Image proofing device, inference device, machine learning device, image proofing method, inference method, and machine learning method

Abstract

To provide an image proofing device that enables simple and appropriate image proofing.SOLUTION: An information processing device 4 includes, as an image proofing function 402, a data acquisition unit 4020 configured to acquire manuscript image data 11 created as a manuscript for an image to be printed by a printing device on a printing surface of a print object, and a data generation unit 4021 configured to input the manuscript image data 11 acquired by the data acquisition unit 4020 into a learning model 16 to generate proofed image data 12 by performing predetermined proofing processing on the manuscript image data 11. The learning model 16 is a trained model that has learned the correlation between the manuscript image data 11 and the proofed image data 12 through machine learning.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to an image proofreading device, an inference device, a machine learning device, an image proofreading method, an inference method, and a machine learning method. [Background technology]

[0002] Conventionally, when an image is printed on a container, an operator proofreads the image based on a sample pattern image submitted by a customer (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-155771 Summary of the Invention [Problem to be solved by the invention]

[0004] Image proofreading is performed using, for example, an image editing program, but because a test print is made based on the proofread image and confirmation by the worker or the customer is required, the image may need to be proofread multiple times. Furthermore, since the image proofreading requires checking various items and proofreading the image, it is highly dependent on the experience and intuition of a skilled worker, making it an extremely difficult task.

[0005] The present invention has been made to solve the above-mentioned problems, and aims to provide an image proofreading device, an inference device, a machine learning device, an image proofreading method, an inference method, and a machine learning method that enable easy and appropriate image proofreading. [Means for solving the problem]

[0006] In order to achieve the above object, an image proofing device according to one aspect of the present invention comprises: a data acquisition unit that acquires original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object; a data generation unit that generates proofread image data by performing a predetermined proofreading process on the original image data by inputting the original image data acquired by the data acquisition unit into a learning model, The learning model is This is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the proof image data. [Effects of the Invention]

[0007] According to one aspect of the image proofreading device of the present invention, by inputting original image data into a learning model, proofread image data is generated by performing a predetermined proofreading process on the original image data, thereby making it possible to proofread images easily and appropriately.

[0008] Problems, configurations, and effects other than those described above will become apparent from the detailed description of the invention that follows. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is an overall schematic diagram showing an example of a printing system 1. FIG. [Figure 2] FIG. 2 is a block diagram showing an example of an information processing device 4. [Figure 3] FIG. 4 is a data configuration diagram showing an example of an image management database 410. [Figure 4] FIG. 2 is a diagram showing an example of document image data 11. [Figure 5] FIG. 2 is a diagram showing an example of proofread image data 12. [Figure 6A] FIG. 4 is a functional explanatory diagram showing an example of a learning function 401. [Figure 6B] FIG. 4 is a functional explanatory diagram showing an example of a learning function 401. [Figure 6C] FIG. 4 is a functional explanatory diagram showing an example of a learning function 401. [Figure 6D] FIG. 4 is a functional explanatory diagram showing an example of a learning function 401. [Figure 7] FIG. 4 is a functional explanatory diagram showing an example of an image proofreading function 402. [Figure 8] FIG. 9 is a hardware configuration diagram showing an example of a computer 900. [Figure 9] 10 is a flowchart showing an example of a machine learning method by the learning function 401. [Figure 10] 10 is a flowchart showing an example of an image proofreading method performed by the image proofreading function 402. [Figure 11] FIG. 10 is a screen configuration diagram showing an example of an image proofreading input screen 17. [Figure 12] FIG. 10 is a screen configuration diagram showing an example of an image proofreading output screen 18. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment for carrying out the present invention will be described with reference to the drawings. The scope necessary for the explanation to achieve the object of the present invention will be schematically shown, and the scope necessary for explaining the relevant part of the present invention will be mainly explained, and the parts that are omitted from the explanation will be based on publicly known techniques.

[0011] 1 is an overall schematic diagram showing an example of a printing system 1. The printing system 1 is a system for performing a process of printing an image 100 on the outer peripheral surface of a container 10 based on proofread image data 12 obtained by performing a predetermined proofreading process on manuscript image data 11 created as a manuscript of the image 100 to be printed on the outer peripheral surface of the container 10. A container processing system that processes raw materials into the container 10 is used in a pre-process of the printing system 1. A container filling system that fills the container 10 with contents is used in a post-process of the printing system 1.

[0012] The container 10 is manufactured by any processing method using a base material such as a metal material (e.g., aluminum, steel, tinplate, etc.), an organic material (e.g., resin, etc.), a ceramic material (e.g., glass, etc.), or a paper material. The container 10 may be any base material covered with a film made of PET, PP, PE, etc., or may be made using an aluminum-deposited film, a white film, a tack label (synthetic paper), etc. Examples of the container 10 include, but are not limited to, cans, bottles, cups, tubes, etc. Examples of contents to be filled into the container 10 include, but are not limited to, beverages, food, cosmetics, detergents, medicines, etc. In this embodiment, the container 10 will be described mainly as a beverage can.

[0013] The printing system 1 includes a printing device 2, a platemaking device 3, an information processing device 4, and a user terminal device 5. Each of the devices 2 to 5 is configured, for example, as a general-purpose or dedicated computer (see FIG. 8 described later), and is connected to a wired or wireless network 6 so that various data can be transmitted and received between them. Note that the number of the devices 2 to 5 and the connection configuration of the network 6 are not limited to the example in FIG. 1 and may be changed as appropriate.

[0014] The printing device 2 is a device that prints an image 100 on the outer peripheral surface of the container 10. The printing device 2 is configured, for example, as a printing machine that performs plate printing using printing plates, or a printing machine that performs plateless printing without using printing plates.

[0015] Plate printing machines use relief printing, lithography, intaglio printing, and stencil printing. Relief printing methods are classified into, for example, flexography and offset printing. Lithography Methods are classified into, for example, water-based offset printing and waterless offset printing. Intaglio printing methods are classified into, for example, gravure printing and intaglio printing. Stencil printing methods are classified into, for example, screen printing and mimeograph printing. Printing methods for plateless printing presses are classified into, for example, inkjet printing and electrophotographic printing.

[0016] The printing device 2 uses process color inks such as cyan (C), magenta (M), yellow (Y), and black (K), but spot color inks other than process colors may be used in addition to or instead of the process color inks. The printing device 2 is not limited to a device that directly prints the image 100 on the outer surface of the container 10, but may also be a device that prints the image 100 on a film and then places the film on the outer surface of the container 10.

[0017] The platemaking device 3 is a device that makes printing plates to be used in the printing device 2 that performs plate-based printing. The platemaking device 3 makes printing plates using a method such as DLE (Direct Laser Engraving) or LAMS (Laser Ablation Masking System).

[0018] The information processing device 4 includes an image management database 410 that can register, in association with each other, manuscript image data 11 and proofread image data 12, etc., when an image 100 was previously printed by the printing device 2, and generates a learning model 16 by performing machine learning based on the manuscript image data 11 and proofread image data 12, etc., registered in the image management database 410. Furthermore, when the information processing device 4 receives newly created manuscript image data 11 from the user terminal device 5, it uses the learning model 16 to generate proofread image data 12 by proofreading the manuscript image data 11, and provides the proofread image data 12 to the printing device 2, the platemaking device 3, and the user terminal device 5.

