JP7059734B2 - Document reader and program - Google Patents

Description

The present invention relates to a document reader and a program.

Patent Document 1 describes a communication system that improves processing efficiency in a communication system that communicates between a terminal that recognizes characters and a central control device. The mobile terminal recognizes the image data including the document ID indicating the type of the document scanned by the scanner by the character recognition unit based on the control signal, and recognizes the document ID by the document ID recognition unit. The character recognition result and the manuscript ID recognition result are transmitted to the central control device by the character recognition result / manuscript ID transmission unit, and the control signal of the image signal is received from the central control device by the threshold value / position information reception unit. The central control device receives the character recognition result and the manuscript ID recognition result from the mobile terminal at the character recognition result / manuscript ID receiving unit, and transmits the image signal control signal to the mobile terminal at the threshold value / position information transmitting unit. Based on the document ID recognition result, the control signal of the image signal is controlled by the threshold value control unit.

Patent Document 2 describes a document management system that makes it possible to determine document attributes from a registration sheet and reduces the work burden on the user. A registration sheet containing document storage destination information according to the document attributes recorded on the document management server is printed, and a paper document containing the registration sheet is scanned to create an electronic image document. Next, after extracting the registration sheet from this image document, OCR (Optical Character Recognition) processing is performed, and the document attribute determination processing is performed using this result. Further, an organizing process for arranging the appearance of the image document for each page corresponding to this document attribute, and the image document are divided and recorded in a predetermined place of the document management server.

Japanese Unexamined Patent Publication No. 11-184966 Japanese Unexamined Patent Publication No. 2008-165386

The so-called zone OCR, which executes OCR processing only on a specific area of a document to extract character information when executing OCR processing on a document, is effective when processing a relatively large amount of documents. On the other hand, setting the reading position for all of two or more documents of the same or similar format to be read is complicated for the user.

An object of the present invention is to solve such a problem.

The invention according to claim 1 is the same as the first document, that is, the document reading unit, the storage unit that reads the first document in which the reading position is set by the document reading unit, and stores the reading position information. Alternatively, the reading position when the document reading unit reads the second document, which has a similar format and the reading position is not set, is specified by using the reading position information stored in the storage unit. The reading position information includes a position specifying portion , the reading position information includes information on a relative position with respect to the specific position in the first document, and the relative position information includes information on position coordinates and a range thereof with respect to the specific position. , A document reader.

The invention according to claim 2 is the document reading device according to claim 1, wherein the reading position information includes position coordinates including a certain number of characters in the first document and information in the range thereof.

The invention according to claim 3 is the document reading device according to claim 1 , wherein the specific position is a position of a specific attribute.

The invention according to claim 4 is the same as the first document, that is, the document reading unit, the storage unit that reads the first document in which the reading position is set by the document reading unit, and stores the reading position information. Alternatively, the reading position when the document reading unit reads the second document, which has a similar format and the reading position is not set, is specified by using the reading position information stored in the storage unit. A position specifying unit is provided, and when the second document constitutes a set consisting of a plurality of sheets, the storage unit reads and reads the first document in correspondence with the set. It is a document reading device that stores as position information.

The invention according to claim 5 is the document reading device according to claim 4 , wherein the plurality of sheets are a front surface and a back surface.

The invention according to claim 6 is the document reading device according to claim 4 , wherein the plurality of sheets constitute a set of forms.

The invention according to claim 7 is the document reading device according to claim 4, wherein the set including the plurality of sheets is the document reading device according to claim 4 , wherein the reading positions are different from each other.

The invention according to claim 8 is the document reading device according to any one of claims 1 and 4 , wherein the setting of the reading position in the first document is by marking.

The document according to claim 9 , wherein the marking is at least one of a specific color enclosure, a specific color fill, a check mark addition, a colorless transparent enclosure, and a colorless transparent fill. It is a reading device.

The invention according to claim 10 is the document reading device according to claim 8 , further comprising a display unit for displaying an image of the first document and an operation unit for marking the image of the first document. ..

The invention according to claim 11 is the same as the first document, that is, the document reading unit, the storage unit that reads the first document in which the reading position is set by the document reading unit, and stores the reading position information. Alternatively, the reading position when the document reading unit reads the second document, which has a similar format and the reading position is not set, is specified by using the reading position information stored in the storage unit. The first document is a document reading device that includes a position specifying unit and is one or more header sheets for the second document.

The invention according to claim 12 is the document reading device according to claim 11 , wherein the number of header sheets is set by the user.

The invention according to claim 13 is the document reading apparatus according to claim 11 , wherein the number of header sheets is determined by the document reading unit.

