JP2012198684A - Information processing device, business form type estimation method, and business form type estimation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To estimate a business form type without preliminarily creating a format definition.SOLUTION: An information processing device 1 to be connected to a feature database 31 for storing feature data related to features of external appearances of business forms for each business form type includes: an image data acquiring section 21 for acquiring image data of a business form subject to processing; a feature data acquiring section 22 for acquiring the feature data related to the business form subject to the processing from the image data; an estimation section 24 for estimating feature data matching or being similar to the feature data related to the business form subject to the processing from the feature data stored in the feature database 31 and estimating a business form type of the business form subject to the processing; and a feature data adding section 25 for, when the estimation section 24 does not estimate the business form type of the business form subject to the processing, adding the feature data related to the business form subject to the processing to the feature database 31.

Description

  The present invention relates to an information processing apparatus, method, and program for estimating a form type.

  Conventionally, a data acquisition unit that acquires input image data, a memory that stores registered image data or a feature amount of a registered image, and a determination process that determines whether the input image data is image data corresponding to the registered image There is an image processing apparatus including a similarity determination processing unit to perform, and the similarity determination processing unit includes an image processing apparatus that changes processing contents of determination processing according to incidental information of input image data (Patent Document 1). See).

  The document type identification processing unit is an image processing apparatus that identifies a document type attribute (form, drawing, etc.) of a registered image, registers the attribute in the image information DB, and registers the registered image in the image DB. When searching for a document, a display screen classified according to the document type is created and displayed on the display device. The user selects a category such as “form” on the display screen using the input device, and the selected “form” is selected. There is an image processing apparatus that displays a list of thumbnails (see Patent Document 2). In addition, various techniques for processing an image obtained from a form or the like have been proposed (see Patent Documents 3 to 6).

JP 2008-301476 A JP 2008-204184 A International Publication No. 97/05561 JP 2000-242855 A JP 2009-238217 A JP 2007-58626 A

  Conventionally, there is a technique of generating image data of a form generated as a paper medium using a scanner or the like, identifying the type of the form, and performing OCR (Optical Character Recognition) processing. In such a technique, in order to identify a form type and perform OCR processing, it is necessary to create a format definition (form definition) for each form type in advance for all form types.

  However, it is difficult to create such a format definition. To create a format definition in advance, it takes time and effort for the user or provider of the form type identification service.

  In view of the above-described problems, an object of the present invention is to estimate a form type without creating a format definition in advance.

The present invention employs the following means in order to solve the above-described problems. That is, the present invention is an information processing apparatus connected to feature data storage means for storing feature data relating to appearance characteristics of a form for each form type, and image data for acquiring image data of a processing target form The feature data acquisition means for acquiring the feature data related to the processing target form from the image data, and the feature data stored in the feature data storage means match the feature data related to the processing target form Or, by specifying similar feature data, an estimation unit that estimates a form type of the processing target form, and a case where the form type of the processing target form is not estimated by the estimation unit, An information processing apparatus comprising: feature data adding means for adding feature data to the feature data storage means.

  Here, the form type is a form type that can be distinguished from the appearance of the form. The feature data accumulating unit accumulates the feature data related to the appearance characteristics of the form for each form type so that the form type can be estimated. The feature data stored by the feature data storage unit or acquired by the feature data acquisition unit includes, for example, ruled lines, the number of ruled lines, feature points, the number of feature points, color information (color condition), and the like. Image feature quantities may be included. The feature data may be any information that can be acquired based on the appearance of the form, and is not limited to the specific example described above.

  According to the present invention, when the form type of the processing target form is estimated by the estimation unit and the estimation of the form type fails, the feature data related to the processing target form is added to the feature data storage unit by the feature data addition unit. This makes it possible to estimate and learn the form type.

  Further, the feature data adding means, when the form type of the processing target form is not estimated by the estimating means, the feature data related to the processing target form as the feature data related to a new form type. You may add to a storage means.

  When the form type estimation fails, the feature data related to the processing target form is added to the feature data storage means as the feature data related to the new form type, so that the form data learned in the subsequent form type estimation Can be used.

