KR102039989B1 - Method and apparatus for extraciting text data from nonlinear text image - Google Patents

Abstract

According to an embodiment of the present invention, a method for providing a work service based on the structured data conversion of unstructured data, comprising: (a) a second user receiving a work request from a first user terminal and corresponding to a receiver of the work request; Transmitting a work request to a terminal; (b) receiving an atypical document image corresponding to a business request from a first user terminal, extracting texts from the received document image, and recognizing a category value and a verification value corresponding to each text; (c) checking whether a list of the category value and the verification value pair previously stored in the verification database of the server and the extracted category value and the verification value correspond to each other; And (d) detecting, according to the correspondence, one of the pre-stored stereotyped templates, and providing the first user terminal with a stereotyped document image in which texts are input into the stereotyped template. The category value is attribute information for a category classified by referring to the meaning of the recognized text, and the verification value is attribute information for a text representation format defined for each text corresponding to the category value.

Description

Methods and apparatus for extracting text data from unstructured text images {METHOD AND APPARATUS FOR EXTRACITING TEXT DATA FROM NONLINEAR TEXT IMAGE}

The present invention relates to an invention for extracting text from an unstructured text image having various formats and unifying it into a document in a standardized format.

In detail, the present invention relates to extracting text from various unstructured documents and applying the same to a standardized template to provide a user with the same document form.

With the recent increase in the number of companies, each company has its own unique document format. When collaborating among these companies, they often used their own document forms and provided them to others. Therefore, in order to classify or archive documents provided by other companies, it is cumbersome to manually classify documents received from various companies or newly fill out the form of the company.

Recently, with the development of digital storage media, existing paper documents are being digitized and stored. In the past, paper documents were often saved as image files simply by scanning or photographing technology, or written to their form using a document writing program. Recently, according to the development of optical character recognition technology that automatically recognizes a character, a technology of automatically recognizing a character in an image and data has been developed. However, the conventional technique simply recognizes a character and various documents There is a limit to classifying them.

Korean Registered Patent Publication No. 10-1834127 (Name of the Invention: Character Recognition Based Fax Service Processing System)

The present invention, in order to solve the above-mentioned problems, in accordance with an embodiment of the present invention, to extract data values from the images of the unstructured document produced in various forms, and to manage them integrated into one standardized document form do.

According to an embodiment of the present invention as a technical means for achieving the above technical problem, in a method of providing a business service based on the structured data conversion of the unstructured data, (a) receiving a business request from the first user terminal and Transmitting a work request to a second user terminal corresponding to a receiver of the work request; (b) receiving an atypical document image corresponding to a business request from a first user terminal, extracting texts from the received document image, and recognizing a category value and a verification value corresponding to each text; (c) checking whether a list of the category value and the verification value pair previously stored in the verification database of the server and the extracted category value and the verification value correspond to each other; And (d) detecting, according to the correspondence, one of the pre-stored stereotyped templates, and providing the first user terminal with a stereotyped document image in which texts are input into the stereotyped template. The category value may be attribute information on a category classified by referring to the meaning of the recognized text, and the verification value may be attribute information on a text expression format defined for each text corresponding to the category value.

In another embodiment, an apparatus for providing a recognition and storage service of unstructured data, the apparatus comprising: a memory in which a program for providing a recognition and storage service of the unstructured data is recorded; And a processor for executing a program, wherein the processor receives the work request from the first user terminal by executing the program, and transmits the work request to a second user terminal corresponding to the receiver of the work request. Receiving an atypical document image corresponding to a business request from a first user terminal, extracting texts from the received document image, recognizing a category value and a verification value corresponding to each text, and pre-stored categories in the verification database of the server Check whether the list of the pair of values and the verification value and the extracted category value and the verification value correspond to each other, and according to the correspondence, detect any one of a plurality of pre-stored stereotype templates, and store texts in the stereotype template. It may be a device for providing the input document image of the form to the first user terminal.

According to the above-described problem solving means of the present invention, it is possible to provide a platform for extracting the category value and the data mapped to it from the unstructured document image created in various formats, apply to the structured document form provided by the server, and deliver to the other party have.