[0019] The user terminal device 5 is a device used by users such as workers performing proofreading work and managers of the printing system 1. An image proofreading program, a production management program, a web browser, etc. are installed on the user terminal device 5, and the user accepts various input operations and outputs various information via a display screen or voice. The user terminal device 5 is used, for example, by a user to input proofreading instructions to the information processing device 4, printing instructions to the printing device 2, platemaking instructions to the platemaking device 3, etc., and to output the processing results of the information processing device 4.

[0020] (Configuration of information processing device 4) 2 is a block diagram showing an example of the information processing device 4. The information processing device 4 includes a control unit 40 configured with a processor or the like, a storage unit 41 configured with an HDD, an SSD, a memory or the like, a communication unit 42 which is a communication interface with the network 6 and external devices, an input unit 43 configured with a keyboard, a mouse or the like, and a display unit 44 configured with a display or the like. Note that the input unit 43 and the display unit 44 may be omitted.

[0021] The storage unit 41 stores an image management database 410, a trained model management database 411, and an information processing program 412, as well as an operating system, other programs, data, etc.

[0022] The control unit 40 executes an information processing program 412 stored in the storage unit 41 to implement an image management function 400, a learning function 401, and an image proofreading function 402. The control unit 40 includes a learning data acquisition unit 4010 and a machine learning unit 4011 as units that implement the learning function 401. The control unit 40 includes a data acquisition unit 4020, a data generation unit 4021, and an output processing unit 4022 as units that implement the image proofreading function 402.

[0023] The following describes the data configuration of each function 400 to 402 and each database 410, 411. explain.

[0024] (Image management function 400) The control unit 40 of the information processing device 4 uses the image management database 410 to implement the image management function 400. By providing the information processing device 4 with the image management function 400, the information processing device 4 operates as an image management device.

[0025] 3 is a data configuration diagram showing an example of the image management database 410. The image management database 410 has multiple records for each image management ID for associating various data handled by the printing system 1. The image management ID is information (product number, model number, etc.) for identifying the image 100. Each record has fields in which, for example, manuscript image data 11, proofread image data 12, proofreading process content data 13 indicating the content of the proofreading process performed on the manuscript image data 11, and performance data 14 when the image 100 was printed based on the proofread image data 12 can be registered.

[0026] Each of the data 11 to 14 registered in the image management database 410 can be referenced from each of the devices 2 to 5. Furthermore, editing operations such as addition, deletion, and correction of each of the data 11 to 14 can be performed on a display screen displayed by the image proofing program or the manufacturing management program of the user terminal device 5. The data format of the manuscript image data 11 and the proof image data 12 may be any data format that can be handled by the printing system 1. For example, the manuscript image data 11 and the proof image data 12 may be a raster data format in which the pixel values ​​of each pixel are recorded, a vector data format in which points and lines are digitized and recorded, or a data format that combines these.

[0027] For example, when the image management function 400 receives each of the data 11 to 14 from the user terminal device 5, it registers the data 11 to 14 in the image management database 410. Furthermore, when the image management function 400 receives a processing result from the image proofreading function 402, it registers the processing result in the image management database 410, and when the image management function 400 receives an operation result from an editing operation, it registers the operation result in the image management database 410.

[0028] 4 is a diagram showing an example of manuscript image data 11. The manuscript image data 11 is generated, for example, by a product manufacturer that manufactures and sells products in which contents are filled into containers 10, using an existing image editing program, and is provided to a user of user terminal device 5, whereby the data is registered in image management database 410.

[0029] The original image data 11 includes a development of an original image 110 to be printed on the outer surface of the container 10, and an original color sample 111 for the original image 110. The original image 110 includes, as multiple object images, a design image 110a including characters and a design representing the product name, a barcode image 110b, a nutrition information image 110c including characters representing nutritional components, a product information image 110d including characters representing the name, ingredients, content volume, expiration date, storage instructions, and seller, a recycle mark image 110e, and a background image 110f.

[0030] The original color sample 111, for example, specifies the colors used in each of the images 110a to 110f included in the original image 110 in association with each of the images 110a to 110f. The original color sample 111 may be specified based on an arbitrary standard of actual colors, or may be specified using CMYK values ​​or RGB values.

[0031] 5 is a diagram showing an example of proofread image data 12. The proofread image data 12 is generated, for example, by a skilled proofreader who performs proofreading work by using an image proofreading program to proofread the manuscript image data 11, and is registered in the image management database 410. The proofread image data 12 is generated, for example, by the image proofreading function 402 performing proofreading processing on the manuscript image data 11 using the learning model 16, and is registered in the image management database 410.

[0032] The proof image data 12 includes a development of a proof image 120, which is obtained by performing a proofing process on the original image 110, and a proof color sample 121, which is obtained by performing a proofing process on the original color sample 111. Like the original image 110, the proof image 120 includes multiple object images, including a design image 120a, a barcode image 120b, a nutritional information image 120c, a product information image 120d, a recycling mark image 120e, and a background image 120f.

[0033] The proof image data 12 may include plate separation data separated for each ink color of the printing device 2. The plate separation data is, for example, separated for each printing plate (i.e., ink color) used in the printing device 2, and may represent the ink pattern for each printing plate, or may represent the shade of each ink color as a collection of halftone dots.

[0034] The proofreading process content data 13 records the content of the proofreading process performed on the manuscript image data 11 when the proofread image data 12 was generated from the manuscript image data 11. The proofreading process content data 13 also includes proofreading support content that records notes regarding the proofreading process, platemaking support content that records notes when making a printing plate, and printing support content that records notes when printing the image 100.

[0035] The proofreading process is a process of editing an object image with respect to at least one of the editing items of the arrangement, shape, size, and color scheme of the object image included in the document image data 11. Specific proofreading processes performed include at least one of the composition limit setting process, ink color selection process, color scheme editing process, character processing process, overlap area editing process, color interval editing process, and halftone dot setting process. However, the contents of the proofreading process are not limited to these.

[0036] The composition limit setting process edits the object image so that it fits within a composition limit 122 that corresponds to the range that can be printed by the printing device 2 on the outer circumferential surface of the container 10. The composition limit 122 is set according to the shape of the container 10, for example, according to the curved shape of the neck portion or the like.

[0037] The ink color selection process selects ink colors to be used for printing the object image from multiple types of color guides that indicate ink colors that can be printed by the printing device 2, based on color samples included in the original image data 11. The color guides are prepared in advance according to the printing method of the printing device 2, for example.

[0038] The color scheme editing process edits the object image so that the color scheme corresponds to the ink colors that can be printed by the printing device 2. In this case, the color scheme editing process edits the object image so that the color scheme corresponds to the ink colors selected in the ink color selection process, and the color scheme after the color scheme editing process is performed is included in the proof image data 12 as a proof color sample 121.

[0039] In the character processing, when an object image includes characters, the characters are processed, for example, to improve visibility or readability.

[0040] In the overlapping area editing process, when there is an overlapping area where object images overlap each other, the object images are edited so that the colors of the object images in the overlapping area are unified. If a pattern is included in the area, the shape is changed so as to cut out the pattern, and the image is changed to the background image 120f.

[0041] In the color-to-color editing process, when a plurality of ink colors are used to print an object image, gaps are set between object images printed in different ink colors.

[0042] When multiple ink colors are used to print an object image, the halftone dot setting process sets at least one of the screen ruling and the screen angle of the halftone dots for each ink color. For example, when multiple ink colors overlap, the halftone dot setting process adjusts the screen ruling and the screen angle of the halftone dots to prevent moire from occurring.

