Information processing device, information processing system, information processing method, and information processing program.

Abstract

The present invention provides an information processing device that reads the values ​​indicated by blurry barcode images. [Solution] The information processing device includes an acquisition means for acquiring a second signal from a sensor, a machine learning model which is trained to output the first page number by inputting the first signal using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal, and a determination means which determines the second page number that corresponds to the second signal using the second signal.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing system, an information processing method, and an information processing program.

Background Art

[0002] In a passbook issued by a financial institution or the like, a barcode corresponding to the page number of the passbook is described. Patent Document 1 describes that an image of the barcode is acquired by controlling a reading device according to the description method of the barcode.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, when the image of the barcode acquired due to external factors or the like is unclear, the reading device of Patent Document 1 cannot read the value indicated by the barcode.

[0005] The present disclosure aims to provide an information processing apparatus that solves the above problems.

Means for Solving the Problems

[0006] According to one aspect of the present disclosure, an acquisition unit that acquires a second signal from a sensor, a first signal acquired in advance, and a first page number acquired in advance corresponding to the first signal are used, and a machine learning model that executes learning to output the first page number by inputting the first signal, and a determination unit that determines a second page number corresponding to the second signal using the second signal are provided, and an information processing apparatus is provided.

[0007] According to one aspect of the present disclosure, an information processing system is provided comprising: a sensor that acquires a second signal from an object; and an information processing device that outputs a second page number corresponding to the second signal, wherein the information processing device includes: acquisition means for acquiring the second signal from the sensor; a machine learning model that performs learning to output the first page number by inputting the first signal using a first signal acquired in advance and a first page number acquired in advance corresponding to the first signal; and determination means for determining the second page number corresponding to the second signal using the second signal.

[0008] According to one aspect of the present disclosure, an information processing method is provided, comprising: an acquisition step of acquiring a second signal from a sensor using an information processing device; a machine learning model that performs learning to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination step of determining a second page number that corresponds to the second signal using the second signal.

[0009] According to one aspect of the present disclosure, an information processing program is provided which performs the following steps using an information processing device: an acquisition step of acquiring a second signal from a sensor; a machine learning model that performs learning to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance and that corresponds to the first signal; and a determination step of determining the second page number that corresponds to the second signal, using the second signal. [Effects of the Invention]

[0010] According to the present disclosure, it becomes possible to provide an information processing device that reads the value indicated by an unclear barcode image. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing an example of a reading device according to one embodiment of the present invention. [Figure 2] This is a block diagram showing the hardware configuration of a reading device according to one embodiment of the present invention. [Figure 3] This is a block diagram showing the hardware configuration of an information processing device according to one embodiment of the present invention. [Figure 4] This is a block diagram showing the functional section of a reading device according to one embodiment of the present invention. [Figure 5] This is a schematic diagram showing the configuration of an object according to one embodiment of the disclosure of this application. [Figure 6] This is a schematic diagram showing the configuration of an object according to one embodiment of the disclosure of this application. [Figure 7] This is a top view showing an example of the object reading process by a reading device according to one embodiment of the present invention. [Figure 8] This is a schematic diagram showing an example of a barcode according to one embodiment of the present invention. [Figure 9] This is a schematic diagram showing an example of a barcode image according to one embodiment of the present invention. [Figure 10] This is a schematic diagram showing an example of a barcode image according to one embodiment of the present invention. [Figure 11] This is a schematic diagram showing an example of a barcode image according to one embodiment of the present invention. [Figure 12] This flowchart shows an example of object learning processing by a reading device according to one embodiment of the present invention. [Figure 13] This flowchart shows an example of the object reading process by a reading device according to one embodiment of the present invention. [Figure 14] This is a block diagram showing a reading device according to one embodiment of the present invention. [Modes for carrying out the invention]

[0012] Hereinafter, a reading device according to an embodiment of the present disclosure will be described with reference to the drawings. In all the drawings, the same or corresponding components are denoted by the same reference numerals, and common descriptions will be omitted.

[0013] [First Embodiment] FIG. 1 is a schematic diagram showing a reading system according to an embodiment of the present disclosure. The reading system 1 includes a reading device 2, an information processing device 3, and a server 7. The reading device 2 is, for example, a passbook printer that reads information described on an object 5 and newly describes information on the object 5. The object 5 is a paper medium having one or more pages. The reading device 2 is connected so as to be communicable with the information processing device 3. The number of reading devices 2 connected to the information processing device 3 may be plural. For example, when five reading devices 2 are installed in a facility where the reading device 2 is installed, the information processing device 3 may be connected so as to be communicable with the five reading devices 2.

[0014] The user can insert the object 5 into the reading device 2 and describe information on the object 5 via an input IF or the like that the reading device 2 has. The reading device 2 describes information in the recording column of the object 5 so that the information described on the object 5 is in date order and the processed order in which it is recorded. When the object 5 has a plurality of pages and the reading device 2 describes a predetermined number of information in the description column of the object 5 on a certain page, the reading device 2 can describe information on other pages of the object 5.