1 is a diagram illustrating a configuration of a system for providing a business service based on a structured data transformation of unstructured data according to an embodiment of the present invention.
2 is a flowchart illustrating a service usage method between users according to an embodiment of the present invention.
3 is a diagram illustrating an example document for explaining a category value and a value value according to an embodiment of the present invention.
4 is a flowchart illustrating a method of recognizing a category value and a value value in an unstructured document image according to an embodiment of the present invention.
5 is an operation flowchart illustrating a process of converting an unstructured document image into a structured document image by a server according to an embodiment of the present invention.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and like reference numerals designate like parts throughout the specification.

Throughout the specification, when a part is "connected" to another part, this includes not only "directly connected" but also "electrically connected" with another element in between. . In addition, when a part is said to "include" a certain component, which means that it may further include other components, except to exclude other components, unless specifically stated otherwise, one or more other features It is to be understood that the present disclosure does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, parts, or combinations thereof.

In the entire specification, the server 100 may refer to a device that provides a business service based on the structured data conversion of the unstructured data.

The following examples are detailed description to aid in understanding the present invention, and do not limit the scope of the present invention. Therefore, the same range of inventions that perform the same functions as the present invention will also fall within the scope of the present invention.

1 is a diagram illustrating a configuration of a system for providing a business service based on a structured data transformation of unstructured data according to an embodiment of the present invention.

Referring to FIG. 1, a system for providing a business service based on a structured data transformation of unstructured data according to an embodiment of the present invention may include a server 100 and a user terminal 200, and the communication network 300. This will be interconnected.

The server 100 converts an atypical document image into a stereotyped document image through a terminal application installed in the user terminal 200, and receives a atypical document image due to a request of the user terminal 200. The terminal 200 may be provided with a standard document image extracted from an atypical document image based on the application interface provided from the server 100.

In this case, the server 100 recognizes the text data extracted from the atypical document image, compares the data with the data previously stored in the verification database, and applies the template to the structured template corresponding to the search result. .

The present invention may be to help smooth and easy work between companies or organizations. In the present invention, the user terminal 200 may include a first user terminal 201 and a second user terminal 202 used by different companies or organizations.

For example, in order to exchange documents between two companies, the first user terminal 201 performs a business request to the server 100, and the server 100 views the document image based on the document image received from the second user terminal 202. It is to proceed the invention.

In addition, the application for providing a business service based on the formal data conversion of the unstructured data may be an application embedded in the user terminal 200 or an application downloaded from an application distribution server and installed in the user terminal 200.

The plurality of user terminals 201 and 202 refer to a communication terminal capable of using a terminal application in a wired / wireless communication environment. In this case, the user terminal 200 may be a device capable of receiving the unstructured text image and providing it to the server 100 and receiving the result information thereof. In FIG. 1, although the user terminal 200 is illustrated as a smart phone, which is a kind of portable terminal, the spirit of the present invention is not limited thereto, and as described above, the terminal capable of mounting a terminal application is not limited thereto. Can be borrowed.

In more detail, user terminal 200 may include any form of a handheld computing device (eg, PDA, email client, etc.), personal computer, or other type of computing or communication platform. However, the present invention is not limited thereto.

Meanwhile, the communication network 300 connects the server 100 and the user terminal 200. That is, the communication network 300 refers to a communication network that provides a connection path for transmitting and receiving data after the user terminals 200 access the server 100. The communication network 300 may be, for example, wired networks such as local area networks (LANs), wide area networks (WANs), metropolitan area networks (MANs), integrated service digital networks (ISDNs), wireless LANs, CDMA, Bluetooth, satellite communications, and the like. Although it may encompass a wireless network, the scope of the present invention is not limited thereto.

2 is a flowchart illustrating a service usage method between users according to an embodiment of the present invention.

2, in order to use the service of the present invention, the server 100 receives a work request from the first user terminal 201 and requests the work to the second user terminal 202 corresponding to the receiver of the work request. Forward (S210).

Here, the work request may mean a work request for exchanging documents required for collaboration between the user of the first user terminal 201 and the user of the second user terminal 202.

For example, when the company to which the first user terminal 201 belongs to request a quotation for a transaction with the company to which the second user terminal 202 belongs, the server 100 firstly associates with the first user terminal 201. Receives a request for a quote of the company, and forwards the request to the company to which the second user terminal 202 belongs.

At this time, the form of the documents exchanged with each other may be composed of a form used by each company or organization alone.

Thereafter, the server 100 receives unstructured data corresponding to the work request from the second user terminal 202 based on the work request, converts it, and delivers the unstructured data to the first user terminal (S220).