[0043] The performance data 14 records various performance values ​​when the printing device 2 performs the process of printing the image 100 on the outer peripheral surface of the container 10 based on the proof image data 12. As shown in FIG. 3 , the performance data 14 includes, for example, the printing method of the printing device 2 selected to print the image 100, the printing cost when the image 100 is printed by the selected printing device 2, the number of prints when the image 100 is printed by the selected printing device 2, the durability of the printing plates when the selected printing device 2 is a printing machine that performs plate printing, the production lead time required to produce the printing plates when the selected printing device 2 is a printing machine that performs plate printing, substrate information regarding the substrate that forms the container 10, formulation information for the ink used by the printing device 2, and sensory information regarding the image 100. Substrate information includes, for example, the material type of the substrate, coating material, or printing primer (base coat), color (white, clear, silver, pearl, etc.), gloss (metallic gloss, matte, etc.), substrate shape (cylindrical, flat, etc.), and molding-related features (metallic streaks, etc.). Ink formulation information includes, for example, the type of base ink, blending ratio, blending amount, application area, film thickness, plate type, and material type of printing primer (base coat). Emotional information includes, for example, the identification information and preferences of the designer who created the original of image 100, the design concept when the original of image 100 was created, and the product image of a product using container 10 on which image 100 is printed.

[0044] Each value included in the performance data 14 may be recorded as a numerical value or as one of a number of stage values. Each value included in the performance data 14 may be a specific value or a value having a specific range such as an upper limit or a lower limit.

[0045] (Learning function 401) 6A to 6D are functional explanatory diagrams showing an example of the learning function 401. A learning data acquisition unit 4010 and a machine learning unit 4011 of the information processing device 4 realize the learning function 401 using an image management database 410 and a trained model management database 411 (trained model storage unit). By providing the learning function 401 in the information processing device 4, the information processing device 4 operates as a machine learning device.

[0046] The learning models 16 stored in the learned model management database 411 are learned models that have been trained by machine learning to learn the correlation between at least the manuscript image data 11 and the proofread image data 12. In this embodiment, a case will be described in which first to eighth learning models 16-1 to 16-8 are stored in the learned model management database 411 as the learning models 16. That is, each of the first to eighth learning models 16-1 to 16-8 receives at least the manuscript image data 11 as input data and outputs at least the proofread image data 12 as output data.

[0047] The first learning model 16-1 receives the manuscript image data 11 as input data and outputs the proofread image data 12 as output data. The second learning model 16-2 receives the manuscript image data 11 as input data and outputs the proofread image data 12 and predicted data as output data. The third learning model 16-3 takes the manuscript image data 11 as input data and outputs the proofread image data 12 and support data as output data. The fourth learning model 16-4 takes the manuscript image data 11 as input data and outputs the proofread image data 12, predicted data, and support data as output data. The fifth learning model 16-5 takes the manuscript image data 11 and requested data as input data and outputs the proofread image data 12 as output data. The sixth learning model 16-6 takes the manuscript image data 11 and requested data as input data and outputs the proofread image data 12 and predicted data as output data. The seventh learning model 16-7 takes the manuscript image data 11 and requested data as input data and outputs the proofread image data 12 and support data as output data. The eighth learning model 16-8 takes the manuscript image data 11 and requested data as input data and outputs the proofread image data 12, predicted data, and support data as output data.

[0048] The request data included in the input data relates to first printing specifications required when printing image 100. The request data includes data similar to performance data 14, and the first printing specifications in the request data are at least one of the printing method, printing cost, number of prints, durability of the printing plate, delivery time for producing the printing plate, substrate information for container 10, ink formulation information, and affective information related to image 100.

[0049] When the requested data includes the printing method, printing cost, number of copies, printing plate durability, or printing plate production lead time, proof image data 12 is output after proofreading to meet the specified contents. When substrate information is requested as the requested data, proof image data 12 is output after proofreading to suit the substrate of the container 10 specified in the substrate information. When ink formulation information is requested as the requested data, proof image data 12 is output after proofreading to suit the ink specified in the ink formulation information. When affective information is requested as the requested data, proof image data 12 is output after proofreading to reflect the designer's preferences, design concept, product image, etc. contained in the affective information.

[0050] The predicted data included in the output data relates to second printing specifications predicted when printing image 100. The predicted data includes data similar to performance data 14, and the second printing specifications in the predicted data are at least one of the printing method, printing cost, number of prints, durability of the printing plate, delivery time for producing the printing plate, substrate information for container 10, ink formulation information, and affective information related to image 100.

[0051] When the printing method, printing cost, number of prints, durability of the printing plate, or production lead time of the printing plate are predicted as the prediction data, the printing method, printing cost, number of prints, durability of the printing plate, or production lead time of the printing plate when printing the image 100 based on the proof image data 12 output from the learning model 16 are predicted. When substrate information is predicted as the prediction data, substrate information suitable for the container 10 when printing the image 100 is predicted based on the proof image data 12 output from the learning model 16. When ink formulation information is predicted as the prediction data, ink formulation information suitable for the printing device 2 when printing the image 100 is predicted based on the proof image data 12 output from the learning model 16. When affective information is predicted as the prediction data, the designer's preferences, design concept, product image, etc. that will be satisfied by the proof image data 12 output from the learning model 16 are predicted.

[0052] In the sixth learning model 16-6 and the eighth learning model 16-8, the first printing specification in the request data is different from the second printing specification in the prediction data. In other words, if the first printing specification in the request data is, for example, a printing method, the second printing specification in the prediction data includes, except for the printing method, the printing method, the printing cost, the number of prints, the number of prints, etc. The information is at least one of the durability of the printing plate, the lead time for producing the printing plate, the substrate information of the container 10, the ink composition information, and the affective information regarding the image 100.

[0053] The support data included in the output data relates to the proofreading process, and includes, for example, precautions for the user to take when checking the proofread image data 12, precautions to take when creating a printing plate based on the proofread image data 12, and precautions to take when printing the image 100 based on the proofread image data 12. In other words, the support data includes data similar to the support content in the proofreading process content data 13, and is at least one of proofreading support content, platemaking support content, and printing support content.

[0054] The first to eighth learning models 16-1 to 16-8 each employ, for example, a neural network structure, and include an input layer 160, an intermediate layer 161, and an output layer 162. Synapses (not shown) that connect the neurons are laid between the layers, and each synapse is associated with a weight. A weight parameter group consisting of the weights of each synapse is adjusted by a machine learning algorithm such as backpropagation.

[0055] The input layer 160 has neurons whose number corresponds to the input data, and each value is input to each neuron. The output layer 162 has neurons whose number corresponds to the output data, and the output data is output as the inference result.

[0056] The learning data acquisition unit 4010 acquires a plurality of sets of first to eighth learning data 15-1 to 15-8, each set consisting of input data including at least manuscript image data 11 and output data including at least proofread image data 12. The first to eighth learning data 15-1 to 15-8 are data used as teacher data (training data), verification data, and test data in supervised learning. Furthermore, characteristic parameters included in the first to eighth learning data 15-1 to 15-8 are data used as correct answer labels in supervised learning.

[0057] For example, the learning data acquisition unit 4010 acquires the first to eighth learning data 15-1 to 15-8 by referring to information registered in the image management database 410 or by receiving an input operation from the user terminal device 5. Note that if information corresponding to the first to eighth learning data 15-1 to 15-8 is stored in an external system (such as a container processing system or a container filling system), the learning data acquisition unit 4010 may acquire the first to eighth learning data 15-1 to 15-8 from the external system.