[0015] When the reading device 2 is, for example, a printer provided in an automatic teller machine (ATM) terminal of a financial institution, the reading device 2 reads information from an object 5 (for example, a passbook or the like) issued by a financial institution or the like and newly describes information on the object 5. The information that the reading device 2 describes on the object 5 may be, for example, the amount and date of deposit into an account, the amount and date of withdrawal from the account, and the account balance.

[0016] If the reader 2 is, for example, a printer installed in a terminal that manages lending history in a library, the reader 2 reads information from the object 5 (for example, a lending record) issued by the library, and writes new information to the object 5. The information that the reader 2 writes to the object 5 may include, for example, the name of the book, the date and time the book was borrowed, and the date and time the book was returned.

[0017] If the reader 2 is, for example, a printer installed in a terminal that manages a user's purchase history of goods at a facility, the reader 2 reads information from the object 5 (for example, a record of goods purchases) issued by the facility and writes new information to the object 5. The information that the reader 2 writes to the object 5 may include, for example, the name of the purchased item, the date and time of purchase, the price of the purchased item, and the remaining balance.

[0018] In this disclosure, this embodiment will be described using the case where the reader device 2 is an ATM terminal of a financial institution as an example.

[0019] The information processing device 3 is a general-purpose computer, and may be, for example, a personal computer, a laptop computer, a tablet computer, a smartphone, a mobile phone, or a workstation. The information processing device 3 transmits the information to be written on the object 5 to the reading device 2. For example, the information processing device 3 can transmit the account balance to be recorded on the object 5 to the reading device 2. The information processing device 3 can receive information to be written on the object 5 and processing requests from the reading device 2. For example, if it is not possible to write information on the object 5 because a predetermined number of pieces of information have been written in the fields of the object 5, the information processing device 3 can receive a request from the reading device 2 to issue a new object 5.

[0020] The information processing device 3 is connected to the server 7 via the network 9 so that they can communicate with each other. If one or more reading devices 2 are installed in multiple facilities, the information processing device 3 installed in one facility may send and receive information about multiple objects 5 from one or more reading devices 2 installed in other facilities via the network 9. For example, the information processing device 3 may receive image data (training data) acquired at other facilities and perform training on a machine learning model at one facility.

[0021] Server 7 is an information processing device, which may be, for example, a personal computer, a workstation, a hardware server, a software server, etc. If Server 7 is a hardware server, it may be, for example, a network server or a cloud server, etc. If Server 7 is a software server, it may be, for example, server software or a server program, etc.

[0022] If the reader 2 is an ATM terminal of a financial institution, the server 7 manages the user's personal information and account information. The server 7 may also perform some or all of the processing that can be performed by the information processing device 3. For example, instead of the information processing device 3, the server 7 may receive image data (training data) from multiple reader devices 2 and perform training on a machine learning model.

[0023] Figure 2 is a block diagram showing the hardware configuration of a reading device according to one embodiment of the present disclosure. The reading device 2 comprises a CPU 21, a ROM 22, a RAM 23, a storage device 24, an input / output interface (IF) 25, a communication interface 26, and a printing unit 27. The CPU 21, ROM 22, RAM 23, storage device 24, input / output interface 25, communication interface 26, and printing unit 27 are interconnected via a bus 29 so as to be able to communicate with each other.

[0024] The CPU 21 is a central processing unit. The CPU 21 controls each part of the read device 2 using application programs. The ROM 22 is read-only memory. The ROM 22 is composed of non-volatile memory and stores application programs for controlling each part of the read device 2. The RAM 23 is random access memory. The RAM 23 provides the memory area necessary for the operation of the CPU 21. The storage device 24 is a mass storage device such as a hard disk drive.

[0025] The input / output IF25 is an input / output interface that transmits and receives data from the user, outputs to the user, and data between the reader 2 and other devices. The input / output IF25 may include a mouse, touch panel, trackball, keyboard, speaker, display, etc. The input / output IF25 may also include multiple displays and multiple speakers. The user can operate the reader 2 via the input / output IF25. The communication IF26 enables data communication between the reader 2 and the information processing device 3 via wired communication and / or wireless communication.

[0026] The printing unit 27 includes a print head, a carriage that mounts the print head and sensors, sensors provided adjacent to the print head, guide rails, and transport rollers. The printing unit 27 may further include hardware configurations found in a typical passbook printer. For example, the printing unit 27 may include an automatic page-turning mechanism. The sensors in the printing unit 27 can acquire barcode images using, for example, a Charge Coupled Device (CCD) system or a laser system. When a CCD system is used for the sensor, the sensor includes a Light Emitting Diode (LED) light source, a mirror, a lens that focuses the reflected light, and a CCD sensor that receives the light from the lens. When a laser system is used for the sensor, the sensor includes a laser light source, a laser light-receiving element, and a movable mirror that reflects the light from the laser light source.

[0027] The printing unit 27 records predetermined information on the object 5, including in the record fields of the object 5. For example, the printing unit 27 can record the amount deposited into the account and the date and time of the deposit, the amount withdrawn from the account and the date and time of the withdrawal, and the account balance. The printing unit 27 also stores images of the object 5 acquired by the sensor in the storage device 24. For example, the printing unit 27 can read the barcode image, page number, and overall image of the object 5 printed on the object 5 and store them in the storage device 24.