In an exemplary embodiment, a plurality of unstructured document images received from a large number of users (or corporations) by the first user terminal 201 are configured by one or more second user terminals 202, which are targets of receiving business requests, from one or more users. Can be converted to a standard document image and received, and vice versa.

In this case, a method of converting the unstructured data into the structured data will be described with reference to FIGS. 3 to 5 to be described later.

3 is a diagram illustrating an exemplary document for describing a category value 410 and a value value 420 according to an embodiment of the present invention.

Referring to FIG. 3, text may be read from a document image (which may mean an atypical document image) scanned or photographed through a service of the present invention, and a category value 410 and a value value for each read text. 420 may be defined in association.

3 illustrates an example of a virtual quotation, and the server 100 may extract and recognize a category value 410 and a value 420 from a document.

First, the category value 410 may refer to text having a meaning for an upper category of a specific word. For example, in the virtual quotation of FIG. 3, the server 100 may designate a category value 410 as a “company representative”, “business status”, “transportation method”, or “transportation condition”.

In addition, the value 420 may refer to text corresponding to one of several sub-concepts belonging to a specific category. For example, the value 420 may be a sub-concept of "P & Psoft" or "service", which is a sub-concept belonging to a specific category of "company manager".

In this case, a detailed method of extracting and recognizing the category value 410 and the value 420 from the atypical document image by the server 100 will be described with reference to FIG. 4.

4 is a flowchart illustrating a method of recognizing a category value 410 and a value value 420 in an unstructured document image according to an embodiment of the present invention.

Referring to FIG. 4, first, an atypical document image received from the user terminal 200 is divided into a text area and another area (S410).

This process can be applied to algorithms for recognizing text in commonly used images, or other area-specific algorithms developed in-house.

The text is extracted from the separated text area (S420), and the extracted text is classified into a category value 410 and a value value 420 based on the text arrangement information of the atypical document image (S430). For example, the text disposed below the category value 410 may be regarded as the value 420. Alternatively, when the sort order of the category values 410 is in the lateral direction, the texts arranged above or below the category value 410 but arranged in the lateral direction may be regarded as the value 420.

Herein, the text arrangement information indicates a method in which the category value 410 and the paired value value 420 are arranged in the document in the atypical document image. In general, if the document has a category value 410 and a value 420 at the same time, the documents are arranged adjacent to each other so that the user can easily recognize them.

For example, in the imaginary quotation of FIG. 2, a plurality of category values 410 such as "sales person" and "business status" are horizontally arranged, and "P & P soft" and "service" paired with the category values 410. Value value 420, such as "" can be found to be located directly below the category value 410.

The server 100 distinguishes the category value 410 and the value value 420 from the extracted text arrangement information or the extracted text through an algorithm for distinguishing the category value 410 and the value value 420.

5 is an operation flowchart illustrating a process of converting an atypical document image into a structured document image by the server 100 according to an embodiment of the present invention.

Referring to FIG. 5, a category value 410 and a value value 420 are recognized from an atypical document image received from the second user terminal 202 according to an embodiment of the present invention (S510).

In this case, the method of recognizing the category value 410 and the value 420 from the atypical document image is the same as the description of FIG. 4 described above, and thus further description thereof will be omitted.

It is determined whether the category value 410 and the value 420 recognized in the step S510 match the list of the category value 410 and the value range pair previously stored in the verification database of the server 100 ( S520).

In detail, the verification database of the server 100 stores various types of formal templates (for example, quotation template, tax invoice template, proposal template, etc.). In addition, a plurality of category values 410 and value ranges 420 mapped thereto are set for each standard template. The category value 410 and the value value 420 mapped thereto may be listed in the verification database and stored in advance.

Thereafter, the recognized category value 410 and the value 420 may be compared with a list of values stored in the verification database. At this time, the matching ratio for each shaping template is calculated and a shaping template having the maximum matching ratio is searched.

On the other hand, in a further embodiment, if the recognized category value 410 and the value 420 exceeds a predetermined ratio, proceed to step S530, and through the step (S530) to search for a template having the maximum matching ratio You may.

In operation S530, the server 100 may recognize a value (category value 410 or value value) recognized in the unstructured document image in the range of the category value 410 and the value value 420 of the plurality of pre-stored template. 420) and then search for a template having a maximum matching ratio (S530).

At this time, if the matching ratio does not exceed the preset ratio, the server 100 recommends a formal template having the maximum matching ratio to the second user terminal 202 (S531) and proceeds to the next step (S540) based on the user's selection. Done.