[0058] The machine learning unit 4011 performs machine learning for each of the first to eighth learning models 16-1 to 16-8 using the multiple sets of learning data 15-1 to 15-8 acquired by the learning data acquisition unit 4010. Then, the machine learning unit 4011 generates the trained first to eighth learning models 16-1 to 16-8 by having the first to eighth learning models 16-1 to 16-8 learn the correlation between input data and output data, respectively.

[0059] When performing machine learning, the machine learning unit 4011 can employ any method, such as online learning, batch learning, mini-batch learning, etc. Furthermore, the machine learning unit 4011 may perform predetermined pre-processing on input data to be input to each of the learning models 16-1 to 16-8, or may perform predetermined post-processing on output data output from each of the learning models 16-1 to 16-8.

[0060] The trained model management database 411 stores the trained first to eighth trained models 16-1 to 16-8 (specifically, adjusted weight parameters) generated by the machine learning unit 4011. The trained learning models 16-1 to 16-8 stored in the trained model management database 411 may be provided to other systems via the network 6, a recording medium, or the like.

[0061] 6A and 6B, a plurality of data configurations with different conditions, such as different machine learning techniques, different input data, different output data, etc., may be employed. In this case, the learning data acquisition unit 4010 acquires a plurality of types of learning data corresponding to a plurality of data configurations with different conditions, and the machine learning unit 4011 performs machine learning using each of the learning data, and the trained learning model 16 is stored in the trained model management database 411.

[0062] For example, when proofread image data 12 that has undergone one of the following proofreading processes is used as output data: composition limit setting process, ink color selection process, color scheme editing process, character processing process, overlap area editing process, color space editing process, and halftone dot setting process, it is possible to generate a learning model 16 that outputs proofread image data 12 that has undergone that one proofreading process. Also, when proofread image data 12 that has undergone two or more proofreading processes that are any combination of the following: composition limit setting process, ink color selection process, color scheme editing process, character processing process, overlap area editing process, color space editing process, and halftone dot setting process, it is possible to generate a learning model 16 that outputs proofread image data 12 that has undergone that two or more proofreading processes.

[0063] (Image proofreading function 402) 7 is a functional explanatory diagram showing an example of the image proofreading function 402. The data acquisition unit 4020, data generation unit 4021, and output processing unit 4022 of the information processing device 4 realize the image proofreading function 402 using the trained model management database 411. By providing the information processing device 4 with the image proofreading function 402, the information processing device 4 operates as an image proofreading device.

[0064] The data acquisition unit 4020 acquires manuscript image data 11 created as a manuscript of the image 100 to be printed by the printing device 2 on the outer peripheral surface of the container 10. For example, the data acquisition unit 4020 acquires the manuscript image data 11 by accepting an input operation from the user terminal device 5 or by referring to the manuscript image data 11 registered in the image management database 410. In this case, the data acquisition unit 4020 may acquire request data related to the first printing specifications along with the manuscript image data 11.

[0065] The data generation unit 4021 inputs the manuscript image data 11 acquired by the data acquisition unit 4020 into the learning model 16, thereby generating proofread image data 12 by performing proofreading on the manuscript image data 11. The learning models 16 according to this embodiment are the first to eighth learning models 16-1 to 16-8 shown in FIGS. 6A and 6B, and therefore the data generation unit 4021 can use the first to eighth learning models 16-1 to 16-8 selectively or in parallel.

[0066] For example, when the data acquisition unit 4020 acquires manuscript image data 11, the manuscript image data 11 may be input to a first learning model 16-1 to generate proofread image data 12 for the manuscript image data 11, or the manuscript image data 11 may be input to a second learning model 16-2 to generate proofread image data 12 and predicted data for the manuscript image data 11, or the manuscript image data 11 may be input to a third learning model 16-3 to generate proofread image data 12 and support data for the manuscript image data 11. Alternatively, the document image data 11 may be input to a fourth learning model 16-4 to generate proofread image data 12, predicted data, and support data for the document image data 11.

[0067] Furthermore, when the data acquisition unit 4020 acquires requested data along with the manuscript image data 11, the manuscript image data 11 and the requested data may be input into a fifth learning model 16-5 to generate proofread image data 12 for the manuscript image data 11, the manuscript image data 11 and the requested data may be input into a sixth learning model 16-6 to generate proofread image data 12 and predicted data for the manuscript image data 11, the manuscript image data 11 and the requested data may be input into a seventh learning model 16-7 to generate proofread image data 12 and support data for the manuscript image data 11, or the manuscript image data 11 and the requested data may be input into an eighth learning model 16-8 to generate proofread image data 12, predicted data, and support data for the manuscript image data 11.

[0068] When the data generation unit 4021 uses a learning model 16 that employs, as output data of the learning model 16, proofread image data 12 that has been subjected to one of a composition limit setting process, an ink color selection process, a color scheme editing process, a character processing process, an overlap area editing process, a color space editing process, and a halftone dot setting process, the data generation unit 4021 may generate proofread image data 12 that has been subjected to the one proofreading process by inputting the manuscript image data 11 to the learning model 16. When the data generation unit 4021 uses, as output data of the learning model 16, proofread image data 12 that has been subjected to any combination of two or more of a composition limit setting process, an ink color selection process, a color scheme editing process, a character processing process, an overlap area editing process, a color space editing process, and a halftone dot setting process, the data generation unit 4021 may generate proofread image data 12 that has been subjected to the two or more proofreading processes by inputting the manuscript image data 11 to the learning model 16.

[0069] The first to eighth learning models 16-1 to 16-8 used by the data generation unit 4021 are the first to eighth learning models 16-1 to 16-8 that have been trained and stored in the trained model management database 411. The data generation unit 4021 may perform predetermined pre-processing on input data to be input to the first to eighth learning models 16-1 to 16-8. The data generation unit 4021 may also perform predetermined post-processing on output data output from the first to eighth learning models 16-1 to 16-8. As post-processing, the data generation unit 4021 may perform arithmetic processing on the prediction data output from the second learning model 16-2 or the sixth learning model 16-6. For example, the data generation unit 4021 may calculate the selling price of the container 10 from the printing cost of the prediction data. Furthermore, as post-processing, the data generation unit 4021 may perform rule-based determination processing on the proof image data 12 output from the first to eighth learning models 16-1 to 16-8. For example, the data generation unit 4021 may determine whether or not a specific color is included in the proof image data 12, and if the result of the determination processing indicates that the proof image data 12 includes the specific color, generate support data including predetermined precautions. Furthermore, the data generation unit 4021 may generate ink formulation information (e.g., the type, formulation ratio, and formulation amount of the original ink) based on the hue, saturation, brightness, and distribution of the colors included in the proof image data 12.

[0070] The output processing unit 4022 performs output processing for outputting the proof image data 12 generated by the data generation unit 4021. For example, as the output processing, the output processing unit 4022 may send display information for displaying the proof image data 12 on the user terminal device 5, may register the proof image data 12 in the image management database 410, or may send the proof image data 12 to the printing device 2 and the platemaking device 3.

[0071] The data generating unit 4021 generates the predicted data and the support data together with the calibration image data 12. When the predicted data and the assistance data are generated, the output processing unit 4022 performs output processing to output the predicted data and the assistance data, similar to the calibration image data 12.

[0072] 8 is a hardware configuration diagram showing an example of a computer 900 that constitutes each device. Each device 2 to 5 in the printing system 1 is constituted by a general-purpose or dedicated computer 900.

