Information processing apparatus, information processing system, recording medium, and information processing method

Abstract

An information processing apparatus, an information processing system, a recording medium, and an information processing method, the information processing apparatus including a processor that performs processing to acquire a plurality of read images obtained by reading a paper medium bundle including a plurality of regular files and a collection of associated files related to the regular files, extract a theme from the plurality of read images, extract an identifier that is common within one regular file bundle and is marked as different between different regular file bundles from the plurality of read images, and divide the plurality of read images into bundles in a manner in which a read image in which the theme is extracted is at the beginning of a bundle, read images in which the identifier is common are included in the same bundle, and read images in which the identifier is different are included in different bundles.

Description

Technical Field

[0001] This invention relates to an information processing device, an information processing system, a recording medium, and an information processing method. Background Technology

[0002] Patent document 1 discloses an image forming apparatus having a scanner function. The image forming apparatus is characterized by comprising: a title detection unit for detecting the title of a file read by the scanner function; a title file creation unit for creating a file with the title string, i.e., the title, as the file name; and a title image data storage unit for storing image data of the file corresponding to the scope of the title in the title file.

[0003] Patent Document 1: Japanese Patent Application Publication No. 2015-102934

[0004] In most cases, paper documents filled out by users are digitized and stored in storage units such as memory devices for management. At the same time, documents with common characteristics are also grouped together for management. Furthermore, when processing digitized documents for various purposes, OCR (Optical Character Recognition) processing, which involves recognizing strings from the document's image, is sometimes performed.

[0005] On the other hand, for various purposes, there are times when it is necessary to divide a group of managed documents into different bundles (hereinafter referred to as "segments"). For example, a bundle (hereinafter referred to as a "regular document bundle") may be formed by combining invoices such as bills (hereinafter referred to as "regular documents") and documents associated with regular documents such as delivery notes (hereinafter referred to as "associated documents"), and a group of documents may be divided into multiple regular document bundles. Currently, techniques are known to extract the headers of documents and divide the documents within the scope of those headers into bundles. However, header extraction does not necessarily achieve the desired segmentation into regular document bundles. Summary of the Invention

[0006] The object of the present invention is to provide an information processing apparatus, information processing system, recording medium, and information processing method that can reduce the number of files that would be included in one set being divided into another set, even when multiple read images obtained by reading a bundle of paper media containing a bundle of multiple regular files and associated files of the regular files are divided according to the bundle.

[0007] The information processing apparatus of the first method includes a processor that performs the following processing: acquiring a plurality of read images obtained by reading a bundle of paper media containing a plurality of regular documents and associated documents related to the regular documents, i.e., a bundle of regular documents; extracting a theme from the plurality of read images; extracting identifiers common within a bundle of regular documents but marked differently between different bundles of regular documents from the plurality of read images; and segmenting the plurality of read images by bundle, with the read image from which the theme has been extracted becoming the beginning of the bundle, read images with common identifiers being contained in the same bundle, and read images with different identifiers being contained in different bundles.

[0008] In the information processing apparatus of the second method, which is the same as that of the first method, the processor performs the following processing: extracting at least one of the items and values ​​corresponding to the identifier pre-specified by the user from the plurality of read images, thereby extracting the identifier.

[0009] In the information processing apparatus of the third method, in the information processing apparatus of the first or second method, the processor sequentially performs the process of extracting the topic and the process of extracting the identifier.

[0010] The fourth type of information processing system includes: a first information processing device, comprising a first processor, the first processor acquiring a plurality of read images obtained by reading a bundle of paper media containing a plurality of regular documents and associated documents related to the regular documents, i.e., a bundle of regular documents; and a second information processing device, comprising a second processor, the second processor performing the following processing: extracting a theme from the plurality of read images received from the first information processing device; extracting identifiers common within a bundle of regular documents but marked differently between different bundles from the plurality of read images; and segmenting the plurality of read images by bundle, with the read image from which the theme has been extracted becoming the beginning of the bundle, read images with common identifiers being contained in the same bundle, and read images with different identifiers being contained in different bundles.

[0011] The fifth type of recording medium records an information processing program for causing a computer to perform the following functions: acquiring a plurality of read images obtained by reading a bundle of paper media containing a plurality of regular documents and associated documents related to the regular documents, i.e., a bundle of regular documents; extracting a subject from the plurality of read images; extracting identifiers common within a bundle of regular documents but marked differently between different bundles of regular documents from the plurality of read images; and dividing the plurality of read images by bundle, with the read image from which the subject has been extracted as the beginning of the bundle, read images with common identifiers being contained in the same bundle, and read images with different identifiers being contained in different bundles.

[0012] The information processing apparatus of the sixth method includes a processor that performs the following processing: acquiring a plurality of read images obtained by reading a bundle of paper media containing a plurality of regular documents and associated documents related to the regular documents, i.e., a bundle of regular documents; extracting a subject from the plurality of read images; extracting an identifier containing a pre-defined character from the plurality of read images; and dividing the plurality of read images into bundles, with the read image from which the subject has been extracted becoming the beginning of the bundle, and read images containing a common identifier being consecutively arranged from the read images from which the subject has been extracted into the same bundle.

[0013] In the information processing apparatus of the seventh method, which relates to the information processing apparatus of the sixth method, the identifier is a character representing an additional file of a read image preceding the read image containing the identifier, or a page number consecutive to the read image preceding the read image containing the identifier.

[0014] The information processing method of the eighth method includes the following steps: obtaining a plurality of read images obtained by reading a bundle of paper media containing a plurality of regular documents and associated documents related to the regular documents, i.e., a bundle of regular documents; extracting a theme from the plurality of read images; extracting identifiers common within a bundle of regular documents but marked with different labels between different bundles of regular documents from the plurality of read images; and segmenting the plurality of read images by bundle, with the read image from which the theme has been extracted as the beginning of the bundle, read images with common identifiers being contained in the same bundle, and read images with different identifiers being contained in different bundles.

