JP2023106065A - Machine learning device, data processing system, printing system, machine learning method, and data processing method - Google Patents

Abstract

To improve the quality of printing.SOLUTION: A machine learning device 100 according to the embodiment performs machine learning processing with input of feature quantity information, medium information, and first control information, and besides, medium color information as state variables, and creates a learned model. An image forming apparatus 1 according to the embodiment stores the learned model, and when a printing failure occurs, uses the learned model to generate third control information and performs re-printing using the information. The image forming apparatus 1 can thus automatically adjust control information suitable for the medium color of a print medium PM.SELECTED DRAWING: Figure 2

Description

The present invention relates to a machine learning device, a data processing system, a printing system, a machine learning method, and a data processing method, and is preferably applied to, for example, an image forming device.

2. Description of the Related Art Conventionally, electrophotographic image forming apparatuses typified by copiers, printers, and multifunction peripherals have been widely used, especially in office environments. In recent years, there has also been an increase in demand for printing specialized for specific tasks in specific industries such as the medical field, manufacturing industry, and distribution industry (hereafter referred to as "industry printing"). An image forming device is used.

Industry printing includes, for example, product package printing, label printing for wine bottles and the like, and wedding invitation printing. As can be seen from the examples, the quality (image quality) of printed matter in industrial printing often greatly affects the value of products and services. Therefore, print quality is of particular importance for industrial printing.

On the other hand, industry printing is different from printing using general paper (A4 plain paper, B4 plain paper, etc.) that is usually performed in offices, etc., and special media such as thick paper and long paper are used depending on the purpose and application. Shakushi, Japanese paper, cardboard, Western paper, films, labels, envelopes, etc. are used. When printing on such a wide variety of media (hereinafter collectively referred to as “printing media”) using a common image forming apparatus, the image forming apparatus requires various settings such as transfer voltage. The value (hereinafter also referred to as control information) needs to be adjusted for the print medium.

An example of an image forming apparatus that performs printing on a special print medium such as that described above is disclosed in Japanese Patent Application Laid-Open No. 2002-200316. In Patent Document 1, machine learning is performed based on paper characteristics (material, thickness, weight, etc.) and control information at the time of printing to generate a learned model, and appropriate control is performed using the learned model. An electrophotographic image forming apparatus is disclosed for setting information.

Japanese Patent Application Laid-Open No. 2021-033236 (Fig. 2, etc.)

However, in the image forming apparatus having such a configuration, even if the above-described learned model is used, there is a problem that appropriate control information cannot necessarily be set when the color of the printing medium itself differs.

The present invention has been made in consideration of the above points, and intends to propose a machine learning device, a data processing system, a printing system, a machine learning method, and a data processing method that can improve printing quality.

In order to solve this problem, in the machine learning apparatus of the present invention, feature amount information in the actual printed material printed by the image forming apparatus, medium information of the printing medium used in the actual printed material, and when the actual printed material is output and the medium color information representing the color of the printing medium as state variables, and the second control information that sets the feature amount information within a predetermined threshold as teacher data. and a trained model generation unit that generates a trained model by performing machine learning based on the information and the training data obtained by the state variable acquisition unit.

Further, in the data processing system of the present invention, the feature amount information of the actual printed material printed by the image forming apparatus, the medium information of the printing medium used for the actual printed material, and the image forming apparatus when the actual printed material was output are provided. an actual printed matter information acquisition unit that acquires first control information, which is control information, and medium color information that represents the color of a print medium; A data processing unit for inputting to the learned model and outputting the third control information, and a control information storage unit for storing the third control information output from the data processing unit are provided.

Further, in the printing system of the present invention, the feature amount information of the actual printed material printed by the image forming apparatus, the medium information of the printing medium used for the actual printed material, and the control of the image forming apparatus when the actual printed material is output are provided. It is generated by performing machine learning based on the first control information that is information, medium color information that represents the color of the print medium, and teacher data that is second control information that makes the feature amount information within a predetermined threshold value. a storage unit for storing the learned model; an actual printed matter information acquisition unit for acquiring actual printed matter information including medium information of a printing medium to be used; inputting the actual printed matter information to the learned model; A data processing unit for output and a print instruction unit for instructing printing on a print medium to be used using the third control information output from the data processing unit are provided.

Furthermore, in the machine learning method of the present invention, a computer is used to perform feature amount information in an actual printed material printed by an image forming apparatus, medium information on a printing medium used in the actual printed material, and when the actual printed material is output. and medium color information representing the color of the print medium as state variables; A third process of generating a learned model by performing machine learning based on the second process obtained as and the processing of

Further, in the data processing method of the present invention, a computer is used to perform the feature amount information of the actual printed material printed by the image forming apparatus, the medium information of the printing medium used for the actual printed material, and the output time of the actual printed material. first data processing for acquiring first control information that is control information of the image forming apparatus and medium color information representing the color of the print medium; acquired feature amount information; medium information; and first control information and medium color information to a learned model generated by the machine learning method according to claim 9, a second data process for outputting third control information, and a second data process for outputting the third control information as and a third data processing of storing.

Since the present invention performs supervised machine learning using medium color information as state variables in addition to feature amount information, medium information, and first control information, it is possible to generate an appropriate secondary transfer voltage according to the medium color. A trained model can be constructed, and by using the trained model, an appropriate secondary transfer voltage can be generated according to the medium color. As a result, the present invention can execute appropriate print processing using control information corresponding to the medium color in the image forming apparatus without requiring adjustment of set values by a professional engineer or the like. .

According to the present invention, it is possible to realize a machine learning device, a data processing system, a printing system, a machine learning method, and a data processing method that can improve printing quality.

1 is a schematic diagram showing the configuration of an image forming apparatus; FIG. 4 is a schematic diagram showing the configuration of a control unit; FIG. 1 is a schematic diagram showing the configuration of a machine learning device; FIG. FIG. 4 is a schematic diagram showing the configuration of a neural network model for supervised learning; 4 is a schematic flow chart showing a machine learning processing procedure; FIG. 5 is a schematic diagram showing the distribution of volume resistance values for each medium color; 4 is a schematic diagram showing the relationship between the secondary transfer voltage and the volume resistance value of the medium; FIG. 4 is a schematic diagram showing the relationship between medium thickness and secondary transfer voltage; FIG. 5 is a schematic diagram showing the relationship between medium thickness and secondary transfer voltage, and a linear regression line for each medium color; FIG. 4 is a schematic flow chart showing a re-learning processing procedure;

EMBODIMENT OF THE INVENTION Hereinafter, the form (it is mentioned as embodiment below) for implementing invention is demonstrated using drawing.