The invention according to claim 14 is a step of reading a first document in which a reading position is set in a computer, a step of generating reading position information about the reading position, and storing the reading position information in a storage unit. After storing the reading position information, the storage unit stores the reading position when reading the second document having the same or similar format as the first document and for which the reading position is not set. It is a program that executes a step to specify using the read position information , the read position information includes information on a relative position with respect to the specific position in the first document, and the information on the relative position is the above. A program that contains information about position coordinates and their range for a particular position.
The invention according to claim 15 has a step of reading a first document in which a reading position is set in a computer, a step of generating reading position information about the reading position, and storing the reading position information in a storage unit. After storing the reading position information, the storage unit stores the reading position when reading the second document having the same or similar format as the first document and for which the reading position is not set. It is a program for executing a step of specifying by using the read position information, and in the step of storing in the storage unit, when the second document constitutes a set consisting of a plurality of sheets, the first document is used. It is a program that reads in correspondence with the above-mentioned set and stores it as read position information.
The invention according to claim 16 is a step of reading a first document in which a reading position is set, a step of generating reading position information about the reading position, and storing the reading position information in a storage unit. After storing the reading position information, the storage unit stores the reading position when reading the second document having the same or similar format as the first document and for which the reading position is not set. It is a program that executes a specific step using the read position information, and the first document is a program that is one or more header sheets for the second document.

According to the inventions of claims 1 to 16, it is possible to save the trouble of setting the reading position for all of two or more documents of the same or similar format to be read.

According to the second aspect of the present invention, further, information on the position coordinates including a certain number of characters and the range thereof in the first document can be used.

According to the inventions of claims 1, 5 and 14, further, information on the position relative to the specific position in the first document can be used.

According to the inventions of claims 4 to 7 , 15, it is possible to further cope with the case where the second document constitutes a set consisting of a plurality of sheets.

According to the inventions of claims 8 to 10 , the reading position in the first document can be further set by marking.

According to the inventions of claims 11 to 13 , 16, the first document may further be composed of one or more header sheets.

It is a system block diagram of Embodiment 1. FIG. It is a functional block diagram of Embodiment 1. FIG. It is explanatory drawing of a zone OCR. It is explanatory drawing (the 1) which shows the process of Embodiment 1. FIG. It is explanatory drawing (the 2) which shows the process of Embodiment 1. FIG. It is explanatory drawing (the 3) which shows the process of Embodiment 1. FIG. It is explanatory drawing (the 4) which shows the process of Embodiment 1. FIG. It is a processing flowchart of Embodiment 1. It is another processing flowchart of Embodiment 1. It is explanatory drawing (the 1) which shows the process of Embodiment 2. It is explanatory drawing (the 2) which shows the process of Embodiment 2. It is a processing flowchart of Embodiment 2. It is explanatory drawing (the 1) which shows the process of Embodiment 3. It is explanatory drawing (the 2) which shows the process of Embodiment 3. It is a reading position explanatory drawing of the modification.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings, taking an image processing device as an example as a document reading device.

<Embodiment 1>
FIG. 1 is a block diagram of a system including an image processing device according to the present embodiment. The system includes a terminal device 10 and an image processing device 12. The terminal device 10 and the image processing device 12 are connected to each other via a communication line 14, and the communication line 14 is a data communication network such as a LAN (local area network). The communication line 14 may be either wired or wireless.

The terminal device 10 is a PC, a smartphone, or the like, and transmits a print job or the like including a document printing command according to a user's instruction.

The image processing device 12 is, for example, a multifunction device that realizes a plurality of functions, and is a ROM 121, a RAM 122, a storage device 123, a control unit 124 composed of one or a plurality of processors, an input / output interface (I / F) 125, and the like. It includes an operation unit 126 and an image forming unit 127.

The control unit 124 composed of one or a plurality of processors receives print job instructions and the like from the terminal device 10 via the input / output I / F 125 according to the processing program stored in the ROM 121 and the storage device 123, and interprets the PDL data. Then, intermediate data is generated, and drawing data (raster data) is further generated from the generated intermediate data. Further, the control unit 124 executes commands of various applications such as copy, scan, fax, and mail received from the operation unit 126.

The image forming unit 127 includes a scanner module 128 as a document reading unit, and reads a document and converts it into electronic data. In addition to the scanner module 128, the image forming unit 127 includes a print module, a fax module, a mail module, a paper feeding module, a document feeding module, and an image processing accelerator.

The print module is a module having a function of outputting an image on paper. For example, it has a known inkjet method configuration and prints drawing data on paper. Liquid or molten solid ink is ejected from a nozzle or the like, and recording is performed on paper, film, or the like. Ink ejection methods include a drop-on-demand method (pressure pulse method) that ejects ink using electrostatic attraction, and ink ejection using the pressure generated by forming and growing bubbles due to high heat. There is a thermal inkjet method and the like. The recording head includes, for example, a head that ejects cyan ink, a head that ejects magenta ink, a head that ejects yellow ink, and a head that ejects black ink, and each head has a width at least equal to the width of the paper. The head is used. Ink droplets of each color are ejected to an intermediate transfer body by a recording head for recording, and then transferred to paper for printing.

