JPH06301821A - Character recognition processing system - Google Patents

Abstract

PURPOSE:To improve the OCR recognition rate and to reduce the burden of verifying work on an operator by performing know-ledge processing based on a dictionary data base at the time of character recognition of each document image. CONSTITUTION:A host computer 1 is provided with a corporation dictionary source generating part to generate a corporation dictionary source as an example of data bases. A knowledge processing device 4 performs the knowledge processing by a knowledge dictionary 5 at the time of character recognition of each document image in a character recognition device 3. The latest dictionary data base for knowledge base to which the corporation dictionary source, which is generated based on an actual result master file 2 by the host computer 1 and consists of proper nouns peculiar to the region like individual names, business type names, corporation names, is converted is stored in the knowledge dictionary 5. This latest dictionary data base is transferred to the knowledge processing device 4 by online, and the knowledge processing is performed in the knowledge processing device 4 by the knowledge dictionary 5 of proper nouns at the time of character recognition of each document image in the character recognition processing device.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition processing system for improving the OCR recognition rate in a character recognition device.

[0002]

2. Description of the Related Art Conventionally, a character recognition device 100 as shown in FIG. 5 is known as an optical character recognition device.

The character recognition device 100 is connected to a host computer 100 via a LAN or a line. The host computer 101 has a result master file 102. This achievement master file 102
Is provided by a related institution of the form read by the character recognition device 100, for example, a financial institution such as a bank, and is used to reduce the number of verification items in the form of a customer with a track record. Further, a verification terminal 103 recognized by character recognition is connected to the host computer 101, and the verification terminal 103 stores an image of characters written on a form and character code data converted by the character recognition device 100. Verify.

According to such a character recognition device 100,
The form image or the OCR form is subjected to so-called plain reading based on a dictionary (not shown) provided in advance and converted into character code data. After that, the verification terminal 103
The image of the form and the character code data converted by the character recognition device 100 are verified by.

[0005]

However, the above-described conventional character recognition device 100 has a problem that the OCR recognition rate is low because it only reads the form image or the OCR form.

[0006] In particular, there is a problem that the OCR recognition rate for proper nouns peculiar to a region such as an individual name, a business type name, a corporate name, etc.

The present invention has been made in view of the above circumstances, and it is an object of the present invention to improve the OCR recognition rate of a character recognition device, thereby reducing the burden of verification work on each operator.

[0008]

In order to achieve the above-mentioned object, a character recognition processing system according to the present invention is a record master created by accumulating record data processed in the past several months in a host computer. Database creation means for editing and creating a dictionary database of proper nouns in the area such as personal name, business category name, corporate name from file
It is characterized by comprising a knowledge processing device that performs knowledge processing when character recognition of each form image in the character recognition device based on the created dictionary database.

[0009]

With the above arrangement, the host computer accumulates and manages the record data processed in the past several months to create a record master file. The latest dictionary database of proper nouns in the area, such as personal name, business category name, and corporate name, is edited and created based on the result data. This latest dictionary database is transferred online to the knowledge processing device. At the time of character recognition of each form image in the character recognition processing device, knowledge processing is performed by the knowledge dictionary of proper nouns in the knowledge processing device. The OCR recognition rate is improved by performing the OCR knowledge processing using this knowledge dictionary.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows the configuration of an embodiment of a character recognition processing system according to the present invention.

This character recognition processing system includes a host computer 1, a record master file 2 connected to the host computer 1, and a host computer 1 as well.
A character recognition device (OCR) 3 connected to the computer via a LAN or a line, a knowledge processing device 4 connected to the character recognition device 3, a knowledge dictionary 5 connected to the knowledge processing device 4, and the host The verification terminal 6 connected to the computer 1 is provided.

The actual result master file 2 is a data file in which the operational results processed by the host computer 1 in the past several months are accumulated. For example, in the case of a bank, FIG.
As shown in, various items related to customers such as bank name, branch name, subject (current account etc.), account number, amount, client, recipient, telephone number, etc. are set, and transaction records are stored in corresponding items. It has become so.

The host computer 1 of this embodiment includes a corporate dictionary source creating section 1A as a knowledge dictionary creating means, and creates a corporate dictionary source as shown in FIG. 3 as an example of a database described later.

