Information processing system, information processing method, and program product

Abstract

An information processing system, an information processing method, and a program product are disclosed. The information processing system includes a processor that generates a generation model by machine learning using learning data that associates a color correction setting value with a temperature and a humidity in a room in which a printed matter printed with an image represented by image data is manufactured, the color correction setting value being used to make a color of the image in the printed matter close to a color shown by the image data, the generation model generating the color correction setting value associated with the input temperature and humidity when the temperature and the humidity included in the learning data are input.

Description

Technical Field

[0001] This invention relates to an information processing system, information processing method, and program product. Background Technology

[0002] Japanese Patent Application Publication No. 2020-195127 discloses an image processing apparatus comprising: an acquisition unit that acquires image data; a first grayscale correction unit that acquires the dark channel value of each region of the image data, estimates the transmittance, and corrects the grayscale; a second grayscale correction unit that corrects the grayscale of the image data using a tone curve; and a control unit that controls which grayscale correction unit is used for grayscale correction. Summary of the Invention

[0003] For example, because the color space in image data differs from the color space that can be reproduced by printing using pigments, the colors shown in the image data cannot always be faithfully reproduced in printed materials. Therefore, printing devices, for example, are equipped with a color correction function that uses color correction settings, represented by hue curves, to make the colors of the printed image approximate the specified colors.

[0004] Furthermore, color perception and preference vary from customer to customer. Therefore, an expert is present at the printing site to visually verify the colors of the printed images obtained from test prints and manually generate color correction settings. Generating such color correction settings requires experience and intuition; it is difficult for someone other than an expert to do so.

[0005] In the printing industry, the experts who generate color correction settings are aging rapidly, and the number of young people entering the printing sector is also decreasing due to declining birth rates. Therefore, there is a possibility that the transfer of technology related to color correction setting generation may not be smooth.

[0006] The purpose of this invention is to provide an information processing system, information processing method, and program product that enables even those who are not experts involved in generating color correction settings to generate such settings.

[0007] According to a first aspect of the present invention, an information processing system is provided, comprising a processor that generates a generative model by machine learning using learning data that correlates color correction setting values ​​with the temperature and humidity of an indoor environment in which printed matter is manufactured using a printing device, the printed matter having an image represented by image data, the color correction setting values ​​being used to make the colors of the image in the printed matter approximate the colors shown by the image data, and the generative model generating the color correction setting values ​​associated with the input temperature and humidity when the temperature and humidity included in the learning data are input.

[0008] According to a second aspect of the present invention, in the information processing system involved in the first aspect, the country name of the printed matter being distributed is added to the learning data, and the processor generates a generative model through machine learning. The generative model generates a color correction setting value associated with the input temperature, humidity, and country name when the temperature, humidity, and country name included in the learning data are input.

[0009] According to a third aspect of the present invention, in the information processing system involved in the first aspect, the ethnicity of the person who will be the recipient of the printed matter is added to the learning data, and the processor generates a generative model through machine learning. The generative model generates a color correction setting value associated with the input temperature, humidity, and ethnicity when the temperature, humidity, and ethnicity included in the learning data are input.

[0010] According to a fourth aspect of the present invention, in the information processing system involved in any of the first to third aspects, the processor uses the learning data of the color correction setting value, including the evaluation value of the color correction setting value reaching or exceeding a preset threshold, for machine learning of the generative model, wherein the evaluation value of the color correction setting value is set based on the similarity between the color of the image in the printed matter and the color shown in the image data.

[0011] According to a fifth aspect of the present invention, in the information processing system involved in the fourth aspect, the evaluation value of the color correction setting value is set by the orderer of the printed matter.

[0012] According to a sixth aspect of the present invention, an information processing system is provided, comprising a processor that uses a generative model to generate color correction settings for manufacturing a new printed matter based on the indoor temperature and humidity when a new printed matter is manufactured using a printing device. The printed matter has an image represented by image data printed on it. The color correction settings are used to make the colors of the image in the printed matter approximate the colors shown by the image data. The generative model is generated by machine learning using learning data. The color correction settings are generated based on the indoor temperature and humidity during the manufacture of the printed matter. In the learning data, the color correction settings are associated with the temperature and the humidity.

[0013] According to a seventh aspect of the present invention, in the information processing system involved in the sixth aspect, the processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at the time of manufacturing the new printed matter and the country name of the country name of the distributed printed matter. The extended generative model is formed by using machine learning with learning data supplemented with the country name of the distributed printed matter to extend the generative model to generate the color correction settings based on the temperature, the humidity, and the country name.

[0014] According to an eighth aspect of the present invention, in the information processing system involved in the sixth aspect, the processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at the time of manufacture of the new printed matter, and the ethnicity of the person to whom the new printed matter is distributed. The extended generative model is formed by machine learning using learning data with the ethnicity of the person to whom the printed matter is distributed, thereby extending the generative model to generate the color correction settings based on the temperature, the humidity, and the ethnicity.

[0015] According to the ninth aspect of the present invention, in the information processing system involved in any of the sixth to eighth aspects, the generative model is a model generated by machine learning using the learning data, wherein the learning data includes a color correction setting value whose evaluation value reaches or exceeds a preset threshold, and the evaluation value of the color correction setting value is set based on the similarity between the color of the image in the printed matter and the color shown in the image data.

[0016] According to a tenth aspect of the present invention, in the information processing system involved in the ninth aspect, the evaluation value of the color correction setting value is set by the orderer of the printed matter.

[0017] According to an eleventh aspect of the present invention, a program product is provided, comprising a program for causing a computer to perform processing, wherein the processing comprises: generating a generative model by machine learning using learning data that correlates color correction setting values ​​with the temperature and humidity of an indoor environment in which printed matter is manufactured using a printing device, the printed matter having an image represented by image data, the color correction setting values ​​being used to make the colors of the image in the printed matter approximate the colors shown by the image data, and the generative model generating the color correction setting values ​​associated with the input temperature and humidity when the temperature and humidity included in the learning data are input.

[0018] According to a twelfth aspect of the present invention, a program product is provided, comprising a program for causing a computer to perform processing, wherein the processing comprises: generating a color correction setting value based on the temperature and humidity of an indoor environment where a printed matter is manufactured using a printing device, by using a generation model generated through machine learning that correlates color correction setting values ​​with the temperature and humidity of the room where the printed matter is manufactured, generating the color correction setting value for manufacturing a new printed matter based on the temperature and humidity at the time of manufacturing the new printed matter, the printed matter having an image represented by image data, the color correction setting value being used to make the color of the image in the printed matter approximate the color represented by the image data.

