Information processing apparatus, control method thereof, program, and information extraction method

Abstract

To properly extract item values to be extracted, in extracting necessary character information from a document image, regardless of positional relationships between item names and item values to be and not to be extracted.SOLUTION: An extraction process in an information processing apparatus is configured to: acquire character string information recognized from a document image by character recognition processing (S202); specify a region to be searched in the document image, based on an item to be extracted of an extraction rule and a positional relationship of an item value of the item; specify a region to be excluded, from the character string information included in the region and information on an item not to be extracted of an exclusion rule and an item value corresponding thereto and add a penalty to the region (S203); and extract an item value of the item to be extracted, on the basis of the extraction rule and the region to be excluded with the penalty added thereto (S204).SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an information processing apparatus that extracts necessary information from a document image, a control method thereof, a program, and an information extraction method. [Background technology]

[0002] Conventionally, there is known a method of searching for character strings corresponding to item names and item values ​​in a scanned image of a receipt or the like read by a scanner, and extracting information based on the relationship between them (Patent Documents 1 and 2). In the technology described in Patent Document 1, extraction is performed by adding a penalty to the distance, but there is a problem in that if an unnecessary item value exists near an item name, that item value is given priority. The technology described in Patent Document 2 has an issue in that when extraction rules are written to ignore unnecessary item values ​​and perform extraction, if a pair of item name and item value to be extracted is adjacent to a pair of item name and item value that is not to be extracted with no gaps between them, the item value not to be extracted will be erroneously extracted. In this way, when unnecessary field values ​​exist near field names, a process must be performed to ignore them and extract the field names. On the other hand, when field name-field value pairs are adjacent with no gaps, the adjacent field values ​​must be extracted. It was difficult to achieve both of these processes with conventional technology. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2011-248609 A [Patent Document 2] Patent Publication No. 2021-043478 Summary of the Invention [Problem to be solved by the invention]

[0004] In consideration of the above problems, the present invention aims to provide a mechanism for appropriately extracting item values ​​that are targets of extraction, regardless of the relative positions of the item names and item values ​​that are targets of extraction and those that are not targets of extraction. [Means for solving the problem]

[0005] The present invention is an information processing device comprising: an acquisition means for acquiring character string information recognized from a document image by character recognition processing; an assignment means for identifying an area to be searched in the document image based on items to be extracted defined as extraction rules and positional relationships related to the item values ​​of the items, and for identifying an exclusion target area based on character string information contained in the identified area to be searched and information related to items not to be extracted defined as exclusion rules and their corresponding item values, and for assigning a penalty to the identified exclusion target area; and an extraction means for extracting item values ​​of the items to be extracted defined by the extraction rule based on the extraction rule and the exclusion target area to which the penalty has been assigned. Effect of the Invention

[0006] According to the present invention, it is possible to appropriately extract the item values ​​of the extraction target regardless of the positional relationship in which the item names and item values ​​of the extraction target and non-extraction target are arranged. [Brief description of the drawings]

[0007] [Figure 1] 1 is a diagram illustrating an entire information processing system according to an embodiment. [Diagram 2] 11 is a flowchart of an extraction process in the information processing apparatus according to the embodiment. [Diagram 3] FIG. 13 is a diagram showing an example of a scanned image (receipt). [Figure 4] 11 is a flowchart illustrating an item removal process according to an embodiment. [Diagram 5] FIG. 13 is a diagram showing an example of extraction / exclusion rules for a receipt. [Figure 6] 11 is a flowchart illustrating an item extraction process according to an embodiment. [Figure 7] 13 is a flowchart showing a search process based on a positional relationship. [Figure 8] 13 is a flowchart showing an item removal process according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. However, the components described in the embodiment are merely examples, and are not intended to limit the scope of the present invention. [First embodiment] FIG. 1 is a diagram showing an example of the overall configuration of an information processing system to which an information processing device according to an embodiment of the present invention can be applied. (Image Processing Device) The image processing device 100 is a multifunction peripheral (MFP) having a scanner 101, a printer 102, and a communication unit 103. In the image processing device 100, a scanner 101 scans a document to generate a scanned image. The printer 102 forms an image on a sheet or the like. The communication unit 103 communicates with an external device (for example, the information processing device 110) via a network. (Information processing device) The information processing device 110 includes a CPU 111, a ROM 112, a RAM 113, a HDD 114, a display unit 115, an input unit 116, a communication unit 117, etc., which are communicatively connected to one another via a bus. The information processing device 110 can be configured, for example, by a personal computer (PC). The CPU 111 reads out a control program stored in the ROM 112 and executes various processes. The RAM 113 is used as a temporary storage area such as a main memory or a work area for the CPU 111 . The HDD 114 stores various data, various programs, etc. Note that other storage devices such as a Solid State Drive (SSD) may be provided instead of or in combination with the HDD (Hard Disk Drive). The functions and processes of the information processing device 110, which will be described later, are realized by the CPU 111 reading out a program stored in the ROM 112, the HDD 114, or the like, and executing this program. The communication unit 117 performs communication processing with an external device (eg, the image processing device 100) via a network. The display unit 115 displays various types of information. The input unit 116 has a keyboard and a mouse, and accepts various operations by the user. The display unit 115 and the input unit 116 may be integrated together, such as a touch panel. The display unit 115 may project an image using a projector, and the input unit 116 may recognize the position of a fingertip relative to the projected image using a camera. In this embodiment, the scanner 101 of the image processing device 100 scans a document such as a receipt and generates an input image (for example, an image such as the scanned image 301 in FIG. 3). The input image is transmitted to the information processing device 110 by the communication unit 103. In the information processing device 110, the input image transmitted from the image processing device 100 is received by the communication unit 117, and is stored in a storage unit such as the HDD 114 or the like.