[0015] Invention Effects

[0016] According to the information processing apparatus of the first method, the information processing system of the fourth method, the recording medium of the fifth method, the information processing apparatus of the sixth method, and the information processing method of the eighth method, even when multiple read images obtained by reading a bundle of paper media containing a set of multiple regular documents and associated documents of the regular documents are divided according to the set, it is possible to reduce the likelihood of documents contained in one set being divided into another set.

[0017] The information processing device according to the second method, compared with the case where the items and values ​​corresponding to the identifiers pre-specified by the user are not used, is able to reflect the inherent conditions of the user in information processing.

[0018] According to the information processing apparatus of the third method, compared with the case where the order of processing for extracting the subject and processing for extracting the identifier is not considered, the identifier extraction processing from the read image after extracting a certain subject to the read image after extracting the next subject can be omitted.

[0019] According to the information processing device of the seventh method, segmentation can be performed with higher precision compared to cases where the characters or page numbers of the attached document are uncertain about the identifier. Attached Figure Description

[0020] The embodiments of the present invention will be described in detail with reference to the following figures.

[0021] Figure 1 In the embodiments, the information processing apparatus involved in each implementation is... Figure 1 (a) is a block diagram representing an example of a hardware structure. Figure 1 (b) is a block diagram representing an example of a functional structure;

[0022] Figure 2 This is a block diagram illustrating an example of the structure of the information processing system involved in each implementation;

[0023] Figure 3 This is a schematic diagram illustrating the process of segmentation involved in the first embodiment;

[0024] Figure 4 (a) to Figure 4 (c) is a schematic diagram illustrating an example of the post-processing of the segmentation process according to the first embodiment;

[0025] Figure 5 It is a flowchart illustrating the information processing flow involved in each implementation method;

[0026] Figure 6 This is a schematic diagram illustrating the process of segmentation involved in the second embodiment;

[0027] Figure 7 (a) to Figure 7 (c) is a schematic diagram illustrating an example of the post-processing of the segmentation process according to the second embodiment;

[0028] Figure 8 This is a schematic diagram illustrating the process of segmentation involved in the third embodiment;

[0029] Figure 9 (a) to Figure 9 (c) is a schematic diagram illustrating an example of the post-processing of the segmentation process involved in the third embodiment.

[0030] Symbol Explanation

[0031] 1-Information processing system, 10-Information processing device, 11-CPU, 12-ROM, 13-RAM, 14-Memory, 15-Receiver, 16-UI, 17-Communication I / F, 18-Bus, 21-Acquisition unit, 22-Identification unit, 23-Extraction unit, 24-Storage unit, 25-Setting unit, 26-Processing unit, 30-Image processing device, 31-Terminal device, 40-Cloud, 41-Network. Detailed Implementation

[0032] Hereinafter, the technical solution for implementing the present invention will be described in detail with reference to the accompanying drawings. Furthermore, the information processing apparatus 10 according to this embodiment will be described as an example of a method for managing a server that reads data such as documents and invoices. However, it is not limited to this. The information processing apparatus 10 may, for example, be mounted on a multifunction printer with functions such as printing, copying, scanning, and faxing, or it may be a terminal such as a personal computer.

[0033] [First Implementation]

[0034] refer to Figures 1 to 5 The information processing apparatus, information processing system, and information processing program involved in this embodiment will be described. For example... Figure 1 As shown in (a), the information processing device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a memory 14, a receiving unit 15, a UI (User Interface) 16, and a communication I / F (interface) 17. The CPU 11, ROM 12, RAM 13, memory 14, receiving unit 15, UI 16, and communication I / F 17 are interconnected via a bus 18. Here, the CPU 11 is an example of a processor according to the present invention.

[0035] The CPU 11 centrally controls the entire information processing device 10. The ROM 12 stores various programs and data, including the segmentation processing program used in this embodiment. The RAM 13 is a memory used as a working area when executing various programs. The CPU 11 performs various information processing by expanding and executing the programs stored in the ROM 12 in the RAM 13. For example, the memory 14 is an HDD (Hard Disk Drive), SSD (Solid State Drive), or flash memory. Additionally, information processing programs may also be stored in the memory 14. The receiving unit 15 receives, for example, multiple read images of page units of a file. For example, the receiving unit 15 is a USB (Universal Serial Bus). For example, the UI 16 is a touch panel-type liquid crystal display, for example, receiving commands issued by the user. The UI 16 may also display image data accompanying the segmentation processing performed by the information processing device 10 (described later). The communication I / F 17 is an interface for connecting to a network (described later), for example, for sending and receiving data with an image processing device. In addition, the memory 14, receiver 15, UI 16 and communication I / F 17 are not essential components of the information processing device 10, and can be selected and included depending on the nature of the information processing device 10.

[0036] Next, refer to Figure 1 (b) describes the functional structure of the information processing device 10. Figure 1 (b) is a block diagram illustrating an example of the functional structure of the information processing apparatus 10 according to this embodiment.

[0037] like Figure 1 As shown in (b), the information processing device 10 includes an acquisition unit 21, an identification unit 22, an extraction unit 23, a storage unit 24, a setting unit 25, and a processing unit 26. The CPU 11 functions as the acquisition unit 21, the identification unit 22, the extraction unit 23, the storage unit 24, the setting unit 25, and the processing unit 26 by executing an information processing program.

[0038] The acquisition unit 21 acquires an image (reading image) of page units of a paper medium containing a plurality of documents. For example, this corresponds to acquiring a reading image via the receiving unit 15 or the communication I / F 17.

