Assistance device, assistance method, and learned model generation device

Abstract

This invention provides an auxiliary device, an auxiliary method, and a learned model generation apparatus that allow a user to determine the maximum ink amount based on their preferences. The processing unit of the auxiliary device performs the following processes: segmentation processing, obtaining multiple segmented test images by segmenting each test image into a predetermined number; prediction value acquisition processing, obtaining a predicted value representing the probability that the ink amount per unit area of ​​each segmented test image is appropriate as the maximum ink amount by executing the learned model using each of the segmented test images as input; and output processing, outputting prediction information indicating whether the ink amount per unit area of ​​each test image is within or outside the appropriate range of the maximum ink amount, based on the ink amount per unit area corresponding to each test image and the multiple predicted values.

Description

Technical Field

[0001] The present invention relates to an apparatus and method for assisting in determining the maximum amount of ink, which is an upper limit of the amount of ink per unit area on a printing medium. Background Technology

[0002] As a printing apparatus, an inkjet printer is known to eject ink droplets from a printhead onto a printing medium. When the amount of ink ejected per unit area of ​​the printing medium is relatively large, phenomena such as ink bleeding can occur, resulting in a color saturation state where the color development remains almost unchanged even if the ink ejection volume increases. Therefore, a maximum ink volume is set as an upper limit to the amount of ink per unit area on the printing medium, and this is used for creating color conversion LUTs (lookup tables).

[0003] Patent Document 1 discloses an information processing apparatus that uses a learned model generated through machine learning to infer the optimal maximum ink amount. The inferred maximum ink amount is a value called the optimal value, which is used in the generation of a color conversion LUT.

[0004] However, even if the user is dissatisfied with the maximum ink amount inferred from the learned model, the color conversion LUT will still be generated based on the inference result. Therefore, a new mechanism is needed for the user to determine the maximum ink amount.

[0005] Patent Document 1: Japanese Patent Application Publication No. 2021-24152 Summary of the Invention The auxiliary device of the present invention is configured to assist in determining a maximum ink amount, which is an upper limit of the ink amount per unit area on a printing medium, and includes: a holding unit that holds multiple test images obtained by reading multiple test color marks with different ink amounts per unit area on the printing medium; and a processing unit capable of executing a learned model that enables a computer to function in such a way that, based on multiple segmented test images obtained by segmenting each of the test images into a predetermined number, it obtains the ink amount per unit area of ​​each of the segmented test images as a result. The processing unit performs the following processes to obtain a predicted value with an appropriate probability of the maximum ink amount: segmentation processing, which obtains multiple segmented test images by segmenting each of the test images into the predetermined number; prediction value acquisition processing, which obtains the predicted value by executing the learned model with each of the segmented test images as input; and output processing, which outputs prediction information indicating whether the ink amount per unit area of ​​each test image is within or outside the appropriate range of the maximum ink amount, based on the ink amount per unit area corresponding to each of the test images and the multiple predicted values.

[0006] Furthermore, the auxiliary method of the present invention is implemented by a computer to assist in determining the maximum ink amount, which is the upper limit of the ink amount per unit area on a printing medium. The computer is capable of executing a learned model that functions to obtain a predicted value representing the probability that the ink amount per unit area of ​​each of the segmented test images is appropriate as the maximum ink amount, based on a plurality of segmented test images. The plurality of test images are images obtained by segmenting each of a plurality of test images obtained by reading a plurality of test color marks with different ink amounts per unit area on the printing medium into a predetermined number. The auxiliary method includes: a segmentation step, obtaining a plurality of segmented test images by segmenting each of the test images into the predetermined number; a prediction value acquisition step, obtaining the predicted value by executing the learned model with each of the segmented test images as input; and an output step, outputting predictive information representing whether the ink amount per unit area of ​​each of the test images is within or outside the appropriate range of the maximum ink amount, based on the ink amount per unit area corresponding to each of the test images and the plurality of predicted values.