  In addition, the information processing apparatus has a predetermined range in which the predetermined parameter included in the feature data is based on the predetermined parameter acquired from the image data from the feature data stored in the feature data storage unit. Extraction means for extracting one or a plurality of feature data included therein, and the estimation means extracts feature data that matches or is similar to the feature data related to the processing target form from the feature data extracted by the extraction means. By specifying, the form type of the processing target form may be estimated.

  That is, the first stage of narrowing is performed by the extracting means, and the second stage of narrowing is performed by the estimating means. At this time, by making the narrowing-down method at any stage lighter than the narrowing-down technique at other stages, it is possible to reduce the overall load related to form type estimation and to speed up the processing. . In particular, in the first-stage narrowing with a large parameter to be narrowed down, a relatively light-weight narrowing-down method such as index search is adopted, and in the second-stage narrowing down, the nearest neighbor search or the like is relatively accurate. By adopting a high narrowing-down method, it is possible to estimate the form type with high accuracy without impairing the processing speed. Note that the stage of the narrowing process is not limited to two stages. The narrowing-down process may be executed by being divided into three or more stages.

  The predetermined parameter includes information capable of specifying the number of feature points acquired from the image data, and the extraction unit converts the feature data stored in the feature data storage unit into the feature data. One or a plurality of feature data in which the number of feature points included is within a predetermined range based on the number of feature points acquired from the image data may be extracted.

The predetermined parameter includes information capable of specifying the number of ruled lines acquired from the image data, and the extraction unit includes the feature data stored in the feature data storage unit. One or more feature data may be extracted in which the number of ruled lines is within a predetermined range based on the number of ruled lines acquired from the image data.

  Here, the predetermined parameter only needs to include information capable of specifying the number of feature points or the number of ruled lines. That is, the predetermined parameter may be the number of feature points or the number of ruled lines themselves, or may be information for each acquired feature point or ruled line. When the predetermined parameter is information for each acquired feature point or ruled line, the number of feature points or the number of ruled lines can be specified by counting the number of information for each feature point or ruled line. I can do it.

  The predetermined parameter includes information capable of specifying a color condition acquired from the image data, and the extraction unit is included in the feature data from the feature data stored in the feature data storage unit. One or a plurality of feature data whose color condition is within a predetermined range based on the color condition acquired from the image data may be extracted.

  Here, the color condition is information indicating a tendency regarding the color of the image data acquired based on the image data. As the color condition, for example, an RGB color histogram can be used.

  The feature data includes a feature amount calculated from the image data, and the estimation unit obtains a feature amount acquired from the image data with respect to a feature amount included in the feature data extracted by the extraction unit. By performing nearest neighbor search using a quantity, the feature data that matches or resembles the feature data related to the processing target form is identified from the extracted feature data, and the form type of the processing target form is estimated. Also good.

  The present invention can also be understood as a method executed by a computer or a program executed by the computer. Further, the present invention may be a program in which such a program is recorded on a recording medium readable by a computer, other devices, machines, or the like. Here, a computer-readable recording medium is a recording medium that stores information such as data and programs by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer or the like. Say.

  According to the present invention, it is possible to estimate a form type without creating a format definition in advance.

It is a figure which shows the outline of a structure of the form classification search system which concerns on embodiment. It is a figure which shows the outline of a function structure of the application server which concerns on embodiment. It is a figure which shows the structure of the table A contained in the characteristic database which concerns on embodiment. It is a figure which shows the structure of the table B contained in the characteristic database which concerns on embodiment. It is a flowchart which shows the flow of the form classification identification and learning process which concerns on embodiment. It is a figure which shows the example of the image data of the process target form which concerns on embodiment. FIG. 6B is a diagram illustrating an example of a grayscale image with only ruled lines extracted from the processing target form illustrated in FIG. 6A. It is a figure which illustrates a mode that the feature point was extracted in the grayscale image only of the ruled line illustrated in FIG. 6B.

  Embodiments of an information processing apparatus according to the present invention will be described below with reference to the drawings. The embodiment described below shows an example for carrying out the present invention, and the present invention is not limited to the specific configuration described below. In practicing the present invention, it is preferable to adopt a specific configuration according to the embodiment as appropriate.