[1. Configuration of Image Forming Apparatus]
First, an electrophotographic image forming apparatus as a learning target of the machine learning device and the machine learning method according to one embodiment of the present invention, and as a processing target of the data processing system, the printing system, and the data processing method. The basic configuration of is briefly described below.

FIG. 1 is a schematic structural diagram of an image forming apparatus 1 used in one embodiment of the present invention. As shown in FIG. 1, the image forming apparatus 1 exemplified here includes a display operation section 10, a printing medium supply section 20, an image forming section 30, a transfer section 40, and a fixing section 20 with respect to the apparatus housing 2. A unit 50, an output unit 60, and a control unit 70 are provided. The image forming apparatus 1 is a so-called intermediate transfer type full-color LED (Light Emitting Diode) printer.

The display/operation unit 10 includes display means such as a liquid crystal panel provided near the front of the upper surface of the device housing 2 and operation means such as operation buttons or a touch panel. Under the control of the control unit 70, which will be described later, the display operation unit 10 issues a predetermined notification to the user who uses the image forming apparatus 1, or accepts an input operation from the user and notifies the control unit 70 of the input operation. do.

The print medium supply unit 20 is a part that supplies the print medium PM into the image forming apparatus 1 , and has a paper tray 21 , a manual feed tray 22 and a plurality of paper feed rollers 23 . The paper tray 21 accommodates a plurality of stacked print media PM. The print media PM accommodated in the paper tray 21 are generally general paper (A4 plain paper, B4 plain paper, etc.).

The manual feed tray 22 is storably provided on the side surface of the main body of the image forming apparatus 1, and mainly feeds the special print medium PM when printing on the special print medium PM different from general paper. Therefore, in the printing medium supply unit 20, the manual feed tray 22 is mainly used when performing industry printing. The plurality of paper feed rollers 23 are rollers arranged at appropriate positions to transport the print medium PM placed in the paper tray 21 or on the manual feed tray 22 to the transport path PL.

The image forming section 30 is for forming a toner image, and has a plurality of (five in this embodiment) image forming units 31 (31C, 31M, 31Y, 31K, and 31S) arranged in parallel. . The plurality of image forming units 31C, 31M, 31Y, 31K, and 31S have basically the same configuration, except that the toner colors are different. It has a head and a toner tank. As for the image forming process by each of these configurations, a process similar to a well-known process can be adopted, and detailed description thereof will be omitted here.

A plurality of image forming units 31C, 31M, 31Y, 31K, and 31S are portions that form toner images of cyan, magenta, yellow, black, and special colors, respectively. The image forming section 30 employs the image forming units 31 for a plurality of colors to enable full-color printing. A toner image formed by each image forming unit 31 is transferred to an intermediate transfer belt 41 of a transfer section 40, which will be described later. As the special color toner, for example, white toner, clear toner, silver toner, or fluorescent color (neon yellow, etc.) toner can be used.

The transfer unit 40 is a portion that transfers the toner image formed by the image forming unit 30 onto the print medium PM, and includes an intermediate transfer belt 41 and a plurality of primary transfer rollers 42 (42C, 42M, 42Y, 42K, 42S). , a backup roller 43 and a secondary transfer roller 44 . The intermediate transfer belt 41 is an endless elastic belt supported by a plurality of rollers including drive rollers and made mainly of a resin material such as rubber. A toner image of each color formed by the image forming unit 31 is transferred (primary transfer) onto the surface of the intermediate transfer belt 41, and the formed toner image is later transferred (secondary transfer) onto the print medium PM. ) is done.

The plurality of primary transfer rollers 42 are portions for transferring the toner images of respective colors formed by the image forming units 31 onto the intermediate transfer belt 41 , and are arranged to face the respective photosensitive drums of the plurality of image forming units 31 . The intermediate transfer belt 41 is sandwiched between the photoreceptor drums. A predetermined primary transfer voltage is applied to the primary transfer roller 42 . The primary transfer voltage is controlled by a control section 70, which will be described later.

The backup roller 43 is one of a plurality of rollers that support the intermediate transfer belt 41, and is arranged at a position facing a secondary transfer roller 44, which will be described later, with the intermediate transfer belt 41 interposed therebetween. The secondary transfer roller 44 is arranged in the middle of the conveying path PL at a position facing the backup roller 43 with the intermediate transfer belt 41 interposed therebetween. functions to transfer a pre-formed toner image to the intermediate transfer belt 41 when the print medium PM is passed between them. A predetermined secondary transfer voltage is applied to the secondary transfer roller 44 . The secondary transfer voltage is controlled by the controller 70, which will be described later.

The fixing section 50 applies heat and pressure to the printing medium PM to which the toner image has been transferred by the transfer section 40 to fix the toner image onto the printing medium PM. and The fixing roller 51 incorporates a heater (not shown) therein, and the toner fixing temperature is controlled by the current value of the current supplied to this heater. The current value of the current supplied to this heater is controlled by a control section 70, which will be described later. The pressure roller 52 is biased against the fixing roller 51 . As a result, a predetermined fixing pressure is applied to the print medium PM passing between the fixing roller 51 and the pressure roller 52 . In this embodiment, the pressure roller 52 is urged against the fixing roller 51 . may be urged against the backup roller.

The output unit 60 outputs the print medium PM on which the toner image is fixed by the fixing unit 50 to the outside of the image forming apparatus 1 as the actual printed matter AP. The discharge tray 61 is formed in the upper part of the image forming apparatus 1, and the actual printed material AP output via the transport path PL is placed thereon. A plurality of transport rollers 62 are provided at various locations along the transport path PL to transport the print medium PM to the discharge tray 61 . Further, cooling means for removing heat generated when the toner image is fixed may optionally be provided at any position of the output section 60 . As the cooling means, for example, a roller having a heat dissipation function is adopted as at least a part of the conveying roller 62, or a well-known heat pipe, heat sink, fan, or the like as the cooling means is arranged at a predetermined position of the output section 60. can be adopted.

The controller 70 is a part that controls each part of the image forming apparatus 1, and includes a well-known CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), and the like. 2, the control section 70 includes an output control section 71, an actual print information acquisition section 72, a medium color information acquisition section 73, a storage section 74, and a data processing section 75. FIG.