[0028] Figure 3 is a block diagram showing the hardware configuration of an information processing device according to one embodiment of the present disclosure. The information processing device 3 comprises a CPU 31, a ROM 32, a RAM 33, a storage device 34, an input / output IF 35, and a communication IF 36. The CPU 31, ROM 32, RAM 33, storage device 34, input / output IF 35, and communication IF 36 are interconnected via a bus 39 so as to be able to communicate with each other.

[0029] The CPU 31 is the central processing unit. The CPU 31 controls each part of the information processing unit 3 using application programs. The ROM 32 is read-only memory. The ROM 32 is composed of non-volatile memory and stores application programs for controlling each part of the information processing unit 3. The RAM 33 is random access memory. The RAM 33 provides the memory area necessary for the operation of the CPU 31. The storage device 34 is a mass storage device such as a hard disk drive.

[0030] The input / output interface 35 is an input / output interface that transmits and receives data from the user, outputs to the user, and data between the information processing device 3 and other devices. The input / output interface 35 may include a mouse, touch panel, trackball, keyboard, speaker, display, etc. The input / output interface 35 may also include multiple displays and multiple speakers. The user can operate the information processing device 3 via the input / output interface 35. The communication interface 36 enables data communication between the information processing device 3 and the reading device 2, and between the information processing device 3 and the server 7, via wired communication and / or wireless communication.

[0031] Figure 4 is a block diagram showing the functional unit of a reading device according to one embodiment of the present disclosure. The functional unit 200 comprises a control unit 201, an acquisition unit 203, a learning unit 205, a determination unit 207, an output unit 209, and a database 211. The functional unit 200 and other functions included in the functional unit 200 can be realized by the CPU 21 executing a program. Note that some of the functions of the functional unit 200 may be realized not only in the reading device 2, but also in the information processing device 3 or the server 7. For example, the functions of the acquisition unit 203 may be realized in the information processing device 3, and the functions of the learning unit 205, the determination unit 207, and the database 211 may be realized in the server 7.

[0032] The control unit 201 controls each function of the functional unit 200. For example, the control unit 201 manages the transmission and reception of data between the functional units of the functional unit 200, the processing order, and the updating of the database 211.

[0033] The acquisition unit 203 acquires an image of the barcode of object 5 from the printing unit 27. Furthermore, if the printing unit 27 reads the page number of object 5, the acquisition unit 203 can acquire the page number of object 5 from the printing unit 27.

[0034] For example, when training a machine learning model in the learning unit 205, which will be described later, the acquisition unit 203 acquires the barcode image of the object 5 read by the printing unit 27 (first image) and the page number corresponding to the barcode image of the object 5 (first page number). When training a machine learning model, the page number corresponding to the barcode image of the object 5 is known. The page number corresponding to the barcode image of the object 5 may be acquired in advance. The page number corresponding to the barcode image of the object 5 may be acquired by taking an image of the object 5 using a camera or the like. That is, when training a machine learning model, the acquisition unit 203 acquires the barcode image of the object 5 and the page number corresponding to the barcode image of the object 5 as training data. The acquisition unit 203 transmits the barcode image and the page number corresponding to the barcode image to the learning unit 205.

[0035] On the other hand, when the determination unit 207, described later, uses a machine learning model to determine the page number corresponding to the barcode image of object 5, the acquisition unit 203 acquires the barcode image of object 5 (second image) read by the printing unit 27. The acquisition unit 203 transmits the barcode image to the determination unit 207.

[0036] The learning unit 205 performs machine learning on a machine learning model. The learning unit 205 receives barcode images and page numbers corresponding to the barcode images as training data from the acquisition unit 203. The machine learning model may be, for example, a neural network having an input layer, hidden layers, and output layers. The learning unit 205 inputs the barcode images into the machine learning model. The machine learning model estimates and outputs the page numbers based on the barcode images. The learning unit 205 calculates the error between the output page numbers and the page numbers received from the acquisition unit 203. The learning unit 205 uses the error to train the machine learning model. For example, the learning unit 205 trains the machine learning model using methods such as the steepest descent method, the least mean squares method, the learning identification method, or the backpropagation method.

[0037] When the machine learning model has been sufficiently trained, for example, when the error between the page number output by the machine learning model and the page number received from the acquisition unit 203 becomes sufficiently small on average, the machine learning model can output the page number corresponding to the barcode image when a barcode image is input. When some or all of the training data received from the acquisition unit 203 has been trained, or when the calculated error is smaller than a predetermined threshold, the learning unit 205 transmits the trained machine learning model to the judgment unit 207.