A fax module is a module that includes a modem and an image processing module for fax, and performs a fax function.

A mail module is a module that executes an e-mail function.

The paper paper feed module is a module for transporting paper from a paper tray to a print module.

The document paper feed module is a module for transporting paper from a document tray to a fax module.

The image processing accelerator is a module that performs compression / decompression processing in conjunction with a scanner module or the like. This image processing accelerator is not essential and may be an additional module.

The operation unit 126 includes a touch panel, and displays application icons corresponding to various functions (copy, scan, fax, mail, OCR, etc.) included in the image processing device 12 in response to a control command from the control unit 124. For example, a copy icon corresponding to a copy function, a fax icon corresponding to a fax, a scan to mail icon corresponding to a function of scanning a document and sending an e-mail, and the like. The user can set the setting information unique to each application by touch-operating the operation unit 126.

The storage device 123 is composed of an HDD (Hard Disk Drive), an SSD (Solid State Drive), or the like, and stores processing programs and document data. Further, in the present embodiment, in particular, the reading position information included in the specific document (first document) is stored.

In addition to these, the image processing device 12 includes a finisher for punching and sorting paper, a USB, an IC card reader, and an authentication unit for authenticating users, a billing unit, a motion sensor, a face camera, and the like. You may prepare. Further, the image processing device 12 may be connected to the Internet and an external server (network server) via the communication line 14, and may be provided with Ethernet (registered trademark) or WiFi. When connecting to an external server via the Internet, an authentication process is required, and the image processing device 12 may include an application for executing the authentication process.

Next, the document reading function (scanning function) of the image processing device 12 as the document reading device will be described.

FIG. 2 shows a functional block diagram of the control unit 124. The control unit 124 controls the operation of the scanner module 128 to capture one or more documents set in the feeder and convert them into electronic data. Then, when the user instructs the execution of the OCR via the operation unit 126, the OCR is executed for the electronic data and the result is output. A known method can be used for the OCR treatment. Generally, the OCR process is composed of a normalization process, a feature extraction process, a matching process, and a knowledge process. In the normalization process, one character to be OCR is converted to a certain size. In the feature extraction process, the normalized character is decomposed into four components in the vertical, horizontal, and diagonal directions, and the four components compressed to, for example, about 7 × 7 pixels are extracted as the features of each character. In the matching process, the extracted features are compared with the standard pattern stored in advance and collated. In the knowledge processing, the language to be recognized, such as Japanese word information and syntax information, is used, and the misread part is corrected by collating the candidate characters obtained in the matching process with the pre-registered word dictionary. ..

The control unit 124 includes a reading position detecting unit 124a, a reading position specifying unit 124b, and an OCR execution unit 124c as functional blocks.

The reading position detection unit 124a detects the reading position by detecting the reading position 21 included in the first document 20 from the electronic data obtained by reading the first document 20. The reading position 21 is set in advance by marking the first document 20, and is set, for example, by filling or enclosing with a specific color marker. When the reading position detection unit 124a detects the reading position of the reading position 21 in the first document 20, the read position detection unit 124a stores the detected reading position information (reading position information) in the storage device 123.

The reading position specifying unit 124b specifies a reading position when performing OCR on a second document 22 having the same or similar format as the first document 20, that is, an OCR target position. Specifically, the reading position specifying unit 124b reads the reading position information stored in the storage device 123 from the first document 20, identifies it as a reading position for the second document 22, and causes the OCR execution unit 124c to specify the reading position information. Output.

When the OCR execution unit 124c executes the OCR for the second document 22, the OCR execution unit 124c executes the OCR targeting the reading position specified by the reading position specifying unit 124b, that is, the zone OCR, and outputs the result. The second document 22 has the same or similar format as the first document 20, but the "same format" means that the same items as the items included in the first document 20 exist at the same position, and "similar". "Format" means that there is an item that is the same as the item included in the first document 20, but the position is different.

FIG. 3 schematically shows the conventional zone OCR processing. When the second document 22 is the target of the zone OCR processing, the user sets the reading position 22 with a marker or the like in advance for all the pages constituting the second document 22. For example, the area "No. 1234" at the top of one page is surrounded by a marker, the area "NO. 5678" at the top of another page is surrounded by a marker, and the area "NO. 5678" at the top of another page is surrounded by a marker. The area of "9990" is surrounded by a marker, and the like. Then, the control unit 124 reads the second document 22 to detect the reading position 23, executes OCR at the reading position 23, and outputs the results to "NO.1234", "No.5678", and "No.9999". "Etc. is output.

