JP3205535B2 - OCR system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an OCR system in which an OCR input paper can be easily designed and prepared, and the operation management such as data input and recognition processing from the prepared input paper is easy.

[0002]

2. Description of the Related Art In recent years, techniques for inputting and processing handwritten characters by OCR have been developed, and it has become possible to recognize handwritten characters to a considerable extent. However, characters can be recognized only when they are written in a predetermined frame. In addition, in order to secure a sufficient recognition rate, the size and type of characters need to be limited. . For this reason, OC
When operating R system, how rational OCR
Designing the input paper is an important factor.

Conventionally, however, the design of this OCR input sheet has been almost entirely dependent on manual work. First, the user and the system provider repeatedly make a meeting, and as a first step, a tentative input item is determined. Where each input item was laid out in the input form was determined. Then, as a second stage, a composition of the input sheet having the layout determined in this way was created and brought to the user.

[0004] Here, the user can grasp the actual condition of the OCR input sheet for the first time. For this reason, as a third step, the user determines whether or not it can actually be used in this edition, and if it is determined that further changes are necessary, the user returns to the first step again. become.

On the other hand, if this layout is judged to be sufficient, it is decided to formally use the composition. Then, as a fourth step, this is printed, a plurality of input sheets are created and delivered to the user, and the actual OCR system is started.

[0006]

However, in the case of a large-scale processing in which one OCR input sheet is used in a very large amount, the above-mentioned steps may be taken. Without this, a great deal of labor and expense would be required for the production of the composition, and practically no convenient OCR system could be used. For this reason, if it is absolutely necessary to use the paper, a general-purpose OCR input paper that has been standardized in advance must be used.

[0007] Furthermore, even if the user intends to change some of them as a result of the use, the above-mentioned many steps have to be taken again.

[0008] From the above points, it has actually been the case that even users who have actually introduced OCR endure the problem and use it.

[0009]

SUMMARY OF THE INVENTION The present invention has been made for the purpose of solving the above-mentioned problems, and has the following arrangement as one means for solving the above-mentioned problems.

That is, the OCR system of the present invention
An OCR system capable of designing a CR input form,
An operation unit for inputting instruction information of a CR input sheet, a display unit for displaying instruction input guidance and an instruction input result on the operation unit, a printer for displaying a layout result of the OCR input sheet, and an OCR input from the OCR input sheet A scanner for reading information and management information from a predetermined control sheet;
It is possible to display the instruction input guidance and the instruction input result on the display unit based on the recognition support data file in which the character characteristics of the writer who will fill in the CR input form are registered and the instruction information from the operation unit. Control unit. Then, the control unit causes the display unit to collectively display a menu screen including items such as specified value management, OCR input paper design, composition preparation, and recognition support.

When the specified value management item is selected from the menu screen, the specified value is specified based on the paper size of the OCR input paper input from the operation unit and its length and width and stored in a memory for management.

When an item of OCR input paper design is selected from the menu screen, automatic layout of the input data items of the OCR input paper is performed based on the specified values stored in the memory. Modify the required part of the specified value.

[0013] When the item of the composition is selected from the menu screen, the printer is instructed based on the result of the automatic layout.
To print the composition of the OCR input paper
Of the OCR input paper output by the printer
When an error between the arrangement of the composition and the arrangement of the prescribed value is input , the composition of the composition of the OCR input sheet is corrected based on the error.

When a recognition support item is selected from the menu screen, control paper and OCR input to the scanner are performed.
An instruction is given to read sheets in bundles, and the control
The management information is recognized from the paper, and the OCR entry information is recognized from the OCR input form based on the character characteristics registered in the recognition support data file.

With these configurations, the present OCR system is configured so that the entire process from the design of the OCR input paper to the operation management can be performed without preparing a program for each item processing.

Further, in the OCR system of the present invention, the menu screen is configured so that data input items can be further displayed. Then, when the data input item is selected from the menu, the control unit controls the control sheet based on the control sheet.
The character characteristics stored in the recognition support data file are selected according to the logical information, and the OCR entry information is recognized for each bundle from the OCR input sheet.

Further, in the OCR system of the present invention, the menu screen is configured so that data maintenance items can be further displayed. Then, when the data maintenance is designated from the display item, the control unit is configured to be able to output a checklist in which recognition results based on the OCR entry information and recognition candidates are arranged in parallel to the printer.