[0073] 8, the computer 900 includes, as its main components, a bus 910, a processor 912, a memory 914, an input device 916, an output device 917, a display device 918, a storage device 920, a communication I / F (interface) unit 922, an external device I / F unit 924, an I / O (input / output) device I / F unit 926, and a media input / output unit 928. Note that the above components may be omitted as appropriate depending on the application of the computer 900.

[0074] The processor 912 is composed of one or more arithmetic processing devices (such as a CPU (Central Processing Unit), MPU (Micro-Processing Unit), DSP (Digital Signal Processor), GPU (Graphics Processing Unit), or NPU (Neural Processing Unit)), and operates as a control unit that controls the entire computer 900. The memory 914 stores various data and programs 930, and is composed of, for example, a volatile memory (such as a DRAM or SRAM) that functions as a main memory, a non-volatile memory (ROM), a flash memory, etc.

[0075] The input device 916 is composed of, for example, a keyboard, a mouse, a numeric keypad, an electronic pen, etc., and functions as an input unit. The output device 917 is composed of, for example, a sound (audio) output device, a vibration device, etc., and functions as an output unit. The display device 918 is composed of, for example, a liquid crystal display, an organic EL display, electronic paper, a projector, etc., and functions as an output unit. The input device 916 and the display device 918 may be integrated into one device, such as a touch panel display. The storage device 920 is composed of, for example, an HDD, an SSD, etc., and functions as a storage unit. The storage device 920 stores various data necessary for executing the operating system and the program 930.

[0076] The communication I / F unit 922 is connected to a network 940 such as the Internet or an intranet (which may be the same as network 6 in FIG. 1) via a wired or wireless connection and functions as a communication unit that transmits and receives data to and from other computers in accordance with a predetermined communication protocol. The external device I / F unit 924 is connected to an external device 950 such as a camera, printer, scanner, or reader / writer via a wired or wireless connection and functions as a communication unit that transmits and receives data to and from the external device 950 in accordance with a predetermined communication protocol. The I / O device I / F unit 926 is connected to an I / O device 960 such as various sensors and actuators and functions as a communication unit that transmits and receives various signals and data, such as detection signals from sensors and control signals to actuators, to and from the I / O device 960. The media input / output unit 928 is composed of a drive device such as a DVD drive or CD drive, a memory card slot, and a USB connector, and reads and writes data from and to media (non-transitory storage media) 970 such as a DVD, CD, memory card, or USB memory.

[0077] In the computer 900 having the above configuration, the processor 912 loads a program 930 stored in the storage device 920 into the memory 914, executes the program, and controls each unit of the computer 900 via the bus 910. The program 930 may be stored in the memory 914 instead of the storage device 920. The program 930 is recorded on the medium 970 in an installable file format or an executable file format, and is The program 930 may be provided to the computer 900 via the media input / output unit 928. The program 930 may be provided to the computer 900 by being downloaded via the communication I / F unit 922 over a network 940. Furthermore, the computer 900 may implement the various functions that are realized by the processor 912 executing the program 930 using hardware such as an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

[0078] The computer 900 is an electronic device of any type, such as a desktop computer or a portable computer. The computer 900 may be a client computer, a server computer, a cloud computer, or an embedded computer such as a control panel or a controller (including a microcomputer, a programmable logic controller, or a sequencer).

[0079] (Printing system 1 operation) Hereinafter, as the operation of the printing system 1, the learning function 401 and the image proofreading function 402 realized by the information processing device 4 will be described.

[0080] (machine learning methods) 9 is a flowchart showing an example of a machine learning method using the learning function 401. The following describes a case where a learning model 16 (one of the first to eighth learning models 16-1 to 16-8) is generated using multiple sets of learning data 15 (one of the first to eighth learning data 15-1 to 15-8).

[0081] First, in step S100, the learning data acquisition unit 4010 acquires a desired number of learning data 15 as a preliminary preparation for starting machine learning, and temporarily stores the acquired learning data 15 in the memory unit 41.

[0082] Next, in step S110, in order to start machine learning, the machine learning unit 4011 prepares a pre-learning learning model 16. The pre-learning learning model 16 prepared here is configured, for example, by a neural network model, and the weight of each synapse is set to an initial value.

[0083] Next, in step S120, the machine learning unit 4011 acquires, for example, one set of learning data 15 at random from the multiple sets of learning data 15 stored in the storage unit 41.

[0084] Next, in step S130, the machine learning unit 4011 inputs the manuscript image data 11 (input data) included in one set of learning data 15 to the input layer 160 of the prepared learning model 16 before learning (or during learning). As a result, proofread image data 12 (output data) is output as an inference result from the output layer 162 of the learning model 16, and this output data was generated by the learning model 16 before learning (or during learning). Therefore, in the state before learning (or during learning), the output data output as an inference result indicates information different from the proofread image data 12 (correct label) included in the learning data 15.

[0085] Next, in step S140, the machine learning unit 4011 performs machine learning by comparing the proofread image data 12 (correct label) included in the set of learning data 15 acquired in step S120 with the proofread image data 12 (output data) output as an inference result from the output layer 162 in step S130, and performing a process of adjusting the weight of each synapse (backpropagation).In this way, the machine learning unit 4011 causes the learning model 16 to learn the correlation between the manuscript image data 11 and the proofread image data 12.

[0086] Next, in step S150, the machine learning unit 4011 determines whether a predetermined learning termination condition has been met, for example, based on the evaluation value of an error function based on the proofread image data 12 (correct label) included in the learning data 15 and the proofread image data 12 (output data) output as the inference result, or the remaining number of unlearned learning data 15 stored in the memory unit 41.

[0087] In step S150, if the machine learning unit 4011 determines that the learning termination condition is not satisfied and that machine learning should continue (No in step S150), the process returns to step S120, and performs steps S120 to S140 multiple times on the learning model 16 being trained using untrained training data 15. On the other hand, in step S150, if the machine learning unit 4011 determines that the learning termination condition is satisfied and that machine learning should end (Yes in step S150), the process proceeds to step S160.

[0088] Then, in step S160, the machine learning unit 4011 stores the trained learning model 16 (adjusted weight parameter group) generated by adjusting the weights associated with each synapse in the trained model management database 411, and ends the series of machine learning methods shown in Fig. 9. In the machine learning method, step S100 corresponds to the training data acquisition step, steps S110 to S150 correspond to the machine learning step, and step S160 corresponds to the trained model storage step.

[0089] As described above, the learning function 401 and machine learning method according to this embodiment can provide a learning model 16 that generates proofread image data 12 from manuscript image data 11. For example, by performing machine learning of the learning model 16 using the manuscript image data 11 and proofread image data 12 obtained when proofreading is performed by a skilled worker as learning data 15, the knowledge of the skilled worker can be reflected in the learning model 16.

[0090] (Image calibration method) 10 is a flowchart showing an example of an image proofreading method by the image proofreading function 402. In the following, it is assumed that the learned model management database 411 stores the first to eighth learned models 16-1 to 16-8 shown in FIGS. 6A and 6B as learned models by the learning function 401.

[0091] First, in step S200, the user terminal device 5 executes an image proofreading program and displays an image proofreading input screen 17 in cooperation with the image proofreading function 402 of the information processing device 4, and receives an input operation of manuscript image data 11 from the user on the image proofreading input screen 17. Then, the user terminal device 5 transmits the manuscript image data 11 to the information processing device 4 based on the result of the input operation by the user. Note that the user terminal device 5 may also receive input operations to specify requested data or whether predicted data or support data is to be generated, along with the manuscript image data 11, and transmit the same to the information processing device 4.

