JP3353999B2 - Image data recognition device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data recognizing apparatus, and more particularly, to an image data recognizing apparatus that corrects received image data so that the received image data can be recognized when the image data cannot be recognized by facsimile communication. About.

More specifically, when character recognition is performed on image data input from an OCR facsimile manuscript received by facsimile communication, a system is desired that corrects characters that cannot be recognized without human intervention and enables recognition. ing.

[0003]

2. Description of the Related Art FIG. 9 is a diagram for explaining a conventional data correction. As shown in the figure, the user enters necessary information on an OCR entry form, and the facsimile apparatus 100
To the character recognition device 120 via the facsimile communication control unit 110. When the character recognition device 120 receives the OCR image data from the facsimile device 100,
The data is scanned by the scanner 221. Scanner 22
1 is passed to the character recognition processing unit 222, analyzed, and the recognized recognition result is output to the user data editing unit 130. The character recognition processing unit 2
The error data that could not be recognized at 22 is manually corrected and re-input to the character recognition device 120.

[0004] In addition, error data that cannot be recognized even after being corrected manually by the character recognition processing unit 122 is notified to the facsimile apparatus 100 that error data is present, re-entered, and transmitted. Prompt.

[0005]

However, in the above-described conventional example, when error data is output from the character recognition device, it is necessary to manually correct the error data so that it is recognized as a normal pattern and re-input the data. . At this time, the data output as the error data can be corrected by the operator in a range that can be determined by the operator, such as a thin marking, a numerical value, etc. being out of a certain frame. If there is a character or the like, it is not possible to determine how the character can be read. In this case, it is necessary to use a facsimile or a telephone to inquire directly what to read.

[0006] In addition, for those which cannot be read, processing such as a request to input again by telephone is required, and there is a problem that a load is imposed on an operator.

The present invention has been made in view of the above problems, and solves the above-mentioned conventional problems. When error data occurs, the error data can be automatically corrected without imposing a load on an operator. It is an object of the present invention to provide a simple image data recognition device.

According to the present invention, there is provided an image data recognizing apparatus for recognizing information entered in a predetermined item column of input image data.
Memory means for storing a correction rule, recognizing means for recognizing the input information as an image data, if the information recognition means recognizes as an error, further determine the error type
A judging means for performing constant, if the result of the determination means is a recognition error, if the re-recognition means for performing re-recognition of the information based on the recognition rule, the determination means results are corrected errors, based on the modified rule Correction means for correcting information.

[0009]

[0010]

According to the present invention, normal character recognition processing of input image data is performed, and if an error is detected in the recognition result, an error type is determined for the error, and the error type is determined. Select a rule, execute any of the rules, and get a modified recognition result.

Further, even when an error is detected as a result of recognition, it is possible to input a plurality of input data by a single input operation such as a continuous form, and correct the error data item with the correctly recognized data. it can.

[0012]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 shows a system configuration of an embodiment of the present invention. 9, the same components as those in FIG. 9 are denoted by the same reference numerals. In the present embodiment, the OCR
The entered image information is transmitted and input to the character recognition processing unit 120 via the facsimile communication control unit 110.
The character recognition processing unit 120 performs a character recognition process on the input image data using the normal recognition pattern 200, and outputs a recognition result. Also, the character recognition processing unit 120
If the recognition result is an error, the error analysis is performed on the recognition pattern in which the error has occurred.
In the recognition error analysis, the type of the error is determined, the rule 300 is searched based on the result, and the rule is executed. As a result, if an error occurs again, an alarm is output and processing for the error is performed. In this case, the operator analyzes the error pattern and creates and adds a rule, or inputs normal data.

The operator generates a recognition pattern properly and registers it in the recognition pattern holding unit 200.

The recognition result recognized by the character recognition unit 120 is input to the user data editing processing unit 130.