As described above, it is complicated to set the reading position for all the pages of the second document 22 with a marker or the like, and it is troublesome especially when the second document 22 is composed of a relatively large number of pages. It can be huge. Further, if the reading position is set with a marker or the like for all pages of the document, the document itself becomes dirty with the marker.

On the other hand, in the present embodiment, if the reading position is set by a marker or the like for the first document 20 having the same or similar format as the second document 22, the reading position is stored in the storage device 123. Since it is used as the reading position of the second document 22, the user only needs to set the reading position for the first document 20, and it is necessary to set the reading position for the second document 22 with a marker or the like. There is no. In addition, stains caused by markers can be suppressed.

The processing of this embodiment will be specifically described.

FIG. 4 specifically shows the first document 20 and the second document 22. Both the first document 20 and the second document 22 are forms (invoices) of the same format, and in addition to the heading of "invoice", each item of "company name", "amount", and "invoice date" is described. Has been done. The item names of the first document 20 and the second document 22 and their positions are the same. The first document 20 corresponds to the first page of the invoice form, and the second document 22 corresponds to the second and subsequent pages (from the second page to the M page) of the invoice form. The scanner module 128 of the image processing apparatus 12 targets the first document 20 as the first page and the second document 22 on the second and subsequent pages as reading targets.

The reading position 21 is set by surrounding or filling the area of each item of "company name", "amount", and "billing date" of the first document 20 as the first page with a marker of a specific color. .. The reading position 21 is not set in the second document 22 as the second and subsequent pages.

FIG. 5 shows a state in which the first document 20 as the first page is read, converted into electronic data, and the reading position 21 is detected from the electronic data. The reading position 21 can be detected by identifying the specific color when it is set by the marker of the specific color. The reading position 21 is an area containing a certain number of characters in the first document 20, and is a reference position of the reading position 21 when a certain position of the first document 20, for example, the upper left is the origin (0,0). It can be defined by two-dimensional coordinates (X, Y) (eg, upper left) and a width and height indicating the range (size) of the area. When the reading position 21 is set in three places in the first document 20, the reading position is detected for each place and the reading position information is generated. The reading position information is
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
And so on. The detected reading position information of the reading position 21 is stored in the storage device 123 in association with the first document 20.

FIG. 6 shows a state in which the second document 22 as the second and subsequent pages is read, converted into electronic data, and zone OCR is executed for the electronic data. Since the storage device 123 stores the reading position information of the reading position 21 in association with the first document 20, the control unit 124 reads these reading position information from the storage device 123 and specifies the reading position information. Zone OCR is performed at the read position to be performed. Therefore, even in the invoice on the second page, OCR is executed at the same position as the reading position 21 set in the invoice on the first page.
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
OCR is performed at the three locations specified in. As a result, as the OCR execution result on the second page, "XX Co., Ltd." is extracted as a character string from the "company name" item, and "¥ 32,400" is extracted as a character string from the "amount" item. "2017 / △△ / △" is extracted as a character string from the item of "billing date".

Similarly, in the invoice on the third page, OCR is executed at the same position as the reading position 21 set in the invoice on the first page.
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
OCR is performed at the three locations specified in. As a result, as the OCR execution result on the third page, "○○ Co., Ltd." is extracted as a character string from the "company name" item, and "¥ 10,800" is extracted as a character string from the "amount" item. , "2017 / XXX / x" is extracted as a character string from the item of "billing date".

When the first page functions as the first document 20, but also functions as the second document 22 as an invoice form, the reading position 21 is detected and the reading position information is stored in the storage device 123. At the same time, the zone OCR may be executed for the first page using the stored read position information, and the result may be output.

In FIG. 4, the first page of the invoice form is the first document 20, and the second and subsequent pages are the second document 22. However, a header sheet is added to the form, and the header sheet is the first document 20. The form itself from the page M to the page M may be the second document 22.

FIG. 7 schematically shows the processing in this case.

It shows how the first document 20 as a header sheet is read, converted into electronic data, and the reading position 21 is detected from the electronic data. The reading position 21 can be detected by identifying the specific color when it is set by the marker of the specific color. When the reading position 21 is set in three places in the first document 20, the reading position is detected for each place and the reading position information is generated.
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
And so on. The detected reading position information is stored in the storage device 123 in association with the first document 20.

After that, the second document 22 as the first and subsequent pages is read, converted into electronic data, and the zone OCR is executed for the electronic data. Since the storage device 123 stores the reading position information in association with the first document 20, the control unit 124 reads the reading position information from the storage device 123 and the reading position specified by the reading position information. Perform zone OCR in. Therefore, even in the invoice on the first page, OCR is executed at the same position as the reading position 21 set in the header sheet, and specifically,
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
OCR is performed at the three locations specified in. As a result, as the OCR execution result on the first page, "XX Co., Ltd." is extracted as a character string from the "company name" item, and "¥ 32,400" is extracted as a character string from the "amount" item. "2017 / △△ / △" is extracted as a character string from the item of "billing date". The same applies to the second to M pages.