[0019] According to a thirteenth aspect of the present invention, an information processing method is provided, which uses learning data that associates color correction settings with the temperature and humidity of an indoor environment where printed matter has been manufactured using a printing device, and generates a generative model by machine learning, wherein the printed matter has an image represented by image data, and the color correction settings are used to make the colors of the image in the printed matter close to the colors shown by the image data, and the generative model generates the color correction settings associated with the input temperature and humidity when the temperature and humidity included in the learning data are input.

[0020] (Effect)

[0021] According to the first, sixth, eleventh, twelfth and thirteenth schemes, the following effect is achieved: even experts who are not involved in the generation of color correction settings can generate color correction settings.

[0022] According to the second and third schemes, a generative model can be generated to produce color correction settings that are more in line with the customer's preferences than a generative model that uses temperature and humidity as input data for machine learning.

[0023] According to the fourth scheme, a generative model that generates the following color correction setting value is achieved: the evaluation value is higher than that of the case where the evaluation value of the machine learning data used for the generative model is selected without considering the evaluation value of the color correction setting value.

[0024] According to the fifth scheme, the following effect is achieved: a generative model is generated that reflects the color correction setting value of the orderer's evaluation of the color of the printed matter.

[0025] According to the seventh and eighth solutions, the effect of generating color correction settings is achieved by producing print colors that better suit the customer's preferences compared to generating color correction settings based on temperature and humidity.

[0026] According to the ninth scheme, the following effect can be achieved: the evaluation value is higher than that generated by a generative model that uses machine learning to generate the color correction setting value without considering the evaluation value of the color correction setting value.

[0027] According to the tenth solution, the following effect is achieved: it is able to generate color correction settings that reflect the orderer's evaluation of the color of the printed matter. Attached Figure Description

[0028] Figure 1 This is a diagram illustrating an example of the functional structure of an information processing system;

[0029] Figure 2 This is a diagram illustrating an example of a tone curve;

[0030] Figure 3 This is a diagram showing an example of the main structural components of the electrical system of an information processing system;

[0031] Figure 4 This is a flowchart illustrating an example of the learning process in a generative model;

[0032] Figure 5 This is a flowchart illustrating an example of the process in generating a tone curve. Detailed Implementation

[0033] The present embodiment will now be described with reference to the accompanying drawings. Furthermore, the same symbols are used to label the same components and processes in all the drawings, and repeated descriptions are omitted. The dimensions in the drawings are exaggerated for ease of explanation and may sometimes differ from the actual dimensions.

[0034] Figure 1 This is a diagram illustrating an example of the functional structure of information processing system 1. For example... Figure 1 As shown, the information processing system 1 includes a learning unit 2 and a generation unit 3.

[0035] Information processing system 1 generates color correction setting values ​​through processing by learning unit 2 and generation unit 3. These color correction setting values ​​are used for color adjustment of image data input into an image forming apparatus (not shown), which is an example of a printing device. The image forming apparatus uses the color correction setting values ​​generated by information processing system 1 to perform color adjustment of the image data. Color adjustment includes, for example, adjustments to hue, chroma, and brightness.

[0036] For ease of explanation, the information processing system 1 will be referred to as "System 1" below. Furthermore, in this invention, the generation of the color correction setting value in the information processing system 1 will be explained using the generation example of the hue curve 5, which is an example of a color correction setting value.

[0037] Learning Unit 2 uses a large amount of pre-prepared learning data 4 to perform machine learning on a generative model M for generating hue curves 5.

[0038] Machine learning refers to the process where a computer learns the causal relationships between data in a latent dataset using given sample data, generating a model that represents these causal relationships. Therefore, the more sample data provided, the more accurately the causal relationships between data in the latent dataset will be learned. The sample data used in machine learning is called training data 4.

[0039] Even if the value of an item that affects the printing color of the recording medium changes, if a hue curve 5 can be obtained such that the printing color of the recording medium is close to the customer's specified color based on the change in value, then the printing color of the recording medium can be close to the customer's specified color regardless of how the value of the item changes.

[0040] However, there are multiple items that need to be considered, which are intricately interconnected. Therefore, it is often difficult to physically formulate the relationship between the items affecting the printed color of the recording medium and the desired tone curve 5. Therefore, in this invention, the relationship between the items affecting the printed color of the recording medium and the desired tone curve 5 is mastered through machine learning using data that correlates the tone curve 5 with the items affecting the printed color of the recording medium as learning data 4. Furthermore, the customer is an example of someone who orders printed materials.

[0041] The tone curve 5 that constitutes the learning data 4 was manually generated by an expert (hereinafter referred to as "skilled worker") who generated the tone curve 5. Specifically, the skilled worker considered the values ​​of items that affect the printed color of the recording medium, i.e., the influencing factors, and generated the tone curve 5 based on experience and intuition.

[0042] Thus, learning data 4 is the data that correlates the influencing factors with the hue curve 5 generated by skilled workers based on those values. In learning data 4, the influencing factors are the input data. Additionally, in learning data 4, the hue curve 5 associated with the influencing factors is the supervision data.

[0043] Thus, the learning unit 2 uses the pre-prepared learning data 4 to perform machine learning on a generative model M, which, when inputted with the influencing factors of each learning data 4, outputs the hue curve 5 included in the learning data 4 with the same input influencing factors. The generative model M is a model that expresses the causal relationship between the influencing factors and the hue curve 5 based on the relationship between the influencing factors included in the learning data 4 and the hue curve 5.

[0044] The method of using learning data 4, which associates the desired supervised data with the input data, to perform machine learning on the generative model M is called "supervised machine learning".

[0045] In supervised machine learning, there are well-known methods such as linear regression, logistic regression, random forests, boosting, support vector machines, deep learning, and autoregressive models. As an example, the learning unit 2 of this invention performs machine learning using a generative model M derived from deep learning. Deep learning is an example of a neural network consisting of multiple layers forming intermediate layers (also called "hidden layers"). For example, a convolutional neural network (CNN) is an example of deep learning. Needless to say, the learning unit 2 can also replace CNN, for example, using well-known supervised machine learning methods such as support vector machines.