  In the present embodiment, the information processing apparatus according to the present invention is implemented as the application server 1 that provides a form type search service in the form type search system 100. However, the application target of the present invention is not limited to the application server 1 as long as it is an information processing apparatus for estimating the form type. For example, the information processing apparatus may be implemented as the user terminal 9 or a so-called cloud or the like.

<Configuration of form type search system>
FIG. 1 is a diagram showing an outline of the configuration of a form type search system 100 according to this embodiment. The form type search system 100 stores a Web server 2 for providing an interface with the user terminal 9, an application server 1 for providing a form type search service, and various types of information such as feature data and image data. Database 3 (feature database 31 and image database 32) is a system connected to each other.

  The Web server 2 is a server for providing the user terminal 9 with an interface for a form type search service. In the form type search system 100 according to the present embodiment, a user transmits image data of a form created using a scanner or the like to the Web server 2. However, when a form is used in business as image data from the beginning, the user does not need to scan a paper medium. The Web server 2 receives the image data transmitted from the user terminal 9 and transmits it to the application server 1.

  The application server 1 receives the image data transmitted from the user terminal 9 via the Web server 2 and executes a form type identification / learning process described later. By executing the form type identification / learning process by the application server 1, the form type of the form related to the image data transmitted from the user terminal 9 is estimated. Then, the application server 1 transmits the estimation result (the form type of the form related to the image data) to the user terminal 9 via the Web server 2. Specifically, when it is estimated that the form type of the form relating to the image data is “AAA company estimate”, the application server 1 notifies the Web server 2 of the estimation result, and the Web server 2 The user is notified of a message such as “A quote from AAA” via the Web interface.

  That is, according to the form type search system 100 according to the present embodiment, the user can know the form type of the form only by transmitting the form image data. Further, by using the form type search system 100 like this, for the user who uses the form type search service, “A AAA company estimate is sent to the purchasing department”, “BBB company delivery confirmation is accepted. Build a business flow such as “Transfer to Department” or “AAA's estimate. OCR is supported by CCC's OCR software, format definition is RX0001, 1 yen per sheet. It is possible to provide additional services such as “Available”.

  The application server 1 includes a CPU (Central Processing Unit) 11, a RAM (Random Access Memory) 12, a ROM (Read Only Memory) 13, an auxiliary storage device 19, and external communication via a network. The network interface 16 is an information processing apparatus electrically connected.

However, regarding a specific hardware configuration of the information processing apparatus, components can be appropriately omitted, replaced, or added according to the embodiment. For example, input / output devices such as a display, a mouse, and a keyboard may be connected to the application server 1 as necessary.

  The CPU 11 is a central processing unit and controls each component provided in the application server 1 such as the RAM 12, the auxiliary storage device 19, and the input / output device by processing instructions and data expanded in the RAM 12, the ROM 13, and the like. To do. The RAM 12 is a main storage device, and is controlled by the CPU 11 to write and read various commands and data. That is, the CPU 11, the RAM 12, and the ROM 13 constitute a control unit of the application server 1.

The auxiliary storage device 19 is a non-volatile storage device, and mainly information to be retained even when the power of the application server 1 is turned off, for example, an OS (Operating System) of the application server 1 loaded into the RAM 12 and a flowchart described later. In addition to the various programs for executing the processing shown in Fig. 5, various data used by the application server 1 are written and read. As the auxiliary storage device 19, for example, an EEPROM (Electrically Erasable Programmable) is used.
ROM), HDD (Hard Disk Drive), etc. can be used.

  The Web server 2, the database 3, and the user terminal 9 are provided with a CPU, RAM, ROM, auxiliary storage device, network interface, etc. (not shown) as a general computer, similar to the application server 1. . The user terminal 9 is, for example, a personal computer, and may be connected to a scanner for acquiring image data of a processing target form.

  FIG. 2 is a diagram illustrating an outline of a functional configuration of the application server 1 according to the present embodiment. In the information processing apparatus according to the present embodiment, the CPU 11 interprets and executes various programs developed in the RAM 12, whereby the image data acquisition unit 21, the feature data acquisition unit 22, the extraction unit 23, the estimation unit 24, and the feature data. It functions as the application server 1 including the adding unit 25. Each of these functional units executes a function corresponding to each unit included in the information processing apparatus according to the present invention. In this embodiment, an example in which all of these functions are executed by the general-purpose CPU 11 is described. However, some or all of these functions are realized by one or a plurality of dedicated processors. May be.