[0018]

As a result, even a general user can manage prescribed values of an OCR input sheet, design an OCR input sheet, create a composition,
Centralized management is possible without preparing a program for each item processing, from the design of the composition such as recognition support, data input and data maintenance, to the operation management by batch management of the actual OCR input paper.
It is possible to provide an easy-to-use OCR system that can easily respond to changes in input paper.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment according to the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a system configuration diagram of an embodiment according to the present invention. The OCR system of this embodiment designs and prints an OCR form as an OCR input sheet,
It is configured to be able to input and process the OCR form in which the handwritten characters are input.

In FIG. 1, reference numeral 1 denotes a CPU for controlling the entire embodiment according to a program described later stored in a memory 2, for example, a memory 2 for recording a program related to an OCR process described later, and 3 a process data and the like. Hard disk drive, which is an external storage device capable of recording a large amount of
Scanner that reads OCR characters etc. written on R form, 5
Is a printer that prints out OCR forms and prints out processing results, etc., 6 is a communication control unit that controls an interface with the network 10, 7 is an operation unit for inputting various instructions, and 8 is processing results and input results. Is a flexible disk (FD) device for displaying, for example, the created form data and the like on an FD medium and reading the stored form data into the system. . A line network 10 is connected to a LAN or another communication device that performs facsimile communication or the like.

Outline OCR of this embodiment having the above configuration
The processing operation will be described below. However, the present invention is not a system that performs only the OCR processing described below, but is a general office computer or personal computer.
Needless to say, the program may be executed by mounting a CR processing program.

FIG. 2 is a flowchart showing the general OCR processing operation of this embodiment. Hereinafter, the outline OCR processing operation of this embodiment will be described with reference to FIG.

First, in steps S1 to S3, various items are selected and necessary numerical values are input according to the guidance display displayed on the display screen of the display unit 8, and a process of designing an OCR form as input paper is performed. .

That is, first, in step S1, referring to the guide display on the display unit 8, the input characters and the size of the input OCR paper are designated and input from the operation unit 7. Specifically, first, the cursor is moved by an instruction input from the operation unit 7 and the display unit 8
Select the paper size (of the form) and the length and width of the input paper to be designed displayed on the display screen of (1). Subsequently, based on the selection, a specified value provided by the system (a control file that determines a positional relationship in which data in a form can be arranged) is confirmed.

If the specified value is sufficient, the display specified value is selected. Here, if there is a prescribed value to be modified, only a necessary portion is modified using the modifying function, and a new user prescribed value file is created.

Subsequently, similarly, in step S2, the input data items of each part are designated and input from the operation unit 7. The above-described designation input states are sequentially displayed on the display screen of the display unit 8, and the designation input is sequentially advanced while checking the designation state with reference to the display screen. Specifically, the business in which the form to be input is used is specified, and the description / title, the definition of the character recognition field, etc. (the size of the input character, the number of input digits, the space between the characters, the recognition character type, the frame type, the input instruction character, etc.) And the like).

When the above instruction is input, the CPU 1 automatically creates an OCR input form based on the designated data in the following step S3. This includes the layout of the input character frame and the input instruction character including the combination of the data items automatically, and the display of the layout result on the display unit 8.

The form created by the first specification is a temporary form, and this created form is displayed on the display unit 8 in a form closer to the actual print output pattern. For this reason, referring to this display, the report created by the printer 5 is printed out if the arrangement or the like is satisfied to some extent. This printout can be performed in the form of a composition or in the form of an actual OCR form. If you want to make further corrections here,
The position of the correction item is moved to an arbitrary position on the display screen of the display unit 8 by the operation unit 7, and the result is displayed again on the screen, or the above-described processing is repeated again with reference to the print result or the like. .

In printing the composition, the print output position can be adjusted in 0.1 mm units for each input item in order to correct the printing accuracy of each printer attached to the system. For this reason, any printer can print out a copy of a form requiring printing accuracy.

When the design of the form is completed in this manner and the printing of the composition is completed, the completed composition can be ordered from the OCR data input sheet printing company as shown in step S101. . Then, step S10
When the OCR sheet printed in step 2 is delivered, in step S5, business processing based on actual OCR input is performed.

When the form design is completed, test data and the like are entered using the printed form in the process of step S5, and read by the scanner 4.
Then, an input recognition process of input data read from the scanner 4 is performed. Then, it is determined whether the form designed in steps S1 to S3 can be used as it is or not, and if there is an improvement, the design processing of the form designed before step S1 is performed again. In the present embodiment, the form can be easily designed and the designed form can be easily corrected by only selecting the item on the display screen of the display unit 8 and inputting a few numerical values. Can be printed out and retested immediately.