[0046] The generation unit 3 generates a tone curve 5 for printing new orders from customers. Specifically, the generation unit 3 inputs the values ​​of influencing factors in manufacturing new printed materials into a generation model M obtained through machine learning by the learning unit 2. The generation model M generates a tone curve 5 based on the input values ​​of the influencing factors, making the printed color of the recording medium approximate the customer's specified color. The generation unit 3 outputs the tone curve 5 generated by the generation model M to the image forming apparatus.

[0047] Figure 2 This is a diagram representing an example of tone curve 5. Figure 2 The horizontal axis represents the color values ​​of the image before correction, i.e., the color values ​​before correction. Figure 2 The vertical axis represents the color values ​​of the corrected image, i.e., the corrected color values. By changing the curve of tone curve 5, the original color values ​​can be corrected to the target color values.

[0048] The image forming apparatus ultimately converts image data into CMYK data for printing on a recording medium using pigments. "C" represents cyan, "M" represents magenta, "Y" represents yellow, and "K" represents black. Therefore, tone curve 5 corrects the CMYK data as an example of color values. Thus, tone curve 5 is generated for each color in CMYK.

[0049] Furthermore, tone curve 5 can also correct color values ​​other than CMYK data. For example, when the image printed on the recording medium is a photograph, the color values ​​of the photograph are mostly represented by RGB data. Therefore, tone curve 5 can also be generated to correct the RGB data before conversion to CMYK data. In this invention, as an example, tone curve 5 is used to correct CMYK data.

[0050] In addition, as an example of an influencing factor in this invention, the temperature and humidity of the room where the printed matter is manufactured, i.e., the temperature and humidity of the room where the image forming apparatus is installed, are used.

[0051] In an image forming apparatus, for example, a recording medium on which an image has been formed using pigments such as toners is heated to fix the image onto the recording medium. Therefore, when the room temperature changes, the melting point of the pigments changes. As a result, the printed color of the recording medium sometimes deviates from the color specified by the customer.

[0052] Furthermore, while paper is commonly used as the recording medium, the amount of moisture absorbed by the recording medium varies with changes in indoor humidity. This variation in moisture absorption can easily lead to uneven fixing of the toner applied to the recording medium. Additionally, when using toner, the image forming apparatus uses electrostatic transfer to transfer the toner onto the recording medium. Changes in indoor humidity cause variations in the charge on the photosensitive element that attaches the toner electrostatically. As a result, the printed color on the recording medium may sometimes deviate from the customer's specified color. Therefore, temperature and humidity are examples of influencing factors.

[0053] The system 1 that performs such processing is, for example, constructed using a computer 10.

[0054] Figure 3 This is a diagram illustrating an example of the main structural components of the electrical system of system 1, which is constructed using computer 10.

[0055] The computer 10 constituting system 1 includes a CPU (Central Processing Unit) 11, which serves as an example of a processor responsible for executing the functions of system 1. Additionally, the computer 10 includes RAM (Random Access Memory) 12, which acts as a temporary working area for the CPU 11. Furthermore, the computer 10 includes non-volatile memory 13 and input / output interfaces (I / O) 14. The CPU 11, RAM 12, non-volatile memory 13, and I / O 14 are connected via a bus 15.

[0056] Non-volatile memory 13 is an example of a storage device that retains stored information even when the power supplied to the non-volatile memory 13 is cut off. For example, semiconductor memory can be used as non-volatile memory 13, but hard disks can also be used. As in the generation model M, information that needs to be stored even when the power supply to the computer 10 is cut off is stored in the non-volatile memory 13.

[0057] In I / O 14, for example, communication unit 16, input unit 17, and display unit 18 are connected.

[0058] The communication unit 16 is connected to a communication line (not shown) and has a communication protocol for sending and receiving data with an external device connected to the same communication line.

[0059] Input unit 17 is a device that accepts user instructions and notifies CPU 11. Input unit 17 may include, for example, buttons, touch panels, mice, keyboards, and pointing devices.

[0060] Display unit 18 is an example of a display device that displays information processed by CPU 11 as an image. For example, a liquid crystal display (LCD) or an organic EL (Electroluminescence) display can be used as display unit 18.

[0061] Furthermore, the units connected to I / O 14 are not limited to communication unit 16, input unit 17, and display unit 18. The required units are selected and connected to I / O 14 based on the functions of System 1.

[0062] <Learning Stage>

[0063] Next, the learning process of the generative model M performed by System 1 will be explained. Figure 4 This is a flowchart illustrating an example of the process in the learning process of the generative model M, which is executed by the CPU 11 of the computer 10 constituting system 1 upon receiving a learning start instruction from the user.

[0064] The information processing program for learning the generative model M is pre-stored in the non-volatile memory 13 of the computer 10. The CPU 11 reads the information processing program stored in the non-volatile memory 13 and executes the learning process for the generative model M. Furthermore, it is assumed that the non-volatile memory 13 stores pre-prepared learning data 4 for each color of the CMYK. The user specifies which color's hue curve 5 in the CMYK generative model M is used for machine learning.

[0065] First of all, Figure 4 In step S10, CPU11 selects a learning data 4 of a color specified by the user. The learning data 4 selected in step S10 is specifically referred to as "selected learning data 4".

[0066] In step S20, CPU 11 acquires the influencing factors and hue curve 5 included in the selected learning data 4. In the example of the present invention, temperature and humidity are acquired as influencing factors.

[0067] In step S30, CPU11 performs machine learning on the generative model M, which generates the same hue curve 5 obtained in step S20 when the temperature and humidity are input.

[0068] In step S40, CPU 11 determines whether there is any unselected learning data 4 among the learning data 4 for the user-specified color that has not yet been selected in step S10. If there is unselected learning data 4, proceed to step S10. Then, select any one of the unselected learning data 4 as the new selected learning data 4. Afterwards, steps S20 to S40 are executed. That is, CPU 11 uses all the learning data 4 pre-prepared for the user-specified color to perform machine learning on the generative model M.

[0069] On the other hand, if the determination process in step S40 determines that there is no unselected learning data 4, then proceed to step S50.

[0070] In step S50, CPU 11 stores the generative model M, which has undergone machine learning in step S30, in non-volatile memory 13. This concludes the process. Figure 4 The learning process of the generative model M is shown.