In detail, the server 100 stores various types of template templates in advance, and each template has a range of a category value 410 and a value value 420 configured according to a purpose. For example, if a template called "quote" can have a range of category values 410, such as "seller", "sold article", "number of articles", "price of article", and so on, a template called "contract" It will have a range of "business name", "contract name", "contract date" and so on. In addition, the value value 420, like the category value 410, will vary in the range configured for each template.

Accordingly, the category value 410 and the value 420 recognized in the unstructured document image are matched with the ranges of the category value 410 and the value value 420 of each template. At this time, the server 100 searches for a template having a maximum matching ratio while the matching ratio exceeds a preset ratio.

The reason for having the preset value in the matching ratio is that the document type of the atypical document image and the document type of the atypical template may be different if the matching ratio is less than or equal to the appropriate line.

If the matching ratio does not exceed the preset value, the server 100 does not exceed the preset value, but recommends the template having the maximum matching ratio to the second user terminal 202 to check whether the use of the template is appropriate. To do the job.

In an alternative embodiment, if the server 100 does not recommend a suitable template, the second user may request, select, and apply the template.

Finally, the recognized category value 410 and the value 420 are applied to the template having the maximum matching ratio.

In this case, the recognized value value 420 is defined in association with the range of the value value 420 registered in the verification database (that is, the category value 410 corresponding to the recognized value value 420). If not satisfied, the value 420 may be modified to satisfy the range of the value 420 of the verification database and may be recommended as the second user terminal 202.

In detail, the range of the value value 420 indicates attribute information about a range and a format that the value 420 may have in the category value 410. For example, if the category value 410 is "up", the value value 420 range may be defined as "Korean", and if the category value 410 is "amount", the value value 420 range is " Number "," least one digit up to fifteen digits ". At this time, in order for the recognized value value 420 to match the range of the value value 420 on the verification database list, one of the syllables constituting the recognized value value 420 is modified or omitted, or a new syllable is added to the syllable. It is recommended to the user.

In addition, if the category value 410 to be mapped has a value such as an amount of money, a date, and a phone number, the minimum and maximum digits of the number consisting of the value 420 may be modified.

The modified value is recommended to the user of the second user terminal 202 that provided the atypical document image, so that the second user can confirm whether the modified value has the same meaning as the original value. As a further embodiment, as the second user desires, if the meaning is different, the second user may repeatedly modify and provide another recommendation word.

Also, as an optional embodiment, even if the value 420 read out from the document image is recommended while the modified value 420 is recommended, the image reading may be incorrect or the value 420 may not be in the range of the value 420. Since the range may be wrong, the value 420 read from the document image may be recommended to the user as it is.

Through this process, when the template search is completed and the recognized value values 420 and the category value 410 are input to the template, the second user terminal 202 confirms the input, and the category value is completed. After storing 410 and the value 420, a structured document image corresponding to the structured template is provided to the first user terminal 201 (S540).

In operation S530, the stereotyped document image is finally confirmed by the second user terminal 202 to determine whether the document is correctly deformed, and is provided to the first user terminal 201.

In this embodiment, the stereotyped document image may be provided in various extensions according to a user's selection to increase user convenience. For example, it is provided as an image extension of pdf, MS word extension of docx, Hancom Office Hangul extension of hwp, etc., so that the first user terminal 201 receiving a standard document image can easily open and modify a file.

The foregoing description of the present invention is intended for illustration, and it will be understood by those skilled in the art that the present invention may be easily modified in other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are exemplary in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is shown by the following claims rather than the above description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present invention. do.

100: server 200: user terminal
300: network
410: category value 420: verification value

Claims (13)