FIG. 8 shows a processing flowchart of the present embodiment. This is a process when a header sheet is used as the first document 20.

First, the user operates the operation unit 126 to input the number of header sheets (S101). If the default number of sheets is 1, and the number of header sheets is 1, the input may be omitted.

Next, the control unit 124 reads the header sheet with the scanner module 128 and detects the reading position 21 as the OCR target area (S102). When the reading position 21 is set by a marker of a specific color, the specific color is detected. The detection of the OCR target area is repeatedly executed for the header sheet until the OCR target area is detected (S103), and all the OCR target areas in the header sheet are detected as the reading position 21.

After all the OCR target areas are detected, the OCR target area list is updated by storing the detected OCR target areas as reading position information in the storage device 123 (S104). For example, as described above, if there are three reading positions 21, they can be associated with the header sheet.
(10, 15, 100, 20)
(10,70,120,10)
(15,100,120,10)
Is created and stored in the storage device 123.

Next, it is determined whether or not the header sheet is the final header sheet for the number of sheets input in S101 (S105). If it is not the final header sheet (NO in S105), the processing after S102 is repeatedly executed for the next header sheet, an OCR target area list is created in association with the header sheet, and is stored in the storage device 123.

When the OCR target area list is stored in the storage device 123 for all the header sheets (YES in S105), the next page of the final header sheet is set as the OCR target page (S106).

Next, the control unit 124 reads and acquires one page of the OCR target area group from the OCR target area list stored in the storage device 123 (S107), and uses the OCR target area group to display the OCR target area on the OCR target page. On the other hand, the zone OCR process is executed (S108). When the zone OCR process is executed and the character string is extracted, the extracted character string is stored in the storage device 123 (S109). It may be stored in the storage device 123 and displayed on the display device, or may be output to the terminal device 10 or the external device (external server) via the communication line 14.

Next, it is determined whether or not it is the final area of the same page (S110), and if it is not the final area, the zone OCR process is repeated. As a result, if there are three OCR target areas, zone OCR processing is executed in all of these three locations, and the result is stored in the storage device 123. The above processing is repeatedly executed for all pages (S111).

In the process of FIG. 8, the user operates the operation unit 126 to input the number of header sheets, but the control unit 124 may detect the number of header sheets.

FIG. 9 shows a processing flowchart in this case.

First, the control unit 124 detects an OCR target area for the input document (S201). As shown in FIG. 8, there is no process in which the user operates the operation unit 126 to input the number of header sheets.

Next, it is determined whether or not the OCR target area is detected (S202), and if it is detected (YES in S202), the detected OCR target area is stored in the storage device 123 to list the OCR target area. Is updated (S203). For example, as described above, if there are three reading positions 21, they can be associated with the header sheet.
(10,15,100,20)
(10,70,120,10)
(15,100,120,10)
Is created and stored in the storage device 123.

Next, it is determined whether or not the page is the final page (S204). If it is not the last page (NO in S204), the processing after S201 is repeatedly executed for the next page, and the OCR target area list is created and stored for all the pages for which the OCR target area, that is, the reading position 21 is set. It is stored in the device 123.

When the processing is completed for all the pages in which the OCR target area, that is, the reading position 21 is set, the OCR target area is no longer detected (S202). Since the second document 22 starts from this page, the control unit 124 reads and acquires one page of the OCR target area group from the OCR target area list stored in the storage device 123 (S205). Zone OCR processing is executed for the OCR target page using the OCR target region group (S206). When the zone OCR process is executed and the character string is extracted, the extracted character string is stored in the storage device 123 (S207). It may be stored in the storage device 123 and displayed on the display device, or may be output to the terminal device 10 or an external device (external server) via the communication line 14.

Next, it is determined whether or not it is the final area of the same page (S208), and if it is not the final area, the zone OCR process is repeated. As a result, if there are three OCR target areas, zone OCR processing is executed in all of these three locations, and the result is stored in the storage device 123. The above processing is repeatedly executed for all pages (S209, S210).

<Embodiment 2>
In the first embodiment, the case where the first document 20 and the second document 22 have the same format has been described, but in the present embodiment, the case where the first document 20 and the second document 22 have similar formats, that is, the items are common but the same. The case where the positions are different will be described.

FIG. 10 shows the first document 20 and the second document 22 in this embodiment. The first document 20 and the second document 22 are different types of forms. The first document 20 includes "customer number" and "expiration date" as items. The second document 22 also includes the "customer number" and the "expiration date" as items, but their positions are different from those of the first document 20. The "customer number" and "expiration date" of the first document 20 are both located at the top, but in one of the second documents 22, the "customer number" is located at the top, but the "expiration date" is at the bottom. In the other one of the second document 22, the "customer number" is located in the central part and the "expiration date" is located in the lower part.