  The image data acquisition unit 21 acquires image data of a processing target form.

  The feature data acquisition unit 22 encodes the image data of the form by acquiring the feature data related to the processing target form from the image data. In the present embodiment, the feature data includes ruled lines, the number of ruled lines, feature points, the number of feature points, color information, and feature quantities (for example, feature quantity vectors) calculated from these pieces of information. Details of the feature data acquisition will be described later using a flowchart.

  The extraction unit 23 performs a first-stage narrowing by index search. More specifically, the extraction unit 23 determines the number of ruled lines, the number of feature points, and the color condition included in the feature data from the feature data stored in the feature database 31 of the ruled lines acquired from the image data. One or a plurality of feature data that are within a predetermined range based on the number, the number of feature points, and the color condition (that is, parameters such as feature points approximate) are extracted.

3 and 4 are diagrams showing the configuration of the feature database 31 according to the present embodiment. The feature database 31 is a database that accumulates feature data related to appearance features of a form for each form type (form type ID). In the present embodiment, the feature database 31 includes a ruled line for each form type. Based on the table A (see FIG. 3) that holds the number, the number of feature points, and the color condition (RGB color histogram) and the image data of the form, the feature amount calculated for the ruled line, the feature point, and the color condition, And a table B (see FIG. 4) held in association with the form type (form type ID) of the form.

  The estimation unit 24 performs second-stage narrowing by nearest neighbor search. More specifically, the estimation unit 24 is extracted by performing a nearest neighbor search using the feature amount acquired from the image data on the feature amount included in the feature data extracted by the extraction unit 23. The feature data matching or similar to the feature data related to the processing target form is identified from the obtained feature data, and the form type of the processing target form is estimated. However, a method other than the nearest neighbor search may be used for estimating the form type as long as it is a method that can identify feature data that is the same or similar.

  The feature data adding unit 25 adds the feature data related to the processing target form to the feature database 31 as the feature data related to the new form type when the estimation unit 24 does not estimate the form type of the processing target form.

<Process flow>
FIG. 5 is a flowchart showing the flow of the form type identification / learning process according to the present embodiment. The processing shown in this flowchart is executed when the image data of the form transmitted from the user terminal 9 is received via the Web server 2. However, the process shown in this flowchart may be batch-processed on the image data of the form accumulated in advance. It should be noted that the specific contents and order of the processing described using the flowcharts in the present embodiment are examples for implementing the present invention. Specific processing contents, order, and the like may be appropriately selected according to the embodiment.

  In step S101, image data is acquired. The image data acquisition unit 21 of the application server 1 acquires the image data of the processing target form transmitted from the user terminal 9 via the Web server 2. Thereafter, the process proceeds to step S102.

  In step S102, ruled lines are extracted. The feature data acquisition unit 22 of the application server 1 extracts ruled lines from the image data acquired in step S101. Since the ruled lines can be extracted by using an image analysis technique that is generally used in the past, detailed description thereof is omitted. Any method may be adopted as a specific method for ruled line extraction. In this embodiment, a part related to a ruled line and a part other than the ruled line (characters, symbols, photographs, etc.) are discriminated by image analysis from image data related to a form, and image data including only the ruled line is generated. The image data of only ruled lines generated here is preferably a grayscale image or a binary image in order to facilitate the specification of ruled lines in this step and the extraction of feature points in step S103.

  FIG. 6A is a diagram illustrating an example of image data of a processing target form according to the present embodiment. FIG. 6B is a diagram illustrating an example of a grayscale image with only ruled lines extracted from the processing target form illustrated in FIG. 6A. According to FIGS. 6A and 6B, it can be seen that in the present embodiment, from the image data related to the form, the part related to the ruled line and the part other than the ruled line are distinguished and image data including only the ruled line is generated.