[0039] The recognition unit 22 performs OCR processing on the read image to recognize the characters contained in the read image. The extraction unit 23 performs title extraction processing on the OCR-processed read image to extract the title (subject) from the read image. Furthermore, it performs key value extraction and extracts identifiers used to determine the correlation between documents. Additionally, the recognition unit 22 does not necessarily need to be installed in the information processing device 10; it can receive the OCR-processed read image from an external source. The processing unit 26 uses the identifiers extracted by the extraction unit 23 from the document bundles (hereinafter referred to as "segmentation candidate bundles") segmented according to the title to reassemble them into a combination of regular documents and related documents, i.e., a regular document bundle. In this embodiment, as described above, the functions of the recognition unit 22, the extraction unit 23, and the processing unit 26 are implemented by software having an information processing program. However, it is not limited to this; it can also be implemented by hardware such as an ASIC (Application Specific Integrated Circuit) or a dedicated LSI.

[0040] The storage unit 24 is implemented, for example, by a memory 14, and stores the results of processing performed in the identification unit 22, the extraction unit 23, and the processing unit 26. The setting unit 25 is implemented, for example, by a UI 16, where the user sets conditions for information processing performed in the information processing device 10.

[0041] Next, refer to Figure 2 The information processing system 1 involved in this embodiment will be described. For example... Figure 2 As shown, the information processing system 1 includes information processing devices 10-1 and 10-2, a cloud 40, a network 41, an image processing device 30, and a terminal device 31. However, the information processing system 1 does not need to include all of these structures, and can be configured by selecting the necessary structures according to the purpose, system conditions, etc.

[0042] Network 41, such as an IP network, is a system that connects various devices to each other. The connection method of network 41 can be wired or wireless, and can also be a local area network (LAN). Cloud 40 is a system that provides various services through network 41, such as an IP network. Information processing devices 10-1 and 10-2 are devices with the same functions as each information processing device 10. Information processing device 10-1 is shown configured on cloud 40, and information processing device 10-2 is shown configured on network 41. That is, information processing device 10-1 shows an example of being connected to cloud 40 via communication I / F 17, and information processing device 10-2 shows an example of being connected to network 41 via communication I / F 17. In this example, as an example, information processing devices 10-1 and 10-2 are implemented by a server. However, this is not a limitation; information processing devices 10-1 or 10-2 can also be used independently.

[0043] Image processing device 30, such as a multifunction printer with scanning (image reading) function, is connected to cloud 40 or network 41. It acquires a plurality of read images of a document bundle and sends the data of the acquired plurality of read images to information processing device 10-1 or 10-2. Alternatively, OCR processing can be performed on each read image before sending it to information processing device 10-1 or 10-2. Terminal device 31, such as a PC (Personal Computer), controls information processing device 10-1 or 10-2 and image processing device 30 in one embodiment of information processing system 1, and receives the results processed by information processing device 10-1 or 10-2.

[0044] However, in document processing, such as combinations of "invoices" and "delivery notes," it is sometimes desirable to group regular documents like invoices and related documents (associated documents) into a single bundle for processing. On the other hand, in routine document processing, there are often situations where regular document bundles are not necessary; for example, they are often processed as a group of documents (hereinafter referred to as "document groups") without distinguishing them by project unit. In such cases, it is conceivable that regular document bundles need to be divided into individual bundles based on deadlines or project completion dates.

[0045] Previously, techniques were known to extract titles from multiple read images that had undergone OCR processing of a file group, and then segment the image into a set of files associated with that title based on the extracted title. However, in these prior art techniques, files are mechanically segmented by title, which may result in the segmentation of files with titles such as "invoice" and "delivery note," which were not originally intended to be segmented. That is, title extraction is suitable for processing single invoices, but sometimes it is not possible to properly segment the read image into a regular bundle of files using only title extraction processing.

[0046] Therefore, the present invention employs the following structure: regarding the reading image obtained from reading a bundle of paper media containing a plurality of regular document bundles, it is first segmented into a plurality of segmentation candidate bundles based on the result of title extraction processing. Then, using the identifiers of the segmentation candidate bundles obtained from key value extraction processing, it is determined whether to re-integrate the segmentation candidate bundles with each other, and finally segmented into regular document bundles. In this embodiment, "identifier" refers to a string that is common within a regular document bundle but is marked differently between different regular document bundles. Thus, it is possible to re-integrate the segmentation candidate bundles mechanically segmented using the title into a set of substantially related documents. That is, when segmenting a bundle of paper media containing a plurality of regular documents and a set of related documents of those regular documents according to this set, it is possible to reduce the likelihood of documents intended to be included in one set being segmented into another set.

[0047] refer to Figures 3 to 5 The information processing apparatus and information processing program involved in this embodiment will be described. Figure 3 This is a schematic diagram illustrating the process of segmentation processing performed by the information processing device 10. Additionally, in Figure 3 In this context, it is assumed that the information processing device 10 has acquired a plurality of read images, i.e., file groups, that have undergone OCR processing, and that the file group is stored as an example in the memory 14.

[0048] Figure 3 <1> shows a group of files before segmentation, i.e., multiple read images that have undergone OCR processing. Figure 3 <2> shows the result of performing header extraction processing on the file group, dividing it into multiple segmentation candidate bundles. Figure 3 <3> shows the state of the identifier obtained by further performing key-value extraction, which is then integrated again into the desired regular file bundle.

[0049] Here, in "Title Extraction," the title (subject) within the read image is extracted using factors such as the size and position of characters. That is, title extraction is performed within a file group, taking into account inherent characteristics of titles, such as the prevalence of titles appearing as large characters on the initial page. Based on the extracted title, the type of document (invoice, delivery note, etc.) of the read image is determined, for example. In other words, the title extraction identifies a single document.

