Information processing device, recording medium, and computer program product

Abstract

The present invention supports a document designer by enabling him to set character string recording rules for document items. The present invention provides an information processing device and a recording medium. The information processing device (10) outputs the extracted character string recording rules for each document item when regularity related to recording of a confirmed character string as a result of confirmation of a document character recognition result by a confirmation and correction person is extracted from the document item.

Description

Technical Field

[0001] The present invention relates to an information processing device and a recording medium. Background Art

[0002] Patent document 1 discloses a method for controlling output of candidate characters of a character recognition device, wherein each character written on a paper medium and read by an optical component is recognized as a character unit as a set of pixels, and a first candidate character group including a plurality of characters that may become a correct character consistent with the character represented by the set of pixels is output according to a sequence of probabilities of being extracted as a correct character arbitrarily set in advance. The method comprises: assigning a candidate character storage component and a second candidate character storage component, wherein the candidate character storage component stores the character codes of the outputted first candidate character group in the sequence, and the second candidate character storage component stores the character codes of the characters stored in the candidate character group in the sequence. The characters extracted as correct characters from the first candidate character group in the character storage component store the total number of times the characters are extracted as correct characters and the number of occurrences corresponding to the order; based on the total number and the number of occurrences stored in the second candidate character storage component, a step is taken to further select a second candidate character group with a high probability of being extracted as the correct character from the candidate character strings stored in the recognition candidate character storage component; a step is taken to extract the correct character manually specified from the selected second candidate character group; and a step is taken to identify the order of occurrence of the correct character in the recognition candidate character storage component, and to correct the number of occurrences of the correct character in the second candidate character storage component corresponding to the order of occurrence and the total number of times the correct character is extracted as the correct character.

[0003] Patent document 2 discloses a text recognition device for storing format control information for reference in order to read text recorded on a paper, etc., and includes: a format control information storage unit for expressing information specifying the type of text in the format control information through a regular expression; a regular expression analysis unit for analyzing the regular expression in the format control information stored in the format control information storage unit; and a reading unit for obtaining a reading result for the text recorded on the paper, etc., based on the analysis result of the regular expression analysis unit.

[0004] Patent document 3 discloses a text recognition device, which includes: a text statistical information production unit for producing text statistical information of a bill; a standard pattern dictionary having a standard pattern representing the characteristics of text; a standard pattern dictionary changing unit for changing the content of the standard pattern dictionary based on the text statistical information; a text recognition unit for comparing the text pattern of the recognition object with the standard pattern of the standard pattern dictionary to perform text recognition of the text pattern; and a recognition result output unit for outputting the result of text recognition.

[0005] [Prior art literature]

[0006] [Patent Document]

[0007] Patent Document 1: Japanese Patent Application Laid-Open No. 03-291777

[0008] Patent Document 2: Japanese Patent Application Laid-Open No. 06-36069

[0009] Patent Document 3: Japanese Patent Application Laid-Open No. 09-35006 Summary of the Invention

[0010] [Problems to be solved by the invention]

[0011] To improve the confidence level of character string recognition results using optical character recognition (OCR), document designers who design the documents that will be read during OCR research what content users enter in the document items and predict whether certain entry rules exist within the character strings representing that content. For example, if a field for age is expected to contain a number, then if a pre-defined entry rule for entering numbers in the age field is set, the OCR process will recognize the age field as a number based on that entry rule. Therefore, even if an ambiguous character string, such as "2" or "Z," is entered, it will still be recognized as the number "2," resulting in higher confidence levels for character string recognition results compared to when no entry rule is set.

[0012] However, it is sometimes difficult to predict what character strings users will enter depending on the item. In such cases, document designers may not be able to fully determine the entry rules that should be set for the document items, and therefore may not set entry rules. This lack of entry rules for document items may also reduce the confidence level of character string recognition results obtained through OCR processing.

[0013] An object of the present invention is to provide an information processing device and an information processing program that can support a form designer in setting character string recording rules for form items even if the form designer cannot predict what character string will be recorded in the form item.

[0014] [Technical means to solve the problem]

[0015] The information processing device of the first mode includes a processor, which, when regularity related to the recording of a character string as a confirmation result of confirming the text recognition result of the bill is extracted in the items of the bill, outputs the recording rules of the extracted character string for each item of the bill.

[0016] The information processing device of the second mode is an information processing device of the first mode, wherein the processor outputs the recording rule of the character string and the degree of change in the number of corrected character strings that changes according to the presence or absence of the setting of the recording rule of the character string, and the corrected character strings are corrected due to misrecognition in text recognition.

[0017] The information processing apparatus according to a third aspect is the information processing apparatus according to the second aspect, wherein the processor outputs a degree of change in the number of corrected character strings that decreases due to setting a description rule for the outputted character strings for the items of the receipt.

[0018] The information processing device of the fourth aspect is the information processing device of the second aspect, wherein the processor outputs, as the degree of change, the degree of the number of the corrected character strings, the corrected character strings being corrected because a description rule for the outputted character strings is not set for the items of the receipt.

[0019] The information processing apparatus of a fifth aspect is the information processing apparatus of any one of the first to fourth aspects, wherein the processor outputs a description rule of the character string for the classification attribute such as the extracted regularity related to the description of the character string.

[0020] The information processing device of the sixth mode is the information processing device of the fifth mode, wherein the processor outputs the recording rule of the character string extracted from the character string of the confirmation result, and the recording rule of the character string is for the classification attribute such that significant differences can be seen in the recording rules of multiple character strings.

[0021] The information processing device of the seventh mode is an information processing device of any one of the first to sixth modes, wherein the processor determines whether to extract regularity related to the recording of the character strings from the character strings of the confirmation results based on the number of character strings of the confirmation results collected for the items of the bill.

[0022] The information processing device of the eighth mode is an information processing device of the seventh mode, wherein when the number of character strings of the confirmation results collected for the items of the bill is greater than a number predetermined as the number of the regularity extracted, the processor outputs a recording rule for the character strings of the items for which the number of character strings of the confirmation results is greater than the predetermined number.

[0023] The information processing device of the ninth mode is an information processing device of the seventh mode, wherein when the number of character strings of the confirmation results collected for the items of the bill is less than a number predetermined as the number of the regularities extracted, the processor does not output the recording rules for the character strings of the items for which the number of character strings of the confirmation results is less than the predetermined number.

[0024] The information processing device of the tenth mode is an information processing device of any one of the first to ninth modes, wherein the processor outputs a change notification urging a change in the recording rules of the character string set in the item of the bill, corresponding to the degree of correction of the character string recorded in the item of the bill.