[0007] Furthermore, the learned model generation apparatus of the present invention is configured to assist in determining a maximum ink amount, which is an upper limit of the ink amount per unit area on a printing medium, and includes: a holding unit that holds multiple teaching images obtained by reading multiple test color marks with different ink amounts per unit area on the printing medium; and a processing unit that generates a learned model that enables a computer to function in such a way that, through machine learning based on a relationship between a label and multiple segmented teaching images obtained by segmenting each of the multiple test images obtained by reading multiple test color marks with different ink amounts per unit area on the printing medium into multiple segmented test images, a predicted value representing the probability that the ink amount per unit area of ​​each segmented test image is appropriate as the maximum ink amount is obtained is obtained based on the multiple segmented test images obtained by segmenting each of the multiple test images obtained by reading multiple test color marks with different ink amounts per unit area on the printing medium into the predetermined number of segmented test images, wherein the label represents whether the ink amount per unit area of ​​each of the teaching color marks is appropriate, exceeds an appropriate range, or is below an appropriate range as the maximum ink amount. Attached Figure Description

[0008] Figure 1 A block diagram illustrating an example structure of an auxiliary system including a learned model generation device and an auxiliary device.

[0009] Figure 2 A diagram illustrating an example of a chart on a printed medium.

[0010] Figure 3 The diagram illustrates examples of teaching and testing chart images.

[0011] Figure 4 A diagram illustrating an example of generating a dataset from multiple teaching images with varying amounts of ink per unit area.

[0012] Figure 5 This is a diagram that schematically illustrates an example of a learned model generated by a learned model generation device and used by an auxiliary device.

[0013] Figure 6 A diagram illustrating examples of predictive information indicating whether the amount of ink per unit area of ​​each test image is within or outside the appropriate range of the maximum ink amount.

[0014] Figure 7 This diagram illustrates an example of a display containing predictive information in an auxiliary screen.

[0015] Figure 8 A flowchart illustrating an example of the learned model generation process.

[0016] Figure 9 A flowchart illustrating an example of auxiliary processing.

[0017] Figure 10 A flowchart illustrating an example of classification processing.

[0018] Figure 11 This diagram illustrates an example of the structure of a color conversion lookup table. Detailed Implementation

[0019] The embodiments of the present invention will now be described. Of course, the following embodiments are merely illustrative examples of the present invention, and not all features shown in these embodiments are necessarily essential to the solution provided by the invention.

[0020] (1) Summary of the methods included in this invention: First, refer to Figures 1 to 11 The examples shown illustrate the general outline of the methods included in this invention. Furthermore, the accompanying drawings are schematic illustrations, and there may be instances where the magnification varies in different directions, and where the individual drawings are not integrated. Of course, the elements of this invention are not limited to the specific examples indicated by symbols. In the "Summary of the Methods Included in this Invention," the words in parentheses indicate supplementary explanations of the preceding terms.

[0021] Furthermore, in this application, the numerical range "Min~Max" refers to the range above the minimum value Min and below the maximum value Max.

[0022] Method 1 like Figure 1 As illustrated, one approach involves an auxiliary device 3 that assists in determining a maximum ink quantity Qm, which is the upper limit of the ink quantity Q1 per unit area in the printing medium ME0, and includes a holding unit (e.g., RAM 113) and a processing unit 110. The holding unit (113) holds multiple test images 141 obtained by reading multiple test color marks PA2 with different ink quantities Q1 per unit area in the printing medium ME0. Figure 5 As illustrated, the processing unit 110 is capable of executing a learned model 300, which enables a computer (e.g., information processing device 100) to function in such a way that it obtains a predicted value (e.g., a predicted value PV1) representing the ink amount Q1 per unit area of ​​each of the segmented test images 142 as the maximum ink amount Qm, based on a plurality of segmented test images 142 obtained by segmenting each of the test images into a predetermined number (e.g., N). Figure 3, Figure 6 , Figure 7 , Figure 9 , Figure 10 As illustrated, the processing unit 110 performs the following processing.

[0023] (a1) A segmentation process is performed to obtain a plurality of segmented test images 142 by dividing each of the test images 141 into the predetermined number (N). Figure 9 Step S206).

[0024] (a2) A prediction value acquisition process (e.g., obtaining the predicted value (PV1) by executing the learned model 300 with each of the segmented test images 142 as input. Figure 9 Step S208).

