Computer and document identification method

Patent metadata

Field	Value
Publication number	US-10783366-B2
Application number	US-201816117198-A
Country	US
Kind code	B2
Filing date	Aug 30, 2018
Priority date	Nov 6, 2017
Publication date	Sep 22, 2020
Grant date	Sep 22, 2020

Abstract

A computer that extracts an attribute which is a text string contained in an predetermined examination target document stores template information for managing a plurality of templates in which a type of attribute is defined, executes a text recognition process on image data of the document, extracts an attribute corresponding to the type of attribute using a result of the text recognition process and the plurality of templates, selects a template based on the extracted attribute, generates output information that includes the attribute extracted using the selected template and is used for the examination; determines a type of confirmation operation performed on the output information, before the examination, based on a comparison result between an evaluation value indicating credibility of the output information and a threshold, and corrects the determined type of confirmation operation based on the text string contained in the document.

First claim

What is claimed is:

1. A computer that extracts an attribute which is a text string contained in a predetermined examination target document, the computer comprising:
a processor; and
a storage device that is connected to the processor,
wherein the storage device stores template information for managing a plurality of templates in which at least one type of attribute is defined,
wherein the template information includes a plurality of entries formed of identification information of the template and identification information indicating the type of attribute, and
wherein the processor
executes a text recognition process on image data of the document,
extracts an attribute corresponding to the type of attribute defined in each of the plurality of templates using a result of the text recognition process and the plurality of templates,
selects a template based on the extracted attribute,
generates output information that includes the attribute extracted using the selected template and is used for an examination,
calculates an evaluation value indicating credibility of the output information,
determines a type of confirmation operation performed on the output information, before the examination, based on a comparison result between the evaluation value and a threshold,
determines whether it is necessary to correct the determined type of confirmation operation, based on a text string contained in the document, and
corrects the determined type of confirmation operation when the processor determines that it is necessary to correct the determined type of confirmation operation.

2. The computer according to claim 1 ,
wherein the storage device stores first management information for managing a risk specifying rule that defines a text string for specifying a type of risk which is likely to occur in the examination or a condition of the text string and second management information for managing a correction rule that defines a correction method for the determined type of confirmation operation for the output information,
wherein the first management information stores information regarding the risk specifying rule corresponding to another type of risk,
wherein the correction rule is formed of a conditional expression defined by a combination of a type of confirmation operation before the correction and the specified type of risk, and a type of confirmation operation after the correction, and
wherein the processor specifies a type of risk which is in the document based on the first management information and the output information,
determines whether there is a corresponding correction rule with reference to the second management information based on the combination of the type of confirmation operation before the correction and the specified type of risk, and
corrects the determined type of confirmation operation based on the corresponding correction rule when the processor determines that there is the corresponding correction rule.

3. The computer according to claim 2 ,
wherein the processor stores learning data in which an attribute extracted from the document is associated with a result of the examination obtained using the output information operated in accordance with the determined type of confirmation operation, in the storage device,
generates display information for displaying a text string in which a predetermined type of risk is likely to occur or a condition of the text string by performing a learning process using the learning data, and
outputs the generated display information.

4. The computer according to claim 2 ,
wherein the first management information includes a risk specifying rule that defines a condition of a magnitude relation of a text string corresponding to a numerical value.

5. A document identifying method performed by a computer that extracts an attribute which is a text string contained in a predetermined examination target document,
wherein the computer includes a processor and a storage device that is connected to the processor,
wherein the storage device stores template information for managing a plurality of templates in which at least one type of attribute is defined,
wherein the template information includes a plurality of entries formed of identification information of the template and identification information indicating the type of attribute, and
wherein the document identifying method comprises:
a first step of executing, by the processor, a text recognition process on image data of the document;
a second step of extracting, by the processor, an attribute corresponding to the type of attribute defined in each of the plurality of templates using a result of the text recognition process and the plurality of templates;
a third step of selecting, by the processor, a template based on the extracted attribute;
a fourth step of generating, by the processor, output information that includes the attribute extracted using the selected template and is used for an examination;
a fifth step of calculating, by the processor, an evaluation value indicating credibility of the output information;
a sixth step of determining, by the processor, a type of confirmation operation performed on the output information, before the examination, based on a comparison result between the evaluation value and a threshold;
a seventh step of determining, by the processor, whether it is necessary to correct the determined type of confirmation operation, based on a text string contained in the document; and
an eighth step of correcting, by the processor, the determined type of confirmation operation when the processor determines that it is necessary to correct the determined type of confirmation operation.

6. The document identifying method according to claim 5 ,
wherein the storage device stores first management information for managing a risk specifying rule that defines a text string for specifying a type of risk or a condition of the text string which is likely to occur in the examination and second management information for managing a correction rule that defines a correction method for the determined type of confirmation operation for the output information,
wherein the first management information stores information regarding the risk specifying rule corresponding to another type of risk,
wherein the correction rule is formed a conditional expression defined by a combination of a type of confirmation operation before the correction and the specified type of risk, and a type of confirmation operation after the correction, and
wherein the seventh step includes a step of specifying, by the processor, a type of risk which is in the document based on the first management information and the output information and a step of determining, by the processor, whether there is a corresponding correction rule with reference to the second management information based on the combination of the type of confirmation operation before the correction and the specified type of risk, and
wherein the eighth step includes a step of correcting, by the processor, the determined type of confirmation operation based on the corresponding correction rule.

7. The document identifying method according to claim 6 , further comprising:
a step of storing, by the processor, learning data in which the attribute extracted from the document is associated with a result of the examination obtained using the output information operated in accordance with the determined type of confirmation operation, in the storage device,
a step of generating, by the processor, display information for displaying a text string or a condition of the text string in which a predetermined type of risk is highly likely to occur by performing a learning process using the learning data,

Assignees

• Hitachi Ltd

Inventors

• Suzuki Yasufumi
• Shinjo Hiroshi
• Odate Ryosuke
• Motobayashi Masahiro

Classifications

• G06V30/12

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

• G06F18/217

  Validation; Performance evaluation; Active pattern learning techniques · CPC title

• G06V30/40 Primary

  Document-oriented image-based pattern recognition · CPC title

• G06V30/10

  Character recognition · CPC title

• G06V10/96

  Management of image or video recognition tasks · CPC title