[0025] The information processing apparatus of an eleventh aspect is the information processing apparatus of the tenth aspect, wherein the processor outputs the change notification when the degree of correction in the item of the receipt increases by a predetermined degree or more from a reference degree.

[0026] The information processing device of the twelfth mode is an information processing device of the tenth mode, wherein the processor outputs the change notification when the degree of correction in the item of the bill after setting the recording rule of the character string is included in a predetermined range from the degree of correction in the same item of the bill before setting the recording rule of the character string.

[0027] The recording medium of the thirteenth mode stores an information processing program, which is used to enable a computer to perform the following processing: when regularity related to the recording of a character string as a confirmation result of confirming the text recognition result of the bill is extracted in the items of the bill, the recording rules of the extracted character string are output for each item of the bill.

[0028] [Effects of the Invention]

[0029] According to the first and thirteenth aspects, there is an effect of being able to provide support in such a manner that, even for a document item in which a document designer cannot predict what kind of character string will be recorded, the document designer can set a character string recording rule for the document item.

[0030] The second aspect provides support in that, even when the document designer does not know which character string recording rule to select, the document designer can select a character string recording rule based on the degree of change in the number of corrected character strings due to setting.

[0031] According to the third aspect, there is an effect that a document designer can select a character string description rule by referring to the actual performance of the number of corrected character strings reduced by setting.

[0032] According to the fourth aspect, there is an effect that a document designer can select a character string description rule by referring to the actual number of character string corrections that occurred due to the lack of settings.

[0033] According to the fifth aspect, there is an effect that, when description rules exist for a plurality of classification attributes, description rules defined from the perspective of different classification attributes can be output for the same character string.

[0034] According to the sixth aspect, there is an effect of being able to output only description rules that are effective for improving the degree of certainty in the recognition result of a character string.

[0035] According to the seventh aspect, there is an effect that the reliability of the extracted description rules can be guaranteed from the aspect of the number of collected character strings.

[0036] According to the eighth method, the following effect is achieved: compared with the case where the recording rules for character strings extracted under the following conditions are set, the number of character strings collected for the items of the bill does not reach a number that only statistically guarantees the reliability of the recording rules, which can improve the confidence in the recognition results of the character strings.

[0037] According to the ninth aspect, there is an effect of preventing a document designer from selecting a character string description rule whose reliability is not statistically guaranteed.

[0038] According to the tenth aspect, there is an effect that a document designer can be notified of a change in the description tendency of the items in the document.

[0039] According to the eleventh aspect, there is an effect of being able to notify the form designer that the degree of correction of the character string has increased.

[0040] According to the twelfth aspect, there is an effect of being able to inform the document designer that the set character string description rule does not contribute to improving the confidence level in the character string recognition result. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a block diagram showing an example of the functional configuration of an information processing device.

[0042] Figure 2 This is a diagram showing an example of a confirmation and correction table.

[0043] Figure 3 This is a diagram showing an example of a stored number table.

[0044] Figure 4 This is a diagram showing an example of a pattern table.

[0045] Figure 5 This is a diagram showing an example of the configuration of main parts of an electrical system in an information processing device.

[0046] Figure 6 This is a flowchart showing an example of extraction processing.

[0047] Figure 7 This is a flowchart showing an example of output processing.

[0048] Figure 8 It is a figure which shows the example of the screen displayed on the display part.

[0049] Figure 9 It is a diagram showing another example of a screen displayed on the display unit.

[0050] Figure 10 It is a diagram showing another example of a screen displayed on the display unit.

[0051] Figure 11 This is a flowchart showing a modified example of the extraction process.

[0052] Figure 12 This is a flowchart showing an example of change notification processing.

[0053] [Explanation of Symbols]

[0054] 2: Confirm the correction form

[0055] 4: Save the number of pieces table

[0056] 6: Mode Table

[0057] 7: Select the notification area

[0058] 8: Dialog Box

[0059] 9: Checkbox

[0060] 10: Information processing device

[0061] 11: Reading unit

[0062] 12: OCR recognition unit

[0063] 13: Confirmation and Correction Department

[0064] 14: Pattern extraction unit

[0065] 15: Output

[0066] 16: Correction information DB

[0067] 20: Computer

[0068] 21: CPU

[0069] 22: ROM

[0070] 23: RAM

[0071] 24: Non-volatile memory

[0072] 25: I / O

[0073] 26: Bus

[0074] 27: Communication unit

[0075] 28: Input unit

[0076] 29: Display unit

[0077] 30: Scanning unit DETAILED DESCRIPTION

[0078] Below, with reference to the attached Figure 1 In all the drawings, the same components and the same processes are denoted by the same reference numerals, and repeated descriptions are omitted.

[0079] Figure 1 This is a block diagram showing an example of the functional structure of an information processing device 10, which confirms and corrects the recognition result of a character string read from an image of a bill generated by optically reading the contents of the bill, stores it in a storage device, and extracts the recording pattern of the character string from the confirmation and correction result of the stored character string and outputs it.

[0080] The so-called "receipt" refers to a document that records information about a specific matter in a predetermined format, such as a record column containing the content of each item recorded by the recorder. The so-called "item" refers to an attribute that represents the content filled in the record column, such as the recorder's address or name, and the item is identified by the title recorded in each record column. The character string recorded in the record column can be handwritten or printed using a printer, etc. Moreover, there is no limit to the type of receipt processed by the information processing device 10. For example, as long as there is a record column for each item, such as an application form, a contract, and a medical questionnaire, the recorder can record the content corresponding to the item.

[0081] Hereinafter, the character string in the entry column of an item that the recorder has entered on the receipt may be referred to as a “character string corresponding to the item.” The term “character string” refers to a sequence of one or more characters.

[0082] like Figure 1 As shown, the information processing device 10 includes functional units such as a reading unit 11 , an OCR recognition unit 12 , a confirmation and correction unit 13 , a pattern extraction unit 14 , and an output unit 15 , and a correction information database (DB) 16 .

[0083] The reading unit 11 optically reads the contents of the receipt written by the writer, for example, using the scanner unit 30, and generates an image of the receipt. The reading unit 11 notifies the OCR recognition unit 12 of the generated receipt image.

[0084] The OCR recognition unit 12 performs OCR processing on the image of the received bill and notifies the recognition result of the character string obtained by the OCR processing, that is, the text recognition result, to the confirmation and correction unit 13. In addition, the OCR recognition unit 12 associates each recognized character string with a confidence level and notifies the correction unit 13.