[0071] If performed for each color of CMYK Figure 4 The learning process of the generative model M shown is used to obtain the generative model M for each color.

[0072] <Generation Phase>

[0073] When a customer commissions the manufacture of new printed materials, the printing company obtains the image data of the printed materials from the customer and prints them using an image forming apparatus. Meanwhile, the temperature and humidity inside the room where the image forming apparatus is located are constantly changing. Therefore, the printing company generates a tone curve 5 each time a printed item is manufactured.

[0074] Below, using via Figure 4 The generative model M, obtained by learning the generative model M shown, explains the generation process of the hue curve 5 corresponding to the values ​​of influencing factors.

[0075] Figure 5 This is a flowchart illustrating an example of the process in which the tone curve 5 is generated. The process of generating tone curve 5 is executed by the CPU 11 of the computer 10 constituting system 1 upon receiving an instruction from the user to start generating tone curve 5.

[0076] CPU 11 reads the information processing program stored in non-volatile memory 13 and executes the generation process of tone curve 5. Furthermore, it is assumed that the non-volatile memory 13 pre-stores the generation model M for each color of the CMYK, generated through the learning process of the generation model M. The user specifies which color of the CMYK tone curve 5 is generated.

[0077] First of all, Figure 5 In step S100, the CPU 11 acquires the temperature and humidity of the room where the image forming apparatus is installed. Specifically, the CPU 11 acquires the temperature and humidity from a sensor connected to a communication line via the communication unit 16. In this case, the CPU 11 preferably acquires the temperature and humidity from the same sensor that measured the temperature and humidity included in the learning data 4. This is because when the temperature and humidity are acquired from a different sensor, the relationship between the temperature and humidity and the hue curve 5 deviates from the relationship shown in the learning data 4. For example, the CPU 11 acquires the temperature and humidity from a sensor built into the image forming apparatus. If no sensor is built into the image forming apparatus, the CPU 11 can acquire the temperature and humidity from a sensor installed indoors.

[0078] In step S110, CPU 11 inputs the temperature and humidity obtained in step S100 into a generation model M corresponding to the color specified by the user. Then, it outputs a hue curve 5 corresponding to the temperature and humidity obtained in step S100 from the generation model M.

[0079] In addition, even when the system 1 and the sensor are not connected via a communication line, the user can still input temperature and humidity through the input unit 17 of the system 1.

[0080] In step S120, CPU 11 stores the output hue curve 5 in non-volatile memory 13.

[0081] In step S130, CPU 11 controls communication unit 16 to send the tone curve 5 output from generation model M to image forming apparatus via communication line. This concludes the process. Figure 5 The generation process of the hue curve 5 shown.

[0082] If performed for each color of CMYK Figure 5 The generation process of the hue curve 5 shown generates a generation model M for each color and sends it to the image forming apparatus. The image forming apparatus uses the hue curve 5 received from system 1 to correct the colors of the printed material.

[0083] <Variation Example 1>

[0084] The above example illustrates the generation of tone curve 5, which incorporates temperature and humidity as influencing factors. However, influencing factors are not limited to temperature and humidity.

[0085] For example, the preferred colors may differ from country to country where the printed materials are distributed. People in Asian countries generally tend to prefer yellowish colors to make their skin appear warmer and healthier. Therefore, skin tones and other colors in printed materials are often adjusted to be more vibrant.

[0086] On the other hand, people in European countries tend to prefer more neutral or natural colors. Additionally, they often prefer strong colors like red or pink tones for skin tones. Therefore, skin tones in printed materials are often adjusted to appear more natural and healthy.

[0087] This national preference for certain colors may be partly due to biological reasons. For example, it is believed that pupil color influences color perception or preference. People with blue pupils, compared to those with brown pupils, have difficulty perceiving changes in contrast and therefore tend to prefer blue or cool-toned colors. On the other hand, people with brown pupils tend to prefer warm colors.

[0088] For this reason, the preferred colors may differ in each country distributing the printed materials. Therefore, the name of the country distributing the printed materials is also an example of an influencing factor.

[0089] The above investigation shows that a color curve 5 can also be generated that takes temperature, humidity, and the country of distribution of the printed materials as influencing factors.

[0090] In this case, it is sufficient to prepare in advance the learning data 4 that correlates the tone curve 5 generated by skilled workers based on temperature, humidity, and the country name of the printed materials with temperature, humidity, and the country name of the printed materials.

[0091] Based on this, Figure 4 In step S20, CPU 11 acquires the temperature, humidity, and country of origin of the printed materials included in the selected learning data 4 as influencing factors. Additionally, in Figure 4 In step S30, CPU 11 performs machine learning on the generative model M, which, given the acquired temperature, humidity, and country of distribution of the printed materials, generates the same color curve 5 included in the selected learning data 4. Thus, a generative model M is obtained that generates the color curve 5 that takes temperature, humidity, and country of distribution of the printed materials as influencing factors. This resulting generative model M is an example of an extended generative model of the present invention.

[0092] In addition, Figure 5In step S100, the CPU 11 acquires the temperature and humidity of the room where the image forming apparatus is located, as well as the country name of the printed materials to be distributed. The country name of the printed materials to be distributed can be input by the user through the input unit 17 of the system 1.

[0093] exist Figure 5 In step S110, CPU 11 inputs the acquired temperature, humidity, and country name of the distributed printed materials into the generation model M obtained in Modified Example 1. Then, the generation model M outputs a color curve 5 corresponding to the temperature, humidity, and country name of the distributed printed materials.

[0094] In this way, in addition to temperature and humidity, the country of distribution of the printed materials can also be added as an influencing factor, and the color tone curve of the printed materials can be generated based on temperature, humidity, and the country of distribution of the printed materials.

[0095] <Variation Example 2>

[0096] As illustrated in Variation 1, influencing factors are not limited to temperature and humidity. The race of the people to whom the printed materials are distributed is also an example of an influencing factor. This is because, as illustrated in Variation 1, different races may have different pupil colors, and their color preferences may also differ.

[0097] Therefore, it can be seen that a color curve 5 can also be generated that takes temperature, humidity, and the ethnicity of the people to whom the printed matter is distributed as influencing factors.