The feature data acquisition unit 22 extracts ruled lines by specifying the ruled lines one by one from the generated image data of only the ruled lines. Furthermore, the feature data acquisition unit 22 counts the number of extracted ruled lines. Generally, there are ruled lines extending in the vertical direction and ruled lines extending in the horizontal direction. The number of the ruled lines is different, but in this embodiment, the total number of ruled lines is not distinguished from the vertical ruled lines and the horizontal ruled lines. Count. This is the step S in the form type identification / learning process according to the present embodiment.
As will be described later in 103, the intersection of the vertical ruled line and the horizontal ruled line is extracted as a feature point. Therefore, in the feature data extraction and form type estimation described later, the weight of the intersection of ruled lines is determined by the ruled line. This is so that the vertical and horizontal numbers and intersections are not considered twice. Thereafter, the process proceeds to step S103.

  In step S103, feature points are extracted. The feature data acquisition unit 22 of the application server 1 extracts feature points from the image data acquired in step S102. Since feature points can be extracted by using an image analysis technique that is generally used in the past, detailed description thereof will be omitted. In the present embodiment, a method for extracting feature points related to ruled lines and their intersections from image data of only ruled lines generated in step S102 is used.

  FIG. 6C is a diagram illustrating a state in which feature points are extracted from the grayscale image having only the ruled lines illustrated in FIG. 6B. In FIG. 6, the extracted feature points are indicated by circles. According to FIGS. 6A and 6C, in the present embodiment, characteristic points (for example, vertical ruled lines and horizontal ruled lines that can be referred to in order to determine whether the image data matches or not from the image data related to the form). It can be seen that the intersection point is extracted. However, the feature point may be a feature point that can be referred to in order to determine whether or not the image data matches, and is not limited to the intersection of ruled lines. As a specific method for extracting feature points, any method may be adopted. Furthermore, the feature data acquisition unit 22 counts the number of extracted feature points. Thereafter, the process proceeds to step S104.

  In step S104, the color condition is acquired. The feature data acquisition unit 22 of the application server 1 acquires the color condition from the image data acquired in step S102. The acquisition of the color condition can be performed by using an image analysis technique that is generally used in the related art, and thus detailed description thereof is omitted. Any method may be adopted as a specific method for obtaining the color condition. In the present embodiment, an RGB color histogram of the image data acquired in step S102 is generated and acquired as a color condition. In the form type identification / learning process according to the present embodiment, ruled lines and feature points are extracted using grayscale image data. Therefore, by obtaining color conditions and using them for narrowing down, which will be described later, It is possible to distinguish forms that have the same or similar feature points but different colors. Thereafter, the process proceeds to step S105.

  In step S105, a feature amount is calculated. The feature data acquisition unit 22 of the application server 1 calculates a feature amount related to the feature data (ruled lines, feature points, and color conditions in this embodiment) acquired in steps S102 to S104. The feature amount can be calculated by using a feature amount calculation technique that is generally used conventionally. The image feature amount includes an image moment, texture, edge amount, and the like. For example, the feature data acquisition unit 22 uses the feature points extracted in step S103 to rotate, translate, enlarge, or reduce the image. It is possible to calculate a feature quantity as an invariable quantity (invariant quantity) with respect to geometric deformation such as. Further, the feature data acquisition unit 22 can calculate a feature amount by performing higher-order local autocorrelation feature calculation on the binary image obtained in step S102. Thereafter, the process proceeds to step S106.

In step S106, a first-stage narrowing is performed. The first stage of narrowing down is performed by index search of the feature database 31 using the number of ruled lines, the number of feature points, and the color condition among the feature data obtained in steps S102 to S104. The extraction unit 23 of the application server 1 searches the feature database 31 using the number of ruled lines, the number of feature points, and the color condition obtained in steps S102 to S104 as search keys. However, in this search, a predetermined allowable range is set for each of the number of ruled lines, the number of feature points, and the color condition, so that a record that does not completely match but includes an approximate value, that is, a value from step S102. Records that are within a predetermined allowable range based on the value obtained in step S104 are also extracted. In this step, only records including values that match or approximate all of the number of ruled lines, the number of feature points, and the color condition are extracted. However, as the extraction condition, different weights may be set for the parameters of the number of ruled lines, the number of feature points, and the color condition. As a result of the first-stage narrowing in this step, one or more records (feature data) are extracted from the feature database 31. Thereafter, the process proceeds to step S107.