The knowledge processing device 4 executes knowledge processing by the knowledge dictionary 5 when the character recognition device 3 performs character recognition of each form image. The knowledge dictionary 5 is created by the host computer 1 based on the result master file 2. Personal name,
The latest dictionary database in which a corporate dictionary source composed of proper nouns unique to the region such as a business name and a corporate name is converted into a knowledge base is stored.

The verification terminal 6 is a terminal device for verifying the image of the character entered on the form and the character code data converted by the character recognition device 3.

Next, with reference to FIG. 4, the operation of this embodiment will be described focusing on the processing procedure of the corporate dictionary creation processing unit 1A.

First, the actual result data processed in the past several months is accumulated in the host computer 1, and the actual result data is managed to create the actual result master file 2.

Next, based on the result data, the latest dictionary database of personal names, business category names, corporate names, and other proper nouns unique to the region is edited by the corporate dictionary creation processing unit 1A of the host computer 1 as follows. -Created.

As shown in FIG. 4, the corporate dictionary creating processor 1A creates a corporate name file creating process 8, a corporate name file 9,
The first sort processing 10, the first work file 11, the second sort processing 12, the second work file 13, and the corporate dictionary source creation processing 14 are provided and the corporate dictionary source 15 is created.

First, by the corporate name file creation process 8,
The business name in the recipient name is excluded by using the achievement master file 2 and the business name file 7, and a corporate name file 9 is created.

Here, the business category name indicates the business mode of a corporation, and the business category name of "corporation A" is "corporation" or "(ka)" and the business category name of "B corporation". Is a "corporation" or "ka", and the business name of the limited company C is "limited company" or "(you)".

First, using the created corporate name file 9,
Sorting is performed in ascending order using the corporate name as a key by the sorting process 10 and is output to the first work file 11. If a record with the same key exists at the time of sorting, the number of records is also counted and output.

The second sort processing 12 using the first work file 11 sorts in descending order of the number of appearances (the number of records) and in ascending order of corporate name (aiueo order). Output to work file 13. Using the second work file 13, the corporate dictionary source creation processing 14 converts the file into a format suitable for OCR character recognition to create a corporate dictionary source as shown in FIG.

The latest knowledge dictionary 5 is created by transferring this latest knowledge dictionary source to the knowledge processing device 5 online and adding it.

As a result, the character recognition device 3 uses the created knowledge dictionary 5 for each OCR form image or OCR.
Knowledge processing is performed by using it when recognizing characters on a form.

By the above method, the OCR character recognition rate in the character recognition device 3 is improved, and the work load of each operator in the verification work is reduced.

In this embodiment, a financial institution such as a bank has been described as an example, but the present invention is not limited to an OCR entry system in a financial institution, and characters of various distribution industries, government offices, etc. OC by recognition device
It is also applicable to the R entry system.

[0028]

As described above, according to the present invention, each form image in the character recognition device is based on the dictionary database of the proper nouns of the region, such as the personal name, business category name, corporate name, etc. created from the performance data. Since the knowledge processing is performed at the time of character recognition of OCR of the character recognition device
The recognition rate is improved, and the burden of the operator's verification work in the OCR entry system is reduced.
The processing efficiency of the entry system is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a character recognition processing system according to the present invention.

FIG. 2 is an explanatory diagram showing an example of a result master file.

FIG. 3 is an explanatory diagram showing an example of a corporate dictionary source.

FIG. 4 is an explanatory diagram showing a processing procedure for creating a corporate dictionary source.

FIG. 5 is a block diagram showing a conventional character recognition processing system.

[Explanation of symbols]

 1 Host computer 2 Actual master file 3 Character recognition device 4 Knowledge processing device 5 Knowledge dictionary 6 Verification terminal 7 Business name file

Claims (1)

[Claims] 1. A dictionary database of proper nouns for a region such as an individual name, a business name, and a corporate name is edited and created from a result master file created by accumulating the result data processed in the past several months on a host computer. A character recognition processing system, comprising: a database creating means for performing the knowledge processing when performing character recognition of each form image in the character recognition apparatus based on the created dictionary database.