[0098] In this case, it is sufficient to prepare in advance the learning data 4 that correlates the tone curve 5 generated by the skilled worker based on temperature, humidity, and the ethnicity of the people who are the recipients of the printed materials with temperature, humidity, and the ethnicity of the people who are the recipients of the printed materials.

[0099] Based on this, Figure 4 In step S20, CPU 11 acquires the temperature, humidity, and ethnicity of the people to whom the printed materials are distributed, included in the selection learning data 4, as influencing factors. Additionally, in Figure 4 In step S30, CPU 11 performs machine learning on the generative model M, which, given the acquired temperature, humidity, and ethnicity of the recipients of the printed materials, generates the same color tone curve 5 included in the selected learning data 4. Thus, a generative model M is obtained that generates the color tone curve 5 by taking temperature, humidity, and ethnicity of the recipients of the printed materials as influencing factors. This resulting generative model M is an example of an extended generative model of the present invention.

[0100] In addition, Figure 5In step S100, the CPU 11 acquires the temperature and humidity of the room where the image forming apparatus is located, as well as the ethnicity of the people to whom the printed matter will be distributed. The ethnicity of the people to whom the printed matter will be distributed can be input by the user through the input unit 17 of the system 1.

[0101] exist Figure 5 In step S110, CPU 11 inputs the acquired temperature, humidity, and ethnicity of the people to whom the printed matter is distributed into the generation model M obtained in Modification 2. Thus, the generation model M outputs a tone curve 5 corresponding to the temperature, humidity, and ethnicity of the people to whom the printed matter is distributed.

[0102] In this way, in addition to temperature and humidity, the ethnicity of the people to whom the printed materials will be distributed can also be added as an influencing factor, and the color tone curve of the printed materials can be generated based on temperature, humidity, and the ethnicity of the people to whom the printed materials will be distributed.

[0103] <Other Influencing Factors>

[0104] Influencing factors are not limited to the temperature, humidity, country of origin of the printed materials, and ethnicity of the recipients of the printed materials, as described above.

[0105] For example, even when colors are specified, each customer's preferred color may differ. Such customer preferences can also vary depending on whether the customer is a business or an individual.

[0106] Furthermore, if the customer has previously commissioned the printing of materials, there may be various requests regarding the colors of the printed materials delivered so far. Therefore, it may be possible to obtain the customer's preferred colors based on past requests, according to the customer's name.

[0107] That is, the way customers express their preferences and the customer's name are also examples of influencing factors.

[0108] Sometimes it's better to adjust the colors of the printed material based on the content. For example, the required colors will differ depending on whether the printed material includes people, landscapes, or food. For instance, to emphasize freshness, food colors may sometimes require a higher saturation compared to the colors of people or landscapes.

[0109] In addition, the required color may vary depending on the contrast of the image or the area of ​​the region covered by the same color, i.e., the amount of solidity.

[0110] That is, the type of subject, contrast, and amount of solidity of the printed content are also examples of influencing factors.

[0111] Sometimes it's better to adjust the colors of the printed material according to its intended use. For example, the required colors will differ depending on whether the printed material is a catalog, poster, or album. For instance, in an album, the colors need to be closer to the customer's specified colors compared to a catalog or poster.

[0112] That is, the intended use of printed materials is also an example of an influencing factor.

[0113] Sometimes it's better to adjust the printed colors according to the user's color correction profile. A color correction profile is data that defines the color correction functions pre-installed in the image forming apparatus.

[0114] Color calibration profiles may include, for example, CMYK input profiles, device link profiles, output profiles, and RGB input profiles.

[0115] A CMYK input profile is a configuration file that defines the CMYK target used to simulate other printing devices. A device link profile is a configuration file that combines the color characteristics of the printing target and the output device. An output profile is a configuration file that defines the color characteristics of the output device. An RGB input profile is a configuration file that defines the RGB color space applied when printing RGB data.

[0116] Each color correction profile consists of multiple types. Users specify the profile to use for each profile individually, based on its name. Because the colors of printed materials vary depending on the profile used, the color correction amount based on tone curve 5 also needs to be varied.

[0117] That is, the color correction profile used by the user, specifically the CMYK input profile, device link profile, output profile, and RGB input profile are also examples of influencing factors.

[0118] Sometimes it's better to adjust the printed color according to the type of recording medium used for printing. For example, if the basis weight or thickness of the paper used as a recording medium varies, the degree of fixation may differ even if the paper is heated for the same amount of time to fix the image. Furthermore, some papers have a glossy finish due to a coating on their surface. The gloss of the paper varies depending on the amount of coating applied; if the amount of coating varies, the degree of fixation may differ even if the paper is heated for the same amount of time to fix the image. Additionally, if the quality of the pulp used for the paper varies, the degree of fixation may differ even if the paper is heated for the same amount of time to fix the image. Moreover, even if the type of paper is the same, the origin of the pulp may vary depending on the manufacturing batch, thus affecting the quality of the pulp. Because the degree of fixation of the image on the paper alters the color of the printed product, it is also necessary to adjust the color correction amount based on tone curve 5.

[0119] That is, the type of recording media used for printing is another influencing factor. The type of recording media is indicated, for example, by the media's basis weight, thickness, gloss, pulp quality, and manufacturing batch. Inherent information related to such recording media is provided by the media's seller.

[0120] Sometimes it is better to change the color of the printed matter depending on the type of ink used for printing. The product number of the ink used in the image forming apparatus is assigned by the seller of the image forming apparatus. However, sometimes even with the same product number, the characteristics of the ink can vary within acceptable limits if the manufacturing batches are different. Because the different characteristics of the ink will change the color of the printed matter, it is also necessary to change the amount of color correction based on tone curve 5.

[0121] That is, the type of ink used for printing is also an example of an influencing factor. The type of ink can be indicated, for example, by the batch number of the ink it was manufactured in.

[0122] Sometimes it is better to adjust the printed color according to the characteristics of the image forming apparatus used for printing. For example, if the model of the image forming apparatus used for printing is different, the color of the printed material may vary even when using the same image data. Furthermore, even if the model is the same, if the individual image forming apparatus used for printing is different, the color of the printed material may vary even when using the same image data. Also, even if the individual is the same, if the number of years the image forming apparatus used for printing has changed, the color of the printed material may vary even when using the same image data. Moreover, even if the individual is the same, if the number of days since the last maintenance or repair has changed, the color of the printed material may vary even when using the same image data. Therefore, it is also necessary to vary the color correction amount based on tone curve 5 according to the characteristics of the image forming apparatus used for printing.

