JPS6214278A - Information processing system - Google Patents

Abstract

PURPOSE:To construct a system which can be applied to a wide scope by connecting an OCR module to a personal computer to travel an incorporated software and offering a data input system by the incorporated software and the OCR. CONSTITUTION:A means to translate a form parameter which the incorporated software has into the data form which the OCR module 50 can process is provided, the form parameter obtained by the same means is transferred to the OCR module 50 and the character written in a form 1 is read. The position of the frame can be assumed by the form parameter, and therefore, the pattern of a frame printed in black and a character pattern can be easily separated. The OCR module returns the read character code to a higher order personal computer, and respective data items are checked by the incorporated software. Thus, the system can be applied to the wide scope and the data input can be executed by recognizing the character.

Description

[Detailed description of the invention]

[Object of the invention] The present invention is a character! ! The present invention relates to an information processing system that manually inputs data based on knowledge, and particularly to an information processing method characterized by directly reading a form output by the system. [Background of the invention] In conventional OCR, the form (slip) on which the characters to be input are written is
It was a special edition printed in dropout color. Therefore, in order to alleviate such restrictions on usage, the same applicant has devised a method to directly read frame-type forms printed in black2). (See patent application No. 59-180517). In such a method, the meaning of each style is extracted by automatically reading the frame structure from the form image and recognizing the characters within the frame. For example, it is estimated that the name of the frame item, the type of data to be entered (name of nine people, date of birth, etc.), two character types, two character counts, etc. That is, format information can be extracted from a blank sample form using pattern recognition technology. Therefore, if the same method is used, OC
It is possible to eliminate the complicated work of creating format information, which has been one of the obstacles to applying R in a wide range of applications. However, since this method requires a large amount of information processing, it is not necessarily suitable for small-scale information processing systems such as those using personal computers. On the other hand, OCR has been miniaturized into an OCR module with a minimum configuration, and is being applied to small-scale information processing systems such as those described in L. However, since the OCR module described in -F has a minimum configuration, the module itself does not have functions such as creating format information and checking the +1 validity of the read data. not present. Therefore, in a limited format・Zoji reading 1~1)

[. However, it is difficult to apply it to a wide range of applications. [Purpose of the invention]