[0038] The determination unit 207 uses a machine learning model to determine the page number corresponding to the barcode image. The determination unit 207 receives a trained machine learning model from the learning unit 205. The determination unit 207 receives the barcode image from the acquisition unit 203. The determination unit 207 inputs the received barcode image into the machine learning model and obtains the output of the machine learning model. The machine learning model outputs the page number (second page number) corresponding to the input barcode image. The determination unit 207 determines that the page number output by the machine learning model is the page number corresponding to the barcode image. The determination unit 207 transmits the determined page number to the output unit 209. The determination unit 207 may also transmit the received barcode image to the output unit 209. The determination unit 207 may also determine the direction of the barcode pattern drawn on the object 5. For example, the determination unit 207 can receive the barcode image from the acquisition unit 203 and determine the direction of the barcode pattern based on the characteristics of the barcode image.

[0039] The output unit 209 outputs the determination result. The output unit 209 receives the determined page number and barcode image from the determination unit 207. The output unit 209 may display the received page number and barcode image on a display or the like via the input / output IF 25. The output unit 209 may also transmit the received page number to the printing unit 27. The printing unit 27 receives the page number from the output unit 209 and can perform printing based on the received page number.

[0040] Database 211 stores training data used in the learning unit 205. For example, database 211 stores barcode images and page numbers corresponding to the barcode images. Database 211 may store barcode images classified into multiple attributes. These classifications may be barcodes of objects 5 issued by different financial institutions, for example, barcodes of objects 5 issued by financial institution X1, financial institution X2, regional bank X3, credit union X4, etc. Furthermore, these classifications may include barcode images read under ideal conditions, barcode images read from soiled objects 5, barcode images read from wet objects 5, etc. Database 211 may also contain barcode printing data used when printing barcodes on objects 5.

[0041] Figures 5 and 6 are schematic diagrams showing the configuration of an object according to one embodiment of the present disclosure. The object 5 is a paper medium having a front cover, a back cover, and one or more pages. The longitudinal direction of the object 5 is the X direction, and the direction perpendicular to the X direction is the Y direction. The direction perpendicular to the plane formed by the X and Y directions is the Z direction. If the object 5 is a passbook issued by a financial institution, the object 5 has a barcode 51, a page number 52, and a record field 53 on one or more pages, as shown in Figures 5 and 6. The barcode 51 is a one-dimensional identifier that shows a numerical value using patterned lines, and is printed on the object 5 so as not to overlap with the record field 53. The page number 52 is the page number corresponding to the numerical value shown by the barcode 51. The reader 2 can record, for example, the amount deposited into the account and the date and time of deposit, the amount withdrawn from the account and the date and time of withdrawal, and the account balance, etc., along the X direction in the record field 53.

[0042] In Figure 5, the pattern of barcode 51A is drawn along the Y direction with predetermined intervals along the X direction. In Figure 6, the pattern of barcode 51B is drawn along the X direction with predetermined intervals along the Y direction. For example, depending on the financial institution that issues object 5, the position of barcode 51 on object 5 and the direction of the pattern of barcode 51 may differ. Similarly, depending on the financial institution that issues object 5, the position of page number 52 on object 5 may also differ. This embodiment will be described assuming that barcode 51 on object 5 is located in the upper left of object 5 and page number 52 is located in the upper right of object 5.

[0043] Figure 7 is a top view showing an example of object reading processing by a reading device according to one embodiment of the present invention. Figure 7 shows the positional relationship between the components of the printing unit 27 and the object 5 when the printing unit 27 of the reading device 2 reads the barcode 51A contained in the object 5. As described above, the printing unit 27 has a print head 271, a carriage 273, a sensor 275, a guide rail 277, and transport rollers 279A and 279B.

[0044] The print head 271 prints the date, numbers, strings of characters, symbols, etc., into the record field 53 of the object 5. The print head 271 may be equipped with an ink tank or an ink ribbon. The carriage 273 reciprocates the print head 271 in the X direction. The carriage 273 is a component that transports the print head 271 and the sensor 275, and is installed on a guide rail 277, and can reciprocate in the X direction along the guide rail 277.

[0045] Sensor 275 is a barcode scanner having an LED light source and a light sensor. In this embodiment, sensor 275 is installed on carriage 273. For example, if barcode 51 is barcode 51A as shown in Figure 5, sensor 275 can read barcode 51A in response to the reciprocating motion of carriage 273. For example, if barcode 51 is barcode 51B as shown in Figure 6, sensor 275 can read barcode 51B in response to the transport of object 5 by transport rollers 279A and 279B.

[0046] When the printing unit 27 uses a CCD system to read the barcode 51, the sensor 275 may include an illumination LED, a CCD sensor, a lens, and a mirror. When the printing unit 27 uses a laser system to read the barcode, the sensor 275 may include a laser light source, a laser light receiving element, and a movable mirror. When the printing unit 27 uses either a CCD system or a laser system, the sensor 275 may be fixedly installed near the guide rail 277.

[0047] The guide rail 277 is a rail for the carriage 273 to reciprocate along the X direction and is provided along the X direction. The guide rail 277 is located above the object 5 that is conveyed by the conveying rollers 279A and 279B. It is desirable that the length of the guide rail 277 in the X direction be sufficiently longer than the length of the object 5 in the X direction.