Here, in the process of recognizing the OCR input data in the work and the like after step S5, step S5
The recognition support data file created in step S52 using the recognition support data in the recognition processing of various OCR inputs, for example, prepared in advance or registered in consideration of the characteristics of the input person as shown in FIG. Device 9
The OCR input is recognized by referring to a recognition support data file using recognition support data in the recognition processing of the various OCR inputs.

In this embodiment, when recognizing handwritten character input, it is possible to preliminarily register the characteristics of the character that will be input by the input user and register it as the recognition support data in step S51. It is possible to easily create a recognition support data file used in the OCR input recognition processing. This support file includes a dictionary for each designated item, a feature dictionary for each person, a private character dictionary, and the like.

In this embodiment, in this OCR processing work, in order to efficiently convert a large number of types and a large number of OCR forms into data without processing errors, a face sheet in which management items are described in batches (formation of forms) is used. This sheet can be set and subjected to input processing, and the subsequent data processing status can be managed using this sheet. For this reason, it is possible to selectively manage any one of the sequential calibration and the batch calibration as needed, or any one of them.

Then, in step S6, a calibration galley is output if necessary, and the calibration process in step S7 can be performed. Proceeding to step S8, the correction data required is displayed on the display unit 8 based on the calibration result, necessary data maintenance is performed,
In step S9, a completion data record file is created. Then, in step S10, the information is registered in the HDD 3.

As described above, according to this embodiment, a single machine is used to design a form as an OCR input sheet, create a copy, enter OCR data, recognize input data,
All business processes based on the R input do not incorporate any computer programs and therefore have special knowledge,
It can be performed without skill.

Next, various support functions are provided so that a design process of an OCR input data sheet (form) specific to the present embodiment during each of the above-described processes and a business using the OCR input data sheet can be smoothly performed. ing. Hereinafter, representative functions of the present embodiment will be described in detail.

In the present embodiment, these functions are sequentially selected in accordance with the guide displayed on the display screen of the display unit 8, and only necessary numerical values are input, whereby the design of the form is automatically completed, and It is configured so that it can be created. That is, first, the process selection menu screen shown in FIG. 3 is displayed,
Various processes shown on the process selection screen are selected and executed.

As shown in FIG. 3, in the present embodiment, a default value file management function, a form design function, a composition creation function,
It has a recognition support function, a data input function, a data maintenance function, and other optional functions.
One of these functions will be selected and executed as needed.

[Specified Value File Management] The specified value file management of this embodiment is performed in various forms of A8 size to A4 size and B6 size to B4 size which can be used in the OCR system of this embodiment. (OCR data sheet) manages a file in which various prescribed values of standard paper space for use are determined. Therefore, when designing a form, the user first uses this management function to determine the character size, character spacing, line spacing, the spacing between fields, and the positional relationship with the outer periphery that can be arranged on the paper of the field. Will be defined.

In this embodiment, the definition of the prescribed value is designed using the same prescribed value in which the presence or absence of the FD, the size of the form, and the direction of the form (vertical direction up and down, horizontal direction up and down) are the same. Business forms can be treated as the same group. The specified value provided by the system can be checked, and the designated selection can be made as needed by moving the cursor from the operation unit 7, and this function is used as appropriate in the case of the next form design.

[Form Design] Form design is a new form of O
This is a subsystem for setting and defining a record item format of an input data item of a new OCR data form when data is input using a CR data sheet. Then, at the same time as defining the record item format, it is possible to optimally arrange where in the OCR data form the record item is arranged according to a simple automatic arrangement rule.

After all the definitions are completed, the automatic arrangement position can be moved to the position desired by the user.

In this embodiment, when this form is designed, a new OCR data sheet that the user wants to create is automatically set in which form and what kind of business data. . Therefore, it is not necessary to record the data in a ledger or the like each time, and data to be read and processed for subsequent data input is automatically managed.

In this form design, the arrangement of the reference marks is determined first, and then the character recognition field area is determined. Then, the relationship between each character recognition field and the frame is determined, and the relationship between the input instruction character and the character recognition field in the character recognition field is determined. Due to this relationship, the arrangement of the character recognition field may be restricted. When displaying the input instruction character next to the character recognition field, the arrangement can be made over a plurality of lines. Is displayed, there is a restriction on the number of arranged rows.

FIGS. 4 and 5 show examples of reference mark arrangement positions and reference marks from the end of the form.
FIG. 5 shows details of the content and value of the position indicated by each character in FIG.