[0123] That is, the characteristics of the image forming apparatus used for printing are also an example of influencing factors. The characteristics of the image forming apparatus are expressed, for example, by the model of the image forming apparatus, individual information, number of years, and number of days since the last maintenance or overhaul.

[0124] Sometimes it's better to adjust the printed colors based on the characteristics of the Digital Front End (DFE) used for image data processing. For example, if different types of DFEs are used for printing, the printed colors may vary even when using the same image data. Furthermore, even if the types are the same, different versions of the DFEs used for printing may cause color changes due to software updates or other factors, even when using the same image data. Therefore, it's necessary to adjust the color correction based on tone curve 5 according to the characteristics of the DFE used for printing.

[0125] That is, the characteristics of the DFE used for printing are also an example of influencing factors. The characteristics of a DFE are, for example, expressed by the type and version of the DFE.

[0126] Sometimes it's better to adjust the printed color based on the characteristics of the RIP (Raster Image Processor) that converts the image, represented by vector data, into raster data understandable by the image forming apparatus. For example, if different types of RIPs are used for printing, the printed color may vary even when using the same image data. Furthermore, even if the types are the same, different versions of the RIPs used for printing may result in different colors due to improvements in processing, even when using the same image data. Therefore, it's also necessary to vary the color correction amount based on tone curve 5 according to the characteristics of the RIP used for printing.

[0127] That is, the characteristics of the RIP used for printing are also an example of influencing factors. RIP characteristics can be represented, for example, by the type and version of the RIP.

[0128] For ease of explanation, factors other than temperature and humidity in the room where such an image forming apparatus is installed will be referred to as "additional influencing factors".

[0129] System 1 can also combine at least one of the additional influencing factors mentioned above, in addition to the temperature and humidity of the room where the image forming apparatus is located, to generate tone curve 5.

[0130] In this case, it is sufficient to prepare in advance the learning data 4, which correlates the color curve 5 generated by skilled workers based on temperature, humidity, and additional influencing factors with temperature, humidity, and additional influencing factors.

[0131] Based on this, Figure 4 In step S20, CPU 11 acquires the temperature, humidity, and additional influencing factors included in the selected learning data 4 as influencing factors. Additionally, in Figure 4 In step S30, CPU 11 performs machine learning on the generative model M, which, when inputted with the acquired temperature, humidity, and additional influencing factors, generates the same hue curve 5 included in the selected learning data 4. Thus, a generative model M is obtained that generates the hue curve 5 with temperature, humidity, and additional influencing factors as influencing factors.

[0132] In addition, Figure 5 In step S100, the CPU 11 acquires the temperature and humidity of the room where the image forming apparatus is located, as well as additional influencing factors. These additional influencing factors can be input by the user through the input unit 17 of system 1. Furthermore, if the influencing factor information containing the additional influencing factors is stored as data, the CPU 11 can also retrieve the influencing factor information from the storage device. For example, if the influencing factor information is stored in the storage device of the image forming apparatus, the CPU 11 controls the communication unit 16 to retrieve the influencing factor information from the image forming apparatus via a communication line.

[0133] exist Figure 5 In step S110, CPU 11 inputs the acquired temperature, humidity, and additional influencing factors into the obtained generation model M. Then, the generation model M outputs a hue curve 5 corresponding to the temperature, humidity, and additional influencing factors.

[0134] <Selection of Tone Curves for Learning Data>

[0135] Thus, the tone curve 5 used for learning data 4 is a tone curve generated by a skilled worker considering the values ​​of various influencing factors and based on experience and intuition. However, even printed materials manufactured based on tone curve 5 generated by a skilled worker may not satisfy all customers regarding the color of the printed materials.

[0136] Therefore, printing companies sometimes conduct interviews and surveys to gather customer feedback on the colors of printed materials. The evaluation of the colors is, for example, represented by a score from 0 to 100. A score of 100 indicates the highest level of customer satisfaction with the colors of the printed material. In other words, the evaluation value is based on the similarity between the colors of the images in the printed material and the colors shown in the image data.

[0137] A high color evaluation value for a printed product indicates a high evaluation of the hue curve 5 used for printing. Conversely, a low color evaluation value for a printed product indicates a low evaluation of the hue curve 5 used for printing. In other words, the color evaluation value of a printed product is also the evaluation value of the hue curve 5.

[0138] To improve the color evaluation value of printed materials that have undergone color correction based on the tone curve 5 generated by system 1, it is preferable to use learning data 4, which includes the tone curve 5 with high evaluation, for machine learning of the generative model M.

[0139] Therefore, it is also possible to select only the tone curve 5 with an evaluation value that reaches or exceeds a preset threshold from the tone curve 5 generated by the skilled worker, and use it as the tone curve 5 of the learning data 4.

[0140] For example, the threshold can be set to 80 points. In this case, CPU11... Figure 4 In step S10 of the learning process of the generative model M shown, only the learning data 4 including the hue curve 5 with an evaluation value of 80 or above is selected, and machine learning of the generative model M is performed.

[0141] Based on this, CPU11 Figure 5 In step S110 of the generation process of the hue curve 5 shown, the influencing factors are input into the generation model M obtained only from the learning data 4, which includes the hue curve 5 with an evaluation value of 80 points or above, to generate the hue curve 5 corresponding to the influencing factors.

[0142] Therefore, compared to the case where learning data 4 is selected for the learning processing of the generative model M without considering the evaluation value of the hue curve 5, a generative model M that generates a hue curve 5 with a higher evaluation value can be obtained. Furthermore, the threshold value is an example; the threshold can be changed by the user.

[0143] The above description illustrates one aspect of System 1 using an embodiment, but the disclosed aspect of System 1 is merely an example. The aspect of System 1 is not limited to the scope described in the embodiment. Various modifications or improvements can be made to the embodiment without departing from the spirit of the invention. The manner in which such modifications or improvements are made is also included within the scope of the disclosed technology. For example, modifications may be made without departing from the spirit of the invention. Figure 4 and Figure 5 The internal processing order of each process is shown.