[0048] The transport roller 279 is a transport mechanism having two rollers that overlap in the Z direction, and can grasp the object 5 with the two rollers and transport the object 5 along the Y direction. Transport rollers 279A and 279B may have similar configurations. Transport rollers 279A and 279B rotate around a rotation axis provided along the X direction, and can transport the object 5 into the reading device 2 and transport the object 5 out of the reading device 2.

[0049] When the user inserts the object 5 into the printing unit 27, the transport roller 279A rotates, moving the object 5 along the Y-direction toward the guide rail 277 and the transport roller 279B. When the upper end of the object 5 reaches the guide rail 277, the carriage 273 reciprocates in the X-direction, and the sensor 275 detects the position of the barcode 51A. Once the sensor 275 detects the position of the barcode 51A, the carriage 273 reciprocates in the X-direction toward the position of the barcode 51A and its vicinity, and the sensor 275 reads the barcode 51A. The image of the barcode 51A read by the sensor 275 (in this case, a one-dimensional image) is transmitted to the acquisition unit 203. In this way, the printing unit 27 reads the barcode 51 of the object 5.

[0050] Figure 8 is a schematic diagram showing an example of a barcode according to one embodiment of the present disclosure. In Figure 8, the structure of a barcode will be explained using barcode 51A as an example. Barcode 51A is a one-dimensional identifier that shows a numerical value using patterned lines. Barcode 51A is represented by a combination of a black bar 51b with width BW and a blank bar 51w with width BW. Bar 51b represents "1" and bar 51w represents "0". For example, if barcode 51A is represented by two bars 51b, one bar 51w, and three bars 51b, the value represented by barcode 51A is "110111". In this way, barcode 51A can represent a value in binary by combining bars 51b and bars 51w. The barcode 51A shown in Figure 8 is "101011011011011010101". Light emitted from the LED light source of sensor 275 causes weak reflection of light at bar 51b and strong reflection at bar 51w. The light sensor of sensor 275 receives the reflected light and can read the barcode image.

[0051] Figures 9, 10, and 11 are schematic diagrams showing examples of barcode images according to one embodiment of the present disclosure. Figure 9 shows an image of a normal barcode 51A, in which bar 51b is not blurred and bar 51w is not blackened. Figure 10 shows an image of a barcode 51A, in which bar 51b is blurred and bar 51w is not blackened. Figure 11 shows an image of a barcode 51A, in which bar 51b is not blurred and bar 51w is blackened.

[0052] In Figure 9, when distinguishing between bars 51b and 51w using thresholds and determining the value of barcode 51A, the value of barcode 51A can be correctly determined whether a threshold THH, which is close to the white level, or a threshold THL, which is close to the black level, is used. For an image of barcode 51A like the example in Figure 9, a threshold can be selected over a wide range from white to black levels.

[0053] On the other hand, in Figure 10, using the threshold THH allows for the correct determination of the value of barcode 51A. However, using the threshold THL, the value of barcode 51A cannot be correctly determined. Similarly, in Figure 11, using the threshold THL allows for the correct determination of the value of barcode 51A. However, using the threshold THH, the value of barcode 51A cannot be correctly determined. In the example in Figure 10, the black level value of bar 51b becomes close to the white level, narrowing the range from white to black. In the example in Figure 11, the white level value of bar 51w becomes close to the black level, narrowing the range from white to black. Thus, it can be seen that when determining barcode 51A using thresholds, the value of barcode 51 may be misdetermined if the image of barcode 51 becomes unclear due to dirt or other reasons.

[0054] Figure 12 is a flowchart showing an example of object learning processing by a reading device according to one embodiment of the present disclosure. In Figure 12, the reading device 2 reads multiple barcodes contained in the object 5, and the user inputs the page number corresponding to the image of the barcode 51, thereby performing training of the machine learning model.

[0055] The user inserts one of the multiple objects 5 into the reading device 2 (step S101). When the user inserts an object 5, the printing unit 27 of the reading device 2 rotates the transport roller 279A as described in Figure 7, and moves the object 5 toward the guide rail 277 and the transport roller 279B.

[0056] When the transport roller 279A moves the object 5 to a predetermined position, the sensor 275 detects the position of the barcode 51 (step S103). In Figure 7, the carriage 273 moves back and forth in the X direction, and the LED light source of the sensor 275 illuminates the barcode 51. The acquisition unit 203 acquires an image of the barcode 51 from the sensor 275 based on the reflected light. The acquisition unit 203 transmits the image of the barcode 51 to the determination unit 207.

[0057] If the image of the barcode 51 based on the acquired reflected light includes areas with and without black lines, as shown in Figure 9, the determination unit 207 determines that the barcode 51 is a barcode 51A as shown in Figure 5. In other words, the determination unit 207 determines that the pattern of the barcode 51 is drawn along the Y direction (see barcode 51A in Figure 5; hereinafter referred to as "vertical barcode 51A").

[0058] On the other hand, if the image of the barcode 51 based on the acquired reflected light includes only parts with black lines, parts without black lines, or if it is not possible to clearly distinguish between parts with black lines and parts without black lines, the determination unit 207 determines that the barcode 51 is a barcode 51B as shown in Figure 6. That is, the determination unit 207 determines that the pattern of the barcode 51 is drawn along the X direction (see barcode 51B in Figure 6; hereinafter referred to as "horizontal barcode 51B"). The determination unit 207 transmits the determination result to the control unit 201. Based on the determination result, the control unit 201 controls the hardware included in the printing unit 27.