[0085] Here, the confidence level of a recognized character string refers to a value indicating the degree of character string recognition accuracy, such as whether the character string contained in the image of the receipt was correctly recognized as it appears on the receipt. For example, a confidence level of 100% indicates that the character string was recognized as it appears on the receipt, while a confidence level of 50% indicates that there is a probability that the recognized character string is different from the one on the receipt at least once.

[0086] For example, if the number "2" is written in a receipt image, the OCR recognition unit 12 outputs the character string that most closely resembles the character string from the characters registered in the dictionary as the character recognition result. However, if the number "2" is handwritten and written in a shape that can also be read as the English letter "Z," the OCR recognition unit 12 may mistakenly output the English letter "Z" as the character recognition result for the number "2." In other words, as the number of character strings similar to the character string being recognized increases, the probability of misrecognizing a character string increases, resulting in a lower confidence level.

[0087] In this way, there is a situation where the character string recognized by the OCR recognition unit 12 is recognized as a character string different from the character string recorded by the recorder in the bill. Therefore, the confirmer compares the bill with the text recognition result of the OCR recognition unit 12 while referring to the confidence level to confirm whether the character string is correctly recognized. If it is not correctly recognized, it is corrected.

[0088] Upon receiving an instruction from the confirmer and corrector that a character string needs to be corrected, the confirmation and correction unit 13 corrects the character string recognized by the OCR recognition unit 12 to the character string specified by the confirmer and corrector. Furthermore, upon receiving an instruction from the confirmer and corrector that a character string does not need to be corrected, the confirmation and correction unit 13 does not correct the character string recognized by the OCR recognition unit 12. The confirmation and correction unit 13 registers the confirmation results of the character strings recognized by the OCR recognition unit 12 for each item on the document in the correction information DB 16 and manages these results using the confirmation and correction table 2. The confirmer and corrector may be the same person as the document designer or another person.

[0089] Figure 2 2 is a diagram showing an example of the confirmation correction table 2. The confirmation correction table 2 is a table including a form name, an item name, a confirmation correction result, a confirmed character string before correction, and whether or not correction has been made.

[0090] In the document name column, the name of the document that is the target of character string confirmation in the confirmation and correction unit 13 is set.

[0091] In the item name column, the title of the item included in the receipt that is the target of the character string confirmation in the confirmation and correction unit 13 is set.

[0092] The confirmed character string confirmed by the confirmation and correction unit 13 is set in the confirmation and correction result column. If the character string has been corrected as a result of the confirmation, the corrected character string is set in the confirmation and correction result column. Furthermore, the confirmed character string confirmed by the confirmation and correction unit 13 is sometimes referred to as a "confirmed character string." The confirmed character string is an example of a character string of the confirmation result in this embodiment. Furthermore, a character string corrected by the confirmation and correction operator within the confirmed character string is sometimes referred to as a "corrected character string."

[0093] In the column for the character string before confirmation and correction, the character string before confirmation, that is, the character string itself recognized by the OCR recognition unit 12 is set.

[0094] In the correction presence / absence column, information indicating whether the character string has been corrected by the confirmation and correction unit 13 is set. For example, if correction has been performed, "yes" is set, and if correction has not been performed, "no" is set.

[0095] Thus, in Confirmation and Correction Table 2, for each item on the document, the character string before confirmation is associated with the character string after confirmation. The collection of information in each column associated with the row direction of Confirmation and Correction Table 2 is referred to as "Confirmation and Correction Information." Furthermore, for confirmation and correction information with "No" set in the Correction Result column, the same character string is set in the Character String Before Confirmation Correction column.

[0096] The confirmation and correction unit 13 then totals the number of confirmation and correction information items registered in the confirmation and correction table 2 for each item of the slip, and manages the total number of items using the stored number table 4 stored in the correction information DB 16 .

[0097] Figure 3 4 is a diagram showing an example of the stored number table 4. The stored number table 4 is a table including receipt names, item names, and the number of stored receipts.

[0098] In the slip name column and the item name column, a slip name and an item name in which the number of confirmed correction information items is totaled are set, respectively.

[0099] In the "Number of Conserved Items" column, enter the number of confirmation and correction information items registered in Confirmation and Correction Table 2 that correspond to the items in the receipt indicated by the contents set in the "Voucher Name" and "Item Name" columns in the same row. The number set in the "Number of Conserved Items" column corresponds to the number of confirmed character strings collected for each receipt item.

[0100] exist Figure 3 In the case of the stored item count table 4 shown in , for example, information processing device 10 stores 100 pieces of confirmation and correction information for the character strings listed in the remarks of the purchase application in confirmation and correction table 2. Thus, stored item count table 4 stores the number of string confirmation results for each item on the receipt.

[0101] The pattern extraction unit 14 refers to the confirmed correction table 2 and the stored number table 4 stored in the correction information DB 16 to extract a character string description rule, that is, a character string description pattern, for each item of each receipt.

[0102] A string entry pattern refers to a common regularity in a string of characters that is recognized across multiple documents. Writers may not record strings in document items according to a predetermined entry pattern. Because the content of entries is limited by item, multiple writers may unintentionally record the same string. The pattern extraction unit 14 identifies potential regularities in the string of characters that appear within the content of the confirmed items and extracts them as string entry patterns.

[0103] The pattern extraction unit 14 registers the description pattern of the extracted character string in the correction information DB 16 and manages it using the pattern table 6 .

[0104] Figure 4 : is a diagram showing an example of the pattern table 6. The pattern table 6 is a table including a document name, an item name, a description pattern, and a similarity ratio.

[0105] In the slip name column and the item name column, the slip name and the item name of the description pattern of the extracted character string are set respectively.

[0106] In the description pattern column, a description pattern extracted from the items of the slip indicated by the contents set in the slip name column and the item name column in the same row is set.

[0107] In the similarity column, a value is set that indicates the degree to which a character string according to a description pattern included in the same row appears in the same item of the same form.

[0108] exist Figure 4 In the case of the pattern table 6 shown in , for example, it indicates that in the remarks of the purchase application, the record pattern "The latter is consistent, transfer to □□□" appears with a similarity rate of 50%. In addition, the expression "□" in the record pattern indicates that an arbitrary character is filled in. Moreover, the so-called rear consistency indicates a record pattern of a character string that is consistent with the specified character string (in the case of the example, "transfer to") when the character string is viewed from the end to the beginning. On the contrary, the so-called front consistency indicates a record pattern of a character string that is consistent with the specified character string when the character string is viewed from the beginning to the end. In addition, the record pattern of the character string is set in a regular expression in the record pattern column, but in Figure 4 In order to facilitate understanding of the description, an example of expressing the contents of a formal expression using an article is shown.