[0009] Processing in information processing devices <Information extraction processing> Fig. 2 is a flowchart showing an example of information extraction processing according to this embodiment. Note that the processing of the flowcharts shown in Fig. 2 and Figs. 4, 6 and 7 described later is realized by the CPU 111 of the information processing device 110 reading and executing a program stored in the ROM 112 or HDD 114, etc. Hereinafter, each step of the processing is represented by "S". In S201, the CPU 111 of the information processing device 110 instructs the image processing device 100 to execute a document scan process using the scanner 101. Then, the CPU 111 acquires a scanned image read by the scan process via the communication unit 117 and stores the image in the HDD 114. Note that the scanner 101 may be configured to scan a document in response to an instruction from an operation unit (not shown) of the image processing device 100 and push-transmit the scanned image to the information processing device 110.

[0010] (Scanned Image) FIG. 3 is a diagram showing an example of a scanned image of a document. Scanned image 301 and scanned image 304 are, as an example, images obtained by scanning receipts. The scanned image 301 includes detailed information such as the unit price and quantity of each product, a total item name 302 indicating the total amount calculated from these, and a total item value 303 indicating the value of the total amount. The scanned image 304 includes, as detailed information, a unit price and quantity for each product, a total item name 305 showing a character string of the total amount calculated from these, a total item value 306 as the value of the total amount, etc. Also included are a tax amount item value 307 as the value of the tax amount, a tax amount item name 308 showing a character string of the tax amount, etc. In this embodiment, the process of extracting the total amount of a receipt will be described. Therefore, for scanned image 301, total item value 303 is set as the item value to be extracted, and for scanned image 304, total item value 306 is set as the item value to be extracted. At this time, the tax amount item name and tax amount item value, which are items not to be extracted, are set as items not to be extracted. Also, in this embodiment, an image of a receipt is used as the document, but documents other than receipts may be used. Returning now to the explanation of the flowchart in FIG. Next, in S202, the CPU 111 performs character recognition processing (OCR; Optical Character Recognition) on the scanned image 301, and stores character candidate information for each character obtained by the character recognition processing in the RAM 113. Furthermore, the CPU 111 obtains character string information based on the obtained character candidate information. The character recognition processing is a processing in which the CPU 111 specifies a character area from a document image including characters, calculates a feature amount of each character image in the character area, compares the feature amount with a character recognition dictionary, and calculates one or more similar character candidates. The processing in which the CPU 111 obtains character string information based on the character candidate information is a processing in which the CPU 111 forms a character string by concatenating character string candidates corresponding to adjacent characters, and configures character string information of one image by a plurality of character strings. Note that it is sufficient that the character strings are grouped into meaningful character units, and they may be combined or separated using morphological analysis, regular expression patterns, dictionaries, or the like. In the present embodiment, the characters include numbers and symbols. Next, in S203, CPU 111 performs an item exclusion process to lower the priority of character strings, among the character strings acquired in S202, that are included in an area specified by an exclusion rule table 510, which will be described later. The item exclusion process will be described in detail below with reference to FIG.

[0011] (Item Exclusion Processing) FIG. 4 is a flowchart showing an example of the item removal process according to the present embodiment. In S401, the CPU 111 causes the RAM 113 to read the extraction rules and the exclusion rules defined for each type of form.