[0092] 11 is a screen configuration diagram showing an example of the image proofreading input screen 17. The image proofreading input screen 17 includes a manuscript image specification field 170 for specifying manuscript image data 11, a requested data specification field 171 for specifying requested data, a predicted data specification field 172 for specifying predicted data, and an assistance data specification field 173 for specifying whether assistance data is to be generated.

[0093] In the request data specification field 171, each print specification to be specified as the first print specification in the request data is entered. Printing specification items can be specified using check boxes, and requirements for each printing specification item can be specified using text boxes, combo boxes, etc. In the example of Figure 11, "relief printing-offset printing" has been specified as the printing method, so the illustration shows a case where proofreading suitable for "relief printing-offset printing" is specified to be performed.

[0094] In the predicted data specification field 172, each printing specification item to be specified as the second printing specification in the predicted data can be specified using a check box. In the example of Fig. 11, "printing cost" and "printing plate durability" are specified, so the illustration shows a case where it is specified to predict the printing cost and printing plate durability when a proofing process suitable for "relief printing-offset printing" is performed and an image is printed.

[0095] In the support data specification field 173, it is possible to specify whether or not support data is to be generated by selecting the "Necessary" or "Not Necessary" selection button. In the example of Fig. 11, the "Necessary" selection button has been pressed, so the example shows a case where support data is specified to be generated when an image is printed after proofreading suitable for "relief printing-offset printing".

[0096] Next, in step S210, the data acquisition unit 4020 receives the document image data 11 sent in step S200, thereby acquiring the document image data 11. At that time, the data acquisition unit 4020 may acquire the designation result of the requested data (in the example of FIG. 11, "relief printing-offset printing" is designated as the printing method), or may acquire the designation result of the predicted data to be generated (in the example of FIG. 11, "printing cost" and "printing plate durability" are designated) or the designation result of the support data (in the example of FIG. 11, "necessary" is designated).

[0097] Next, in step S220, the data generation unit 4021 inputs the manuscript image data 11 acquired in step S210 into the learning model 16, thereby generating proofread image data 12 by performing proofreading processing on the manuscript image data 11. In this embodiment, the data generation unit 4021 selects a learning model 16 to input the manuscript image data 11 into from the first to eighth learning models 16-1 to 16-8 in response to an input operation on the image proofreading input screen 17, and inputs the manuscript image data 11 into that learning model 16.

[0098] Specifically, when requested data is not specified and predicted data and assistance data are not specified as data to be generated, the data generation unit 4021 inputs the manuscript image data 11 to the first learning model 16-1 to generate proofread image data 12. When requested data is not specified and predicted data is specified as data to be generated, the data generation unit 4021 inputs the manuscript image data 11 to the second learning model 16-2 to generate proofread image data 12 and predicted data. When requested data is not specified and assistance data is specified as data to be generated, the data generation unit 4021 inputs the manuscript image data 11 to the third learning model 16-3 to generate proofread image data 12 and assistance data. When requested data is not specified and predicted data and assistance data are specified as data to be generated, the data generation unit 4021 inputs the manuscript image data 11 to the fourth learning model 16-4 to generate proofread image data 12, predicted data, and assistance data.

[0099] Furthermore, when requested data is specified and predicted data or assistance data is not specified as the data to be generated, the data generation unit 4021 inputs the manuscript image data 11 and requested data into a fifth learning model 16-5 to generate proofread image data 12. When requested data is specified and predicted data is specified as the data to be generated, the data generation unit 4021 inputs the manuscript image data 11 and requested data into a sixth learning model 16-6 to generate proofread image data 12 and predicted data. When requested data is specified and assistance data is specified as the data to be generated, the data generation unit 4021 inputs the manuscript image data 11 and requested data into a seventh learning model 16-7 to generate proofread image data 12 and assistance data. When requested data is specified, the data generation unit 4021 inputs the manuscript image data 11 and requested data into a fifth learning model 16-5 to generate proofread image data 12. When requested data is specified and predicted data is specified as the data to be generated, the data generation unit 4021 inputs the manuscript image data 11 and requested data into a sixth learning model 16-6 to generate proofread image data 12 and assistance data. When the eighth learning model 16-8 is set and predicted data and support data are specified as the data to be generated, the manuscript image data 11 is input to the eighth learning model 16-8 to generate proofread image data 12, predicted data, and support data.

[0100] 11, the request data specifies "relief-offset printing" as the printing method, the prediction data specifies "printing cost" and "printing plate durability," and the support data is specified as "required," so the data generation unit 4021 selects the eighth learning model 16-8 from the first to eighth learning models 16-1 to 16-8.The data generation unit 4021 then inputs the manuscript image data 11 and request data acquired via the image proofreading input screen 17 into the eighth learning model 16-8, thereby generating proofread image data 12, prediction data, and support data.

[0101] Next, in step S230, the output processing unit 4022 executes output processing to output the proof image data 12 generated in step S220. For example, if display information for displaying the proof image data 12 is transmitted to the user terminal device 5, in step S240 the user terminal device 5 displays the image proof output screen 18 based on the display information, thereby presenting the proof image data 12 to the user. Note that if prediction data or assistance data is generated along with the proof image data 12, output processing is performed on the prediction data or assistance data.

[0102] 12 is a screen configuration diagram showing an example of the image proofreading output screen 18. The image proofreading output screen 18 includes an original image display field 180 that displays the file name of the original image data 11 specified on the image proofreading input screen 17, a requested data display field 181 that displays the requested data specified on the image proofreading input screen 17, a confirm / correct button 182 for confirming or correcting the proofread image data 12 generated by the data generation unit 4021, a predicted data display field 183 that displays the predicted data generated by the data generation unit 4021, an assistance data display field 184 that displays the assistance data generated by the data generation unit 4021, a platemaking button 185 for sending the proofread image data 12 generated by the data generation unit 4021 to the platemaking device 3, a print button 186 for sending the proofread image data 12 generated by the data generation unit 4021 to the printing device 2, and a back button 187 for returning to the image proofreading input screen 17.

[0103] When the Confirm / Modify button 182 is pressed, for example, a proof image display screen for displaying the proof image data 12 is displayed, and the user can visually check the proof image 120 and the proof color sample 121 on the proof image display screen. Furthermore, on the proof image display screen, the user can perform editing work to modify the proof image 120 and the proof color sample 121. In the example of FIG. 11, since "relief printing-offset printing" is specified as the printing method in the request data, when the Confirm / Modify button 182 in FIG. 12 is pressed, proof image data 12 that has been proofed to suit "relief printing-offset printing" is displayed.

[0104] The predicted results for each printing specification item specified as the second printing specification in the predicted data are displayed in predicted data display field 183. The example in Fig. 12 shows the case where the predicted results of "printing cost" and "printing plate durability" are displayed when image 100 is printed based on proof image data 12 that has been proofread suitable for "relief printing-offset printing."

[0105] At least one of proofreading support content, platemaking support content, and printing support content is displayed as the support content included in the support data in the support data display field 184. The example in Fig. 12 shows a case where proofreading support content, platemaking support content, and printing support content are displayed as support content for proofread image data 12 that has been proofread suitable for "relief printing-offset printing."

[0106] When the platemaking button 185 or the print button 186 is pressed, the output processing unit 4022 performs output processing by transmitting the proof image data 12 to the platemaking device 3 or the printer 2. Note that the output processing unit 4022 may transmit the proof image data 12 to the platemaking device 3 or the printer 2 regardless of the platemaking button 185 or the print button 186.