[0144] Furthermore, in the above implementation method, as an example, the implementation is carried out by software. Figure 4 The learning process of the generative model M shown, and Figure 5 The generation process of the hue curve 5 shown has been explained. However, it is also possible to perform the same processing as shown in the flowchart using hardware. In this case, compared to the case where each process is implemented in software, a higher processing speed is achieved.

[0145] In this embodiment, each process is executed by any computer. Alternatively, any computer may execute these processes via a processor as hardware, a program as software, or a combination thereof. In this case, the processor is configured to cooperate with the program to execute the various processes of this embodiment, and may also function as a unit or means of this embodiment. Furthermore, the execution order of the processes performed by the processor is not limited to the order described and may be appropriately varied. Any computer may also be a general-purpose computer, a computer for a specific purpose, a workstation, or other system capable of executing the processes.

[0146] A processor can be composed of one or more hardware components, and the type of hardware is not limited. For example, a processor can be composed of programmable logic devices such as CPUs, MPUs (Micro Processing Units), FPGAs (Field Programmable Gate Arrays), dedicated circuits for performing specific processes such as ASICs (Application Specific Integrated Circuits), GPUs (Graphics Processing Units), or NPUs (Neural Processing Units). Furthermore, the hardware can also be a combination of different types of hardware. When multiple hardware components are configured to execute one or more processes of a processor, these multiple hardware components can exist in physically separate devices or in the same device. Additionally, in any embodiment, the order of the processes performed by the processor is not limited to the above order and can be appropriately varied. Moreover, the hardware can be composed of circuits composed of semiconductor elements and other circuit components.

[0147] Furthermore, the program can also be firmware, microcode, or other software. Additionally, the program can be, for example, a group of program modules, each of whose functions can be implemented by a processor configured to perform those functions. The program can also be program code or multiple code segments stored on one or more non-transitory computer-readable media (e.g., storage media or other storage devices). The program can also be segmented and stored on multiple non-transitory computer-readable media existing in physically separate devices. Program code or code segments can represent steps, functions, subroutines, routines, subroutines, modules, software packages, categories, or commands, data structures, or any combination of program statements. Program code or code segments can also be connected to other code segments or hardware circuitry by sending and receiving information, data, arguments, parameters, or the contents of memory. Furthermore, the program of this application can also be provided as a program product.

[0148] Furthermore, System 1 can also download information processing programs from external devices through communication unit 16 and store the downloaded information processing programs in non-volatile memory 13.

[0149] (Postscript) ((1))

[0151] An information processing system, wherein,

[0152] Equipped with a processor

[0153] The processor

[0154] Using learning data that correlates color correction settings with the temperature and humidity of an indoor environment where printed materials are manufactured using a printing device, a generative model is generated via machine learning. The printed materials contain images represented by image data, and the color correction settings are used to make the colors of the images in the printed materials approximate the colors shown by the image data. The generative model generates the color correction settings associated with the input temperature and humidity, given the temperature and humidity included in the learning data. ((2))

[0156] According to the information processing system described in (1), wherein,

[0157] The names of the countries from which the printed materials were distributed are appended to the learning data.

[0158] The processor generates a generative model through machine learning. When the generative model is input with the temperature, humidity, and country name included in the learning data, it generates a color correction setting value associated with the input temperature, humidity, and country name. ((3)))

[0160] According to the information processing system described in (1), wherein,

[0161] The ethnicity of the people who will be the recipients of the printed materials is added to the learning data.

[0162] The processor generates a generative model through machine learning. When the generative model is input with the temperature, humidity, and ethnicity included in the learning data, it generates color correction settings associated with the input temperature, humidity, and ethnicity. ((4))

[0164] The information processing system according to any one of ((1)) to ((3)) is wherein,

[0165] The processor uses the learning data, including the color correction setting value whose evaluation value reaches or exceeds a preset threshold, for machine learning of the generative model. The evaluation value of the color correction setting value is set based on the similarity between the color of the image in the printed matter and the color shown in the image data. ((5))

[0167] According to the information processing system described in (4), wherein,

[0168] The evaluation value of the color correction setting is set by the orderer of the printed matter. (6))

[0170] An information processing system, wherein,

[0171] Equipped with a processor

[0172] The processor

[0173] Using a generative model, color correction settings are generated for the manufacture of the new printed matter based on the indoor temperature and humidity when the printing device is used to manufacture the new printed matter, the printed matter having an image represented by image data, the color correction settings being used to make the colors of the image in the printed matter approximate the colors shown in the image data, the generative model being generated using machine learning with learning data, the color correction settings being generated based on the indoor temperature and humidity in which the printed matter was manufactured, the color correction settings being associated with the temperature and humidity in the learning data. ((7))

[0175] According to the information processing system described in (6), wherein,

[0176] The processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at which the new printed matter is manufactured and the country name in which the new printed matter is distributed. The extended generative model is formed by machine learning using learning data with the country name in which the printed matter is distributed, thereby extending the generative model to generate the color correction settings based on the temperature, humidity, and country name. (8))

[0178] According to the information processing system described in (6), wherein,

[0179] The processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at which the new printed matter is manufactured, and the ethnicity of the people to whom the new printed matter is distributed. The extended generative model is formed by machine learning using learning data that has been supplemented with the ethnicity of the people to whom the printed matter is distributed, thereby extending the generative model to generate the color correction settings based on the temperature, the humidity, and the ethnicity. (9))

[0181] The information processing system according to any one of (6) to (8) is wherein,

[0182] The generative model is generated by machine learning using the learning data, which includes color correction settings whose evaluation values ​​reach or exceed a preset threshold. The evaluation values ​​of the color correction settings are set based on the similarity between the colors of the images in the printed matter and the colors shown in the image data. ((10))

[0184] According to the information processing system described in (9), wherein,

[0185] The evaluation value of the color correction setting is set by the orderer of the printed matter. ((11))

[0187] An information processing program is a program that causes a computer to perform processing, wherein the processing is:

[0188] Using learning data that correlates color correction settings with the temperature and humidity of an indoor environment where printed materials are manufactured using a printing device, a generative model is generated via machine learning. The printed materials contain images represented by image data, and the color correction settings are used to make the colors of the images in the printed materials approximate the colors shown by the image data. The generative model generates the color correction settings associated with the input temperature and humidity, given the temperature and humidity included in the learning data. ((12))

[0190] An information processing program is a program that causes a computer to perform processing, wherein the processing is:

[0191] A generative model, generated using machine learning that correlates color correction settings with indoor temperature and humidity during the manufacture of printed matter using a printing device, generates color correction settings for the manufacture of new printed matter based on the temperature and humidity at which the printed matter is manufactured, wherein the printed matter contains an image represented by image data, and the color correction settings are used to make the colors of the image in the printed matter approximate the colors represented by the image data.