[0109] A specific method of extracting a description pattern of a character string in the pattern extraction unit 14 will be described in detail later.

[0110] The output unit 15 outputs the form designated by the form designer to the display unit 29, etc. When the form designer selects certain items for the output form, the output unit 15 refers to the pattern table 6 stored in the correction information DB 16 and outputs the description pattern of the character string corresponding to the selected item.

[0111] When the document designer selects at least one character string pattern from the input character string pattern, the OCR recognition unit 12 assigns the character string pattern selected by the document designer to the selected document item. Subsequently, when performing OCR processing on an image of a received document, the OCR recognition unit 12 recognizes the character string by referring to the character string pattern assigned to the document item.

[0112] Next, a configuration example of a main portion of the electrical system in the information processing device 10 will be described.

[0113] Figure 5 1 is a diagram showing a configuration example of a main part of an electrical system in the information processing device 10. The information processing device 10 is configured using a computer 20, for example.

[0114] The computer 20 includes a central processing unit (CPU) 21 as an example of a processor that serves as each functional unit of the information processing device 10, a memory that allows the computer 20 to function as Figure 1 The CPU 21 includes a read-only memory (ROM) 22 for an information processing program that enables the functional units shown in FIG2 to function, a random access memory (RAM) 23 used as a temporary work area for the CPU 21, a nonvolatile memory 24, and an input / output (I / O) interface 25. The CPU 21, ROM 22, RAM 23, nonvolatile memory 24, and I / O 25 are connected via a bus 26.

[0115] Nonvolatile memory 24 is an example of a storage device that maintains stored information even when power to nonvolatile memory 24 is cut off. For example, a semiconductor memory may be used, but a hard disk may also be used. Nonvolatile memory 24 does not necessarily need to be built into computer 20. For example, it may be a removable storage device such as a memory card that can be installed in and removed from computer 20.

[0116] The I / O 25 is connected to, for example, the communication unit 27 , the input unit 28 , the display unit 29 , and the scanner unit 30 .

[0117] The communication unit 27 is connected to a communication line (not shown) and includes a communication protocol for performing data communication with an external device connected to the communication line (not shown).

[0118] The input unit 28 receives instructions from the reviewer and the document designer and notifies the CPU 21. For example, buttons, a touch panel, a keyboard, and a mouse can be used. When instructions are given by voice, a microphone may be used as the input unit 28.

[0119] The display unit 29 is a device that outputs information processed by the CPU 21 , and for example, a liquid crystal display or an organic electroluminescence (EL) display can be used.

[0120] The scanning unit 30 optically reads a receipt on which a recorder has written information and generates an image of the receipt. The scanning unit 30 is not necessarily required in the information processing device 10. The information processing device 10 may also obtain an image of a receipt read by a scanning device connected to a communication line (not shown) via the communication unit 27.

[0121] The units connected to I / O 25 are not limited to Figure 5Each unit shown in , for example, may also be connected to other units such as an image forming unit that forms an image on a recording medium. Furthermore, for example, a semiconductor memory such as a memory card or a Universal Serial Bus (USB) memory may be used to obtain the image of the bill.

[0122] Next, the operation of the information processing device 10 for extracting the description pattern of the character string based on the confirmation correction table 2 will be described.

[0123] Figure 6 This flowchart illustrates an example of the extraction process executed by CPU 21 of information processing device 10 when extracting a character string pattern from an item on a receipt. An information processing program that specifies the extraction process is pre-stored in ROM 22 of information processing device 10, for example. CPU 21 of information processing device 10 reads the information processing program stored in ROM 22 and executes the extraction process.

[0124] In addition, there is no restriction on the execution time of the extraction process, and the CPU 21 can execute the extraction process at any time. For example, the CPU 21 can also execute the extraction process every time the image of the bill is OCR-processed. Here, as an example, the extraction process is executed every predetermined period, such as one month. Figure 6 Before the extraction process shown in , all pattern information is deleted from the pattern table 6.

[0125] Figure 6 The extraction process shown in FIG is an example of extracting a character string pattern for any item of a receipt. Figure 6 By performing the extraction process shown in , the description pattern of the character string is extracted for each item of all the receipts on which the OCR process has been performed.

[0126] In step S10 , the CPU 21 acquires all confirmation and correction information for items of selected slips (hereinafter referred to as “selected items”) from the confirmation and correction table 2 .

[0127] In step S20, CPU 21 extracts confirmed character strings from the confirmation and correction result column of each confirmation and correction information acquired in step S10 and classifies each confirmed character string using character codes. CPU 21 then groups the classified confirmed character strings based on the viewpoints of front-end consistency and back-end consistency.

[0128] Specifically, CPU 21 sequentially checks the classified confirmed character strings from the front to the back, groups confirmed character strings having the same number of consecutive characters from the front into the same group, and totals the number of confirmed character strings included in each group.

[0129] Next, the CPU 21 sequentially checks the classified confirmed character strings from the end to the beginning, groups confirmed character strings having the same number of consecutive characters from the end into the same group, and totals the number of confirmed character strings included in each group.

[0130] In step S30, the CPU 21 selects an unselected group that has not yet been selected from the groups generated in step S20. The group selected in step S30 is referred to as a "selected group."

[0131] In step S40 , the CPU 21 extracts a description pattern of the character strings based on the matching status of the character strings in the selected group.

[0132] For example, if the selected group is a string of characters with three identical characters starting from the beginning, and the matching characters are "AAA", the entry pattern of the string represented by a regular expression such as "^A{3}" is extracted. Furthermore, if the selected group is a string of characters with four identical characters starting from the end, and the matching characters are "Deee", the entry pattern of the string represented by a regular expression such as "De{3}$" is extracted.

[0133] Then, the CPU 21 calculates the number of confirmed character strings included in the selected group relative to the number of confirmed correction information acquired in step S10 as a similarity.

[0134] In step S50 , the CPU 21 registers pattern information in the pattern table 6 , in which the invoice name and item name from which the character string description pattern is to be extracted are associated with the character string description pattern extracted in step S40 and the calculated similarity.

[0135] In step S60, CPU 21 determines whether any unselected groups exist among the groups compiled in step S20 that were not selected in step S30. If any unselected groups exist, the process jumps to step S30, where any group is selected from the unselected groups. The process of steps S30 through S60 is repeated until no unselected groups exist, thereby setting a multiple character string recording mode for the selected item.

[0136] On the other hand, when it is determined in the determination process of step S60 that there is no unselected group, the process ends. Figure 6 extraction processing.