[0050] exist Figure 3 In the example shown in <2>, the file group is segmented into four segmentation candidate bundles: 1, 2, 3, and 4, as a result of title extraction. That is, the title "Bill" is extracted to generate segmentation candidate bundle 1, the title "Delivery Note" is extracted to generate segmentation candidate bundle 2, the title "Bill" is extracted to generate segmentation candidate bundle 3, and the title "Quotation" is extracted to generate segmentation candidate bundle 4.

[0051] However, split candidate bundle 1 and split candidate bundle 2 share the same document number, and the "delivery note" of split candidate bundle 2 is an appendix to the "invoice" of split candidate bundle 1. That is, since split candidate bundle 1 is a regular document and split candidate bundle 2 is an associated document, split candidate bundle 1 and split candidate bundle 2 are not intended to be split and are instead set as a single bundle. Furthermore, split candidate bundle 3 contains the mark "Remarks, Quotation Number 3456," and split candidate bundle 4 is a "Quotation" with "Quotation Number 3456." That is, since split candidate bundle 3 is a regular document and split candidate bundle 4 is an associated document, split candidate bundle 3 and split candidate bundle 4 are not intended to be split and are instead set as a single bundle.

[0052] Therefore, in this embodiment, key-value extraction is further performed. In "key-value extraction," items (keys) are searched within the file, and the values ​​corresponding to the found items are extracted. Figure 3 In the example shown in <2>, in segmentation candidate bundles 1 and 2, "No." is identified as the key and "1234" is identified as the value. Furthermore, in segmentation candidate bundles 3 and 4, "quotation number" is identified as the key and "3456" is identified as the value. In this embodiment, the string extracted through key-value extraction is called the "identifier," and these are then integrated with segmentation candidate bundles that share the same identifier to generate a regular file bundle.

[0053] exist Figure 3In the example shown in <3>, the identifier "No.1234" is common (consistent), therefore, segmentation candidate bundle 1 and segmentation candidate bundle 2 are integrated again to generate regular file bundle 1. That is, segmentation candidate bundle 1 is set as the beginning, and segmentation candidate bundle 2 is made continuous with segmentation candidate bundle 1 to generate regular file bundle 1. Furthermore, the identifier "quotation number 3456" is common (consistent), therefore, regular file bundle 2 is generated. That is, segmentation candidate bundle 3 is set as the beginning, and segmentation candidate bundle 4 is made continuous with segmentation candidate bundle 3 to generate regular file bundle 2. In other words, the file group is segmented into regular file bundles that combine regular files and associated files for the purpose set forth in this embodiment.

[0054] Furthermore, from an easy-to-understand perspective, this embodiment has been described by illustrating the execution in the order of title extraction and key-value extraction, but it is not limited to this. It can be executed in the order of key-value extraction and title extraction, or simultaneously. However, the method of executing in the order of title extraction and key-value extraction has the following effects: Firstly, by performing title extraction first, the pages from the file marked with title A to the file marked with title B can be processed as a file bundle associated with title A, thus omitting the key-value extraction process regarding which bundle the file bundle belongs to. Secondly, when the cover pages of title A and title B have a common identifier, it can determine whether the common identifier exists within the file bundle or not, thus eliminating the need for matching.

[0055] Next, refer to Figure 4 The post-processing involved in this embodiment will be described. In this embodiment, "post-processing" refers to a method for saving regular file bundles after a file group has been divided into regular file bundles. In this embodiment, three types of post-processing are provided as an example. However, post-processing is not limited to these, and other appropriate post-processing may be performed depending on the purpose, etc. Figure 4 of (a), Figure 4 (b) and Figure 4 (c) shows processing 1, processing 2 and processing 3 as post-processing, respectively.

[0056] The contents of processing steps 1, 2, and 3 are as follows. Furthermore, as described later, post-processing can be specified by the user each time the segmentation process involved in this embodiment is performed, or pre-set processes can be executed automatically.

[0057] <Process 1>

[0058] about Figure 3 Each of the shown regular file bundles 1 and 2 is assigned a filename and saved by combining the headers of the split candidate bundles. Specifically, regarding regular file bundle 1, as... Figure 4As shown in (a), candidate bundles 1 and 2 are summarized and saved with the filename "Bill_Delivery Note_No1234". For regular file bundle 2, candidate bundles 3 and 4 are summarized and saved with the filename "Bill_Quotation_No6789".

[0059] <Process 2>

[0060] about Figure 3 For each of the regular file bundles 1 and 2 shown, create a folder with a string representing its respective characteristic (e.g., an identifier name) as its name, and store the segmentation candidate bundles in this folder. Specifically, as... Figure 4 As shown in (b), regarding regular document bundle 1, create a folder named "No_1234" and store candidate bundles 1 and 2 in this folder. Then, assign the filename "Bill_No1234" to candidate bundle 1 and "Delivery Note_No1234" to candidate bundle 2. Regarding regular document bundle 2, create a folder named "No6789" and store candidate bundles 3 and 4 in this folder. Then, assign the filename "Bill_No6789" to candidate bundle 3 and "Quotation_No3456" to candidate bundle 4.

[0061] <Process 3>

[0062] When the results of segmentation into candidate bundles are extracted from the title, and a key value referencing another candidate bundle is extracted from a specific candidate bundle, a folder for the reference source candidate bundle is created, and the reference source candidate bundle is stored in this folder. Further, a folder for the reference destination candidate bundle is created, and the files of the reference destination candidate bundle are stored in this folder. Specifically, as follows... Figure 4 As shown in (c), regarding regular file bundle 1, create a folder for regular file bundle 1 (illustration omitted), and save the split candidate bundle 1 in this folder with the filename "Bill_No1234". Further create a folder named "No1234_Delivery Note", and save the split candidate bundle 2 in this folder with the filename "Delivery Note_No1234". Regarding regular file bundle 2, create a folder for regular file bundle 3 (illustration omitted), and save the split candidate bundle 3 in this folder with the filename "Bill_No6789". Further create a folder named "No6789_Quotation", and save the split candidate bundle 4 in this folder with the filename "Quotation_No3456".