When the positions are different in this way, attention is paid to the attribute of the item and its value, and the attribute and its value are used as a set as the reading position information. For example, in the first document 20,
"Customer number 01234"
If there is an item with, the attribute is "customer number", the value is "01234", and the value is located on the right side of the attribute. Therefore, paying attention to the relative positional relationship between the attribute and its value,
Attribute: "Customer number"
Positional relationship with value attribute: right side Value width: 100
High value: 20
The reading position information is generated by such means. The width and height of the value specifies the size of the area where the value is described.

Here, the positional relationship with the attribute of the value may be defined as right side, left side, upper side, lower side, lower right side, etc., but relative position coordinates from the reference position of the attribute, for example, (10, 10), etc. May be specified in. The right side, the lower side, etc. can be said to be one expression of the position coordinates.

FIG. 11 specifically shows the reading position 21 set in the first document 20. The control unit 124 detects the reading position 21 set by the marker of a specific color or the like, and detects the attribute, the positional relationship with the attribute of the value, the width of the value, and the height of the value. That is, with respect to the reading position 21 set for the item of "customer number 01234".
Attribute: "Customer number"
Positional relationship with value attribute: right side Value width: 100
High value: 20
Is detected. Further, regarding the reading position 21 set for the item of "expiration date 2018/3/10", since the value of the expiration date is located below the expiration date, the attribute: "expiration date".
Positional relationship with value attribute: lower Value width: 80
High value: 30
Is detected. The control unit 124 stores the detected attribute and the set data of the value in the storage device 123 as read position information.

In the first embodiment, the reading position 21 is defined by the two-dimensional coordinates (X, Y) and the area size in the first document 20, but in the second embodiment, the reading position 21 is in the first document 20. Since it is defined by the relative positional relationship between the attribute and its value, in order to detect the corresponding attribute in the second document 22, first, OCR is executed for the entire page of the second document 22.

FIG. 12 shows a processing flowchart of the present embodiment.

First, the user operates the operation unit 126 to input the number of header sheets (S301). If the default number of sheets is 1, and the number of header sheets is 1, the input may be omitted.

Next, the control unit 124 reads the header sheet with the scanner module 128 and detects the reading position 21 as the OCR target area (S302). When the reading position 21 is set by a marker of a specific color, the specific color is detected. The detection of the OCR target area is repeatedly executed for the header sheet until the OCR target area is detected (S303), and all the OCR target areas in the header sheet are detected as the reading position 21.

After all the OCR target areas are detected, the OCR target area list is updated by storing the detected OCR target areas in the storage device 123 (S304). For example, as described above, if there are two reading positions 21, the header sheet may be associated with the reading position 21.
Attribute: "Customer number"
Positional relationship with value attribute: right side Value width: 100
High value: 20
Attribute: "Expiration date"
Positional relationship with value attribute: lower Value width: 80
High value: 30
Is created and stored in the storage device 123.

Next, it is determined whether or not the header sheet is the final header sheet for the number of sheets input in S101 (S305). If it is not the final header sheet (NO in S305), the processing after S302 is repeatedly executed for the next header sheet, an OCR target area list is created in association with the header sheet, and is stored in the storage device 123.

When the OCR target area list is stored in the storage device 123 for all the header sheets (YES in S305), the next page of the final header sheet is set as the OCR target page (S306).

Next, the control unit 124 reads and acquires a set of attributes and values for one page from the OCR target area list stored in the storage device 123 (S307). Then, OCR is executed for the entire OCR target page, an area matching the read attribute is detected, and a character string of the value of the detected attribute is extracted (S308). For example, the read attributes and values are
Attribute: "Customer number"
Positional relationship with value attribute: right side Value width: 100
High value: 20
If so, the control unit 124 executes OCR for the entire OCR target page and extracts the character string “customer number”. Then, using the extracted "customer number" as a reference, an attribute value such as "01234" is acquired by extracting a character string from the OCR processing result of the area having a width of 100 and a height of 20 located on the right side thereof. The attribute is extracted from the OCR processing result of the entire page, and the value of the attribute is acquired by the zone OCR for a specific area defined by the relative positional relationship with the attribute. When the OCR process is executed to extract the attribute and its value, the extracted attribute and its value are stored in the storage device 123 (S309). It may be stored in the storage device 123 and displayed on the display device, or may be output to the terminal device 10 or the external device (external server) via the communication line 14.

Next, it is determined whether or not it is the final area of the same page (S310), and if it is not the final area, the process is repeated. As a result, if there are two OCR target areas, the OCR process is executed in all of these two places, and the result is stored in the storage device 123. The above processing is repeatedly executed for all pages (S311).