[0137] exist Figure 6In step S10, the character string record pattern is extracted based on the matching status of the confirmed character string. However, the viewpoint for extracting the character string record pattern is not limited to the matching status of the confirmed character string. CPU 21 refers to all the confirmed correction information obtained in step S10, analyzes the characteristics of the confirmed character string from the perspective of various classification attributes, and determines whether the character string record pattern can be found.

[0138] The classification attribute refers to a category that is focused on in order to extract the description pattern of a character string from a confirmed character string. In addition to the matching status of the confirmed character string, the appearance status of character types is an example of the classification attribute.

[0139] The term "character type" refers to the representation format of characters used in the confirmed character string, and includes, for example, numbers, uppercase letters, lowercase letters, hiragana, and katakana. In particular, when the confirmed character string is printed by a printer, numbers, uppercase letters, lowercase letters, and katakana may be full-width or half-width.

[0140] When extracting the record pattern of the character string by focusing on the appearance of the character type, the CPU 21 only needs to Figure 6 In step S20 , confirmed character strings are extracted from the respective confirmation correction information acquired in step S10 , and confirmed character strings having the same appearance of character types in the confirmed character strings are grouped together.

[0141] Specifically, CPU 21 checks confirmed character strings from the front to the back, groups the same number of confirmed character strings of the same character type that are continuous from the front into the same group, and counts the number of confirmed character strings included in each group.

[0142] Next, CPU 21 checks the confirmed character strings in order from the end to the beginning, groups the same number of confirmed character strings of the same character type that are continuous from the end into the same group, and counts the number of confirmed character strings included in each group.

[0143] Afterwards, in Figure 6 In step S40 , the CPU 21 extracts a description pattern of a character string based on the appearance of the character types in the selection group.

[0144] For example, if the selected group is a confirmed string of characters whose first three characters are of the same character type, and the same character type is half-width uppercase, the entry pattern of the string represented by a regular expression such as "^[AZ]{3}" is extracted. Furthermore, if the selected group is a confirmed string of characters whose first five characters are of the same character type, and the characters from the first to the third character are half-width uppercase, and the fourth and fifth characters are half-width lowercase, the entry pattern of the string represented by a regular expression such as "^[AZ]{3}[az]{2}" is extracted.

[0145] Therefore, in Figure 6 In step S50, the CPU 21 only needs to register pattern information in the pattern table 6, wherein the pattern information establishes a correspondence between the invoice name and item name to be extracted from the character string recording pattern, the extracted character string recording pattern, and the calculated similarity.

[0146] For a specific document item, if the similarity rates among all extracted character string patterns are similar, any character string pattern will appear in the document item with the same probability. In this case, it is difficult to say that the extracted character string pattern is a representative confirmed character string pattern for the document item in question.

[0147] Therefore, CPU 21 may register only the character string record patterns in the classification attribute that show significant differences in the extracted character string record patterns in pattern table 6. Here, "significant differences in the character string record patterns" refers to a situation where the difference in similarity between character string record patterns is greater than a predetermined threshold, indicating that the record pattern is a characteristic character string that is more likely to be used by the recorder than other character string record patterns. Furthermore, "approximate similarity" in character string record patterns refers to a situation where the difference in similarity between character string record patterns is less than the predetermined threshold.

[0148] Moreover, CPU21 is Figure 6 When the recording mode of the character string is registered in the pattern table 6 in step S50, the degree of change in the number of corrected character strings that has changed due to setting the recording mode of the character string in the item of the bill can also be registered in the pattern table 6, and the corrected character string has been corrected by the confirmer due to misrecognition in the text recognition in the OCR processing.

[0149] Specifically, for each character string recording pattern registered in pattern table 6, CPU 21 sets the character string recording pattern in the invoice item. Then, CPU 21 registers the number of character strings that have been misrecognized in character recognition during OCR processing and for which it has been determined that the corrector does not need to correct the character strings, in pattern table 6. Thus, the number of corrected character strings that has decreased due to setting the character string recording pattern in the invoice item is registered in pattern table 6.

[0150] In this case, the number of corrected character strings that are corrected because the character string description pattern is not set in the item of the receipt for each character string description pattern registered in the pattern table 6 is registered in the pattern table 6 .

[0151] If the recording pattern of the character string is set in the item of the bill, the number of characters that can be corrected without the correction of the character string by the corrector is confirmed, that is, the number of characters that need to be corrected because the recording pattern of the character string is not set in the item of the bill is represented by the number of corrected characters in the group from which the recording pattern of the character string has been extracted.

[0152] Furthermore, in the above description, the number of corrected character strings in each item of the receipt that changes depending on whether a character string entry pattern is set is registered in the pattern table 6. However, the ratio of corrected character strings that change may also be registered. The ratio of corrected character strings that change may be represented, for example, by the ratio of the number of corrected character strings to the number of confirmed character strings included in the group from which the character string entry pattern has been extracted.

[0153] exist Figure 6 In the extraction process shown in , for each item of the bill, all confirmation correction information registered in the confirmation correction table 2 corresponding to the item is used to extract the record pattern of the character string. However, for example, if the extraction process is performed every predetermined period (for example, one month), Figure 6 In the case of the extraction process shown in , CPU 21 may also acquire only the confirmed correction information registered in confirmation correction table 2 during a predetermined period, and acquire the character string record pattern, similarity, and the number or ratio of corrected character strings that vary depending on whether the character string record pattern is set at each predetermined period. In this case, information indicating the period during which the character string record pattern was extracted is also included in the pattern information and managed using pattern table 6.

[0154] In addition, when extracting a character string record pattern at predetermined intervals, if Figure 6 If the pattern information is not deleted from the pattern table 6 before the extraction process shown in , the transition of the change of the pattern information in each period can be obtained.

[0155] Figure 7 This flowchart illustrates an example of output processing performed by CPU 21 of information processing device 10 when a document designer selects a document item displayed on a screen using a mouse or other device to set a character string entry mode for the document item. An information processing program defining the output processing is pre-stored in ROM 22 of information processing device 10, for example. CPU 21 of information processing device 10 reads the information processing program stored in ROM 22 and executes the output processing.

[0156] In addition, in the pattern table 6, there are registered Figure 6 The extraction process shown in extracts pattern information of the description pattern of the character string.

[0157] on the other hand, Figure 8 Yes means through Figure 7 FIG. 1 is a diagram showing an example of a screen displayed on the display unit 29 by the output processing shown in FIG. Figure 8 , while Figure 7 Description of the output processing shown in .

[0158] In step S100 , the CPU 21 acquires the character string description pattern corresponding to the item of the form selected by the form designer, ie, the selected item, from the pattern table 6 , and displays the acquired character string description pattern on the screen of the display unit 29 .