FIG. 2 shows a configuration of a character recognition processing unit according to one embodiment of the present invention. The character recognition processing unit 120 shown in FIG.
Data input unit 121, feature extraction unit 122, pattern matching unit 123, recognition pattern holding unit 124, error analysis unit 125, alarm generation unit 126, and rule 127
It is composed of

The data input unit 121 receives image data written in OCR by facsimile communication. The feature extraction unit 122 performs feature extraction into a format that can handle an input character string by pattern matching. The pattern matching unit 123 performs pattern matching on the data from which the feature has been extracted, with reference to a pattern of the recognition pattern holding unit 124 prepared in advance. As a result of the pattern matching, if the extracted feature matches the pattern of the recognition pattern holding unit 124, the recognition result is output. Also,
If they do not match, the error is output to the error analysis unit 125. The error analysis unit 125 determines the error based on information such as the sharpness of the image data, the communication state at that time, the state of the input terminal, a user's operation / entry error, the OCR sheet, the recognition field attribute, and the relationship between the preceding and following data. And analyze. The rule is executed by referring to the rule 127 based on the analysis result. There are rules instructing calculation of numerical values, rules instructing re-recognition processing, and the like. The error analysis unit 125 determines, for data that results in an error even when the rule 127 is applied,
The occurrence of an error is notified to the alarm generation unit 126.

The alarm generation unit 126 includes an error analysis unit 1
If there is data output as an error as a result of performing the reading recognition process using a specific pattern in 25,
Generate and output an alarm.

FIG. 3 is a flowchart showing the operation of one embodiment of the present invention.

First, when image data is input to the data input unit 121, features are extracted by the feature extraction unit 122, and input to the pattern matching unit 123, the pattern matching unit 123 (Step 101). Here, if the recognition is successful, the recognition result is output, and the process ends. On the other hand, no recognition result is obtained, that is, the input data is
4 does not match the pattern in the error analysis unit 1
25 analyzes the error pattern and selects an error type (step 103). At this time, the error analysis unit 1
Step 25 determines whether there is a rule corresponding to the input data in which the error has occurred, based on the error type (step 10).
4) If not, the process proceeds to step 112 (step 112). The type of the rule is determined again whether it is a recognition rule for performing recognition or a correction rule for correcting the recognition result by executing the rule (step 105). When the recognition rule is used, a specific pattern is searched. (Step 106), and performs recognition processing (step 107). If the recognition result is an error, step 1 is executed.
The process proceeds to step S12, and if it is normal, a recognition result is output (step 108).

When a correction rule is used, the rule is executed (step 109), and the recognition result is corrected (step 110).

If there is no rule corresponding to the error type, or if the recognition result shows an error even if the recognition process is performed again, the alarm generation unit 126 generates and outputs an alarm (step 112). Recognition processing is performed again based on the input image data (step 11).
3) Based on the recognition result, the operator performs processing such as creating a normal recognition result, creating a new pattern and registering it in the recognition pattern holding unit 124, and changing / adding rules (step). 114). Finally, a pattern reflection process is performed on the user data editing processing unit 130 (step 115).

FIG. 4 shows the configuration of the error analysis unit according to one embodiment of the present invention. The error analysis unit 125 shown in FIG. 8 is for re-recognizing or correcting the recognition result in which an error has occurred in the pattern matching unit 123.

The error analysis unit 125 examines the error type of the error recognition result for each factor and selects an error type. The error type analysis unit 1251 selects an error type based on the error type. A rule selection unit 1252 for selecting a rule to be applied to the result; a rule storage unit 1255 for storing rules to be applied; and a specific pattern storage unit 1256 for storing specific patterns other than those stored in the recognition pattern storage unit 124 , Rule selection unit 125
Execute the rule selected by 2 on the recognition result,
The rule execution unit 1253 corrects the recognition result, and the recognition result recognition determination unit 1254 determines whether the corrected recognition result is an error.

The error type analysis unit 1251 determines the reception environment such as the presence of noise or the like during communication with respect to the sharpness of the image data as a whole, the presence or absence of noise or the like during communication, the presence or absence of an abnormality during transmission by the OCR-input transmitting side, and the like. Whether the correct format is used for the OCR sheet or the type of the recognition result in error is determined based on information such as the attribute of the recognition field in error and the data relevance before and after.

The rule selecting section 1252 selects a rule corresponding to the error type from the rule holding section 1255.