[0107] Then, in step S250, when the platemaking device 3 receives the proof image data 12, it creates a printing plate. Also, when the printing device 2 receives the proof image data 12, it uses the printing plate created by the platemaking device 3 to print an image 100 on the outer peripheral surface of the container 10 based on the proof image data 12.

[0108] In this manner, the series of image proofreading methods shown in Fig. 10 is completed. In the above image proofreading method, step S210 corresponds to a data acquisition step, step S220 corresponds to a data generation step, and step S230 corresponds to an output processing step. Note that the series of image proofreading methods can be executed at any timing as long as the information processing device 4 can acquire the document image data 11.

[0109] As described above, according to the image proofreading function 402 and image proofreading method of the information processing device 4 of this embodiment, by inputting the original image data 11 into the learning model 16, proofread image data 12 is generated in which a predetermined proofreading process has been performed on the original image data 11, so that the image 100 can be proofread easily and appropriately.

[0110] Furthermore, by inputting required data regarding the first printing specifications into the learning model 16 along with the manuscript image data 11, proofreading processing is performed on the manuscript image data 11 to satisfy the required data, thereby generating proofread image data 12, thereby enabling easy and appropriate proofreading of the image 100 that satisfies the first printing specifications.

[0111] Furthermore, by inputting the manuscript image data 11 into the learning model 16, predicted data regarding the second printing specifications is generated along with proofread image data 12, which is obtained by performing a proofreading process on the manuscript image data 11.Therefore, the usefulness of the proofread image data 12 can be confirmed by referring to the predicted results of the second printing specifications.

[0112] Furthermore, by inputting the manuscript image data 11 into the learning model 16, proofread image data 12, which is obtained by performing proofreading on the manuscript image data 11, is generated along with support data indicating the support content for the proofreading process, so that the proofread image data 12 can be used effectively by referring to the support content.

[0113] (Other embodiments) The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and scope of the present invention, all of which are included in the technical concept of the present invention.

[0114] In the above embodiment, the printing surface on which the image 100 is printed by the printing device 2 is described as the outer peripheral surface of the container 10, but it is not limited to the outer peripheral surface of the container 10 and may be the printing surface of any printing object. The printing object may be an object having a three-dimensional outer shape such as a cylinder or a rectangular tube, or an object having a planar outer shape such as a flat plate, film, or paper.

[0115] In the above embodiment, the multiple functions of the information processing device 4 are described as being realized by one device, but each function may be distributed among multiple devices (computers) to be realized by multiple devices. For example, an image management device that realizes the image management function 400 and a learning device may be used. The machine learning device that realizes the function 401 and the image proofreading device that realizes the image proofreading function 402 may be configured as different devices.

[0116] In the above embodiment, the information processing device 4 is described as having a communication unit 42 and transmitting and receiving various data between the information processing device 4 and the user terminal device 5 via the communication unit 42, but the information processing device 4 may also operate as a standalone device.

[0117] In the above embodiment, the case where the information processing device 4 operates according to the flowcharts shown in Figures 9 and 10 has been described, but some of the steps (units) may be omitted, or other steps may be added. In this case, the omitted steps (units) may be executed by an external system.

[0118] In the above embodiment, a case has been described in which the learning function 401 of the information processing device 4 generates the first to eighth learning models 16-1 to 16-8, and the image proofreading function 402 uses the first to eighth learning models 16-1 to 16-8 to generate proofread image data 12 from the original image data 11. In contrast, the learning model 16 generated by the learning function 401 and used by the image proofreading function 402 may be any one of the first to eighth learning models 16-1 to 16-8, or any combination of two or more of them.

[0119] In the above embodiment, the image proofing function 402 of the information processing device 4 accepts a specification of a first printing specification in the requested data and a specification of a second printing specification in the predicted data on the image proofing input screen 17. Alternatively, the image proofing function 402 may accept, for example, a specification of proofing processes (which may be some or all) to be performed on the manuscript image data 11 from among a composition limit setting process, an ink color selection process, a color scheme editing process, a character processing process, an overlap area editing process, a color-to-color editing process, and a halftone dot setting process on the image proofing input screen 17. In this case, the data generation unit 4021 may input the manuscript image data 11 to a trained learning model 16 corresponding to the specified proofing process, thereby generating proofed image data 12 that has undergone the specified proofing process.

[0120] In the above embodiment, a case has been described in which a neural network is used as the learning model 16 for realizing machine learning by the learning function 401, but other machine learning models may also be used. Examples of other machine learning models include tree types such as decision trees and regression trees, ensemble learning such as bagging and boosting, and neural network types (including deep learning) such as recurrent neural networks, convolutional neural networks, and LSTM. hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, k-means, etc. Examples include multivariate analyses such as staring type, principal component analysis, factor analysis, and logistic regression, as well as support vector machines.

[0121] (Inference device, inference method and inference program) The present invention can be provided not only in the form of the information processing device 4 (information processing method or information processing program) according to the above embodiment, but also in the form of an inference device (inference method or inference program) used to infer proofread image data 12. In this case, the inference device (inference method or inference program) can include a memory and a processor, and the processor executes a series of processes. The series of processes includes a data acquisition process (data acquisition step) for acquiring manuscript image data 11, and an inference process (inference step) for inferring proofread image data 12 by performing a predetermined proofreading process on the manuscript image data 11 once the manuscript image data 11 has been acquired in the data acquisition process.

[0122] Various aspects of the present disclosure are summarized below as appendices.

[0123] (Appendix 1) a data acquisition unit that acquires original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object; a data generation unit that generates proofread image data by performing a predetermined proofreading process on the original image data by inputting the original image data acquired by the data acquisition unit into a learning model, The learning model is A trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the proof image data. Image proofreader.

[0124] (Appendix 2) The calibration process includes: a process of editing the object image included in the document image data with respect to at least one editing item among the layout, shape, size, and color scheme of the object image; 10. An image proofing device as described in Appendix 1.

[0125] (Appendix 3) The calibration process includes: a composition limit setting process for editing the object image so that the object image falls within a composition limit corresponding to a printable range by the printing device on the printing surface; an ink color selection process for selecting ink colors to be used for printing the object image from a plurality of types of color guides that indicate ink colors printable by the printing device, based on color samples included in the document image data; a color scheme editing process for editing the object image so that the color scheme corresponds to the ink colors printable by the printing device; If the object image includes text, a text processing process is performed to process the text. an overlapping area editing process for editing the object images in the overlapping area so as to unify the colors of the object images in the overlapping area, if there is an overlapping area where the object images overlap each other; a color editing process for setting a gap between the object images printed in different ink colors when a plurality of ink colors are used to print the object images; and a halftone dot setting process for setting at least one of the screen ruling and the screen angle of the halftone dots corresponding to the ink colors for each of the ink colors when a plurality of ink colors are used to print the object image; At least one of 1. An image proofing device as described in Appendix 2.

[0126] (Appendix 4) The data acquisition unit acquiring, together with the document image data, request data relating to a first printing specification required when printing the image; The data generation unit generating proofread image data by performing the proofreading process on the original image data so as to satisfy the required data by inputting the original image data and the required data acquired by the data acquisition unit into a learning model; The learning model is A trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the requested data, and the proof image data. 4. An image proofing device according to any one of claims 1 to 3.

[0127] (Appendix 5) The first printing specification is a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 10. An image proofing device as described in Appendix 4.