[0059] In step S103, if the barcode 51 is a vertical barcode 51A, the carriage 273 reciprocates in the X direction, and the sensor 275 reads the image of the barcode 51. In step S103, if the barcode 51 is a horizontal barcode 51B, the carriage 273 remains at the position of the barcode 51B, the transport roller 279A rotates, transporting the object 5 toward the guide rail 277 and the transport roller 279B, and the sensor 275 reads the image of the barcode 51 (step S105). The read barcode 51 image is transmitted to the acquisition unit 203. The reading device 2 may also display the page number entered by the user via the input / output IF 25 along with the image of the barcode 51 when it reads the image of the barcode 51.

[0060] The user inputs the page number 52 corresponding to the image of the barcode 51 to the reader 2 via the input / output IF 25 or the like (step S107). Alternatively, the page number 52 corresponding to the barcode image may be acquired in advance by imaging the object 5 using a camera or the like. Or, the reader 2 may be connected to another device via the input / output IF 25 or communication IF 26, and the page number 52 may be automatically input by the other device. The input page number 52 is transmitted to the acquisition unit 203. The acquisition unit 203 associates the image of the barcode 51 with the page number 52 entered by the user and transmits it to the learning unit 205.

[0061] When the learning unit 205 receives the image of the barcode 51 and the page number 52 from the acquisition unit 203, it inputs the image of the barcode 51 into the machine learning model. The machine learning model calculates weights in the input layer, hidden layer, and output layer, estimates the page number according to the image of the barcode 51, and outputs it. The learning unit 205 calculates the error between the output page number and the page number 52 received from the acquisition unit 203. The learning unit 205 uses the calculated error to train the machine learning model (step S109). When training the machine learning model, for example, based on the above error, the steepest descent method, least mean squares method, learning identification method, or backpropagation method can be used.

[0062] For each page included in object 5, the process from step S103 to step S109 is repeated. Once the process from step S103 to step S109 is completed for one object 5, the process from step S103 to step S109 is executed for the other objects 5. In this way, the learning unit 205 trains the machine learning model for multiple objects 5.

[0063] The learning unit 205 performs training of the machine learning model on a predetermined number of objects 5, evaluates the performance of the machine learning model, and determines whether the machine learning model has been sufficiently trained. For example, the learning unit 205 inputs an image of a barcode 51 having representative features into the machine learning model. The learning unit 205 calculates the error between the page number output by the machine learning model and the page number 52 corresponding to the image of the barcode 51 (step S111). A barcode 51 having representative features includes, for example, a newly issued barcode that is free from damage, a barcode with dirt on the printed area, a barcode with a low black level (e.g., the printing is faded), and a barcode with a small difference between the black level and the white level (e.g., the printed area is darkened).

[0064] If the calculated error is smaller than a predetermined threshold (YES in step S111), the learning unit 205 determines that the machine learning model has been sufficiently trained and transmits the trained machine learning model to the determination unit 207. On the other hand, if the calculated error is smaller than a predetermined threshold (NO in step S111), the learning unit 205 determines that the machine learning model has not been sufficiently trained. The learning unit 205 may also request the user to insert the object 5 into the reading device 2 via the input / output IF 25. In step S111, the learning unit 205 determines whether or not the machine learning model has been sufficiently trained, but the determination of whether or not the training has been sufficiently performed may be made by a method other than the one described above. For example, the learning unit 205 may make the determination based on the number of object 5 used to train the machine learning model. For example, the number of object 5 may be 50, 100, or 1000 or more. Alternatively, the learning unit 205 may make the determination based on the type of object 5 used for machine learning, etc. For example, the types of object 5 may include object 5 issued by multiple different financial institutions, object 5 with uneven surfaces due to water damage, object 5 with dirt on the barcode 51, object 5 with a barcode 51 that is unevenly printed, etc.

[0065] The learning unit 205 may train a machine learning model under conditions such as when the light intensity of the LED light source of sensor 275 is less than a predetermined light intensity, when the light intensity of the LED light source of sensor 275 is greater than a predetermined light intensity, or when dirt or other substances are attached to the light sensor. Alternatively, the acquisition unit 203 may read a training dataset that has been acquired in advance and contains multiple sets of barcode 51 images and corresponding page numbers 52, and train a machine learning model.

[0066] Figure 13 is a flowchart illustrating an example of the object reading process by a reading device according to one embodiment of the present disclosure. In Figure 13, the reading device 2 reads multiple barcodes contained in the object 5, and the determination unit 207 determines the page number corresponding to the image of the barcode 51. The flowchart in Figure 13 will be explained assuming that the barcode 51 on the object 5 is a vertical barcode 51A.