[0192] According to the information processing system ((1)), the information processing system ((6)), the information processing program ((11)), and the information processing program ((12)), the following effect is achieved: even experts who are not involved in the generation of color correction settings can generate color correction settings.

[0193] According to the information processing system of ((2)) and ((3)), the generative model can generate the following color correction setting value: the printed color is more in line with the orderer's preference than the generative model that uses temperature and humidity as input data for machine learning.

[0194] According to the information processing system ((4)), it has the effect of generating a generative model that generates the following color correction setting value: the evaluation value is higher than that of the case where the evaluation value of the machine learning data used for the generative model is selected without considering the evaluation value of the color correction setting value.

[0195] According to the information processing system ((5)), it has the following effect: it can generate a color correction setting value that reflects the orderer's evaluation of the color of the printed matter.

[0196] According to the information processing system of ((7)) and ((8)), it has the effect of generating the following color correction setting value: achieving a printing color that is more in line with the orderer's preference than generating a color correction setting value based on temperature and humidity.

[0197] According to the information processing system ((9)), it has the effect of obtaining the following color correction setting value: the evaluation value is higher than that of the generative model that uses machine learning to generate the color correction setting value without considering the evaluation value of the color correction setting value.

[0198] According to the information processing system ((10)), it has the following effect: it can generate color correction settings that reflect the orderer's evaluation of the color of the printed matter.

Claims

1. An information processing system, characterized in that, Equipped with a processor The processor Using learning data that correlates color correction settings with the temperature and humidity of an indoor environment where printed materials are manufactured using a printing device, a generative model is generated via machine learning. The printed materials contain images represented by image data, and the color correction settings are used to make the colors of the images in the printed materials approximate the colors shown by the image data. The generative model generates the color correction settings associated with the input temperature and humidity, given the temperature and humidity included in the learning data.

2. The information processing system according to claim 1, wherein, The names of the countries from which the printed materials were distributed are appended to the learning data. The processor generates a generative model through machine learning. When the generative model is input with the temperature, humidity, and country name included in the learning data, it generates a color correction setting value associated with the input temperature, humidity, and country name.

3. The information processing system according to claim 1, wherein, The ethnicity of the people who will be the recipients of the printed materials is added to the learning data. The processor generates a generative model through machine learning. When the generative model is input with the temperature, humidity, and ethnicity included in the learning data, it generates color correction settings associated with the input temperature, humidity, and ethnicity.

4. The information processing system according to any one of claims 1 to 3, wherein, The processor uses the learning data, including the color correction setting value whose evaluation value reaches or exceeds a preset threshold, for machine learning of the generative model. The evaluation value of the color correction setting value is set based on the similarity between the color of the image in the printed matter and the color shown in the image data.

5. The information processing system according to claim 4, wherein, The evaluation value of the color correction setting is set by the orderer of the printed matter.

6. An information processing system, characterized in that, Equipped with a processor The processor Using a generative model, color correction settings are generated for the manufacture of the new printed matter based on the indoor temperature and humidity when the printing device is used to manufacture the new printed matter, the printed matter having an image represented by image data, the color correction settings being used to make the colors of the image in the printed matter approximate the colors shown in the image data, the generative model being generated using machine learning with learning data, the color correction settings being generated based on the indoor temperature and humidity in which the printed matter was manufactured, the color correction settings being associated with the temperature and humidity in the learning data.

7. The information processing system according to claim 6, wherein, The processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at which the new printed matter is manufactured and the country name in which the new printed matter is distributed. The extended generative model is formed by machine learning using learning data with the country name in which the printed matter is distributed, thereby extending the generative model to generate the color correction settings based on the temperature, humidity, and country name.

8. The information processing system according to claim 6, wherein, The processor uses an extended generative model to generate color correction settings for the manufacture of the new printed matter based on the temperature and humidity at which the new printed matter is manufactured, and the ethnicity of the people to whom the new printed matter is distributed. The extended generative model is formed by machine learning using learning data that has been supplemented with the ethnicity of the people to whom the printed matter is distributed, thereby extending the generative model to generate the color correction settings based on the temperature, the humidity, and the ethnicity.

9. The information processing system according to any one of claims 6 to 8, wherein, The generative model is generated by machine learning using the learning data, which includes color correction settings whose evaluation values ​​reach or exceed a preset threshold. The evaluation values ​​of the color correction settings are set based on the similarity between the colors of the images in the printed matter and the colors shown in the image data.

10. The information processing system according to claim 9, wherein, The evaluation value of the color correction setting is set by the orderer of the printed matter.

11. A program product comprising a program that causes a computer to perform a process, characterized in that the process is: Using learning data that correlates color correction settings with the temperature and humidity of an indoor environment where printed materials are manufactured using a printing device, a generative model is generated via machine learning. The printed materials contain images represented by image data, and the color correction settings are used to make the colors of the images in the printed materials approximate the colors shown by the image data. The generative model generates the color correction settings associated with the input temperature and humidity, given the temperature and humidity included in the learning data.

12. A program product comprising a program that causes a computer to perform a process, characterized in that the process is: A generative model, generated using machine learning that correlates color correction settings with indoor temperature and humidity during the manufacture of printed matter using a printing device, generates color correction settings for the manufacture of new printed matter based on the temperature and humidity at which the printed matter is manufactured, wherein the printed matter contains an image represented by image data, and the color correction settings are used to make the colors of the image in the printed matter approximate the colors represented by the image data.

13. An information processing method, characterized in that, Using learning data that correlates color correction settings with the temperature and humidity of an indoor environment where printed materials are manufactured using a printing device, a generative model is generated via machine learning. The printed materials contain images represented by image data, and the color correction settings are used to make the colors of the images in the printed materials approximate the colors shown by the image data. The generative model generates the color correction settings associated with the input temperature and humidity, given the temperature and humidity included in the learning data.

Images