The output control section 71 is a section that transmits various control signals to each section of the image forming apparatus 1 and receives various detection values. The output control unit 71 controls, for example, the secondary transfer voltage applied to the secondary transfer roller 44, the toner temperature of the fixing roller 51 (specifically, the current value of the current supplied to the heater inside the fixing roller 51), and the like. It has the function of controlling

The actual printed material information acquisition unit 72 is a part that acquires various information related to the actual printed material AP including printing defects, and includes a feature amount information acquisition unit 81 (details will be described later), a medium information acquisition unit 82, and a 1 control information acquisition unit 83 . Further, the actual printed material information acquiring section 72 associates the acquired various pieces of information with each other as one data set, and causes the storage section 74 to store the data set.

The medium information acquisition unit 82 acquires medium information, which is information about the print medium PM. This medium information is various information about the print medium PM, and includes information such as presence/absence of coating on the print medium PM, material, thickness, weight, density, and the like. Such medium information can be acquired by, for example, having the user input a predetermined product code assigned in advance to the print medium PM via the display operation unit 10 (FIG. 1).

As for the material among the information included in the medium information, it is sufficient to specify the main material used in the print medium PM, and it is possible to omit the additionally included material. Regarding the weight, various weights can be used as long as they exhibit properties related to the weight of the print medium. You can use what is used. Further, the density can be specified by calculation if the values of thickness and weight (basis weight) can be specified.

The first control information acquisition unit 83 acquires control information set when the actual printed material AP is output (hereinafter referred to as first control information). This first control information includes, for example, the secondary transfer voltage applied to the secondary transfer roller 44 and the toner fixing temperature of the fixing roller 51 . This first control information is information stored in a predetermined memory provided in the output control section 71 or in the storage section 74 in which the control information of the output control section 71 is stored. Therefore, the first control information acquisition section 83 can acquire the first control information from the output control section 71 or the storage section 74 via the internal bus.

Since the image forming apparatus 1 employs an intermediate transfer method in which a toner image is transferred from the intermediate transfer belt 41 to the print medium PM, the secondary transfer voltage is used as the first control information. However, in a tandem-type image forming apparatus that does not have an intermediate transfer belt and directly transfers from a photoreceptor drum to a print medium, the toner from the photoreceptor drums provided in each of a plurality of image forming units with different toner colors onto the print medium PM. A transfer voltage for transferring the toner image can be included in the first control information.

The medium color information acquisition unit 73 is a part that acquires information (hereinafter referred to as medium color information) for identifying the medium color (also called paper color) of the print medium PM. This medium color information can be obtained by, for example, inputting a predetermined product code pre-assigned to the print medium PM from the user via the display operation unit 10 (FIG. 1), similarly to the medium information described above. In addition, the names of representative colors roughly classified, such as "red", "blue", and "green", may be displayed on the display operation unit 10, and the user may select one of them. The medium color information may be obtained by

Alternatively, for example, image data of the print medium PM may be obtained using an external image scanner SC, and medium color information may be obtained from the image data. In this case, for example, the average pixel value within a predetermined range in the image data of the print medium PM can be used as the medium color information. As the medium color information, the average pixel value of a predetermined area outside the print area in the image data of the actual printed material AP read by the external image scanner SC may be used.

The medium color information acquisition unit 73 is a part that acquires information (hereinafter referred to as medium color information) regarding the color of the print medium PM (hereinafter referred to as medium color). The medium color information acquisition unit 73 associates the acquired medium color information with various information acquired by the actual printed material information acquisition unit 72 and stores the information in the storage unit 74 .

The storage unit 74 is a storage area for storing various information in the image forming apparatus 1, and is configured by, for example, an SSD (Solid State Drive) or HDD (Hard Disk Drive). The storage unit 74 includes a learned model storage unit 91, a data set storage unit 92, and a third control information storage unit 93 (details will be described later).

The data processing section 75 is a section that calculates the secondary transfer voltage applied to the secondary transfer roller 44 and the toner fixing temperature of the fixing roller 51 by performing predetermined arithmetic processing based on various information. described later).

[2. Adjustment of control information by machine learning]
By the way, in the conventional image forming apparatus, when performing industry printing, it is difficult to sufficiently adjust the control parameters as described above, and as a result, the rate of low-quality printing (printing defects) is high. rice field. When such a printing defect occurs, in the conventional image forming apparatus, a professional engineer comprehensively considers the state of the actual printed matter actually printed, the control information at the time of printing, etc. Furthermore, based on past experience and accumulated know-how, it is conceivable to perform work such as deriving optimum control parameters to obtain desired printing results. However, this series of operations takes a long time and occupies a technician for each printing failure that occurs, resulting in a very high cost.

Therefore, in the present embodiment, adjustment of control information (control parameters) of the image forming apparatus 1 is automated by using a machine learning apparatus 100 and a learned model generated by a machine learning method described below. .

In the image forming apparatus 1 described above, when printing is executed, the user adjusts various control parameters in order to obtain an optimum print result. As a result of verification of these various control parameters by the present inventors, among the various control parameters, the secondary transfer voltage applied to the secondary transfer roller 44 and the toner fixing temperature of the fixing roller 51 (that is, the fixing roller 51 It has been found that the current value of the current supplied to the internal heater is the control information that directly affects the printing quality. In other words, it was found that these secondary transfer voltage and toner fixing temperature are control information that has a high correlation with printing defects that occur on the printing surface.

When the relationship between these two pieces of control information and printing failure was examined in more detail, the following relationship was found. In other words, a high secondary transfer voltage mainly causes dust (printing defects that cause white spots), and a low secondary transfer voltage causes blurring (printing defects that cause colors to fade). Become. Also, if the toner fixing temperature is high, it mainly causes spots (printing defects in which a speckled pattern occurs). This can cause misalignment (printing defects that result in areas of low density).

Therefore, in the machine learning device 100 described below, the control information to be adjusted is the secondary transfer voltage and the toner fixing temperature, and a learned model for adjusting these two pieces of control information is created.

[2-1. Configuration of machine learning device]
FIG. 3 is a schematic block diagram of machine learning device 100 according to one embodiment of the present invention. The machine learning device 100 includes a state variable acquisition unit 110 , a teacher data acquisition unit 120 , a trained model generation unit 130 and a storage unit 140 . This machine learning device 100 is a device that generates a trained model by so-called supervised learning. Note that FIG. 3 illustrates an example in which the machine learning device 100 is built in a computer (a server device, a PC (Personal Computer), etc.) separate from the image forming device 1 for the purpose of facilitating understanding. However, the machine learning device 100 may be built in the image forming apparatus 1 .