The rule holding unit 1255 sets rules for each type of recognition result in which an error has occurred. For example, in the case of a delivery slip, what should be entered for each column of the delivery slip, the rules of entry, etc. are described.If a certain column is a numerical item, an error occurs due to calculation. The rules for correcting the data in the column being stored are accumulated.

The rule executing section 1253 is provided with a rule selecting section 1
Based on the rule selected in step 252, correction of an error recognition result, re-recognition processing, and the like are performed.

The recognition determining unit 1254 includes the rule executing unit 12
Based on the recognition result output from 53, it is determined whether the error is a recognition result that has not been removed yet or an accurate recognition result. If the error has not been removed yet, the process of the error analysis unit 125 is performed again. Or the alarm generator 12
6 is notified that there is an error recognition result. Here, if the recognition result by the specific pattern holding unit 1256 is correct, the recognition pattern used for the recognition processing may be added to the recognition pattern holding unit 124.

FIG. 5 is a flowchart showing the operation of the error analysis unit according to one embodiment of the present invention. Error analysis unit 12
When the recognition result determined to be an error is input to the error type analysis unit 1251 of No. 5 (step 201), the error type analysis unit 1251 determines the type of the error in the recognition result. For example, the form name of the data with the recognition error,
The type of the error is determined based on the item name, whether the item is a character item or a numeral item (step 202). Rule selector 1252
Selects a corresponding rule from the rule holding unit 1255 based on the type of error determined by the error type determination unit 1251 (step 203). For example, if the low definition of the image data (error code: 10) is the cause of the error, the rule is “increase the definition again and re-scan” (rule code: A-2). Content. For such association, a rule code of the rule holding unit 1255 is searched for from an error code as shown in the table below.

[0031]

[Table 1] The rule executing unit 1253 corrects the recognition result in error and performs pattern matching based on the rule selected by the rule selecting unit 1252 (step 204). If the recognition result that resulted in an error due to the execution of the rule has been corrected or the error has been removed (Step 205, Yes), the recognition result is output. If the rule cannot be corrected or the error is not removed by executing the rule (step 206, N
o) Notify the alarm generation unit 126 that manual intervention is required (step 207).

Next, the detailed operation of the error analyzer 125 will be described using a specific example.

Error type analysis unit 1 of error analysis unit 125
251 analyzes the cause of the error based on the recognition result, and outputs an error code to the rule selection unit 1252.
The rule selection unit 1252 searches for a corresponding rule code based on the error code. (1) For example, pattern matching is performed when the rule “refer to the column at the same position on the previous page” is selected by the rule selecting unit 1252 based on the relationship between the preceding and following data. There are columns for delivery destination name, delivery destination code, delivery source name, delivery date, delivery item, delivery quantity, unit price, total amount, and the like. Here, a case in which the delivery destination name cannot be read due to unclearness and an error occurs will be described.

In the case of this delivery slip, it is assumed that a plurality of (continuous report) delivery slips are input from a trader of the same delivery source.

For example, in the case of a continuous form, almost the same contents are statistically often written in the previous and next pages. Therefore, if the contents of the previous page are correctly recognized, the page is referred to. I do. In order to refer to the previous page, the character of the column of the previous page at the same position as the character of the column of the recognition result in error is registered as a specific pattern in the specific pattern holding unit 1256, and the registered pattern and the error The pattern matching with the pattern of is performed.

Or, when inputting a plurality of slips at the time of input, and inputting by one input operation, if the following result is obtained, the first data has been recognized. The second data has been recognized. Data error 4th data error 5th data When the recognition is completed, the 3rd data is an error. In this case, 3 data out of the 2nd data input earlier Correct the eye data by referring to the same item as the item in which the error occurred. The data of the fourth sheet is also an error, but at this time, the data of the immediately preceding third sheet has an error, so the data is corrected with reference to the data of the second sheet.

It should be noted that, in a slip, fixed items can be corrected by referring to such a method. However, the numerical values, such as the number and the amount of money, are the probability that different numerical values are entered for each slip. May be configured to determine whether or not this method can be applied depending on the item.