In the present embodiment, if the relative positional relationship between the attribute and its value is the same, even if the position of the attribute itself changes for each document, it can be dealt with.

In the present embodiment, by setting an allowable range in the relative positional relationship between the attribute and its value as the data of the reading position 21, not only the position of the attribute itself changes for each document, but also the attribute and its value. It is also possible to deal with the case where the relative positional relationship of is changing. For example, as reading position information Attribute: "Customer number"
Positional relationship with value attribute: right or lower Value width: 100
High value: 20
And so on. In this case, in S308, the zone OCR processing on the right side and the lower side of the detected attribute is executed to extract the character string of the value.

<Embodiment 3>
In the first and second embodiments, the case where the second document 22 is a form in units of one page has been described, but the second document 22 may be a set of forms with a plurality of pages. Alternatively, the second document 22 may be a front and back double-sided document.

In such a case, the first document 20 may have a plurality of pages as one set corresponding to the page structure of the second document 22, or both the front and back sides may have one set.

FIG. 13 specifically shows the first document 20 and the second document 22 in this embodiment. The first document 20 and the second document 22 are both forms (contracts) of the same format, and the front surface and the back surface are configured as one set. In addition to the heading of "contract", each item of "company name" and "person in charge" is described on the front side, and the item of "contract date" is described on the back side. The item names of the first document 20 and the second document 22 and their positions are the same. The first document 20 corresponds to the front surface 20a and the back surface 20b of the first page of the contract form, and the second document 22 is the second and subsequent pages (from the second page to the M page) of the contract form. It corresponds to the front surface 22a and the back surface 22b. The scanner module 128 of the image processing apparatus 12 targets the front surface 20a and the back surface 20b of the first document 20 as the first page, and the front surface 22a and the back surface 22b of the second document 22 on the second and subsequent pages.

The area of each item of "company name", "person in charge" on the front side 20a of the first document 20 as the first page, and "contract date" on the back side 20b shall be surrounded or filled with a marker of a specific color. The reading position 21 is set with. The reading position 21 is not set in the second document 22 as the second and subsequent pages.

FIG. 14 shows a state in which the first document 20 as the first page is read, converted into electronic data, and the reading position 21 is detected from the electronic data. The reading position 21 can be detected by identifying the specific color when it is set by the marker of the specific color. The reading position information includes the two-dimensional coordinates (X, Y) on the upper left of the reading area when a fixed position of the first document 20, for example, the upper left is the origin (0,0), and the width and height indicating the size of the reading area. Can be specified by. When three reading positions 21 are set on the front surface 20a of the first document 20 and one reading position is set on the back surface 20b, the reading position information is generated for each location.
Surface 20a:
(X, Y, width, height)
= (10,15,100,20)
= (10,70,120,10)
= (15,100,120,10)
Back side 20b:
(X, Y, width, height)
= (15,100,140,10)
And so on. The detected reading position information is stored in the storage device 123 in association with the first document 20.

In this embodiment, the reading position information is the two-dimensional coordinates (X, Y) on the upper left of the reading area when the origin (0,0) is a fixed position of the first document 20 as in the first embodiment. Although it is specified by the width and height indicating the size of the reading area, it may be specified by the relative positional relationship between the attribute and its value and the width and height indicating the size of the value area as in the second embodiment. .. for example,
Surface 20a:
Attribute: "Customer number"
Positional relationship with value attribute: right side Value width: 100
High value: 20
Back side 20b:
Attribute: "Expiration date"
Positional relationship with value attribute: lower Value width: 80
High value: 30
And so on.

Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications are possible. Hereinafter, a modified example will be described.

<Modification 1>
In the embodiment, the reading position information is generated and stored in the storage device 123 by focusing on the set of the attribute and its value in the first document 20, but the attributes include "customer number" and "expiration date". In addition to the specific character string of, a specific color or a specific shape may be used.

<Modification 2>
In the embodiment, as markings for the first document 20, enclosing and filling with a marker of a specific color, adding a check mark, and the like are exemplified. May detect the colorless and transparent marking of the first document 20 by irradiating the first document 20 with a black light and generate reading position information. When surrounded by a marker of a specific color, it is desirable to use a color other than the ruled line color in order to distinguish it from the ruled line of the specific color constituting the table.

Further, as the marking for the first document 20, not only one type but also a plurality of types may be mixed. For example, a combination of a box and a fill with a marker of a specific color, a combination of a box and a check mark with a marker of a specific color, and the like.

FIG. 15 shows an example of marking in which a plurality of types are mixed. The reading position 21 of the first document 20 may be set by enclosing, filling, and checking marks with a specific color (eg, yellow or red) marker.

Further, the marking on the first document 20 may be performed by the user directly on the first document 20, but the electronic image obtained by converting the first document 20 into electronic data by the scanner module 128 is used as an operation unit. It may be displayed on 126 and the user may operate the operation unit 126.