[0063] Next, refer to Figure 5 The segmentation process performed by the information processing device 10 will be described. This segmentation process involves dividing a group of files into regular bundles of files. Figure 5This is a flowchart illustrating the processing flow of the segmentation procedure. This segmentation procedure is stored in memory units such as ROM 12 and memory 14. The CPU 11 reads it from these memory units and expands it into RAM 13, etc., to execute the program. Furthermore, this segmentation procedure can also be supplied externally via the receiving unit 15 or communication I / F. Additionally, in... Figure 5 In this example, the information processing device 10 acquires a plurality of read images (file groups) that have undergone OCR processing from an external device such as the image processing device 30.

[0064] like Figure 5 As shown, in step S100, a plurality of read images are acquired.

[0065] In step S101, title extraction processing and key value extraction processing are performed on the entire plurality of read images. Furthermore, the order in which the title extraction processing and key value extraction processing are performed is arbitrary.

[0066] In step S102, it is determined whether the title has been extracted from the plurality of read images. If the determination is positive, proceed to step S103; otherwise, if the determination is negative, proceed to step S104.

[0067] In step S103, the page with the title extracted is stored. A page with the title extracted refers to, for example... Figure 3 The initial pages of the candidate bundles 1, 2, 3, and 4 in <2>.

[0068] In step S104, it is determined whether the current page is the final page of a plurality of read images. If the determination is negative, the process returns to step S102 and continues to determine whether the title has been extracted. On the other hand, if the determination is positive, the process proceeds to step S105.

[0069] In step S105, it is determined whether a segmentation candidate bundle exists. The existence of a segmentation candidate bundle is determined based on whether there are multiple read images with the title extracted. When the determination is positive, proceed to step S106; otherwise, when the determination is negative, proceed to step S113.

[0070] In step S106, the identifiers generated from the key value extraction results are compared between the segmentation candidate bundles.

[0071] In step S107, it is determined whether the identifiers of the candidate segments are consistent. If the determination is positive, the process proceeds to step S108; otherwise, if the determination is negative, the process proceeds to step S109. Furthermore, in this embodiment, the example of setting the file number as the key value has been described, but the key value is not limited to this and can also be a pre-defined string (e.g., "attached file," page number, etc., described later). Moreover, the pre-defined key value can be pre-stored in a storage unit such as ROM12. It can also be set to a method input by the user via UI16 before this segmentation process.

[0072] In step S108, the combination of storage segmentation candidate bundles is the regular file bundle.

[0073] In step S109, it is determined whether the combination of candidate bundles is the final combination. If the determination is positive, proceed to step S110; otherwise, if the determination is negative, return to step S107 and continue determining whether the identifiers match.

[0074] In step S110, it is determined whether a regular file bundle exists. If the determination is positive, proceed to step S111; otherwise, if the determination is negative, proceed to step S113.

[0075] In step S111, it is determined whether there is a user-specified post-processing requirement. If the determination is positive, the process proceeds to step S112; otherwise, if the determination is negative, the process proceeds to step S113. That is, in this embodiment, when there is no user-specified post-processing requirement, process 1 is executed. However, it is not limited to this; process 2 or process 3 may also be executed.

[0076] In step S112, it is determined whether the user specified for post-processing is process 1. If the determination is positive, the process proceeds to step S113 to execute process 1 and end this segmentation process. On the other hand, if the determination is negative, the process proceeds to step S114.

[0077] In step S114, it is determined whether the user specified for post-processing is process 2. If the determination is positive, the process proceeds to step S115 to execute process 2 and end this segmentation process. On the other hand, if the determination is negative, the process proceeds to step S116.

[0078] In step S116, post-processing 3 is performed, and the current segmentation process is terminated.

[0079] <First Modification of the First Embodiment>

[0080] In the first embodiment described above, a method for automatically extracting general key values ​​was illustrated, but this embodiment describes a method for user-defined (specification-changing) key values. For example, when files related to the main device and files of sub-devices that constitute part of the main device coexist in a file group, it is sometimes desirable to use these files as a regular file bundle. Hereinafter, we will assume that the following files exist in the file group for explanation.

[0081] (1) Regarding sub-device 1, the document titled "Quotation" and with document number "Quotation No. 1234".

[0082] (2) Regarding sub-device 2, the document titled "Quotation" and with document number "Quotation No. 2345".

[0083] The files (1) and (2) have different identifiers in the usual key-value extraction process, so they will be split.

[0084] In this case, for example, a key that can determine the main device (e.g., the model number of the main device) can be associated with a value corresponding to the sub-device (e.g., the model number of the sub-device) and pre-registered. When this value is associated with the key, the files associated with sub-devices 1 and 2 can be considered as a splitting candidate bundle and integrated again, along with the files associated with the main device. The above functionality can also be customized by each user through the aforementioned splitting process. That is, the splitting process can prepare an option to "not split if the value extracted from a certain key is 'same' or 'same under XX conditions'" so that the user can specify it in advance. Furthermore, when a user's specification exists, it can be stored, for example, in a storage unit such as memory 14, and given higher priority compared to other splitting rules.

[0085] <Second variation of the first embodiment>

[0086] In the first embodiment described above, the example was given when the extracted key value was a single key value. However, this embodiment describes the case where the extracted key value is a plurality of keys. For example, suppose that the title extraction result generates the following segmentation candidate bundle.