The purpose of the present invention is to solve the problems described in L.
The object of the present invention is to provide an information processing method that can be applied to a wide range of applications and that can input data through character recognition. (Summary of the Invention) In recent years, integrated software has been developed for business use in small information processing systems such as personal computers. Although these software systems mainly handle numerical data, they are expected to have a wide range of applications, as they integrate functions such as manual data editing, database functions, 9 charting and tabulation functions, and axis 91 document creation functions. However, human data input relies on manual input from a keyboard. The method of the present invention is to construct an information processing system that can be applied to a wide range of areas by providing the integrated software described in L and a data input method using OCR. Integrated software usually consists of numerical data, and when inputting that data, an input/output format is required.
It has the function of printing (forms) and the function of displaying a frame as a guide for input operations on the terminal screen. Furthermore, the fields on the above forms and screens (
For example, 0 has a function to check the validity of data input in the frame (corresponding to the frame -)
You can check the character type, 2-digit number, data type, etc., and check whether the relationships defined between fields are satisfied. One example is a sum check that determines whether the sum of fields is correct. Also, write down the data to be input into the input/output form,
Input from a keyboard is also commonly performed. In this case, the system internally stores parameters related to the form (field positions, data type, number of digits, conditions to be satisfied between fields, etc.). Therefore, connect the OCR module to a personal computer running integrated software such as -F.
By implementing new methods such as those described below, it is possible to construct an overall efficient information processing system. That is, a means is provided for translating form parameters possessed by the integrated software into a data format that can be processed by the OCR module, and the form parameters obtained by the means are transferred to the OCR module to read characters written on the form. At that time, the form does not have to be a special form printed in dropout color, it can be created with a normal printer, and the order of the frame can be estimated from the form parameters.
The frame pattern printed in black and the character pattern can be easily separated. Furthermore, the OCR module returns the read character code to the host personal computer, and each data item is checked by the integrated software. [Examples of the Invention] The present invention will be described below based on Examples. FIG. 1 shows an information processing system 1 which is an embodiment of the method of the present invention.
FIG. 1 is a configuration diagram of No. 1. The system is a CRT display w20. Magnetic disk (file) 30. It consists of a personal computer processing device 10 having a printer 40 and an OCR module 50. Further, the processing bag WI 110 can be connected to the large computer 100 through the communication line 101. Data input to the system is performed using the keyboard 21 according to the instructions on the displayed screen (RT 201) and the form number printed by the printer 40:
It has a function to read written characters from OCR 50 thousand joules. FIG. 2 is an example of a data input form output by the printer 40. With the 0CR50, characters can be read in a frame-like form printed in black as shown in the figure. Of course, the conventional m-sheet for OCR may also be used. FIG. 3 shows the main functional blocks of software that controls this system 11. The software consists of a business integrated software 60 and a character recognition human control module 7, and runs on the processor 10. The main processing flow is as follows. First, using the input/output cafform parameter definition function 62, a new form is defined, for example, as shown in FIG. 2, and the defined parameters are stored in the file 31. In addition to defining the table (frame structure), this definition also includes information such as which frames are for human data input, the type of data to be filled in each frame (field) (numeric value or alphanumeric character code, etc.), the maximum number of characters (digits), etc. (number), relationship between fields (for example, the result of vertical addition is the total value), etc., and these are also stored in the file 31 as parameters. Based on the defined form parameters, it is possible to print the input form as shown in Figure 2 or to create a report print based on the data in the file 32. Yes, you can create the input form by printing a large number of copies using the same print function, or by creating a large number of copies from a single print.Data input can be done using the form created as described in 1-1. This can be done by entering data and having the 0CR50 read it.The 0CR50 is activated from the terminal 20.21 via the data input editing program 64.Inside the program 64, data is input from the terminal 20.21 side. ka, O
It has a software switch to select whether to perform it from the CR side. If the switch is set to the OCR side, instead of inputting data from the keyboard 21,