[0159] Figure 8 The example shown in Figure 1 shows a situation where the document designer has selected the "Remarks" column for a purchase application. In this case, CPU 21 retrieves pattern information from pattern table 6, where the document name is "Purchase Application" and the item name is set to "Remarks," and displays dialog box 8 on the screen, showing the entry patterns and similarity scores of the character strings contained in the pattern information. If multiple matching pattern information exists, CPU 21 displays the entry patterns and similarity scores of the character strings contained in each matching pattern information in dialog box 8. CPU 21 can display the entry patterns of the character strings using regular expressions, or it can convert the meaning of the regular expressions into text for display. Figure 8 The "(blank)" in the dialog box 8 is an example of a regular expression "￥s" which is a description pattern of a character string expressed in a sentence.

[0160] When displaying the character string entry patterns in dialog box 8, CPU 21 may refer to the similarity and rearrange the character string entry patterns into descending order of similarity or ascending order of similarity, and then display them in dialog box 8. Furthermore, CPU 21 may refer to stored number table 4 and display the number of confirmed character strings collected so far for the selected item in dialog box 8. Furthermore, CPU 21 may also display the number of confirmed character strings collected so far within a predetermined period (e.g., the most recent month) among the confirmed character strings collected so far. For example, CPU 21 may include the date and time information of the confirmation and correction of the text recognition results obtained by the OCR process by the confirmation and correction personnel in the confirmation and correction information and manage this information using confirmation and correction table 2. Alternatively, CPU 21 may aggregate the number of confirmed character strings collected for each item of the receipt at predetermined intervals and manage this information using stored number table 4.

[0161] The document designer selects the desired character string recording mode from the character string recording modes displayed in dialog box 8 and confirms the selection by pressing an OK button (not shown). Dialog box 8 includes a check box 9 for selecting a character string recording mode. For example, the check box 9 corresponding to the selected character string recording mode is filled in black.

[0162] The CPU 21 displays the selected character string description pattern in, for example, the selection notification area 7 provided in the dialog box 8. When multiple character string description patterns are selected, the CPU 21 represents the combination of the selected character string description patterns using a regular expression and displays it in the selection notification area 7. Figure 8 In the example, since "transfer to the Human Resources Department", "transfer to the General Affairs Department", and "(blank)" are selected, a regular expression such as "transfer to the Human Resources Department|transfer to the General Affairs Department|￥s" is displayed in the selection notification area 7.

[0163] In step S110, CPU 21 determines whether the document designer has selected a character string recording mode. If no character string recording mode has been selected, the CPU 21 repeats the determination process of step S110 to monitor the document designer's selection of a character string recording mode. On the other hand, if at least one character string recording mode has been selected, the CPU 21 proceeds to step S120.

[0164] In step S120, the CPU 21 sets the recording mode of the selected character string in the selection item. Figure 7 The output processing shown in .

[0165] In addition, in the dialog box 8, various displays can be performed in accordance with the description pattern of the character string corresponding to the selected item.

[0166] For example, Figure 9 As shown, the record mode of the string can also be divided into the record mode that is consistent at the beginning and the record mode that is consistent at the end to display, such as Figure 10 As shown, when there is a character string description pattern extracted based on the appearance of the character type, after displaying "character type", the meaning represented by the regular expression corresponding to the character string description pattern may be displayed using text.

[0167] Furthermore, if a character string record pattern with a similarity exceeding the baseline exists, CPU 21 may change the representation of the character string record pattern with a similarity exceeding the baseline to that of another character string record pattern when displayed in dialog box 8 in order to distinguish the character string record pattern with a similarity exceeding the baseline from the other character string record patterns. Specifically, CPU 21 changes at least one of the following: the text color, the background color, the text size, and the font.

[0168] Furthermore, CPU21 can also display other information registered in the pattern table 6 for the recording mode of each character string, such as the number of characters that can be corrected if the recording mode of the character string is set in the item of the bill. In other words, the number of corrected characters that need to be corrected because the recording mode of the character string is not set in the item of the bill.

[0169] In this way, according to the information processing device 10 of this embodiment, the recording pattern of the character string is extracted from the confirmed character string of each item of the bill confirmed by the confirmer, and when the bill designer wants to set the recording pattern of certain character strings in the items of the bill, the recording pattern of the character string corresponding to the item of the bill selected by the bill designer is output.

[0170] Therefore, the document designer can save the trouble of considering the recording pattern of the character strings to be set in the document items. Furthermore, since the information processing device 10 generates the recording pattern of the character strings in the form of regular expressions, even if the document designer does not understand regular expressions, for example, by viewing an article explaining the content of regular expressions displayed in dialog box 8 and selecting the recording pattern of the character strings to be set in the document items, the regular expression corresponding to the selected content will be set in the document items.

[0171] Furthermore, there are cases where a document designer, judging from the content of an item, assumes that a character string pattern does not exist in the content recorded by the recorder and, therefore, insists on suggesting a character string pattern for an item for which a character string pattern is not set. Consequently, a character string pattern may be set for a document item. Furthermore, there are cases where the document designer is prompted by the information processing device 10 for a character string pattern that the designer has not yet discovered. If the suggested character string pattern is one that is expected to increase the confidence level of the character string recognized by OCR processing compared to the character string pattern already set for the document item, the document designer can save the effort of researching a character string pattern that has the effect of increasing confidence level.

[0172] <Variation 1>

[0173] exist Figure 6 In the extraction process shown in , regardless of the number of confirmed character strings collected for a document item, a character string record pattern is extracted from the collected confirmed character strings. However, if, for example, the number of confirmed character strings collected for a document item from which the character string record pattern is to be extracted is only one, it is impossible to determine whether the character string record pattern extracted from the confirmed character strings represents the record pattern of the document item from which the character string record pattern is to be extracted.

[0174] Therefore, in this modification, an information processing device 10 is described that determines whether or not character string description pattern extraction is possible based on the number of confirmed character strings collected for items of a receipt from which character string description pattern extraction is to be performed.

[0175] Figure 11 This is a flowchart showing a modified example of the extraction process executed by the CPU 21 of the information processing device 10 when extracting a description pattern of a character string described in an item of a receipt. Figure 11 The extraction process shown in Figure 6 The difference between the extraction process shown in FIG and FIG is that steps S2 and S4 are added. The other processes are the same as those in FIG. Figure 6 Therefore, the following description will focus on the processing of step S2 and step S4.

[0176] In step S2 , the CPU 21 refers to the stored number table 4 and acquires the stored number of confirmed character strings corresponding to the selected item.