[0087] (1) Segment candidate bundle 1 (Total pages 3)

[0088] Page 1: Title "Quotation Sheet", Quotation No. "1234", Name "YY"

[0089] (2) Segment candidate bundle 2 (Total pages 2)

[0090] Page 1: Title "Quotation Sheet", Quotation No. "2345", Name "YY"

[0091] (3) Segmenting candidate bundles 3 (Total pages 4)

[0092] Page 1: Title "Delivery Note", Quotation No. "2345", Name "YY"

[0093] (4) Segment candidate bundle 4 (Total pages 2)

[0094] Page 1: Title "Delivery Note", Quotation No. "3467", Name "ZZ"

[0095] In the example above, there are two possible key values: combinations of quote No. "1234", quote No. "2345", and quote No. "3467", and combinations of name "YY" and name "ZZ". In this case, it's necessary to pre-determine which one to use as the identifier. However, for example, a key value with more pages in a regular document bundle could also be used. In the example above, the key value "YY" has a total of 9 pages (the total number of pages in split candidate bundles 1, 2, and 3), and the key value quote No. "2345" has a total of 6 pages (the total number of pages in split candidate bundles 2 and 3). Therefore, using the name "YY" as the identifier, the first 9 pages are grouped into regular document bundle 1, and the last 2 pages are also grouped into regular document bundle 1.

[0096] [Second Implementation]

[0097] refer to Figure 6 and Figure 7 The information processing apparatus, information processing system, and information processing program involved in this embodiment will be described. This embodiment describes the method used in the first embodiment described above when additional files are extracted by key-value extraction. Therefore, the information processing apparatus, information processing system, and information processing program involved in this embodiment are basically the same as those in the first embodiment described above, and should be referred to as needed. Figure 1 , 2 5. Detailed explanations are omitted. Additionally, in this embodiment, "attached file" refers to a file attached to a regular file.

[0098] Figure 6 The flowchart of the segmentation process involved in this embodiment is shown. Figure 6 <1> shows the file group, i.e., multiple read images, before segmentation. Figure 6 <2> shows the state of four segmentation candidate bundles generated by the header extraction of the file group. Figure 6 <3> shows the state where the results of further key-value extraction are integrated again into two regular file bundles. Furthermore, from an easy-to-understand point of view, this embodiment is illustrated by performing title extraction and key-value extraction in that order, but it is not limited to this; it can be performed in the order of key-value extraction and title extraction, or it can be performed simultaneously.

[0099] like Figure 6 As shown in <2>, the title extraction results in four candidate splitting bundles: 1, 2, 3, and 4. Candidate splitting bundle 1 is a quotation with the document number "No. 1234" and the tag "Additional Information: XX Specification". Candidate splitting bundle 3 is a quotation with the document number "No. 6789". Candidate splitting bundle 2 contains the tag "XX Specification" and is adjacent to candidate splitting bundle 1; therefore, candidate splitting bundle 1 and candidate splitting bundle 2 are considered a regular document and an associated document, and therefore should not be split. Candidate splitting bundle 4 is a "Budget Sheet" marked "Additional Information" and is adjacent to candidate splitting bundle 3; therefore, candidate splitting bundle 3 and candidate splitting bundle 4 are considered a regular document and an associated document, and therefore should not be split.

[0100] Therefore, in this embodiment, it is assumed that when the key-value extraction result fully meets specific conditions, an additional file is considered to have been found, and this additional file is considered an associated file and integrated again with the previous segmentation candidate bundle. Specifically, the key-value pair "Additional Information: XX Specification" ("Additional Information" is the key, "XX Specification" is the value) of segmentation candidate bundle 1, and the value "XX Specification" is the same as the title of segmentation candidate bundle 2, and segmentation candidate bundle 2 is continuous with segmentation candidate bundle 1. In this case, segmentation candidate bundle 2 is considered an additional file, and segmentation candidate bundle 1 and segmentation candidate bundle 2 are integrated again to form regular file bundle 1. That is, segmentation candidate bundle 1 is set as the beginning, and segmentation candidate bundle 2 is continuous with segmentation candidate bundle 1 to generate regular file bundle 1. Furthermore, in this case, "XX Specification" is considered an identifier. On the other hand, in segmentation candidate bundle 4, the value "Additional Information" is extracted, and it is obvious that it is continuous with the previous segmentation candidate bundle 3. In this case, segmentation candidate bundle 4 is considered an additional file, and segmentation candidate bundle 3 and segmentation candidate bundle 4 are integrated again to form regular file bundle 2. That is, by setting segment candidate bundle 3 as the beginning, segment candidate bundle 4 is made consecutive with segment candidate bundle 3 to generate regular file bundle 2. Furthermore, in this case, "additional data" is treated as an identifier. Alternatively, in the example of regular file bundle 2, the term "additional data" can also be set to display the value of the additional file and registered in the pre-ROM 12, etc.

[0101] Similar to the first embodiment, post-processing can also be performed after the segmentation process in this embodiment. Figure 7 The post-processing involved in this embodiment is illustrated. Figure 7 of (a), Figure 7 (b) and Figure 7(c) shows the results of processes 1, 2, and 3, respectively. The idea and order of the post-processing are the same as in the first embodiment, so only the results are briefly described here.

[0102] <Process 1>

[0103] Regarding regular file bundle 1, such as Figure 7 As shown in (a), candidate bundles 1 and 2 are summarized and saved with the filename "Quotation_XX Specification_No1234". For regular file bundle 2, candidate bundles 3 and 4 are summarized and saved with the filename "Quotation_Budget_No6789".