  In step S107, second-stage narrowing is performed. The second-stage narrowing is performed by nearest neighbor search using the feature amount calculated in step S105. The estimation unit 24 of the application server 1 performs nearest neighbor search on the record extracted in step S106, using the feature points calculated in step S105. According to the form type identification / learning process shown in this flowchart, the records are narrowed down by the index search with a relatively light processing load by the first-stage narrowing described in step S106. For this reason, the record (feature data) to be subjected to the nearest neighbor search in step S107 is smaller than the total size of the feature database 31, and is compared with the case where the nearest neighbor search is performed on the entire feature database 31. Thus, the processing load is reduced. By the nearest neighbor search, a record having the feature data closest to the feature data of the image data acquired in step S101 is specified. Then, the estimation unit 24 estimates that the form type related to the identified record is the form type of the form related to the image data acquired in step S101. Thereafter, the process proceeds to step S108.

  In step S108, the presence or absence of an estimation result is determined. In step S107, the control unit of the application server 1 determines whether one record (feature data) is identified as the feature data of the image data acquired in step S101. In other words, the form type estimation result is obtained. It is determined whether or not it has been. If it is determined that the estimation result of the form type has not been obtained, the process proceeds to step S109. On the other hand, when it is determined that the estimation result of the form type is obtained, the process proceeds to step S110.

  In step S109, registration in the feature database 31 is performed. The feature data adding unit 25 of the application server 1 stores the number of ruled lines, the number of feature points, the color condition, and the feature amount acquired in steps S102 to S105 in the feature database 31 as feature data relating to a new form type. Accumulate. However, the acquired feature data (feature points, ruled lines, etc.) may also be stored in the feature database 31. The feature data adding unit 25 stores the image data acquired in step S101 in the image database 32 in association with the stored feature data. That is, according to the form type identification / learning process according to the present embodiment, image data and feature data of a form for which an estimation result has not been obtained are learned as image data and feature data related to a new form type. The feature data learned here is referred to for identifying the form type when the form type identifying / learning process is executed from the next time onward. Thereafter, the processing shown in this flowchart ends.

In step S110 and step S111, the estimation result is presented to the user, and whether the estimation result is correct or not is determined. When it is determined in step S108 that the estimation result of the form type is obtained, the control unit of the application server 1 acquires the form type related to the estimation result from the feature database 31, and transmits it to the user terminal 9 via the Web server 2. (Step S110). The form type related to the estimation result received by the user terminal 9 is presented to the user by being displayed by the user terminal 9. The user refers to the estimation result displayed on the user terminal 9 and returns to the Web server 2 whether or not a correct result is obtained. According to the content of the reply from the user terminal, the application server 1 determines whether or not the estimation result is a correct result, that is, the form type related to the estimation result is the correct form type of the form related to the image data acquired in step S101. It is determined whether or not it is shown (step S111).

  If it is determined that the estimation result is not correct, the process proceeds to step S109. That is, when it is determined that the estimation result is not a correct result, the obtained information is stored in the feature database 31 and the image database 32 as information related to a new form type. Details of the processing in step S109 are as described above, and thus the description thereof is omitted. Thereafter, the processing shown in this flowchart ends. On the other hand, when it is determined that the estimation result is a correct result, the process proceeds to step S112.

  In step S112 and step S113, an OCR process is performed and the result is transmitted. The control unit of the application server 1 acquires a format definition related to the estimated form type from a format definition database (not shown) in which format definitions for OCR processing are accumulated. Then, the control unit of the application server 1 performs OCR processing on the image data acquired in step S101 according to the format definition (step S112). When the OCR process is completed, the control unit of the application server 1 transmits the result of the OCR process including the character information included in the image data of the form to the user terminal 9 via the Web server 2 (Step S1). S113). At this time, the control unit may also transmit the form type to the user terminal 9 together. Thereafter, the processing shown in this flowchart ends.

<Effect>
According to the form type search system 100 according to the present embodiment, it is possible to estimate the form type without creating a format definition in advance. Further, it is possible to learn about an unknown form type and use the learned form data in the subsequent form type estimation.