The state variable acquisition unit 110 acquires, as state variables, parameter information necessary for generating a trained model capable of adjusting control information of the image forming apparatus 1 . In machine learning, state variables are the most important factor that determines the accuracy of the generated trained model. If the combination of information acquired as state variables is different, naturally the generated trained model is also different. Therefore, it should be especially noted in the present invention that the combination is also an extremely important factor.

In the present embodiment, the state variable acquisition unit 110 acquires information such as feature amount information in the actual printed material AP, medium information of the printing medium PM used in the actual printed material AP, first control information used during printing, and medium color information. Get four pieces of information as state variables. The method of acquiring the state variables can be arbitrarily set according to the connection state between the machine learning device 100 and the image forming device 1. Alternatively, it is acquired via any storage medium. Then, these four pieces of acquired information are stored in the storage unit 140 as one data set.

The feature amount information in the actual printed material AP includes information about printing defects occurring in the actual printed material AP. The information about printing failure is information about the degree of printing failure. Here, the feature amount information actually input to the machine learning apparatus 100 may be image data information obtained by reading the actual printed matter AP using a known scanner. There is no need to specify the type or degree in advance.

This is because the learning method of the machine learning device learns the control information as a suitable output result for the state variables regardless of the type and degree of the printing defect, and it is necessary to determine the type and degree of the printing defect. This is because there is no Nonetheless, for example, a preprocessing process is adopted to adjust the information of the image data in advance into information suitable for inputting to the input layer of the machine learning device (for example, information such as the type and degree of printing defects). It is of course possible. For a specific preprocessing process, a person skilled in the art can easily understand that a method commonly used in the technical field of image recognition can be adopted, so the explanation thereof is omitted here.

The teacher data acquisition unit 120 (FIG. 2) obtains control information (hereinafter referred to as second control information) in which the feature amount information, which is the information about the printing failure, is within a predetermined threshold in the actual printed matter AP including the printing failure. , are obtained as training data. This second control information is, simply put, improved control information that provides print results free of print defects.

This second control information is, for example, control information derived by the engineer EN based on the output result of the actual printed matter AP and the control information at that time. Therefore, the above-mentioned "predetermined threshold value" does not necessarily indicate a specific value. It can be said that "feature amount information is within a predetermined threshold". Also, the second control information is acquired by the engineer EN by directly inputting it to the teacher data acquiring unit 120 or by transmitting it via various communication means or an arbitrary storage medium. Note that this second control information is stored in the storage unit 140 in association with the corresponding data set acquired by the state variable acquisition unit 110 and stored in the storage unit 140 .

Based on the data set consisting of four pieces of information as state variables acquired by the state variable acquisition unit 110 and the teacher data acquired by the teacher data acquisition unit 120 and associated with the corresponding data set, the trained model generation unit 130 machine learning to generate a trained model. A specific method of machine learning performed here will be described in detail below.

The machine learning device 100 according to the present invention employs supervised learning using a neural network model as its learning method.

FIG. 4 is a diagram showing an example of a neural network model for supervised learning implemented in machine learning device 100. As shown in FIG. The neural network in the neural network model shown in FIG. 4 has k neurons (x1 to xk) in the input layer, m neurons (y11 to y1m) in the first hidden layer, and n neurons in the second hidden layer. , and two neurons (z1, z2) in the output layer.

The first intermediate layer and the second intermediate layer are also called hidden layers, and the neural network may have a plurality of hidden layers in addition to the first intermediate layer and the second intermediate layer. Alternatively, only the first intermediate layer may be a hidden layer. Between the input layer and the first intermediate layer, between the first intermediate layer and the second intermediate layer, and between the second intermediate layer and the output layer, nodes connecting neurons between the layers are provided. , and each node is associated with a weight wi (i is a natural number).

In the neural network in the neural network model according to the present embodiment, learned model generation unit 130 uses a learning data set to determine the correlation between the control information of the image forming apparatus and the printed matter output by the image forming apparatus. learn. That is, the trained model generation unit 130 associates each state variable with a neuron in the input layer, and calculates the value of each neuron in the output layer by using a general neural network output value calculation method.

Specifically, for example, for the input layer and the first intermediate layer, the trained model generation unit 130 converts the value of each neuron on the output side (first intermediate layer) to the input side (input layer) connected to the neuron. and multiple weights wi associated with nodes connecting the neurons on the output side (first intermediate layer) and the neurons on the input side (input layer). Calculated as the sum of The same is true for the first intermediate layer and the second intermediate layer, and the same is true for the second intermediate layer and the output layer.

In this way, the trained model generation unit 130 can calculate the value of each neuron in the output layer by performing such operations on all neurons other than the neurons in the input layer. When the state variables are associated with the neurons of the input layer, the form in which the information obtained as the state variables is associated can be appropriately set in consideration of the accuracy of the generated trained model. For example, when associating the image data of the actual printed material AP as the feature amount information in the actual printed material AP with the input layer, the image data is divided into predetermined regions, and then the color values (for example, RGB values) of each divided region are calculated. Information can be associated with each input layer.

Then, the trained model generation unit 130 compares the calculated values of the two neurons z1 and z2 in the output layer with the values of the two teacher data t1 and t2 to obtain an error. Here, the values of the neurons z1 and z2 are the toner fixing temperature of the fixing roller 51 and the secondary transfer voltage applied to the secondary transfer roller 44 in this embodiment. The values of the teacher data t1 and t2 are the toner fixing temperature of the fixing roller 51 and the secondary transfer voltage applied to the secondary transfer roller 44, which are associated with the data sets. Then, the trained model generation unit 130 repeats adjusting the weight wi associated with each node (back promotion) so that the obtained error becomes smaller.

Then, when a predetermined condition is satisfied, such as repeating the above-described series of steps a predetermined number of times or the above-described error is smaller than the allowable value, the trained model generation unit 130 terminates learning, The neural network model is stored in the storage unit 140 as a learned model. In this way, the trained model generating unit 130 generates a trained model including information about all weights wi associated with each node of the neural network model.

The learned model stored in the storage unit 140 is applied to a real system upon request via communication means such as the Internet or a storage medium. A specific application of the trained model to the actual system (data processing system) will be detailed later.

[2-2. machine learning processing]
Next, machine learning processing performed by the machine learning device 100 will be described with reference to the flowchart of FIG. This machine learning processing is realized by using a computer (also called an information processing device or an arithmetic processing device), and various computers can be applied. Specifically, for example, a component that constitutes the control unit 70 in the image forming apparatus 1, a computer device that is locally connected to the image forming apparatus 1, or a server device that is arranged on a network can be used.