[0104] <Process 2>

[0105] like Figure 7 As shown in (b), regarding regular file bundle 1, create a folder named "No_1234" and store candidate splitting bundles 1 and 2 in this folder. Then, assign the filename "Quotation_No1234" to candidate splitting bundle 1 and "XX Specification_No1234" to candidate splitting bundle 2. Regarding regular file bundle 2, create a folder named "No6789" and store candidate splitting bundles 3 and 4 in this folder. Then, assign the filename "Quotation_No6789" to candidate splitting bundle 3 and "Budget Sheet" to candidate splitting bundle 4.

[0106] <Process 3>

[0107] like Figure 7 As shown in (c), regarding regular file bundle 1, create a folder for regular file bundle 1 (illustration omitted), and save the splitting candidate bundle 1 in this folder with the filename "Quotation_No1234". Further create a folder named "Quotation_No1234_Additional Information", and save the splitting candidate bundle 2 in this folder with the filename "XX Specification_No1234". Regarding regular file bundle 2, create a folder for regular file bundle 3 (illustration omitted), and save the splitting candidate bundle 3 in this folder with the filename "Quotation_No6789". Further create a folder named "Quotation_No6789_Additional Information", and save the splitting candidate bundle 4 in this folder with the filename "Budget Sheet".

[0108] [Third Implementation]

[0109] refer to Figure 8 and Figure 9The information processing apparatus, information processing system, and information processing program involved in this embodiment will be described. This embodiment is the method used when page numbers are extracted by key-value extraction in the first embodiment described above. Therefore, the information processing apparatus, information processing system, and information processing program involved in this embodiment are basically the same as those in the first embodiment described above, and should be referred to as needed. Figure 1 , 2 5, and omitting detailed explanations.

[0110] Figure 8 The flowchart of the segmentation process involved in this embodiment is shown. Figure 8 <1> shows the file group, i.e., multiple read images, before segmentation. Figure 8 <2> shows the state of two segmentation candidate bundles generated by the header extraction of the file group. Figure 8 <3> shows the state where the results of further key-value extraction are again assembled into a regular file bundle. Furthermore, from an easy-to-understand point of view, this embodiment is illustrated by performing title extraction and key-value extraction in that order, but it is not limited to this; it can be performed in the order of key-value extraction and title extraction, or it can be performed simultaneously.

[0111] like Figure 8 As shown in <2>, the title extraction results in the file group being divided into two candidate splitting bundles, splitting bundle 1 and splitting bundle 2. Candidate splitting bundle 1 is a quotation with the document number "No. 1234", and candidate splitting bundle 2 is a specification document with the title "XX Specification". Here, the page number "1 / 3" is marked in the lower column of candidate splitting bundle 1, and the page number "2 / 3" is marked in the lower column of candidate splitting bundle 2. Therefore, candidate splitting bundle 1 and candidate splitting bundle 2 are considered to be a continuous block of documents, and thus we do not want to split them.

[0112] Therefore, in this embodiment, it is assumed that when the key value extraction result fully meets specific conditions, the page number is considered to have been found. The file at the beginning (in this example, segmentation candidate bundle 1) and the files with consecutive page numbers are considered as associated files and are integrated again with the previous segmentation candidate bundle. Specifically, the page number is considered as one way of using the key value. That is, in the case of segmentation candidate bundle 1, " / " is considered as the key and "1, 3" is considered as the value; in the case of segmentation candidate bundle 2, " / " is considered as the key and "2, 3" is considered as the value. The key values ​​in this way can also be pre-registered in a storage unit such as ROM 12. Thus, segmentation candidate bundle 1 and segmentation candidate bundle 2 with consecutive page numbers are integrated again and set as regular file bundle 1. In addition, the format of the page number key value is not limited to "Y / X" (Y and X are integers), for example, it can also be "Y / ", "Y:X", etc. Furthermore, the determination of the page number can also be limited to the bottom or top of the page. When limited to the bottom or top, the page number can be only a number.

[0113] Similar to the first embodiment, post-processing can also be performed after the segmentation process in this embodiment. Figure 9 The post-processing involved in this embodiment is illustrated. Figure 9 of (a), Figure 9 (b) and Figure 9 (c) shows the results of processes 1, 2, and 3, respectively. The idea and order of the post-processing are the same as in the first embodiment, so only the results are briefly described here.

[0114] <Process 1>

[0115] like Figure 9 As shown in (a), the file name is assigned to "Quotation_No1234", and candidate bundles 1 and 2 are summarized and saved.

[0116] <Process 2>

[0117] like Figure 9 As shown in (b), assign the filename "Quotation_No1234" to candidate bundle 1 and the filename "XX Specification_No1234" to candidate bundle 2, and save both separately. Alternatively, "No1234" can be omitted from each filename.

[0118] <Process 3>

[0119] like Figure 9 As shown in (c), create a folder for regular file bundle 1 (illustration omitted), and save the candidate bundle 1 in this folder with the filename "Quotation_No1234". Further create a folder named "Additional Information_XX Specifications", and save the candidate bundle 2 in this folder with the filename "XX Specifications".

[0120] Furthermore, in the above embodiments, the method of performing a comparison of key values ​​for the entire file group after extracting the title has been described, but it is not limited to this; the scope of files whose key values ​​are compared can also be limited. That is, the comparison of the results of extracting key values ​​for a plurality of segmentation candidate bundles is limited to the key values ​​marked in the pages marked with titles. In other words, the key value extraction results of pages whose titles cannot be extracted are not used for comparison. Thus, the key value comparison between segmentation candidate bundles is limited, and therefore an improvement in processing speed can be expected.