The input request signal is sent to the OCR control program 1 by software. If it is set on the keyboard side, input data from the keyboard is accepted. When you enter a command to set the above switch to the OCR side from the terminal 20.21, the system displays the types of input forms that have already been defined as a menu on the CRT 20, and displays the form you are trying to read. Prompt the user for instructions. When the form . OCR format parameter creation progera! , 72 receives input form parameters from the file 31, and sends position information of each character frame, 9 character type information, digit number (maximum number of characters) information, etc. to the character cutting program in the OCR 50. Here, since OCR format parameters and form definition parameters generally have different description formats, it is necessary to interpret the form definition parameters and convert (translate) them into the OCR format parameter format. The program 72 transfers the above parameters to 0CR50 with the same format conversion. When the switch for selecting the data input device is set to the OCR side, the program 64 presses a predetermined key to end input from the OCR. A request is issued to the OCR control program 71. The OCR control program 71 is 0 in its initial state.
A thousand paper feed 11 request is issued to the 0R device 50, the first field is read, and the reading result is Progera l'=
71 & is sent back to the program 64 via &. When the reading of each field is completed, the progera 1171 sends back the reading result, and also checks whether the reading of all the fields of the doujin kafoom is finished using the parameters obtained from the progera 1172. When reading of all fields is completed, the form reading completion flag is transferred to the program 64, and the form reading completion flag is transferred to the program 64. , (form) from the OCR reading station is output to the OCR 50, and then a request is made to feed the paper form. When the data input editing program 64 receives the read result for each field, it stores it in a predetermined buffer, and when it receives the form read end flag described in , it checks the data in the buffer using programs 65 and 66. conduct. This check is recorded in file:31. This is done based on the form definition parameters for in-person attendance. If the reading result shows that there is no iEL by the same check, display the image of the form and the binding (pronunciation change) on the CRT 20, and further highlight the incorrect i1 on the blinking etc. do. The operator can correct the error according to the display. (:
The form image displayed on Ri'20 is r) r:
When Rtx is not used, the frame structure can be the same as that used to guide data input from the keyboard. Generally, it is desirable that the frame structure (form) printed on #A is the same as the frame structure displayed on the CRT. Next, the input/output cuff form parameter definition will be explained using the form shown in FIG. 2 as an example. The form in the figure can be defined as follows. DEFINE FORM-^; (1) 1'1EF IIEAD)! R'Deposit slip' Cl-NT
):R; (2) DEF ROWS C
AB, old, R2, R3, R4; (:(
)1) EF COL11MNS^(10), B(10)
, C(10), n(10); (4) rlEF
FTEl, T) LAB(A)='product name code';
(5) 11FF F position r) CAB(B)=
'lN valence', (6) Kano F FT
El, D 1. AB((',)='quantity',
(7) DPF FIE! 1. rl CAR(
D) = 'amount', (8) DEF
FTF! 1. DR4(A)='total',
(9) OFF FIHLDS R1=TNP
tlT(N, 5); (10)nE
FFTF! 1. Ds R2=INpHT(N, 5):
(11) nFF FTElnS
R3=INPIIT(N, 5); (1
2) nEF FTHl, r) R4(11)='[NP
IIT (N, 5); (13) 11F
F C(lNnTTTON old(11)*R1(C)=R1(n);
(14) D) iF C0NrlTTION R2(R) Homare R2(r,)=R2(n);
(15) DEF C0NIIITION R3(B) * R3(C)=R3(['l):
(lf'1)DEF C0N111TT
ON Old (D) + R2 ([1) + R3 (n) = R4 (I)):
(17)nllFCnNnlTTON ■(1)<g'11. E(PConF);
(18) rlF! F C0N
DTTTON R2(A)<FTLH(PCOl)H);
(19) DHF C0NIIITT(IN R3(^)<FTl,f!(PCOnF);
(20) DEF HEIGIIT I, Al
l 1.5; (21
)DFF HIETG)IT R11,5:
(22) DEF HETGIIT R21,5;
(23) DEF H
HIGHT R31,5;
(24) Dr! FHEIG)IT R4],
5: (25) END
FORM-A; (2
6) In the above definition statement, lines 1.26 declare the start and end of the definition. Line 2 defines the beginning of the form and declares that it will be printed in the center. Row 3 defines four! ,
At the same time, we define that the rows in (i.e. table) consist of five rows, and at the same time we write L A R, R1, R1,
They have been given tentative names R2, R3, and R4. Row 4 similarly defines the columns of the table. Lines 5 to 9 define fields and specify that constants on the right side of the isochrones are to be embedded. Rows 10 to 12 define fields collectively. For example, in row 10, all fields in row R1 of the table are fields for data input1), and the nine character types that describe input data are numbers. (N) and the maximum number of digits is 5. Similarly, field R4(1)) in row 13 is for manual input, meaning that a maximum of 5 digits can be input by 5 [1 person. Here, fields are specified by row names and column names enclosed in parentheses. Lines 14-20 define the conditions that the data between each field must satisfy. For Internal Medicine 14-17, four arithmetic operations are a condition. Rows 18-20 are fields R1(A), R2
(A), R3(A) data is file PCO1'IF
It specifies the condition that it is a member of a data set defined in (a file that describes part codes). For rows 21 to 25, each row of the table corresponds to 1.5 character lines of the printer.
It stipulates that the height be equal to the pitch. The width of the columns in the table is specified by the column definition statement, and in line 4 it is specified that each column is 10 printer character pitches. The four I explained below! , parameters (constants) are created by the program 62 and stored in the file 31. The OCR format parameter configuration program [1gram 72 reads the F-form parameters from the file 31 and converts them into a data format that can be interpreted by OCR.
Parameters regarding the terms and conditions of the agreement are stored. For example, the attributes such as the width, height, horizontal pitch, and vertical pitch of a single character (font) of the printer 40 are 0.1■