When executing supervised learning as the machine learning method according to the present invention, the machine learning device 100 starts the machine learning processing procedure RT1, moves to the first step SP1, and prepares a pre-learning model with initial weights. After preparation, move to the next step SP2. In step SP2, the machine learning device 100 uses the state variable acquisition unit 110 (FIG. 3) to obtain the feature amount information of the actual printed material AP actually printed by the image forming apparatus 1, the medium information of the printing medium PM, and the first The control information and the medium color information of the print medium PM are acquired as state variables, and the process proceeds to the next step SP3.

In step SP3, the machine learning device 100 acquires teacher data corresponding to the state variables acquired in step S12 by the teacher data acquiring unit 120 (FIG. 3), and proceeds to the next step SP4. In step SP4, the machine learning device 100 causes the learned model generation unit 130 (FIG. 3) to associate the state variables acquired in step SP2 with the input layer (FIG. 4) of the pre-learning model, so that the output layer (FIG. 4) and go to the next step SP5.

Here, since the control information forming the output layer generated in step SP4 is generated by the pre-learning model, it is usually not the control information that produces the print result that satisfies the user's request. Therefore, the machine learning device 100 performs machine learning using the control information forming the teacher data acquired in step SP3 and the control information forming the output layer generated in step SP4 by the trained model generation unit 130. and move to the next step SP6. The machine learning performed here compares the control information that constitutes the teacher data and the control information that constitutes the output layer, detects the error between the two, and performs learning so as to obtain an output layer that minimizes this error. Adjusting each weight wi (Fig. 4) in the previous model.

In step SP6, the machine learning device 100 determines whether or not it is necessary to perform further machine learning. If a negative result is obtained here, the machine learning device 100 continues machine learning by returning to step SP2 and repeating the series of processes. If the machine learning is to be continued here, the machine learning device 100 will perform steps SP2 to SP5 a plurality of times. In general, the machine learning device 100 can increase the accuracy of the finally generated learned model in proportion to the number of times of machine learning. On the other hand, when a positive result is obtained in step SP6, the machine learning device 100 finishes machine learning and proceeds to the next step SP7.

In step SP7, the machine learning device 100 stores the neural network generated by adjusting each weight wi (FIG. 4) through a series of steps in the storage unit 140 as a trained model. After that, the machine learning device 100 proceeds to the next step SP8 and ends the machine learning processing procedure RT1. The learned models stored here are applied and used in various data processing systems, the details of which will be described later.

[2-3. Relationship Between Medium Color and Secondary Transfer Voltage]
Next, attention was paid to the relationship between the print medium PM and the medium color information, and various investigations and experiments were conducted. First, the distribution of the volume resistance value [Ω], which is the electrical resistance value per volume, was examined for each medium color of the print medium PM, specifically for each of black, blue, and red. The results shown were obtained. From FIG. 6, regarding the print medium PM, the volume resistance value in black is often lower than the volume resistance value in other colors, and the volume resistance value in blue is often higher than the volume resistance value in other colors. It can be seen that the distribution is as follows. In other words, it has been clarified that the volume resistivity distribution of the print medium PM differs depending on the medium color.

On the other hand, for the print medium PM, it is known that there is a relatively strong correlation between the volume resistance value and the optimum secondary transfer voltage for printing. Therefore, by using the medium color information as part of the state variables in the above-described machine learning, it becomes possible to learn medium colors with different distributions of volume resistance values, thereby improving the prediction accuracy of the secondary transfer voltage. can be expected.

Here, reference information will be shown for understanding the correlation between the volume resistance value of the print medium PM and the optimum secondary transfer voltage for printing. FIG. 7 shows a scatter diagram of the volume resistivity of the print medium PM and the optimum secondary transfer voltage for printing. In FIG. 7, the correlation coefficient between the volume resistance value and the secondary transfer voltage is approximately 0.91, indicating a relatively strong correlation.

In general, obtaining the volume resistance value of the print medium PM requires specialized knowledge and an appropriate measuring instrument. Therefore, if a volume resistance value is used as a state variable for machine learning, even if a service using the result of the machine learning is provided, the users who can use the service are limited.

On the other hand, the medium color information of the print medium PM can be obtained relatively easily by a user's selection operation or using an image scanner. Therefore, if medium color information is added to the state variables of machine learning, even if a service using the result of the machine learning is provided, it can be used widely without limiting the users who can use it. be. In this way, the use of medium color information in the above-described machine learning can be expected to improve the prediction accuracy of the secondary transfer voltage without limiting opportunities for using learning results.

Next, when examining the correlation between the optimum secondary transfer voltage and the thickness of the printing medium PM, which is one of the medium information, the results are shown in FIGS. A scatter plot as shown was obtained. FIG. 8A is a scatter diagram in which black, blue and red data are mixed. 8(B), (C) and (D) are scatter diagrams using data of only black, only blue, or only red, respectively.

In FIG. 8A, when three colors are mixed, the distribution of each data is relatively uneven, and the correlation coefficient is about 0.43. In the case of black in FIG. 8B, the distribution was such that the optimum secondary transfer voltage for printing was concentrated at about 1.6 [kV] regardless of the thickness of the medium. In blue in FIG. 8(C) and red in FIG. 8(D), the distribution of data is relatively consolidated compared to FIG. 8(A), and the respective correlation coefficients are about 0.48 and about 0. 0.61, which is higher than the correlation coefficient of about 0.43 in FIG.

FIG. 9 shown below shows linear regression straight lines LE (blue) and straight lines LR (red), respectively, based on the distribution of blue (FIG. 8 (C)) and red (FIG. 8 (D)) data. It is what I did. FIG. 9 shows that the slopes of the linear regression lines LE and LR are different, that is, the linear regression lines are different for each medium color. For this reason, in the machine learning described above, it is better to perform learning processing separately for each medium color than to perform learning processing in a state in which a plurality of medium colors are mixed. It is expected that the prediction accuracy of the next transfer voltage is likely to be improved.

Based on the above relationship, in machine learning, it is possible to learn the characteristics of each medium color by using medium color information as a state variable. That is, the medium color information in machine learning can be said to be a useful state variable for improving the prediction accuracy of the secondary transfer voltage.

[3. Use of trained model]
Next, application examples of the trained model generated by the machine learning device and the machine learning method described above will be described. Here, it is assumed that the learned model generated through the above-described machine learning process is applied to the image forming apparatus 1 (FIGS. 1 and 2).