[0121] Furthermore, while the above embodiments exemplify an information processing apparatus that receives data from an external source for a plurality of read images that have undergone OCR processing, the invention is not limited to this. For example, it could be an overall information processing apparatus that incorporates an image reading device and performs segmentation processing. Alternatively, it could be an apparatus that incorporates OCR functionality and receives a plurality of read images that have not undergone OCR processing.

[0122] Furthermore, the structure of the information processing device 10 described in the above embodiment is an example, and it can be changed according to the situation without departing from the main idea.

[0123] Furthermore, the processing flow of the procedure described in the above embodiments is also an example. Without departing from the main idea, unnecessary steps can be deleted, new steps can be added, or the processing order can be changed.

[0124] Furthermore, while the above embodiments describe the information processing program being pre-stored (installed) in ROM 12 or memory 14, this is not a limitation. The program may also be provided as a recording medium such as CD-ROM (Compact Disc Read-Only Memory), DVD-ROM (Digital Versatile Disc Read Only Memory), or USB memory. Additionally, the program may be configured to be downloaded from an external device via a network.

[0125] In the above embodiments, "processor" refers to a processor in a broad sense, including general-purpose processors (e.g., CPU: Central Processing Unit, etc.) and special-purpose processors (e.g., GPU: Graphics Processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic devices, etc.). Furthermore, the actions of the processor in the above embodiments can be executed by a single processor or by multiple processors located in physically separate locations working together. Moreover, the order of the processor's actions is not limited to the order described in the above embodiments and can be appropriately modified.

[0126] The embodiments of the present invention described above are provided for illustrative purposes. Furthermore, these embodiments do not encompass the entirety of the invention, nor do they limit the invention to the disclosed methods. It will be apparent to those skilled in the art that various modifications and variations will be readily understood. These embodiments were chosen and described to most readily explain the principles and applications of the invention. Thus, those skilled in the art can understand the invention through various modifications that are assumed to be optimized for specific uses of various embodiments. The scope of the invention is defined by the foregoing claims and their equivalents.

Claims

1. An information processing device comprising a processor, The processor performs the following processing: Acquire multiple read images by reading a bundle of paper media containing multiple bundles of regular documents, wherein each bundle of regular documents is a set containing regular documents and associated documents related to the regular documents; Extract the subject from the plurality of read images; Extract common elements from the plurality of read images within a regular file bundle, while marking them with different identifiers between different regular file bundles; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images with common identifiers are included in the same bundle, and read images with different identifiers are included in different bundles. The plurality of read images are segmented by bundle.

2. The information processing apparatus according to claim 1, wherein, The processor performs the following processing: The identifier is extracted by performing an extraction process on the plurality of read images, which is performed on at least one of the items and values ​​corresponding to the identifier pre-specified by the user.

3. The information processing apparatus according to claim 1 or 2, wherein, The processor sequentially performs the process of extracting the topic and the process of extracting the identifier.

4. An information processing system comprising: A first information processing apparatus includes a first processor that acquires a plurality of read images obtained by reading a bundle of paper media comprising a plurality of bundles of regular documents, wherein each bundle of regular documents is a set comprising regular documents and associated documents related to the regular documents; and The second information processing device includes a second processor, which performs the following processing: Extracting the subject from the plurality of read images received from the first information processing device; Extract common elements from the plurality of read images within a regular file bundle, while marking them with different identifiers between different regular file bundles; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images with common identifiers are included in the same bundle, and read images with different identifiers are included in different bundles. The plurality of read images are segmented by bundle.

5. A recording medium having recorded an information processing program for causing a computer to perform the following functions: Acquire multiple read images by reading a bundle of paper media containing multiple bundles of regular documents, wherein each bundle of regular documents is a set containing regular documents and associated documents related to the regular documents; Extract the subject from the plurality of read images; Extract common elements from the plurality of read images within a regular file bundle, while marking them with different identifiers between different regular file bundles; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images with common identifiers are included in the same bundle, and read images with different identifiers are included in different bundles. The plurality of read images are segmented by bundle.

6. An information processing device comprising a processor, The processor performs the following processing: Acquire multiple read images by reading a bundle of paper media containing multiple bundles of regular documents, wherein each bundle of regular documents is a set containing regular documents and associated documents related to the regular documents; Extract the subject from the plurality of read images; Extract identifiers containing predefined characters from the plurality of read images; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images that are consecutively arranged from the read images from which the topic has been extracted and contain a common identifier are included in the same bundle. The plurality of read images are segmented by bundle.

7. The information processing apparatus according to claim 6, wherein, The identifier is a character representing an additional file of a read image preceding the read image containing the identifier, or a page number consecutive to the read image preceding the read image containing the identifier.

8. An information processing method, comprising the following steps: Acquire multiple read images by reading a bundle of paper media containing multiple bundles of regular documents, wherein each bundle of regular documents is a set containing regular documents and associated documents related to the regular documents; Extract the subject from the plurality of read images; Extract identifiers that are common within a regular file bundle but marked differently between different regular file bundles from the plurality of read images; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images with common identifiers are included in the same bundle, and read images with different identifiers are included in different bundles. The plurality of read images are segmented by bundle.

9. A computer program product, characterised in that, Includes programs for enabling the computer to perform the following processes: Acquire multiple read images by reading a bundle of paper media containing multiple bundles of regular documents, wherein each bundle of regular documents is a set containing regular documents and associated documents related to the regular documents; Extract the subject from the plurality of read images; Extract identifiers containing predefined characters from the plurality of read images; Based on the extraction results of the topic, the plurality of read images are segmented into a plurality of segmentation candidate bundles. Each segmentation candidate bundle begins with a read image from which the topic has been extracted. Read images that are consecutively arranged from the read images from which the topic has been extracted and contain a common identifier are included in the same bundle. The plurality of read images are segmented by bundle.

Images