FIG. 6 shows an example of correction of the error analysis unit according to one embodiment of the present invention. FIG. 7A shows a recognition result in which an error has occurred in the pattern matching processing, in which the customer name is recognized as “Alpha? Kinkabashikaisha” and the department name “Ayo? Hombu”. Here, if the “?” Mark is a part that cannot be recognized, the recognition result of the data one page before shown in FIG. As a result, the characters "de" and "u" at the same position
Is used to correct the error recognition result (A).

(2) When the rule “correct the numerical value by calculating the vertical, horizontal and total sums” based on the recognition field attribute is selected by the rule selection unit 1252: OCR already filled in as shown in FIG. When a sheet is input and the result recognized by the pattern matching unit 123 is as shown in FIG. 11B, there are three “?” Columns, and all columns are recognized. Since it has not been performed, it is output to the error analysis unit 125 as an error. In such a case, the error analysis unit 125 uses the rule 127 regarding the order slip as a pattern to be recognized. As the rule 127, the rule used at this time is “use a rule that checks the total length and width and corrects the data”. The rule 127 first determines the recognition result for each row and corrects data.

First, for the first column, 1 + 2 + a
= 6, so the part a of the "?" For the second column, 0 + 0 + 1 = 1, which is consistent. For the third column, since 0 + b + 3 = c, and no solution can be obtained, it is undecided here. 4th
For the column, since 1 + 2 + d = 10, “?”
Part 7 becomes 7.

Next, the recognition result is determined from the vertical sum and the data is corrected. The first vertical line is 1 + 0 + 0 = 1, which indicates that the recognition is accurate. The second line is 2 + 0
Since + b = 3, b of “?” Becomes 0. Then the third
In the line, a + 1 + 3 = d, which is undecided. In the fourth row, since 6 + 1 + c = 10, c of “?” Is 3. 1) 1 + 2 + a = 6 a = 3 2) 0 + 0 + 1 = 1 3) 0 + b + 3 = c b =? , C =? 4) 1 + 2 + d = 10 d = 7 Vertical rows 1) 1 + 0 + 0 = 1 2) 2 + 0 + b = 2 b = 0 3) a + 1 + 3 = da =? , D =? 4) From the above result of 6 + 1 + c = 10 c = 3, a = 3, b = 0, c = 3, d = 7. As described above, when the numerical value is not known, the data is corrected from the total in the vertical and horizontal directions. The corrected result is output as a recognition result.

In order to correct the error data as described above, the rules are as follows:
A resolution rule may be set for every form received. In the case of the above delivery slip, the text items are the delivery destination name, delivery source name, and delivery items, and the numeric items are the delivery date, delivery destination code, delivery quantity, unit price, total price, etc. Column. Of the numerical items, the items calculated by the calculation include the delivery quantity, the unit price, and the total price. That is, when making corrections to this delivery note, the following rules are applied.

FIG. 8 is a flowchart showing an example of a rule according to an embodiment of the present invention.

First, the rule selecting section 1252 determines whether the location of the recognition result in error is a character item or a numeric item. In the above example, if an error has occurred in the items of the delivery destination name, the delivery source name, and the delivery item (step 301), the process proceeds to step 302 because the error is in the character item. If the location of the recognition result in error is the delivery date, delivery destination code, delivery quantity, unit price, or total amount (step 302), the process proceeds to step 306.

If it is a character item, reference is made to the item at the same location corresponding to the error position of the designated page (eg, the previous page, the first input page, etc.) (step 302). If there is no data at the corresponding location (Step 303, No), the corresponding location on the page one page before is referred to (Step 304).
If there is data at the corresponding same location (Step 303, Yes), the data referred to the error location is applied (Step 305).

If the item in error is a numerical item (step 301), it is determined whether the item is a calculable item (step 306). If the item cannot be counted, the process proceeds to step 302. The process proceeds to perform the same processing as for the character item. If the item is a countable item, the value of an error location is calculated by a calculation rule. For example, vertical and horizontal calculations are performed as in the above (2) (step 307). The obtained value is set as an error location (step 308).

The type of rule may be created for each input form, or may be created only for general items. Further, as a rule, for example, when the sharpness of the image data is lower than a predetermined value, the sharpness (resolution) is increased to a predetermined value, and the re-recognition processing is performed. In a series of processing of extracting and normalizing an area to be detected based on an input character pattern, performing a feature extraction process, and performing pattern recognition, for example, in the first recognition process, only normalization of a character position is performed. In such cases, normalization of character inclination and normalization of character line width are also performed, and pattern matching processing is performed again. If error data still occurs, processing for each item is performed. There may be.