[3-1. Detailed configuration of control unit]
First, the detailed configuration of the feature amount information acquisition unit 81, the storage unit 74, and the data processing unit 75 of the actual printed material information acquisition unit 72 provided in the control unit 70 (FIG. 2) of the image forming apparatus 1 will be described. Incidentally, the control section 70 is connected to an external image scanner SC via an input/output interface (not shown) so that image data of the actual printed material AP can be obtained.

The feature amount information acquisition unit 81 is a part that acquires image data of the actual printed matter AP as feature amount information. Like the machine learning device 100 described above, the feature amount information acquisition unit 81 acquires image data of the actual printed material AP read by the external image scanner SC as feature amount information.

The storage unit 74 has the learned model storage unit 91, the data set storage unit 92, and the third control information storage unit 93, as described above. The learned model storage unit 91 performs machine learning using the learned model generated by the above-described machine learning method, that is, the four pieces of information of the feature amount information, the medium information, the first control information, and the medium color information as state variables. and the generated trained model is stored.

The data set storage unit 92 stores the feature amount information, medium information, first control information, and medium color information related to the common actual printed material AP acquired by the actual printed material information acquiring unit 72 as one data set. . The third control information storage unit 93 stores the third control information output by the data processing unit 75, which will be described later.

The data processing unit 75 uses the data set related to the specific actual printed material AP stored in the data set storage unit 92 and the learned model stored in the learned model storage unit 91, and uses the input layer of the learned model as Enter various information that constitutes the dataset. Thereby, the data processing section 75 outputs the secondary transfer voltage applied to the secondary transfer roller 44 and the toner fixing temperature of the fixing roller 51 as predetermined control information (third control information) to the output layer.

The secondary transfer voltage and the toner fixing temperature, which constitute the third control information, are control information capable of improving printing defects that occur in the actual printed matter AP. That is, for the third control information, a desired printed matter can be obtained by using the printing medium PM used in this actual printed matter AP and performing print output by the image forming apparatus 1 using this third control information as a control parameter. It is the control information adjusted by the data processing unit 75 so that it can be performed.

[3-2. Reprint process]
Next, reprint processing performed by the image forming apparatus 1 will be described with reference to the flowchart of FIG. This reprinting process is a process of adjusting the control information and performing the printing process again when a printing error occurs in the image forming apparatus 1 after the printing process has been performed on the printing medium PM. In this case, it is assumed that the image forming apparatus 1 has executed the print process on the print medium PM using the first control information.

When performing print processing, the data processing unit 75 (FIG. 2) provided in the control unit 70 of the image forming apparatus 1 starts the reprint processing procedure RT2 (FIG. 10) and proceeds to the first step SP11. In step SP11, the data processing unit 75 determines whether or not a printing defect has occurred. In this case, whether or not a printing defect has occurred is determined based on a report by the user.

Specifically, it is determined by making a request to improve the printing defect via the display operation unit 10 of the image forming apparatus 1, or to a customer service or the like that manages the image forming apparatus 1. . In the present embodiment, various configurations are incorporated so that printing failure can be resolved by the image forming apparatus 1 alone. For this reason, for example, it is preferable in terms of convenience if the image forming apparatus 1 is equipped with a specific operation mode such as a print defect elimination mode. In this case, when the user selects the operation mode via the display/operation unit 10, the following processing for eliminating printing defects is executed.

If a negative result is obtained here, this means that there is no problem with the print result and that there is no need to perform the print processing again. At this time, the data processing unit 75 returns to step SP11 and waits until the next printing failure occurs.

On the other hand, when a positive result is obtained in step SP11, the data processing section 75 proceeds to the next step SP12. At step SP12, the data processing unit 75 requests the user for actual printed material information and medium color information necessary to resolve the printing failure via a predetermined user interface such as the display operation unit 10. FIG.

More specifically, the data processing unit 75, for example, instructs the user to input medium information and medium color information, and reads the actual printed material AP in which the printing defect has occurred with the image scanner SC (FIG. 2). Requests an operation to input feature amount information and medium color information in the AP. The actual print information and the medium color information input by the user based on this request are stored in the data set storage unit 92 as one data set by the actual print information acquisition unit 72 . The data processing unit 75 acquires the actual printed matter information and the medium color information by referring to the data set storage unit 92, and proceeds to the next step SP13.

Further, as a request to the user for the actual printed material information and the medium color information, various information can be sent to the user by displaying a method of obtaining various information on the display/operation unit 10, or by outputting a navigation voice through voice output means (not shown). It is possible to use a method such as prompting the input work of

In this example, when a printing error occurs, the feature amount information and the medium color information of the actual printed matter AP in which the printing error occurred are input. However, the present invention is not limited to this. It is also possible to input medium color information.

In step SP13, the data processing unit 75 refers to the learned model storage unit 91 (FIG. 2), acquires the learned model (FIG. 4), and proceeds to the next step SP14. In step SP14, the data processing unit 75 generates the third control information as an output layer by inputting the actual printed matter information and the medium color information acquired in step SP12 to the input layer of the acquired learned model. to step SP15. In step SP15, the data processing unit 75 temporarily stores the generated third control information in the third control information storage unit 93 (FIG. 4), and proceeds to the next step SP16.

In step SP16, the data processing unit 75 determines whether or not a reprint instruction has been issued by the user's operation on the display operation unit 10. FIG. If a negative result is obtained here, the data processing section 75 waits for a reprint instruction by repeating step SP16. On the other hand, if an affirmative result is obtained in step SP16, the data processing section 75 applies to the secondary transfer roller 44 based on the third control information in the third control information storage section 93 instead of the first control information. After adjusting the secondary transfer voltage and the toner fixing temperature of the fixing roller 51, the printing process is executed again. After that, the data processing unit 75 proceeds to the next step SP18 and ends the reprint processing procedure RT2.

The printed material that is printed again and output through the above steps is output based on the third control information that has been adjusted based on the previous printing result. will generally be eliminated.

As described above, according to the present embodiment, no manpower such as an engineer EN is required from the occurrence of printing defects until the user obtains printed matter in which the printing defects have been eliminated. Therefore, in the data processing system of the present invention, data processing can be realized at reduced cost.

In most cases, printed matter reprinted through the above series of steps can be printed with printing defects resolved, but the possibility of printing defects occurring even after this process is very small. be. In such a case, the steps SP11 to SP17 may be performed again.