[0067] The user inserts one of the multiple objects 5 into the reading device 2 (step S201). The process in step S201 is the same as the process in step S101 in Figure 12. When the transport roller 279A moves the object 5 to a predetermined position, the sensor 275 detects the position of the barcode 51 (step S203). The process in step S203 is the same as the process in step S103 in Figure 12. The carriage 273 moves back and forth in the X direction, and the sensor 275 reads the image of the barcode 51 (step S205). The image of the read barcode 51 is transmitted to the acquisition unit 203, and from the acquisition unit 203 it is transmitted to the determination unit 207.

[0068] When the determination unit 207 receives an image of the barcode 51 from the acquisition unit 203, it inputs the image of the barcode 51 into a trained machine learning model (step S207). The machine learning model outputs a page number 52 corresponding to the image of the barcode 51, and the determination unit 207 acquires the output of the machine learning model as the determination result (step S209). The determination unit 207 transmits the output of the machine learning model as the determined page number 52 to the output unit 209. The output unit 209 outputs the page number 52 received from the determination unit 207 to the control unit 201 or the CPU 21 of the reader 2 (step S211). For example, if the account balance etc. is to be written on the third page of the object 5, the process from step S203 to step S211 is repeated until the third page of the object 5 is reached. After that, the printing unit 27 performs the printing process (step S213).

[0069] In step S209, if the determination unit 207 is unable to determine page number 52, it may temporarily suspend the determination of page number 52. For example, if the determination unit 207 is unable to determine page number 52 corresponding to barcode 51 on the second page of object 5, it sends a message to the control unit 201 stating that it was unable to determine page number 52 and that it will proceed with processing the next page (corresponding to the processing from step S203 to step S211). The printing unit 27 then processes the next page (i.e., the third page), and the determination unit 207 determines page number 52. Based on the determination result for page number 52 (the third page number) on the third page, the determination unit 207 may determine the suspended page number 52 (the second page number). In other words, if the determination unit 207 is unable to determine page number 52 on a particular page, the printing unit 27 can determine the undetermined page number 52 based on page number 52 on other pages.

[0070] In this embodiment, the training of the machine learning model shown in Figure 12 and the determination using the machine learning model shown in Figure 13 are performed separately. However, the machine learning model may be trained using the image of the barcode 51 acquired by the acquisition unit 203 while the determination using the machine learning model is being performed. Here, the image of the barcode 51 acquired by the acquisition unit 203 and the page number 52 determined by the determination unit 207 are transmitted to the information processing device 3 or server 7, and after a sufficient amount of training data has been accumulated, the information processing device 3 or server 7 may perform the training of the machine learning model.

[0071] In this way, according to one embodiment of the present invention, a machine learning model is trained using a barcode image and the page number corresponding to the barcode image, and the reader determines the page number corresponding to the barcode image using the trained machine learning model. This makes it possible for the reader to read the page number corresponding to the barcode image even if the barcode image is unclear and contains noise or other imperfections. Furthermore, by using objects with different specifications as training data for the machine learning model, it becomes unnecessary to incorporate software to control the reader for each specification.

[0072] [Other embodiments] Figure 14 is a block diagram showing a reading device according to one embodiment of the present disclosure. The reading device 1000 comprises an acquisition means 1001 and a determination means 1003. The acquisition means 1001 acquires a second signal from a sensor. The determination means 1003 uses a first signal acquired in advance, a first page number acquired in advance that corresponds to the first signal, a machine learning model that has been trained to output a first page number when the first signal is input, and a second signal to determine a second page number corresponding to the second signal.

[0073] Furthermore, the processing method of recording a program that operates the configuration of each embodiment to realize the functions of each embodiment on a recording medium, reading the program recorded on the recording medium as code, and executing it on a computer is also included in the scope of each embodiment. In other words, computer-readable recording media are also included in the scope of each embodiment. Moreover, not only the recording medium on which the above-mentioned computer program is recorded, but also the computer program itself is included in each embodiment.

[0074] Examples of recording media that can be used include floppy disks, hard disks, optical disks, magneto-optical disks, CD-ROMs (Compact Disc-Read Only Memory), magnetic tapes, non-volatile memory cards, and ROMs. Furthermore, the scope of each embodiment is not limited to programs that perform processing on the recording media alone, but also includes programs that operate on an OS (Operating System) in cooperation with other software and the functions of expansion boards to perform processing.

[0075] Although the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and detailed description of the present disclosure can be made that will be understood by those skilled in the art within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0076] Some or all of the above embodiments may also be described as follows, but are not limited to the following:

[0077] (Note 1) An acquisition means for acquiring a second signal from a sensor, An information processing device comprising: a machine learning model that performs learning to output the first page number by inputting the first signal, using a first signal acquired in advance and a first page number acquired in advance corresponding to the first signal; and a determination means that determines the second page number corresponding to the second signal using the second signal.

[0078] (Note 2) The information processing apparatus according to Appendix 1, wherein the determination means determines the second page number based on the output of the machine learning model obtained by inputting the second signal to the machine learning model.

[0079] (Note 3) By inputting the first signal to the machine learning model, an output is obtained from the machine learning model. The error between the output of the machine learning model and the first page number is calculated. An information processing device as described in Appendix 1, which performs training of the machine learning model in such a way as to minimize the aforementioned error.