By using the rules as described above,
It is possible to automatically correct an item for which the recognition result is erroneous.

Next, when a notification indicating that the recognition result is erroneous recognition is input from the error analysis unit 125, the alarm generation unit 126 generates an alarm signal and turns on the alarm LED on the device. Sound the alarm. As a result, the operator confirms the recognition result in error, for example, by displaying it on a display, and corrects the recognition result that can be determined by the operator.

Further, by learning the character pattern in which the error has occurred a plurality of times, the character pattern learned a predetermined number of times can be registered in the recognition pattern holding unit 124. For example, the character (A1) having an error and the correct recognition result (A2) are made to correspond to each other, and after performing a feature extraction process on the input character (A3), the character (A1) having an error first Is performed. Such processing is repeated a predetermined number of times, and if the input character (A3) matches the character (A1), the character (A
It is also possible to register 1) in the recognition pattern holding unit 124.

If the recognition result in which the error occurred is not corrected by the error analysis unit 125, a new rule is added to the rule storage unit 1252, or a recognition pattern is created and stored in the recognition pattern storage 124. I do.
As a result, even if the subsequent data has a similar pattern, it is not erroneously recognized from the next time.

The present invention is not limited to the above embodiments, but can be variously modified within the scope of the claims.

[0053]

As described above, according to the present invention, a rule which is an error logic is searched for data having an error in the character recognition processing, and the rule is applied to the data having an error. Can be automatically corrected. For a pattern that cannot be corrected, it is possible to generate an alarm and create a rule that will not cause an error in the next recognition process by the operator.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

FIG. 2 is a configuration diagram of a character recognition processing unit according to an embodiment of the present invention.

FIG. 3 is a flowchart showing the operation of one embodiment of the present invention.

FIG. 4 is a configuration diagram of an error analysis unit according to one embodiment of the present invention.

FIG. 5 is a flowchart illustrating an operation of an error analysis unit according to one embodiment of the present invention.

FIG. 6 is a diagram (part 1) illustrating an example of correction of an error analysis unit according to an embodiment of the present invention.

FIG. 7 is a diagram (part 2) illustrating an example of correction of the error analysis unit according to the embodiment of the present invention;

FIG. 8 is a flowchart illustrating an example of a rule according to an embodiment of the present invention.

FIG. 9 is a diagram for explaining conventional data correction.

[Explanation of symbols]

 REFERENCE SIGNS LIST 100 facsimile device 110 facsimile communication control unit 120 character recognition device 121 data input unit 122 feature extraction unit 123 pattern matching unit 124 recognition pattern holding unit 125 error analysis unit 126 alarm generation unit 127 rule 130 usage data editing unit 200 recognition pattern 221 scanner 222 Character recognition processing unit 300 Rule 400 Specific pattern 1251 Error type analysis unit 1252 Rule selection unit 1253 Rule execution unit 1254 Recognition result determination unit 1255 Rule holding unit 1256 Specific pattern holding unit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06K 9/03 G06K 9/62 620 G06K 9/72

Claims (3)

(57) [Claims] 1. An image data recognizing apparatus for recognizing information entered in a predetermined item column of input image data, comprising: storage means for storing a recognition rule or a correction rule for each type of the item; recognition means for recognizing information inputted as data, if said information in said recognition means recognizes as an error, is further determining means for determining the error type, the result is recognized error in the determination means A re-recognition unit that re-recognizes the information based on the recognition rule; and a correction unit that corrects the information based on the correction rule when a result of the determination unit is a correction error. An image data recognizing device characterized in that: (2)The correcting means includes: Based on the calculation results using the numerical values in the information,
4. The image data recognition method according to claim 1, wherein the error is corrected.
Sense device. (3)The correcting means includes: The total value for each numerical row in the information and the total value for each numerical column in the information
4. The method of claim 1, wherein the error is corrected based on a total value.
5. The image data recognition device according to 4.