[0025] (a3) Output processing of prediction information 400, based on the ink amount Q1 per unit area corresponding to each of the test images 141 and a plurality of the predicted values ​​(PV1), indicating whether the ink amount Q1 per unit area of ​​each of the test images 141 is within or outside the appropriate range of the maximum ink amount Qm (e.g., Figure 9 Steps S210 to S214).

[0026] When the learned model 300 is executed with multiple segmented test images 142 obtained by segmenting each of multiple test images 141 obtained by reading different ink amounts Q1 per unit area from multiple test color patches PA2, multiple predicted values ​​(PV1) representing the probability that the ink amount Q1 per unit area of ​​each segmented test image 142 is appropriate as the maximum ink amount Qm are obtained. Thus, for each test color patch PA2, detailed information can be obtained regarding whether the predicted ink amount Q1 per unit area as the maximum ink amount Qm is appropriate or inappropriate. The output prediction information 400 is based on the ink amount Q1 per unit area corresponding to each test image 141 and multiple predicted values ​​(PV1). The prediction information 400 is not limited to a single recommended value, but rather shows whether the ink amount Q1 per unit area of ​​each test image 141 is within or outside the appropriate range of the maximum ink amount Qm. Therefore, the user can incorporate their own preferences into the determination of the maximum ink amount Qm while referencing the predicted appropriate range with a margin of error. Therefore, the above method can provide an auxiliary device that allows users to determine the maximum ink volume based on their own preferences.

[0027] Various examples can be listed in the above methods.

[0028] Although ink is usually a liquid containing coloring materials such as pigments and dyes, it can also be a powdered solid, like tinting ink.

[0029] Multiple test color marks can be read by a scanner or by a camera. Therefore, multiple test images can be read images obtained by a scanner or captured images obtained by a camera.

[0030] Color scales, including test color scales and teaching color scales described later, can also include patterns such as line graphs, or solid color scales for uniformly recording concentration. (Refer to...) Figure 1 , Figure 2 To explain further, the recording density (denoted as RD) refers to the proportion (including percentage) of the number of dots DT0 formed by ink droplets 237 relative to a predetermined number of pixels PX0 on the printing medium ME0, and it refers to the ratio converted to the largest dot (e.g., a large dot) when dots of different sizes are formed. Pixel PX0 is the smallest element constituting an image that can be independently assigned color. Although in Figure 2 The diagram shows 25 pixels PX0, but when Nd large dots are formed relative to 100 pixels PX0, the recording density RD becomes Nd%. The ink amount Q1 per unit area refers to the amount of ink ejected from the print head 230 to a unit area of ​​the printing medium ME0, which is equivalent to the amount of ink used to form the color mark PA0 for recording density RD on the printing medium ME0, and is substantially equal to the recording density RD.

[0031] Cases where the ink amount per unit area of ​​each test image is outside the appropriate range of the maximum ink amount include those where the ink amount per unit area of ​​each test image is above the appropriate range, and those where the ink amount per unit area of ​​each test image is below the appropriate range. Therefore, the prediction information can also indicate whether the ink amount per unit area of ​​each test image is within, exceeds, or falls below the appropriate range of the maximum ink amount.

[0032] The output of the prediction information can be displayed, printed, or voice-output.

[0033] Of course, the additional notes described above will also apply to the following methods.

[0034] Method 2 like Figure 6 , Figure 7 , Figure 9 , Figure 10As illustrated, the processing unit 110 may also be configured to perform statistical processing on the plurality of predicted values ​​(PV1) obtained by executing the learned model 300 for each of the test images 141, thereby calculating an appropriateness index P representing the probability that the amount of ink Q1 per unit area corresponding to the test image 141 is appropriate. In the output processing, the processing unit 110 may generate the prediction information 400 based on the appropriateness index P regarding the amount of ink Q1 per unit area for each of the test images 141, and may also output the prediction information 400.

[0035] In the above circumstances, preferred examples can be provided for generating predictive information.

[0036] Here, statistical processing can be either averaging the arithmetic mean of multiple predicted values ​​(PV1) or extracting the median value when multiple predicted values ​​(PV1) are arranged in order (ascending or descending). This additional explanation will also apply to the following methods.