[0080] (Note 4) The information processing device described in Appendix 3, wherein the machine learning model is trained using one of the following methods: gradient descent, least mean squares, learning identification, or backpropagation.

[0081] (Note 5) The information processing device described in Appendix 1 comprises a light source that emits light and a light-receiving element that receives reflected light.

[0082] (Note 6) The information processing apparatus described in Appendix 1 comprises a light source for irradiating light, a Charge Coupled Device (CCD) sensor, a lens, and a mirror.

[0083] (Note 7) The sensor is an information processing device as described in Appendix 1, comprising a laser light source, a light-receiving element, and a movable mirror.

[0084] (Note 8) The first signal is a signal obtained by reading a one-dimensional barcode contained in the first object, and the first page number indicates the page number of the first object. The information processing device described in Appendix 1, wherein the second signal is a signal obtained by reading a one-dimensional barcode contained in the second object, and the second page number indicates the page number of the second object.

[0085] (Note 9) If the determination means is unable to determine the second page number on one page of the second object, the determination means determines the third page number on another page of the second object. The information processing device described in Appendix 8, which determines the second page number based on the third page number.

[0086] (Note 10) The information processing device described in Appendix 8, wherein the first object and the second object are one of the following: a passbook issued by a financial institution, a book lending record issued by a library, or a purchase record of an item issued by an institution.

[0087] (Note 11) A sensor that acquires a second signal from the object, The system comprises an information processing device that outputs a second page number corresponding to the second signal, The aforementioned information processing device is An acquisition means for acquiring the second signal from the sensor, An information processing system comprising: a machine learning model that performs training to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination means that determines the second page number that corresponds to the second signal, using the second signal.

[0088] (Note 12) By an information processing device, An acquisition step to obtain a second signal from the sensor, An information processing method comprising: a machine learning model that is trained to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination step that determines the second page number that corresponds to the second signal, using the second signal.

[0089] (Note 13) By an information processing device, An acquisition step to obtain a second signal from the sensor, An information processing program that performs the following steps: a machine learning model that is trained to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination step that determines the second page number that corresponds to the second signal, using the second signal. [Explanation of symbols]

[0090] 1: Reading System 2: Reading device 21: CPU 22 :ROM 23: RAM 24:Storage device 25: Input / Output Interface 26: Communication Interface 27:Printing Department 201: Control Unit 203: Acquisition Department 205: Learning Department 207: Judgment section 209: Output section 211: Database 271: Print head 273: Carriage 275: Sensor 277: Guide rail 279: Conveyor roller 3: Information Processing Device 5: Object 51: Barcode 52: Page number 53: Record section 7: Server 9: Network

Claims

1. An acquisition means for acquiring a second signal from a sensor, An information processing device comprising: a machine learning model that performs learning to output the first page number by inputting the first signal, using a first signal acquired in advance and a first page number acquired in advance corresponding to the first signal; and a determination means that determines the second page number corresponding to the second signal using the second signal.

2. The information processing apparatus according to claim 1, wherein the determination means determines the second page number based on the output of the machine learning model obtained by inputting the second signal to the machine learning model.

3. By inputting the first signal to the machine learning model, an output is obtained from the machine learning model. The error between the output of the machine learning model and the first page number is calculated. The information processing apparatus according to claim 1, wherein the machine learning model is trained in such a way that the error is minimized.

4. The information processing apparatus according to claim 3, wherein the machine learning model is trained using one of the following methods: gradient descent, least mean squares, learning identification, or backpropagation.

5. The first signal is a signal obtained by reading a one-dimensional barcode contained in the first object, and the first page number indicates the page number of the first object. The information processing device according to claim 1, wherein the second signal is a signal obtained by reading a one-dimensional barcode contained in the second object, and the second page number indicates the page number of the second object.

6. If the determination means is unable to determine the second page number on one page of the second object, the determination means determines the third page number on another page of the second object. The information processing apparatus according to claim 5, which determines the second page number based on the third page number.

7. The information processing device according to claim 5, wherein the first object and the second object are one of the following: a passbook issued by a financial institution, a book lending record issued by a library, or a purchase record of an item issued by an institution.

8. A sensor that acquires a second signal from the object, The system comprises an information processing device that outputs a second page number corresponding to the second signal, The aforementioned information processing device is An acquisition means for acquiring the second signal from the sensor, An information processing system comprising: a machine learning model that performs learning to output the first page number by inputting the first signal, using a first signal acquired in advance and a first page number acquired in advance corresponding to the first signal; and a determination means that determines the second page number corresponding to the second signal using the second signal.

9. By an information processing device, An acquisition step to obtain a second signal from the sensor, An information processing method comprising: a machine learning model that is trained to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination step that determines the second page number that corresponds to the second signal, using the second signal.

10. By an information processing device, An acquisition step to obtain a second signal from the sensor, An information processing program that performs the following steps: a machine learning model that is trained to output the first page number by inputting the first signal, using a first signal that has been acquired in advance and a first page number that has been acquired in advance corresponding to the first signal; and a determination step that determines the second page number that corresponds to the second signal, using the second signal.

Images