[0012] The extraction rules and the exclusion rules will be described below with reference to FIG. 5 is a diagram showing an example of extraction rules and exclusion rules for receipts in this embodiment. The extraction rules and exclusion rules are assumed to be stored in advance in the HDD 114, for example. (Extraction rule table) 5(a) shows an extraction rule table 501. The extraction rule table 501 is a simplified table for explaining the definition of the extraction rules for the items to be extracted. In this embodiment, as an example, the rules for extracting the total amount written on a receipt are described. The label column 502 defines the name of the search content. The search type column 503 defines the search method. In the definition column 504, a definition value for each search method is defined. Line 505, labeled Amount A, defines an amount with a currency symbol using regular expressions to search for a string of numbers, commas, and dots, preceded by a "\" symbol and followed by a "yen" symbol. Line 506, labeled Amount B, defines a regular expression to search for a string of numbers, commas, and dots as an amount without a currency symbol. Row 507, whose label is the total item name, defines a dictionary search to search for character strings that match "total," "amount including tax," and "amount charged." Row 508 with the label Total Item Value A defines a search based on location to search for the amount with a currency symbol that is closest to the right of the "Total Item Name" label. In row 509, whose label is Total Item Value B, a search based on location is defined to search for the amount without a currency symbol that is located furthest to the right of the "Total Item Name" label. It also specifies that the higher the row, the higher the priority during extraction. Therefore, it is defined that the search result of row 508 whose label is total item value A has priority over row 509 whose label is total item value B. (Exclusion rule table) FIG. 5( b ) shows an exclusion rule table 510 . The exclusion rule table 510 is a simplified table for explaining the definition of the exclusion rule for excluding items that are not subject to extraction. In this embodiment, the rule for excluding the tax amount written on the receipt is described as an example. The label column 511 defines the name of the search content. The search type column 512 defines the search method. In the definition column 513, a definition value for each search method is defined. Line 514, whose label is amount, defines a regular expression to search for an amount without a currency symbol, a string of numbers, commas, and dots. In line 515, whose label is the tax item name, a dictionary search is performed to search for a character string that matches "tax amount." In row 516, where the label is a tax item value, a search based on location is defined to search for an amount with a currency symbol that is closest to the right side of the "total item name" label.

[0013] Returning now to the explanation of the flowchart in FIG. In S402, the CPU 111 extracts an area to be searched for a search based on a positional relationship from among the loaded rules. When a positional relationship is specified as the search type in the extraction rule table 501, the specified position is set as the area to be searched. For example, when the positional relationship is specified as "(right, recent)", the area to be searched is based on an area included in a rectangle that has the left side of a rectangle containing the item name as one side and is in contact with the right side of the scanned image. Note that the rule may specify enlargement or reduction as necessary. In S403, CPU 111 checks whether there are any unconfirmed non-target items in the search area among the non-target items defined in exclusion rule table 502. Non-target items here refer to candidates for item names that are not targets for extraction, and a check is performed in S402 to see whether there are any unconfirmed candidates for item names that are not targets for extraction in the search area. Here, if there are any unconfirmed candidates for items that are not targets for extraction (Yes in S403), CPU 111 proceeds with the process to S404, and if there are no items that are not targets for extraction (No in S403), this process ends. In S404, when an item name that is not subject to extraction exists as a result of S403, the CPU 111 identifies an exclusion target area in which an item value corresponding to the item name that is not subject to extraction exists, based on the exclusion rules described in the exclusion rule table 510. For example, in row 516 in FIG. 5B where the label is the tax item value, the positional relationship is specified as "(right, recent)", and the right side of the item name as a result of S403 above becomes the exclusion target area. In this case, the area included in the rectangle that has the left side of the rectangle containing the item name as one side and is in contact with the right side of the scanned image becomes the reference, and for example, in the scanned image 304 in FIG. 3B, the exclusion target area is the area containing the tax item name 308 and the tax item value 307. In S405, the CPU 111 assigns a penalty to the region to be excluded identified in S404 above, and stores the information in the RAM 113. Note that assigning a penalty refers to a process of removing the corresponding region from the extraction target when extracting items to be extracted, which will be described later, or a process of weighting each region to reduce the likelihood of it being a target item, but any method may be used.

[0014] Returning now to the explanation of the flowchart in FIG. Next, in S204, the CPU 111 executes an item extraction process to extract item values ​​after applying a penalty to the exclusion target area extracted in S203. The item extraction process is a process for extracting item values ​​of extraction targets according to a defined extraction rule. The item extraction process will be described in detail below with reference to FIG. 6.