In the method for providing a business service based on the structured data conversion of the unstructured data performed by the server,
(a) receiving a work request from a first user terminal and forwarding the work request to a second user terminal corresponding to a recipient of the work request;
(b) receiving an atypical document image corresponding to the business request from a second user terminal, extracting texts from the received document image, and classifying the texts into category and value value pairs;
(c) checking whether the extracted value satisfies the value range by referring to a list of category values and value range pairs previously stored in the verification database of the server; And
(d) detecting, according to the confirmation, any one of a plurality of pre-stored stereotyped templates, and providing the first user terminal with a stereotyped document image inputting the texts into the stereotyped template;
Including,
The category value is text having a meaning for an upper category of a specific word.
The value is text corresponding to one of several sub-concepts belonging to a specific category,
The value range is information on the numerical range or format of the value,
Step (d)
(d1) detecting a structured template of any one of a plurality of pre-stored structured templates when a matching ratio between the list of the verification database and the extracted category value and value range exceeds a preset ratio; And
(d2) inputting the texts into the formal template;
Step (d1) is
For each of the above template, a category value and a value range are defined.
When a category value and a value value for the extracted texts are compared with a category value and a value value range for each standard template registered in the verification database,
If a matching template is extracted from a template that exceeds a preset ratio, but a template that exceeds the preset ratio is not found, recommend a template that is the maximum matching ratio to the second user terminal. And receiving the confirmation result of a second user terminal and storing the category value and the value value based on a search result. The method of claim 1,
In step (b),
(b1) separating a text area and other areas from the document image;
(b2) extracting text from the text area; And
(b3) separating the category value and the value value by referring to the arrangement information of the extracted texts;
The method comprising providing a business service based on the structured data conversion of unstructured data. delete delete delete The method of claim 1,
Step (d2),
(d2-1) When a specific value does not match a value range registered in the verification database, a value value range corresponding to a category value of the specific value value among the value values stored in the verification database and within a preset range Recommending a value having a difference or recommending the specific value itself to the second user terminal; And
(d2-2) receiving a selection of the second user terminal for the recommended information and inputting it into the formal template;
A method of providing a business service based on the structured data conversion of unstructured data, including. The method of claim 6,
Step (d2-1),
If a particular value does not match the value range registered in the verification database, modify or omit at least one of the syllables constituting the particular value so as to match the value range of the verification database, or add a new syllable. Recommend to the second user terminal to provide a business service based on the structured data conversion of the unstructured data. The method of claim 7, wherein
Step (d2-1),
When the category value corresponds to any one of an amount, a date, and a phone number, the value range includes any one of a minimum and maximum number of numbers constituting a recognized value value and an arrangement form of the number. How to provide business services based on structured data transformation of unstructured data. The method of claim 1,
In step (d),
(d3) providing the structured template in which text input is completed to the second user terminal, and when the confirmation input of the second user terminal is completed, transmitting the structured template to the first user terminal. A method of providing business services based on structured data transformation of unstructured data. The method of claim 9,
Step (d3),
And transmitting the document image received at the second user terminal together with the structured template. The method of claim 1,
The first user terminal and the second user terminal is a terminal owned by users belonging to different companies,
The document image provided by the second user terminal is an image about unstructured data configured in a form that is used solely by the company of the first user for the work request. How to give it. An apparatus for providing a recognition and storage service of unstructured data,
A memory in which a program for providing a recognition and storage service of unstructured data is recorded; And
A processor for executing the program;
The processor, by executing the program,
Receiving a work request from a first user terminal, forwarding the work request to a second user terminal corresponding to a recipient of the work request, receiving an atypical document image corresponding to the work request from a first user terminal, Extracts text from the received document image, recognizes a category value and a value value range corresponding to each text, and lists the category value and value range pair previously stored in the verification database of the device and the extracted category. Determine whether the value and the value value range correspond to each other, and according to the correspondence, detect one of the pre-stored stereotyped templates and input the texts into the stereotyped template. 1 provided to the user terminal,
The category value is text having a meaning for an upper category of a specific word, the value value is text corresponding to one of several sub-concepts belonging to a specific category, and the value range is a numerical range of the value value or Information about the format,
In the step of providing the document image to the first user terminal
When a matching ratio between the list of the verification database and the extracted category value and the value range exceeds a preset ratio, the stereotype template of any one of a plurality of pre-stored stereotyped templates is detected and the texts are stored in the stereotyped template. In,
In the step of inputting the text into the formal template
For each of the above template, a category value and a value range are defined.
When a category value and a value value for the extracted texts are compared with a category value and a value value range for each standard template registered in the verification database,
If a matching template is extracted from a template that exceeds a preset ratio, but a template that exceeds the preset ratio is not found, recommend a template that is the maximum matching ratio to the second user terminal. And receiving a confirmation result of a second user terminal and storing the category value and the value value based on a search result. 12. A computer readable recording medium having recorded thereon a program for providing a method for providing a business service based on a structured data conversion of unstructured data according to any one of claims 1, 2 and 6.

Images