He is remembered for his hard work. Or, the row position where the eldest is printed, the first row in the table]
The line position, etc., where the next line will be printed is memorized. Using this information and form parameters, the program 72 can calculate the position, size, and number of characters of all input fields in Jt format, and this information is transferred to 0CR50. Ru. Further, the number of input fields is passed to the 00R control program 71. 0CR50 is constructed using known technology. What is different from the conventional method is the means for searching the position of the frame (front) on the form using a pattern, and the means for cutting out the image of the area inside the frame pattern. Regarding these means, separate inventions (Japanese Patent Application No. 59-180517, Patent Application No. 58-2342)
48, Japanese Patent Application No. 59-1918), so a detailed explanation will be omitted, but the line structure is extracted from the image input from the scanner, and a reference table is created from the coordinates of the four corners of the rectangular pattern with the maximum outline. It is possible to determine the location of Once the criteria is determined, the detailed location of each field is determined by L
It can be accurately determined by the method described below. Extracting a group of character patterns corresponding to a field from an accurately defined partial region of an image, and extracting each character pattern from le! This can also be easily achieved using conventional technology. The result code of the 18 characters identified as described above is sent to the data manual editing program 64 via the OCR control program 71. The programmer 1564 refers to the form parameters, reads lines 10 to 1; It is checked that this is true and that the four arithmetic expressions in definition statements 15 to 17 are satisfied. Also, for the reading results for the product name code column, read the file named pconF,
Check whether it is included in the set of product name codes defined in the file. If the conditions are not satisfied, the operator is prompted to make corrections as described above. In the embodiment described above, the data processing and database functions are provided in the integrated software 60 running on the computer 10, but the present invention is not limited to the same form. Large functional host] 00-, h
In that case, the personal computer 10 functions as a terminal of the host 100, and the contents of the file 30 are stored in a file system owned by the host 100. Even in this case, the essence of the invention does not change. Furthermore, in this embodiment, the attribute information of each field is recorded in the phono and definition parameters (file 31).
) BMs), apart from the form definition, 4 is recorded as a data attribute in the database. Did you?
), in the case of a system that is realized in combination with a general-purpose database management programmer, the form described in item 9, in which the form definition and data definition are managed in separate files, is also included in the system of the present invention. Furthermore, although the OCR module of this embodiment directly scans the form and reads the characters, it may also be used for historical recognition by separately scanning and inputting image information stored in advance on an optical disk or the like. This case is also included in the present invention. [Effects of the Invention] As described above, according to the method of the present invention, it is no longer necessary to create format parameters (format information), which is a complicated task required by conventional OCR, and data processing and database functions can also be improved. An efficient information processing system can be constructed by easily combining software and OCR. Furthermore,
With this system, there is no need to separately design input forms specific to OCR. , itself becomes a nf function, and a system rich in economy and quick response can be constructed. Brief explanation of the drawings Fig. 1 is a system configuration diagram, Fig. 7M is an example of a manual for manual use, and Fig. 3 is a rough diagram of functional blocks 1 to 11)) Fig. F. 10...Personal computer, 2 rl -(:
RT display device, 230...File device t, 40 Printer, 50...OCr<module, 1-11!4 votes (
Four! , ), 100・To large Hoss 1.

Claims (1)

[Claims] 1. A first means for storing form parameters that define a data input/output form, a second means for converting the form parameters into OCR format parameters, and a first means for storing form parameters that define a data input/output form; a second means for converting the form parameters into OCR format parameters; and a third means for extracting a character pattern from the image information to perform character recognition, the third means extracting a character pattern according to a format parameter output by the second means. An information processing method characterized by character recognition. 2. In the information processing method according to claim 1, the information processing system further includes fourth means for storing data attributes of each field on the form and constraint conditions defined between the fields. The information processing method is characterized in that the result of character recognition by the third means determines whether or not the result satisfies a condition to be stored by the fourth means. 3. The information processing method according to claim 1 further comprises a fifth means for defining the data input/output form, and a sixth means for outputting the form according to form parameters defined by the means. An information processing method characterized in that the form created by the sixth means is recognized by the third means.