[0177] In step S4, the CPU 21 determines whether the number of stored items acquired in step S10 is a predetermined reference number N. A The so-called "reference number N" A"" refers to the minimum number of confirmed character strings required to guarantee the reliability of the recording pattern of the character string extracted from the confirmed character string, and is an example of a number determined in advance as the number of regularities of the confirmed character strings to be extracted. The reference number N A For example, it is set in advance from a statistical point of view and stored in the non-volatile memory 24. A Modified according to instructions from the bill designer, etc.

[0178] If the number of confirmed character strings for the selected item is N, A The above, the reliability of the recording mode of the character string extracted from the confirmed character string is guaranteed, so jump to step S10, execute Figure 6 The extraction process described in .

[0179] On the other hand, if it is determined in the determination process of step S4 that the number of confirmed character strings for the selected item is less than the reference number N, A When the reliability of the recording pattern of the character string extracted from the confirmed character string is in doubt, the recording pattern of the character string is not extracted, and the process ends. Figure 11 The extraction process shown in .

[0180] Of course, when extracting a character string record pattern from confirmed character strings collected at predetermined intervals, the reference number N is not the total number of stored confirmed character strings collected in each period, but the number of stored confirmed character strings collected in a single period. A In the above case, the description pattern of the character string is extracted.

[0181] <Variation 2>

[0182] Sometimes, even if a string entry pattern has already been set for a document item, it's necessary to review the set entry pattern. For example, if a document item is "Part Number," the part number is entered in the entry field. However, if the numbering system for the part number, which previously started with numerals, is changed to alphabetical, the string entry pattern set before the change becomes incompatible with the new part numbering system and needs to be reviewed. However, document designers are not necessarily informed of situations such as changes in the numbering system that affect string entry patterns. As a result, they may continue to use a string entry pattern that is inappropriate for the actual content being recorded.

[0183] Therefore, in this modification, information processing device 10 is described that detects a situation where it is desirable to change the description mode of a character string set in an item of a form and outputs a change notification urging the form designer to change the description mode of the character string.

[0184] Figure 12 1 is a flowchart showing an example of a change notification process executed by the CPU 21 of the information processing device 10. The CPU 21 can execute the change notification process at any time. Here, as an example, the CPU 21 executes the change notification process every predetermined period. Figure 6 or Figure 11 The extraction process shown in is executed, and the change notification process is executed in conjunction with the execution of the extraction process. For the sake of convenience, the period that will be the target of the change notification process is referred to as the "target period".

[0185] in addition, Figure 12 The change notification process shown in FIG is an example of determining the necessity of a change notification for any item of a bill. By executing the change notification process for each item of each bill, Figure 12 The necessity of change notification is determined for each item of all documents on which OCR processing has been performed, according to the change notification processing shown in .

[0186] In step S200, CPU 21 calculates the correction rate for the target period. The correction rate refers to the proportion of corrected character strings among the confirmed character strings collected during the target period, and is an example of the degree of correction. For example, if the predetermined period is one month, the correction rate for one month is calculated.

[0187] In step S210, CPU 21 determines whether the correction rate for the target period calculated in step S200 is higher than the correction rate calculated for a period earlier than the target period (referred to as the comparison period). To determine the necessity of a change notification based on recent changes in the correction rate as much as possible, the comparison period is preferably set to a period adjacent to the target period. For example, if the target period is August, the comparison period is set to July. If the correction rate for the target period is higher than the correction rate for the comparison period, the process proceeds to step S220.

[0188] In step S220, the CPU 21 calculates the rate of increase of the correction rate in the target period using the correction rate in the comparison period as a reference. That is, the correction rate in the comparison period is an example of a reference level.

[0189] In step S230, the CPU 21 determines whether the increase rate calculated in step S220 is the reference increase rate N. B The so-called "baseline rise rate N B", means the minimum rate of increase at which it is considered that it is best to consider reconsidering the recording mode of the character string set in the selection item if it becomes above this value. The standard rate of increase N B For example, it is stored in advance in the nonvolatile memory 24 and is modified according to instructions from a document designer or the like.

[0190] For example, if the content in the entry field of an item changes, such as when the part numbering system is changed, the character string entry pattern corresponding to the new content has not yet been set in the item, resulting in an increase in the correction rate compared to before the change. Therefore, by monitoring the rate of increase in the correction rate, it is possible to determine whether the character string entry pattern set for the selected item needs to be reviewed.

[0191] When the rising rate calculated in step S220 is the reference rising rate N B If above, jump to step S240.

[0192] In this case, since the rising rate becomes the reference rising rate N B Therefore, it is considered that it is best to explore the reconsideration of the recording mode of the character string set in the selection item. Therefore, in step S240, the CPU 21 outputs the change notification and ends Figure 12 The change notification process shown in . As long as the form designer can be notified of the change notification, there is no limitation on the method of outputting the change notification. Information prompting the change in the character string description format may be displayed on the screen of the display unit 29 or sent to an email address assigned to a mobile device such as a smartphone carried by the form designer.

[0193] On the other hand, if the correction rate of the target period is determined to be lower than the correction rate of the comparison period in the determination process of step S210, or if the increase rate of the correction rate in the target period is determined to be less than the reference increase rate N in the determination process of step S230, B When the change notification is not output, the end Figure 12 Change notification processing as shown in .

[0194] Furthermore, if the string entry pattern set by the document designer is an invalid string entry pattern that, even if set, does not affect the confidence level of the recognized string, there is no need to forcibly set the string entry pattern in the document entry. Furthermore, if such an invalid string entry pattern is set as is in the document entry, it may be unclear which string entry pattern is effective for improving the confidence level.

[0195] Therefore, the CPU 21 may compare the correction rates before and after the character string recording pattern is set in the receipt item. If the difference in correction rate is within a predetermined range, the CPU 21 may output a change notification to the receipt designer, urging the designer to delete the character string recording pattern that resulted in a correction rate change that was only within the predetermined range before and after the setting. In this case, the CPU 21 may also include the invalid character string recording pattern in the change notification and output it.

[0196] Thus, the information processing device 10 of this variation determines whether a change notification is necessary based on the degree of change in the correction rate and outputs a change notification as needed. This provides a document designer who may have been unaware of changes in the content of a document item with an opportunity to reconsider the character string recording pattern. The information processing device 10 also presents a character string recording pattern indicating the trend of the confirmed character string after the change in the recorded content. Therefore, the document designer simply selects the desired recording pattern from the presented character string recording patterns to complete the reconsideration of the character string recording pattern.