[4. effects, etc.]
In the above configuration, the machine learning device 100 according to the present embodiment inputs medium color information as state variables in addition to feature amount information, medium information, and first control information, and outputs secondary transfer voltage and toner fixing temperature. Perform machine learning processing such as to generate a learned model. At this time, in the machine learning device 100, by specifying the state variables input to the input layer as the four pieces of information described above, a highly accurate trained model that can be applied to print output on various special print media can be developed with high accuracy. can be generated efficiently.

Therefore, the machine learning device 100 can generate a highly accurate learned model that can accurately improve printing defects that occur when printing on various printing media, such as industrial printing. . Therefore, by applying this learned model to an actual system, it is possible to improve printing defects with high accuracy and obtain desired printing results.

In particular, industry printing is more likely to use print media PM of various colors compared to general office or home printing. Therefore, the machine learning device 100 can generate a learning model that can adjust the secondary transfer voltage suitable for the color of the print medium PM by including the medium color information in the state variables.

Further, in the machine learning device 100, the feature amount information includes information about printing defects in the actual printed matter AP. As a result, the machine learning apparatus 100 can generate a learned model that can resolve printing defects that occur in the actual printed material AP in machine learning, and can generate a learned model that can perform data processing desired by the user.

Furthermore, in the machine learning device 100, the medium information includes information on whether the print medium PM is coated, material, thickness, weight, and density. As a result, the machine learning device 100 considers not only the thickness of the print medium PM but also various medium information about the print medium PM during machine learning, so that it is possible to generate a learned model that fully considers the medium information. .

In the machine learning device 100, the second control information includes, for example, information about the toner fixing temperature and transfer voltage (secondary transfer voltage in this example) in the electrophotographic image forming apparatus 1 shown in FIG. . The second control information is composed of toner fixing temperature and transfer voltage, which are control parameters commonly used in electrophotographic image forming apparatuses. Therefore, the generated learned model can be applied to various types of electrophotographic image forming apparatuses.

In addition, the image forming apparatus 1 stores in the storage unit 74 a learned model generated using the medium color information in addition to the feature amount information, the medium information, and the first control information. The learned model is used to generate the third control information, which is used for reprinting. As a result, since the image forming apparatus 1 can automatically adjust the control information necessary for printing, it is possible to realize a low-cost data processing system without manpower required every time a printing error occurs. In addition, due to this cost reduction, users can easily try multiple special print media for industrial printing without being conscious of the cost. Therefore, for the user, the degree of freedom in selecting a print medium is improved.

Here, as experimental examples, a trained model in which the state variables include medium color information and a trained model in which the medium color information is not included were prepared, respectively, and the prediction accuracy of each trained model was confirmed. In this experiment, the trained model including the medium color information in the state variables improved the prediction error of the secondary transfer voltage by about 8% compared to the trained model without the medium color information. That is, it was confirmed that the medium color information is important information for predicting the secondary transfer voltage. Since such medium color information is information that can be easily input by the user, the user does not need to have specialized knowledge, and there is no need to incorporate special control into the image forming apparatus. can be used for

According to the above configuration, machine learning apparatus 100 according to the present embodiment inputs medium color information as state variables in addition to feature amount information, medium information, and first control information, and performs machine learning processing. Generate a model. Further, the image forming apparatus 1 according to the present embodiment stores this learned model, and when a printing failure occurs, uses this learned model to generate the third control information, and uses this to reprint. I do. As a result, the image forming apparatus 1 can automatically adjust the control information suitable for the medium color of the print medium PM.

[5. Other embodiments]
In the above-described embodiment, the medium information includes the presence or absence of coating, the material, the thickness, the weight, and the density of the print medium PM. However, the present invention is not limited to this. For example, some of these may be omitted, or various values such as smoothness, volume resistance, and surface resistance may be added.

In the above-described embodiment, the case where the toner fixing temperature of the fixing roller 51 and the secondary transfer voltage applied to the secondary transfer roller 44 are included in the first control information has been described. However, the present invention is not limited to this, and for example, the toner fixing temperature may be omitted from the first control information. Also, various control parameters actually set when the actual printed material AP is output may be added to the first control information. Furthermore, the first control information may include not only the secondary transfer voltage but also the primary transfer voltage applied to the primary transfer roller 42 .
The same applies to the second control information and the third control information.

Furthermore, in the above-described embodiment, a case has been described in which the toner fixing temperature of the fixing roller 51 and the secondary transfer voltage applied to the secondary transfer roller 44 are used as the third control information. However, the present invention is not limited to this, and for example, the toner fixing temperature may be omitted from the third control information. Alternatively, various control parameters that directly affect print quality may be added to the third control information.

Furthermore, in the above-described embodiment, a case has been described in which the secondary transfer voltage of the third control information, which is calculated using the learned model based on the medium color information and the like, is used as it is to perform the reprinting process. However, the present invention is not limited to this, and secondary transfer voltages to be used in reprint processing may be further calculated based on secondary transfer voltages calculated based on a plurality of pieces of medium color information. Specifically, for example, when the medium color is purple and the ratio of the red component and the blue component is known, the secondary transfer voltage calculated based on the red medium color information and the blue medium color information A secondary transfer voltage obtained by calculating a weighted average according to the ratio of the red component and the blue component may be used to perform the reprinting process. This means that the data are located in the area sandwiched between the two straight lines LE and LR in FIG.

Furthermore, in the above-described embodiment, a case has been described in which the secondary transfer voltage of the third control information, which is calculated using the learned model based on the medium color information and the like, is used as it is to perform the reprinting process. However, the present invention is not limited to this, and correction processing may be performed according to the density of the medium color based on the secondary transfer voltage calculated based on the medium color information.

Furthermore, in the above-described embodiment, a case has been described in which machine learning is performed using medium color information separately from medium information and applied to a trained model. However, the present invention is not limited to this. For example, medium color information may be included in part of the medium information, machine learning may be performed, and the machine learning may be applied to the learned model.

Furthermore, in the above-described embodiment, the machine learning device 100 (FIG. 3) repeatedly performs machine learning processing for one neural network (pre-learning model) a plurality of times, thereby improving the accuracy and learning. We described the case of generating a finished model. However, the present invention is not limited to this, and for example, a plurality of trained models that have undergone machine learning a predetermined number of times may be stored in the storage unit 140 as one candidate. In this case, the data set for validity judgment is input to a group of trained models to generate the output layer (the value of the neuron), and the accuracy of the control information specified in the output layer is compared and examined. A single best trained model can be selected to apply to the data processing system. The validity judgment data set may have the same state variables as the learning data set used for learning, but may consist of different data.