Similarly, FIG. 6 and FIG. 7 show examples of the position of each character recognition field and the distance between each feed from the end of the form.
Shown in FIG. 7 shows details of the content and value of the position indicated by each character in FIG.

FIG. 8 shows an example of specifying the positions of the character recognition field and the frame. Nine types of frames shown in FIG. 9 are prepared, and the frame layout can be easily controlled simply by designating each line type.

It is to be noted that each of the above arrangements can be specified with the precision in parentheses shown as respective values, and only by inputting a numerical value corresponding to the position indicated by each alphanumeric character,
Each part can be automatically arranged. However, the thickness of the frame line can be specified by the number of dots.

Along with the designation of this frame, the arrangement of the input instruction character (instruction of the type of data to be written / written in the corresponding character recognition field) at the time of OCR input is specified. As an example of the arrangement of the input instruction characters, any one of upper, lower, left, and right of the character recognition field, or an interval can be designated.

FIG. 10 shows a print example of the OCR input sheet on which the form is designed as described above. Here, an example is shown in which the frame is a square with a line type of 7, and the input instruction character is arranged in the frame above the character recognition field.

[Preparation of Composition] In this embodiment, the function of printing a data sheet on paper and the purpose of printing the OCR data sheet form defined in the form design described above in preparing the composition are defined. An item for displaying data on the display screen of the display unit 8 when the data sheet is read is defined.

In order to perform block printing, first, a model chart of the paper size and orientation to be used is output, and subsequently, measured values are registered, and a block copy, for example, a drop-out plate, a black plate, or a completed plate is prepared. And print the desired number of copies. At this time, as an auxiliary function, the printout position can be corrected in units of (1/10) mm to correct the error so that the composition can be output regardless of the printing accuracy of the printer 5. It is configured as follows. This correction function is prepared for each size and direction of the composition paper.

Therefore, when creating a composition from sheets of the same size, the correction value once specified can be used as it is. These correction values may be stored in the FD or may be stored in the HDD 3.

The black ink plate of the composition is positioned on the dropout color plate or the completed plate with the register mark [+], and the specified data is written in the frame, and the mark is accurately cut to the specified paper size with the register mark and read. After performing a test and confirming that the operation is correct, the display is automatically displayed on the display screen of the display unit 8, and items to be displayed here are selected. Items that do not need to be displayed can be omitted.

For example, when the data items of the master-slave relationship that need to be recognized appear in the sixth and seventh rows of one screen, the division is automatically prohibited and the next screen is displayed. Control such as lowering is performed.

[Recognition support] This is particularly effective when the writer (written person) of handwritten characters has a habit of writing unique characters and the recognition rate cannot be increased by ordinary recognition. Demonstrate. In such a case, in this embodiment, a handwriting is registered exclusively for the creator, and a recognition correction file is created and registered so that recognition can be performed as accurately as possible. When registering, it is a troublesome OC
There is no need to create an R form, and a free format data sheet can be used. In the present embodiment, the automatic registration of the external character font can be performed by the dot code using the function of reading the character image by the OCR. For this purpose, a dedicated external character font registration sheet is prepared.

[Data Entry] This is a function for reading actual business data using the form created as described above.
It is configured so that batch lot management can be performed by creating a bunch of a large amount of data for each given number. In this case, for the batch management, a method of writing an instruction on the face control sheet or a method of inputting control instruction data from the operation unit 7 can be selected.

The present embodiment is not limited to the batch processing, and it is also possible to select a job directly from the screen and input data without performing batch lot management. Furthermore, at the time of data input, it is also possible to set a recognition file in which personal handwriting is registered and perform data recognition. Furthermore, as a data input method, 1
Sequential input that corrects instantly while displaying on the screen one by one,
It is possible to select only the data input first and the batch input to make a new correction after creating the file.

[Data Maintenance] Data maintenance provides various functions for performing data update management for an OCR data sheet record file.
It has the following functions.

(1) A checklist function for arranging and printing out the read image information and text code information strings of recognition results (recognition candidates) (an example of printout of this checklist is shown in FIG. 11).

(2) Function for correcting read data.

(3) Function of creating completed data in a data file in the MS-DOS format and recording the data in the FD.

(4) A function of displaying the status of the read data on the display unit 8 and displaying the batch status for grasping the status, and changing and correcting the status.

In the present embodiment, any access can be easily performed on a per-work-batch basis.