[0015] (Item extraction process) FIG. 6 is a flowchart showing an example of details of the item extraction process in S204 of FIG. 2 according to this embodiment. In S601, the CPU 111 causes the RAM 113 to read the extraction rules shown in FIG. In S602, the CPU 111 performs an item search using a dictionary or regular expressions for the character string recognized in S202 of FIG. 2 in accordance with the extraction rule acquired in S601. In S603, the CPU 111 searches for the items acquired in S602 above based on the positional relationship specified in the extraction rule, and extracts item values ​​corresponding to the items to be extracted. Note that the penalty given in the item exclusion process in Fig. 4 is taken into consideration. Details will be described later (Fig. 7). Then, this process ends.

[0016] (Search processing based on location) FIG. 7 is a flowchart showing an example of details of the search process (S603) in this embodiment for searching for an extraction item value based on its positional relationship with an item to be extracted that is specified in an extraction rule. In S701, CPU 111 searches for an amount with a currency symbol closest to the right of the "Total Item Name" label using the extraction rule specified in row 508, whose label is Total Item Value A. If there is no amount with a currency symbol to the right of the "Total Item Name" label (No in S701), CPU 111 proceeds to S703. If there is an amount with a currency symbol (Yes in S701), CPU 111 proceeds to S702. In S702, the CPU 111 extracts the amount with a currency symbol. When performing this process, as a process taking into consideration a penalty, the area of ​​the extraction result of the non-target item in S203 is excluded from extraction, or the likelihood of the item value in the area is reduced. In S703, the CPU 111 searches for an amount without a currency symbol that is located furthest to the right of the “Total Item Name” label according to the extraction rule specified in row 509 of the label Total Item Value B, and if an amount without a currency symbol is present, extracts the value and ends the search process. At this time, the CPU 111 performs processing that takes into account penalties, as in S702. In the case of the scanned image 304 in FIG. 3(b), the amount without the currency symbol is extracted with penalties applied to the areas of the tax item name 308 and the tax item value 307, so that the total item value 306 is appropriately extracted. The extraction rule table 501 and the exclusion rule table 510 described in this embodiment are examples for explaining this embodiment, and the definition contents are not limited to these. Vocabulary may be expanded according to the use case, and conditions such as various constraints may be described to improve accuracy. As described above, by defining areas that are not to be extracted, it is possible to achieve both the process of extracting item names by excluding unnecessary item values ​​near the item names, and the process of extracting item values ​​when item name / item value pairs are adjacent with no gaps, thereby making it possible to appropriately extract item values.

[0017] [Second embodiment] In the second embodiment, only the differences from the first embodiment described above will be described, and the parts not specifically mentioned have the same configuration / procedure as the first embodiment. In the above-mentioned first embodiment, when performing the item exclusion process, the exclusion target area to which a penalty is to be applied is determined using the item name information of the excluded items. In the second embodiment, the same purpose is achieved by extracting the item values ​​corresponding to the item names of the excluded items and applying a penalty to the character string information of the extracted item names and item values. (Item Exclusion Processing) Fig. 8 is a flowchart showing an example of the item removal process of S203 in Fig. 2 in the second embodiment. Note that the process contents of S401 to S403 are the same as those in the flowchart of Fig. 4 in the first embodiment, and therefore a description thereof will be omitted. Note that the process of the flowchart shown in Fig. 8 is realized by the CPU 111 of the information processing device 110 reading and executing a program stored in the ROM 112, the HDD 114, or the like. In S801, if an item name that is not a target for extraction is found as a result of S403, the CPU 111 extracts an item value corresponding to the item name that is not a target for extraction. Here, extracting an item value corresponding to an item name that is not a target for extraction means a process of extracting a corresponding item value according to the exclusion rules described in the above-mentioned exclusion rule table 510. As an example, in the case of the scanned image 304 in Fig. 3, the tax amount item name 308 is set as the item name, and the tax amount item value 307, which is an item value candidate that exists at the shortest distance to the right, is extracted. In S802, the CPU 111 assigns a penalty to the item value extracted in S801 and the character string information corresponding to that item value. Note that assigning a penalty refers to a process of excluding the part of the character string information to which the penalty is assigned from the extraction process when extracting the item to be extracted defined by the extraction rule and its item value, or a process of weighting the character string information to reduce the likelihood of it being an extraction target, but any method may be used. As explained above, by excluding item names and their corresponding item values ​​that are not to be extracted and then extracting item values, it becomes possible to extract appropriate item values ​​even in more complex cases, such as when a tax amount item name and a tax amount item value exist between a total item name and a total item value.