  Furthermore, according to the form type search system 100 according to the present embodiment, index retrieval with a relatively light load is adopted in the first stage narrowing by the extraction unit 23, and relatively accurate in the second stage narrowing by the estimation unit 24. By adopting a highly proximate nearest neighbor search, it is possible to perform highly accurate form type estimation without impairing the processing speed.

DESCRIPTION OF SYMBOLS 1 Application server 2 Web server 3 Database 9 User terminal 21 Image data acquisition part 22 Feature data acquisition part 23 Extraction part 24 Estimation part 25 Feature data addition part 31 Feature database 32 Image database

Claims (9)

An information processing apparatus connected to feature data storage means for storing feature data relating to appearance features of a form for each form type,
Image data acquisition means for acquiring image data of a processing target form;
Feature data acquisition means for acquiring feature data related to the processing target form from the image data;
Estimation means for estimating the form type of the processing target form by identifying feature data that matches or is similar to the characteristic data related to the processing target form from the feature data stored in the feature data storage means;
Feature data adding means for adding feature data related to the processing target form to the feature data accumulating means when the form type of the processing target form is not estimated by the estimating means;
An information processing apparatus comprising: If the form type of the processing target form is not estimated by the estimating means, the feature data adding means uses the feature data relating to the processing form as feature data relating to a new form type. To add to the
The information processing apparatus according to claim 1. From the feature data stored in the feature data storage means, one or more features in which a predetermined parameter included in the feature data is within a predetermined range based on the predetermined parameter acquired from the image data It further comprises extraction means for extracting data,
The estimation means estimates the form type of the processing target form by specifying feature data that matches or is similar to the characteristic data related to the processing target form from the feature data extracted by the extraction means.
The information processing apparatus according to claim 1 or 2. The predetermined parameter includes information capable of specifying the number of feature points acquired from the image data,
The extraction unit is configured such that, from the feature data stored in the feature data storage unit, the number of feature points included in the feature data is within a predetermined range based on the number of feature points acquired from the image data. Extract one or more feature data in
The information processing apparatus according to claim 3. The predetermined parameter includes information capable of specifying the number of ruled lines acquired from the image data,
The extraction means includes, based on the feature data stored in the feature data storage means, the number of ruled lines included in the feature data is within a predetermined range based on the number of ruled lines acquired from the image data. Extract one or more feature data;
The information processing apparatus according to claim 3 or 4. The predetermined parameter includes information that can specify the color condition acquired from the image data,
The extraction unit is configured such that, from the feature data stored in the feature data storage unit, the color condition included in the feature data is within a predetermined range based on the color condition acquired from the image data. Extract multiple feature data,
The information processing apparatus according to any one of claims 3 to 5. The feature data includes a feature amount calculated from the image data,
The estimation unit performs a nearest neighbor search using the feature amount acquired from the image data with respect to the feature amount included in the feature data extracted by the extraction unit, thereby obtaining the feature amount from the extracted feature data. , Identifying feature data that matches or resembles the feature data related to the processing target form, and estimates the form type of the processing target form.
The information processing apparatus according to any one of claims 1 to 6. A computer connected to the feature data storage means for storing the feature data relating to the appearance features of the form for each form type,
An image data acquisition step for acquiring image data of a processing target form;
A feature data acquisition step of acquiring feature data related to the processing target form from the image data;
An estimation step for estimating a form type of the processing target form by specifying feature data that matches or is similar to the characteristic data related to the processing target form from the feature data stored in the feature data storage unit;
A feature data adding step of adding feature data relating to the processing target form to the feature data storage means when the form type of the processing target form is not estimated in the estimating step;
Form type estimation method that executes. To the computer connected to the feature data storage means for storing the feature data relating to the appearance features of the form for each form type,
An image data acquisition step for acquiring image data of a processing target form;
A feature data acquisition step of acquiring feature data related to the processing target form from the image data;
An estimation step for estimating a form type of the processing target form by specifying feature data that matches or is similar to the characteristic data related to the processing target form from the feature data stored in the feature data storage unit;
A feature data adding step of adding feature data relating to the processing target form to the feature data storage means when the form type of the processing target form is not estimated in the estimating step;
Form type estimation program to execute.

Images