Furthermore, in the above-described embodiment, when a printing failure occurs in the image forming apparatus 1, the user is prompted to input medium information and medium color information, or the actual printed matter AP is read by the image scanner SC (FIG. 2). , the case of acquiring feature amount information and medium color information. However, the present invention is not limited to this, and the feature amount information and the medium color information may be acquired using various other methods. Specifically, for example, the feature amount information acquisition unit 81 reads using an imaging means other than a scanner (for example, a camera function built into a smartphone, an image reading sensor installed at an outlet of an image forming apparatus, etc.). The feature amount information and the medium color information may be acquired by receiving the image data of the actual printed material AP obtained via the Internet or the like. Also, for example, the user may be prompted to input a product code assigned in advance to the print medium PM, and based on this, the feature amount information and the medium color information may be obtained. The same applies to the case where the machine learning device 100 acquires the feature amount information and the medium color information of the print medium PM.

Furthermore, in the above embodiment, the case where the storage unit 74 and the data processing unit 75 are provided inside the control unit 70 in the image forming apparatus 1 has been described. However, the present invention is not limited to this, and for example, at least one of the storage unit 74 and the data processing unit 75 may be provided in a server device or the like connected to the image forming apparatus 1 via a network. In this case, for example, the image forming apparatus 1 transmits feature amount information, medium color information, and the like to the server apparatus, and the server apparatus generates the third control information using the learned model, and the third control information is transmitted to the image forming apparatus 1. should be sent to

Furthermore, in the above-described embodiment, the case where the image forming apparatus 1 is an intermediate transfer type full-color LED printer has been described. However, the present invention is not limited to this, and for example, a tandem system or a rotary system in which the image is directly transferred from the photoreceptor drum to the printing medium may be employed, or monochrome may be employed instead of full color. Further, instead of a full color type having a special color toner, a full color type not having a special color toner may be employed, and a laser head may be employed instead of an LED head as an exposure means. good. In addition, the present invention may employ a copier or a facsimile machine in place of the image forming apparatus (so-called printer), or employ a digital multi-function machine in which the functions of a printer, an image scanner, a communication processing circuit, etc. are combined. Also good. In this case, an image scanner may be used to obtain feature amount information from the actual printed material AP, and medium color information may also be obtained.

Furthermore, the present invention is not limited to the embodiments described above and other embodiments. In other words, the scope of the present invention extends to embodiments obtained by arbitrarily combining part or all of each of the above-described embodiments and other embodiments described above, and to embodiments in which a part is extracted. is.

INDUSTRIAL APPLICABILITY The present invention can be used in an electrophotographic image forming apparatus.

1 image forming apparatus 10 display operation section 30 image forming section 70 control section 71 output control section 72 actual print information acquisition section 73 medium color information acquisition Unit 74 Storage unit 75 Data processing unit 81 Feature amount information acquisition unit 82 Medium information acquisition unit 83 First control information acquisition unit 91 Learned model storage unit , 92 Data set storage unit 93 Third control information storage unit 100 Machine learning device 110 State variable acquisition unit 120 Teacher data acquisition unit 130 Trained model generation unit , 140...storage unit, AP...actual printed matter, PM...printed medium.

Claims (10)

Feature amount information in an actual printed material printed by an image forming apparatus, medium information of a print medium used in the actual printed material, first control information that is control information when the actual printed material is output, and the printing a state variable acquisition unit that acquires medium color information representing the color of the medium as a state variable;
a teacher data acquisition unit that acquires, as teacher data, second control information in which the feature amount information is within a predetermined threshold;
A machine learning device comprising: a trained model generation unit that generates a trained model by performing machine learning based on the information acquired by the state variable acquisition unit and the teacher data. The machine learning device according to claim 1, wherein the feature amount information includes information about printing defects in the actual printed matter. The image forming apparatus is an electrophotographic type,
3. The machine learning device according to claim 1, wherein the first control information includes transfer voltage information in the image forming apparatus. The image forming apparatus is an electrophotographic type,
wherein the second control information includes transfer voltage information in the image forming apparatus;
The machine learning device according to any one of claims 1 to 3. 5. The machine learning device according to claim 3, wherein the learned model generation unit performs the machine learning on the correlation between the thickness of the printing medium and the transfer voltage in the image forming apparatus. 6. The machine according to claim 5, wherein the learned model generation unit performs the machine learning on a correlation between the thickness of the print medium and the transfer voltage in the image forming apparatus for a plurality of colors of the print medium. learning device. A first control that is characteristic amount information of an actual printed material printed by an image forming apparatus, medium information of a printing medium used for the actual printed material, and control information of the image forming apparatus when the actual printed material is output. an actual print information acquisition unit for acquiring information and medium color information representing the color of the print medium;
Data for inputting the information acquired by the actual printed matter information acquisition unit into a trained model generated by the machine learning device according to any one of claims 1 to 6, and for outputting third control information. a processing unit;
A data processing system comprising: a control information storage unit that stores the third control information output from the data processing unit. A first control that is characteristic amount information of an actual printed material printed by an image forming apparatus, medium information of a printing medium used for the actual printed material, and control information of the image forming apparatus when the actual printed material is output. A trained model generated by performing machine learning based on information, medium color information representing the color of the print medium, and teacher data that is second control information that makes the feature amount information within a predetermined threshold. a storage unit that stores
an actual printed material information acquisition unit that acquires actual printed material information including the medium information of the print medium to be used;
a data processing unit that inputs the actual printed matter information to the learned model and outputs third control information;
A printing system comprising: a print instruction unit that uses the third control information output from the data processing unit to issue a print instruction to print on the print medium to be used. A first control that is feature amount information of an actual printed material printed by an image forming apparatus using a computer, medium information of a printing medium used for the actual printed material, and control information when the actual printed material is output. a first process of acquiring information and medium color information representing the color of the print medium as state variables;
a second process of acquiring, as teacher data, second control information in which the feature amount information is within a predetermined threshold;
a third process of generating a learned model by performing machine learning based on the feature amount information, the medium information, the first control information, the medium color information, and the teacher data; machine learning methods including; Feature amount information of an actual printed material printed by an image forming apparatus using a computer, medium information of a printing medium used for the actual printed material, and control information of the image forming apparatus when the actual printed material is output and first data processing for acquiring first control information and medium color information representing the color of the print medium;
The obtained feature amount information, the medium information, the first control information, and the medium color information are input to a learned model generated by the machine learning method according to claim 9, and a third control is performed. a second data process that outputs information;
and a third data processing for storing the outputted third control information.

Images