[0128] (Appendix 6) The data generation unit The manuscript image data acquired by the data acquisition unit is input into a learning model to generate proofread image data obtained by performing the proofreading process on the manuscript image data and prediction data regarding second printing specifications predicted when the image is printed; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the proof image data, and the predicted data; 4. An image proofing device according to any one of claims 1 to 3.

[0129] (Appendix 7) The second printing specification is a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 10. An image proofing device as described in Appendix 6.

[0130] (Appendix 8) The data acquisition unit acquiring, together with the document image data, request data relating to a first printing specification required when printing the image; The data generation unit By inputting the manuscript image data and the required data acquired by the data acquisition unit into a learning model, proofread image data in which the proofreading process is performed on the manuscript image data so as to satisfy the required data, and predicted data regarding second printing specifications predicted when the image is printed are generated; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the requested data, and the proof image data and the predicted data; 4. An image proofing device according to any one of claims 1 to 3.

[0131] (Appendix 9) The first printing specification and the second printing specification are different, The first printing specification and the second printing specification are: a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 9. An image proofing device as described in Appendix 8.

[0132] (Appendix 10) The data generation unit The manuscript image data acquired by the data acquisition unit is input into a learning model to generate proofread image data obtained by performing the proofreading process on the manuscript image data and support data indicating support details for the proofreading process; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the proofread image data, and the support data; 4. An image proofing device according to any one of claims 1 to 3. [Explanation of symbols]

[0133] 1...printing system, 2...printing device, 3...plate making device, 4...information processing device, 5...user terminal device, 10...container (printing object), 11...original image data, 12...Proofreading image data, 13...Proofreading processing content data, 14...Results data, 15-1~15-8...Training data, 16, 16-1~16-8...Training model, 17...Image proofreading input screen, 18...Image proofreading output screen, 40...control unit, 41...storage unit, 42...communication unit, 43...input unit, 44...display unit, 100...image, 110...original image, 111...original color sample, 120...proof image, 121...proof color sample, 400...Image management function, 401...Learning function, 402...Image proofreading function, 410...Image management database, 411...Model management database, 412...information processing program, 4010...learning data acquisition unit, 4011...machine learning unit, 4020...data acquisition unit, 4021...data generation unit, 4022...output processing unit

Claims

1. a data acquisition unit that acquires original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object; a data generation unit that generates proofread image data by performing a predetermined proofreading process on the original image data by inputting the original image data acquired by the data acquisition unit into a learning model, The learning model is A trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the proof image data. Image proofreader.

2. The calibration process includes: a process of editing the object image included in the document image data with respect to at least one editing item among the layout, shape, size, and color scheme of the object image; 2. The image proofing device of claim 1.

3. The calibration process includes: a composition limit setting process for editing the object image so that the object image falls within a composition limit corresponding to a printable range by the printing device on the printing surface; an ink color selection process for selecting ink colors to be used for printing the object image from a plurality of types of color guides that indicate ink colors printable by the printing device, based on color samples included in the document image data; a color scheme editing process for editing the object image so that the color scheme corresponds to the ink colors printable by the printing device; If the object image includes text, a text processing process is performed to process the text. an overlapping area editing process for editing the object images in the case where there is an overlapping area where the object images overlap each other so as to unify the colors of the object images in the overlapping area; a color editing process for setting a gap between the object images printed in different ink colors when a plurality of ink colors are used to print the object images; and a halftone dot setting process for setting at least one of the screen ruling and the screen angle of the halftone dots corresponding to the ink colors for each of the ink colors when a plurality of ink colors are used to print the object image; At least one of 3. The image proofing device of claim 2.

4. The data acquisition unit acquiring, together with the document image data, request data relating to a first printing specification required when printing the image; The data generation unit generating proofread image data by performing the proofreading process on the original image data so as to satisfy the required data by inputting the original image data and the required data acquired by the data acquisition unit into a learning model; The learning model is A trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the requested data, and the proof image data.

4. An image proofing device according to claim 1.

5. The first printing specification is a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 5. The image proofing device of claim 4.

6. The data generation unit The manuscript image data acquired by the data acquisition unit is input into a learning model to generate proofread image data obtained by performing the proofreading process on the manuscript image data and prediction data regarding second printing specifications predicted when the image is printed; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the proof image data, and the predicted data; 4. An image proofing device according to claim 1.

7. The second printing specification is a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 7. The image proofing device of claim 6.

8. The data acquisition unit acquiring, together with the document image data, request data relating to a first printing specification required when printing the image; The data generation unit The manuscript image data and the requested data acquired by the data acquisition unit are input into a learning model, whereby proofread image data in which the proofreading process is performed on the manuscript image data so as to satisfy the requested data, and predicted data relating to second printing specifications predicted when the image is printed are generated; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the requested data, and the proof image data and the predicted data; 4. An image proofing device according to claim 1.

9. the first printing specification and the second printing specification are different, The first printing specification and the second printing specification are: a printing method of the printing device; the printing cost when the image is printed by the printing device; the number of prints when the image is printed by the printing device; the durability of the printing plates used by said printing device; the lead time for producing a printing plate to be used by the printing device; substrate information relating to a substrate on which the printing object is formed; Information about the ink formulation used by the printing device; and Emotional information about the image At least one of 9. The image proofing device of claim 8.

10. The data generation unit The manuscript image data acquired by the data acquisition unit is input into a learning model to generate proofread image data obtained by performing the proofreading process on the manuscript image data and support data indicating support details for the proofreading process; The learning model is a trained model that has been trained by machine learning to determine the correlation between the manuscript image data, the proofread image data, and the support data; 4. An image proofing device according to claim 1.

11. An inference device comprising a memory and a processor, The processor: a data acquisition process for acquiring document image data created as a document of an image to be printed by a printing device on a printing surface of a printing object; When the document image data is acquired by the data acquisition process, an inference process is executed to infer proofread image data obtained by performing a predetermined proofreading process on the document image data. Reasoning device.

12. a learning data acquisition unit that acquires multiple sets of learning data each consisting of input data and output data; a machine learning unit that uses the plurality of sets of learning data acquired by the learning data acquisition unit to train a learning model by machine learning to learn a correlation between the input data and the output data; a learned model storage unit that stores the learned model in which the correlation is learned by the machine learning unit, The input data is Original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object, The output data is The document image data is proofread image data obtained by performing a predetermined proofreading process on the document image data. Machine learning device.

13. 1. A computer-implemented method for image proofing, comprising: a data acquisition step of acquiring original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object; a data generating step of generating proofread image data by performing a predetermined proofreading process on the original image data by inputting the original image data acquired in the data acquiring step into a learning model, The learning model is A trained model that has been trained by machine learning to determine the correlation between the manuscript image data and the proof image data. Image calibration method.

14. An inference method executed by an inference device having a memory and a processor, The processor: a data acquisition process for acquiring document image data created as a document of an image to be printed by a printing device on a printing surface of a printing object; When the document image data is acquired by the data acquisition process, an inference process is executed to infer proofread image data obtained by performing a predetermined proofreading process on the document image data. Reasoning method.

15. 1. A computer-implemented machine learning method comprising: a learning data acquisition step of acquiring a plurality of sets of learning data each composed of input data and output data; a machine learning step of causing a learning model to learn a correlation between the input data and the output data by machine learning using the plurality of sets of learning data acquired by the learning data acquisition step; a learned model storage step of storing the learned model, which has learned the correlation through the machine learning step, in a learned model storage unit; The input data is Original image data created as an original of an image to be printed by a printing device on a printing surface of a printing object, The output data is The document image data is proofread image data obtained by performing a predetermined proofreading process on the document image data. Machine learning methods.

Images