[0018] <Other embodiments> The present invention can also be realized by a process in which a program for implementing one or more of the functions of the above-described embodiments is supplied to a system or device via a network or a storage medium, and one or more processors in a computer of the system or device read and execute the program. The present invention can also be realized by a circuit (e.g., ASIC) that implements one or more of the functions. Furthermore, the present invention may be applied to a system made up of a plurality of devices, or to an apparatus made up of a single device.

[0019] The disclosure of this embodiment includes the following configuration, method, and program. (Configuration 1) An acquisition means for acquiring character string information recognized by character recognition processing from a document image; an assignment means for specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying an exclusion target area based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified exclusion target area; an extraction means for extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the excluded region to which the penalty is assigned; 13. An information processing device comprising: (Configuration 2) An acquisition means for acquiring character string information recognized by character recognition processing from a document image; an assignment means for specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction means for extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. An information processing device comprising: (Method 1) An acquisition step of acquiring character string information recognized by character recognition processing from the document image; a step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to the item values ​​of the items, specifying an exclusion target area based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and their corresponding item values, and assigning a penalty to the specified exclusion target area; an extraction step of extracting item values ​​of the items to be extracted defined by the extraction rules based on the extraction rules and the excluded area to which the penalty is assigned; 13. A method for controlling an information processing apparatus comprising the steps of: (Method 2) An acquisition step of acquiring character string information recognized by character recognition processing from the document image; an assignment step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and their corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. A method for controlling an information processing apparatus comprising the steps of: (Program 1) A program for causing a computer to execute the control method for an information processing device according to the first or second method. (Method 3) An information extraction method executed by an information processing device, comprising: An acquisition step of acquiring character string information recognized by character recognition processing from the document image; a step of identifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, identifying an exclusion target area based on character string information included in the identified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the identified exclusion target area; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the excluded area to which the penalty is assigned; 13. An information extraction method comprising: (Method 4) An information extraction method executed by an information processing device, comprising: An acquisition step of acquiring character string information recognized by character recognition processing from the document image; an assignment step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. An information extraction method comprising: [Explanation of symbols]

[0020] 100 Image processing device 101 Scanner 110 Information processing device 111 CPU 112 ROM 113 RAM 114 HDD

Claims

1. An acquisition means for acquiring character string information recognized by character recognition processing from a document image; an assignment means for specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying an exclusion target area based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified exclusion target area; an extraction means for extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the excluded region to which the penalty is assigned; 13. An information processing device comprising:

2. An acquisition means for acquiring character string information recognized by character recognition processing from a document image; an assignment means for specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction means for extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. An information processing device comprising:

3. An acquisition step of acquiring character string information recognized by character recognition processing from the document image; a step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to the item values ​​of the items, specifying an exclusion target area based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and their corresponding item values, and assigning a penalty to the specified exclusion target area; an extraction step of extracting item values ​​of the items to be extracted defined by the extraction rules based on the extraction rules and the excluded area to which the penalty is assigned; 13. A method for controlling an information processing apparatus comprising the steps of:

4. An acquisition step of acquiring character string information recognized by character recognition processing from the document image; an assignment step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and their corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. A method for controlling an information processing apparatus comprising the steps of:

5. 5. A program for causing a computer to execute the method for controlling an information processing device according to claim 3.

6. An information extraction method executed by an information processing device, comprising: An acquisition step of acquiring character string information recognized by character recognition processing from the document image; a step of identifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, identifying an exclusion target area based on character string information included in the identified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the identified exclusion target area; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the excluded area to which the penalty is assigned; 13. An information extraction method comprising:

7. An information extraction method executed by an information processing device, comprising: An acquisition step of acquiring character string information recognized by character recognition processing from the document image; an assignment step of specifying a search target area in the document image based on items to be extracted defined as extraction rules and positional relationships related to item values ​​of the items, specifying character string information to be excluded based on character string information included in the specified search target area and information related to items not to be extracted defined as exclusion rules and corresponding item values, and assigning a penalty to the specified character string information to be excluded; an extraction step of extracting an item value of the item to be extracted defined by the extraction rule based on the extraction rule and the character string information of the excluded object to which the penalty is assigned; 13. An information extraction method comprising:

Images