As described above, according to this embodiment, the OC
A series of software development from the design of the R form to the operation management can be easily performed only by selecting the items displayed on the display screen of the display unit 8 and inputting simple numerical values.

In particular, in designing a form, a graphical layout diagram can be obtained by inputting specified values for the form design in accordance with a guide from the display unit 8.
The result can also be output to a printer. For this reason, O
It can also be used as a composition for CR forms.

[0069]

As described above, according to the present invention, the OC
A series of software developments from R form design to operation management can be easily performed only by selecting items displayed on the display screen and inputting simple numerical values. Therefore, from the design of the OCR input paper to the operation management can be centrally managed without preparing a program for each item processing.

[Brief description of the drawings]

FIG. 1 is a block diagram of an OCR system according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a form design and OCR recognition process according to the embodiment.

FIG. 3 is a diagram illustrating a display example of a process selection menu on a display unit in the embodiment.

FIG. 4 is a diagram for explaining the basics when arranging reference marks according to the embodiment.

FIG. 5 is a diagram for explaining the basics when arranging reference marks according to the embodiment.

FIG. 6 is a diagram for explaining the basics when arranging a character field area according to the embodiment;

FIG. 7 is a diagram for explaining the basics when arranging a character field area according to the embodiment;

FIG. 8 is a diagram illustrating an example of specifying a position of a frame with respect to a character recognition field according to the present embodiment.

FIG. 9 is a diagram for explaining an example of specifying a position of a frame with respect to a character recognition field according to the embodiment;

FIG. 10 is a diagram illustrating an example of a form design in the present embodiment.

FIG. 11 is a diagram illustrating an example in which read image information and text code information strings of recognition results (recognition signs) are printed out side by side.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 CPU 2 Memory 3 Hard disk drive HDD 4 Scanner 5 Printer 6 Communication control unit 7 Operation unit 8 Display unit 9 Flexible disk (FD) device 10 Network

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-59-165187 (JP, A) JP-A-3-74776 (JP, A) JP-A-3-240880 (JP, A) JP-A-62 200483 (JP, A) JP-A-63-147286 (JP, A) JP-A-64-91288 (JP, A) JP-A-3-88086 (JP, A) JP-A-63-108429 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06K 9/00-9/82 G06F 19/00

Claims (3)

(57) [Claims] 1. An OCR system capable of designing an OCR input sheet, comprising: an operation unit for inputting instruction information of the OCR input sheet; and a display unit for displaying instruction input guidance and an instruction input result on the operation unit. A printer for displaying a layout result of the OCR input form, a scanner for reading OCR entry information from the OCR input form and management information from a predetermined control form, and a character characteristic of a writer who will fill in the OCR input form. A registered recognition support data file, and a control unit capable of displaying the instruction input guidance and the instruction input result on the display unit based on the instruction information from the operation unit. , Prescribed value management , OCR input paper design , composition
A menu screen including items such as creation and recognition support is collectively displayed on the display unit, and when the item of the specified value management is selected from the menu screen, a sheet of OCR input paper input from the operation unit is displayed. managed and stored in the memory to identify the specified value by the size and aspect, when the menu item of the OCR input sheet designed from the screen is selected, the input data item of the OCR input sheet based on the specified value If the automatic layout requires correction, correct the necessary portion of the specified value, and when the item of the composition is selected from the menu screen, the automatic layout of the automatic layout is performed for the printer. Conclusion
To print the composition of OCR input paper based on the result
And the OCR input output by the printer.
When an error between the layout of the printing paper and the layout of the specified value is input, the OCR input paper is set based on the error.
When the item of recognition support is selected from the menu screen, the control paper and the OCR input paper are read by the scanner for each bundle.
To take the management information from the control sheet.
By recognizing the OCR input information from the OCR input paper based on the character characteristics registered in the recognition support data file, thereby processing each item from design to operation management of the OCR input paper. An OCR system that can be centrally managed without preparing a program for each system. 2. The menu screen is further configured to be capable of displaying data entry items, and the control unit is configured to, when the data entry item is selected from the menu, display a previous data entry item.
The OCR system according to claim 1, wherein character characteristics stored in the recognition support data file are selected in accordance with management information based on the control paper, and the OCR entry information is recognized for each bundle from the OCR input paper. 3. The menu screen is further configured to display data maintenance items, and the control unit includes:
2. The OCR system according to claim 1, wherein when the data maintenance is designated from the display item, a checklist in which recognition results based on the OCR entry information and recognition candidates are arranged in parallel can be output to the printer.