storing a program causing a computer to execute a process for recognizing a character, the process comprising:

outputting a plurality of evaluation values for each of a plurality of character recognition programs which are different from each other, each of the plurality of evaluation values indicating a degree to which a character pattern which has been input corresponds to each of a plurality of character codes to be recognized using the character recognition program;

generating feature information for the character pattern which has been input, the feature information including, as elements, the plurality of evaluation values;

learning classifications for feature information on a character-code-by-character-code basis on the basis of feature information generated for a character pattern for which a character code is specified in advance; and

determining a character code of an unknown character pattern on the basis of which classification among the learned classifications for feature information on a character-code-by-character-code basis includes feature information generated for the unknown character pattern, the unknown character pattern being a character pattern whose character code is unknown.

6. A character recognition method comprising:

outputting a plurality of evaluation values for each of a plurality of character recognition programs which are different from each other, each of the plurality of evaluation values indicating a degree to which a character pattern which has been input corresponds to each of a plurality of character codes to be recognized using the character recognition program;

generating feature information for the character pattern which has been input, the feature information including, as elements, the plurality of evaluation values;

learning classifications for feature information on a character-code-by-character-code basis on the basis of feature information generated for a character pattern for which a character code is specified in advance; and

determining a character code of an unknown character pattern on the basis of which classification among the learned classifications for feature information on a character-code-by-character-code basis includes feature information generated for the unknown character pattern, the unknown character pattern being a character pattern whose character code is unknown.