[0197] Furthermore, since invalid character string description patterns are also prompted, the form designer can organize character string description patterns already set in form items simply by deleting the suggested character string description patterns.

[0198] While the above embodiment illustrates an example in which the information processing device 10 prompts the document designer for a character string entry pattern, it is also possible to select an appropriate character string entry pattern from the character string entry patterns already extracted by the information processing device 10 and set it in the document entry. For example, an appropriate character string entry pattern may be selected, for example, to include a character string entry pattern whose similarity exceeds a baseline similarity, and a character string entry pattern for which no correction is required if set in the document entry. Furthermore, the information processing device 10 may autonomously reconsider the character string entry pattern without waiting for instructions from the document designer.

[0199] Moreover, as Figure 1As described in , this embodiment is described using an information processing device 10 as an example, including the functional units of the reading unit 11, the OCR recognition unit 12, the confirmation and correction unit 13, the pattern extraction unit 14, and the output unit 15, as well as the correction information DB 16. However, the processing of this embodiment can be implemented even when using an information processing device 10 that only includes the pattern extraction unit 14 and the output unit 15. Specifically, an external device is provided with the functional units of the reading unit 11, the OCR recognition unit 12, and the confirmation and correction unit 13, as well as the correction information DB 16, and communicates with the external device via the communication unit 27. The pattern extraction unit 14 only needs to refer to the confirmation and correction table 2 and the number of stored items table 4 contained in the correction information DB 16 included in the external device, and to set and refer to the pattern table 6.

[0200] While the present invention has been described above using the embodiments, the present invention is not limited to the scope described in the embodiments. Various modifications or improvements may be made to the embodiments without departing from the spirit of the present invention, and such modifications or improvements are also included within the technical scope of the present invention. For example, the order of processing may be changed without departing from the spirit of the present invention.

[0201] In this embodiment, as an example, the extraction process, output process, and change notification process are implemented by software. However, the same may also be applied to the process. Figure 6 、 Figure 7 、 Figure 11 ,and Figure 12 The equivalent processing in each flowchart shown in the figure can be implemented in hardware, such as an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a programmable logic device (PLD). In this case, the processing speed can be increased compared to when the confirmation and correction processing is implemented in software.

[0202] In this way, the CPU 21 may be replaced with a dedicated processor focused on specific processing, such as an ASIC, FPGA, PLD, Graphics Processing Unit (GPU), or Floating Point Unit (FPU).

[0203] The operation of the CPU 21 in the embodiment may be performed by a single CPU 21 or by a plurality of CPUs 21. Furthermore, the operation of the CPU 21 in the embodiment may be performed by cooperation of CPUs 21 in a plurality of computers 20 located in physically separate locations.

[0204] Moreover, in the above embodiment, the form of installing the information processing program in the ROM 22 is described, but it is not limited to this. The information processing program of the present invention can also be provided in the form of being recorded in a storage medium that can be read by a computer. For example, the information processing program of the present invention can also be provided in the form of being recorded in an optical disc such as a compact disc (CD)-ROM or a digital versatile disc (DVD)-ROM. Furthermore, the information processing program of the present invention can also be provided in the form of being recorded in a semiconductor memory.

[0205] Furthermore, the information processing device 10 may obtain the information processing program of the present invention from an external device via a communication line (not shown).

Claims

1. An information processing device comprising a processor, Extracting a character string record pattern related to a character string record of a confirmation result of each item of the bill based on a result of confirming a character recognition result of the bill; Based on the character string recording mode, outputting a recording rule of the extracted character string for each item of the bill; and When the degree of correction corresponding to the character string already recorded in the item of the form increases by a predetermined degree or more from a reference degree, a change notification urging a change in a recording rule of the character string already set in the item of the form is output.

2. An information processing device according to claim 1, wherein the processor outputs the recording rules of the character strings and the degree of change in the number of corrected character strings that changes according to the presence or absence of the setting of the recording rules of the character strings, and the corrected character strings are corrected due to misrecognition in text recognition.

3. The information processing device according to claim 2, wherein the processor outputs a degree of change in the number of corrected character strings that decreases due to setting a description rule for the output character strings for the items of the receipt.

4. The information processing device according to claim 2, wherein the processor outputs, as the degree of change, the degree of the number of the corrected character strings, the corrected character strings being corrected because a recording rule for the outputted character strings is not set for the items of the receipt. 5 . The information processing apparatus according to claim 1 , wherein the processor outputs a description rule of a character string for a classification attribute such as a regularity associated with description of the character string that is extracted.

6. The information processing device according to claim 5, wherein the processor outputs a recording rule of a character string extracted from the character string of the confirmation result, wherein the recording rule of the character string is for the classification attribute in which a significant difference can be seen in the recording rules of a plurality of character strings.

7. The information processing device according to any one of claims 1 to 4, wherein the processor determines whether to extract regularity related to the description of character strings from the character strings of the confirmation results based on the number of character strings of the confirmation results collected for the items of the bill.

8. An information processing device according to claim 7, wherein when the number of character strings of the confirmation results collected for the items of the bill is greater than a number predetermined as the number of the regularity extracted, the processor outputs the recording rule of the character strings of the items for which the number of character strings of the confirmation results is greater than the predetermined number.

9. An information processing device according to claim 7, wherein when the number of character strings of the confirmation results collected for the items of the bill is less than a number predetermined as the number of the regularities extracted, the processor does not output the recording rules for the character strings of the items for which the number of character strings of the confirmation results is less than the predetermined number.

10. An information processing device according to claim 1, wherein when the degree of correction in the item of the bill after setting the recording rule of the character string is included in a predetermined range from the degree of correction in the same item of the bill before setting the recording rule of the character string, the processor outputs the change notification.

11. A recording medium storing an information processing program for causing a computer to execute the following processing: Extracting a character string record pattern related to a character string record of a confirmation result of each item of the bill based on a result of confirming a character recognition result of the bill; Based on the character string recording mode, outputting a recording rule of the extracted character string for each item of the bill; and When the degree of correction corresponding to the character string already recorded in the item of the form increases by a predetermined degree or more from a reference degree, a change notification urging a change in a recording rule of the character string already set in the item of the form is output.

12. A computer program product for causing a computer to perform the following processing: Extracting a character string record pattern related to a character string record of a confirmation result of each item of the bill based on a result of confirming a character recognition result of the bill; Based on the character string recording mode, outputting a recording rule of the extracted character string for each item of the bill; and When the degree of correction corresponding to the character string already recorded in the item of the form increases by a predetermined degree or more from a reference degree, a change notification urging a change in a recording rule of the character string already set in the item of the form is output.

Images