<Modification 3>
In the first embodiment, the position coordinates, the width, and the height are defined as the reading position information as shown in FIG. 5, and in the second embodiment, the attributes and the relative positional relationship with the attributes are defined as the reading position information as shown in FIG. , Width and height are specified, but these may be combined. That is, one reading position 21 has the format shown in FIG. 5, another reading position 21 has the format shown in FIG. 11, and the like.

<Modification example 4>
In the embodiment, the reading position of the second document 22 is specified by using the reading position information generated by detecting the reading position 21 of the first document 20, but the reading position is exactly the same. In addition to the position to be used, the latest position within the allowable range may be included. Therefore, when the character string does not exist at the position specified by the reading position information in the second document 22 (for example, in the case of a check mark), the character string closest to the position specified by the reading position information is targeted for OCR. do it.

<Modification 5>
In the first embodiment, the zone OCR is executed for each page of the second document 22 and the OCR result is output. However, when the second document 22 has the same format, all the pages of the second document 22 are read. The reading positions specified by the position information may be collectively read, and then OCR may be executed for each page and the result may be output.

10 terminal device, 12 image processing device, 14 communication line, 20 first document, 22 second document, 123 storage device, 124 control unit, 128 scanner module.

Claims (16)

Document reader and
A storage unit that reads the first document for which a reading position is set by the document reading unit and stores it as reading position information, and a storage unit.
The reading position information stored in the storage unit is the reading position when the document reading unit reads the second document which has the same or similar format as the first document and the reading position is not set. The reading position specifying part specified by using
Equipped with
The reading position information includes information on a position relative to a specific position in the first document.
The relative position information includes position coordinates with respect to the specific position and information on the range thereof.
Document reader. The document reading device according to claim 1, wherein the reading position information includes information on position coordinates including a certain number of characters in the first document and a range thereof. The document reading device according to claim 1 , wherein the specific position is a position of a specific attribute. Document reader and
A storage unit that reads the first document for which a reading position is set by the document reading unit and stores it as reading position information, and a storage unit.
The reading position information stored in the storage unit is the reading position when the document reading unit reads the second document which has the same or similar format as the first document and the reading position is not set. The reading position specifying part specified by using
Equipped with
When the second document constitutes a set consisting of a plurality of sheets, the storage unit reads the first document in correspondence with the set and stores it as reading position information.
Document reader. The document reading device according to claim 4 , wherein the plurality of sheets are a front surface and a back surface. The document reading device according to claim 4 , wherein the plurality of sheets form a set of forms. The document reading device according to claim 4 , wherein the set including the plurality of sheets has different reading positions from each other. The document reading device according to any one of claims 1 and 4 by marking the setting of the reading position in the first document. The document reading device according to claim 8 , wherein the marking is at least one of enclosing with a specific color, filling with a specific color, adding a check mark, enclosing with colorless and transparent, and filling with colorless and transparent. A display unit that displays the image of the first document, and
An operation unit for marking the image of the first document,
The document reading apparatus according to claim 8 . Document reader and
A storage unit that reads the first document for which a reading position is set by the document reading unit and stores it as reading position information, and a storage unit.
The reading position information stored in the storage unit is the reading position when the document reading unit reads the second document which has the same or similar format as the first document and the reading position is not set. The reading position specifying part specified by using
Equipped with
The first document is one or more header sheets for the second document.
Document reader. The document reading device according to claim 11 , wherein the number of header sheets is set by the user. The document reading device according to claim 11 , wherein the number of header sheets is determined by the document reading unit. On the computer
The step of reading the first document for which the reading position is set, and
A step of generating reading position information about the reading position, and
The step of storing the read position information in the storage unit,
After storing the reading position information, the reading position when reading the second document which is the same as or similar to the first document and the reading position is not set is stored in the storage unit. It is a program that executes a step of specifying using position information , and the read position information includes information of a position relative to a specific position in the first document.
The relative position information is a program including information on the position coordinates and the range thereof with respect to the specific position. On the computer
The step of reading the first document for which the reading position is set, and
A step of generating reading position information about the reading position, and
The step of storing the read position information in the storage unit,
After storing the reading position information, the reading position when reading the second document which is the same as or similar to the first document and the reading position is not set is stored in the storage unit. Steps to identify using location information
In the step of storing in the storage unit, when the second document constitutes a set consisting of a plurality of sheets, the first document is read in correspondence with the set, and the reading position information is read. A program to remember as. On the computer
The step of reading the first document for which the reading position is set, and
A step of generating reading position information about the reading position, and
The step of storing the read position information in the storage unit,
After storing the reading position information, the reading position when reading the second document which is the same as or similar to the first document and the reading position is not set is stored in the storage unit. Steps to identify using location information
The first document is one or more header sheets for the second document.

Images