[0037] Method 3 like Figure 6 , Figure 10 As illustrated, the processing unit 110 may also be configured to compare the calculated suitability index P with the threshold TH1 during the output processing, and determine that the amount of ink Q1 per unit area where the suitability index P exceeds the threshold TH1 is within the suitability range. This auxiliary device 3 may also include an operation unit (e.g., input device 115) for accepting operations that change the threshold TH1.

[0038] In the above situations, users can change the correction range according to their own wishes. Therefore, the above method improves the convenience of determining the maximum ink volume.

[0039] Method 4 like Figure 7As illustrated, the processing unit 110 may also be configured to enable the display unit (e.g., display device 116) to display a plurality of display color marks 510 corresponding to the plurality of test color marks PA2 respectively. Alternatively, the plurality of display color marks 510 may include a plurality of appropriate range color marks 511 whose corresponding ink amount Q1 per unit area is within the appropriate range, and a plurality of inappropriate range color marks 512 that are not within the appropriate range color marks 511. Alternatively, the processing unit 110 may, in the output processing, include display information 515 as the prediction information 400 that distinguishes the plurality of appropriate range color marks 511 from the plurality of inappropriate range color marks 512, and enable the display unit (116) to display the plurality of display color marks 510.

[0040] In the above situation, since the user can visually confirm the appropriate range of color marks 511 among the multiple display color marks 510 corresponding to the multiple test color marks PA2, it is easy to confirm which test color mark PA2 is within the appropriate range among the multiple test color marks PA2. Therefore, the above method can improve the convenience of determining the maximum ink amount.

[0041] Method 5 like Figure 1 , Figure 7 As illustrated, the auxiliary device 3 may also be configured to include an operation unit (115) for handling operations on the plurality of display color marks 510 displayed on the display unit (116). Figure 9 As shown, the processing unit 110 can also perform the following processes.

[0042] (a4) When the operation unit (115) receives an operation for any one of the plurality of display color marks 510, it sets the ink amount Q1 per unit area corresponding to the operated display color mark 510 to the maximum ink amount Qm as a maximum ink amount setting process (e.g., Figure 9 Step S216).

[0043] In the above case, since the maximum ink quantity Qm can be set by operating the display color mark 510, the convenience of determining the maximum ink quantity can be further improved.

[0044] Method 6 Instead, the processing unit 110 may also perform the following processing.

[0045] (a5) When the operation unit (115) is prohibited from accepting operations on the plurality of inappropriate range color marks 512 and the operation unit (115) accepts an operation on any one of the plurality of appropriate range color marks 511, the maximum ink amount Q1 per unit area corresponding to the appropriate range color mark 511 that has been operated on is set to the maximum ink amount Qm.

[0046] In the above case, since the ink quantity Q1 per unit area that is determined to be the maximum ink quantity Qm becomes an appropriate range, the convenience of determining the maximum ink quantity can be further improved.

[0047] Method 7 like Figure 6 , Figure 10 As illustrated, the processing unit 110 may also be configured to determine the recommended value 410 of the maximum ink amount Qm in the output processing based on the ink amount Q1 per unit area corresponding to each of the test images 141 and a plurality of predicted values ​​(PV1). Figure 7 As illustrated, the processing unit 110 may also be configured to output recommendation information (e.g., recommendation color code 520) representing the recommendation value 410 in addition to outputting the prediction information 400 during the output processing.

[0048] Since recommendation information (520) indicating a recommended value 410 for the maximum ink quantity Qm is also output under the above circumstances, the user can also select the recommended value 410 according to the recommendation information (520). Therefore, the above method can improve the convenience of determining the maximum ink quantity.

[0049] Method 8 Another approach involves an auxiliary method that uses a computer (100) to assist in determining the maximum ink quantity Qm, which is the upper limit of the ink quantity Q1 per unit area on the printing medium ME0. The computer (100) is capable of executing the learned model 300. Figure 3 , Figure 6 , Figure 7 , Figure 9 , Figure 10 As illustrated, this auxiliary method includes the following steps.

[0050] (b1) Segmentation step ST1, which obtains a plurality of segmented test images 142 by segmenting each of the test images 141 into the predetermined number (N).

[0051] (b2) Prediction acquisition step ST2, which obtains the prediction value (PV1) by taking each of the segmented test images 142 as input to execute the learned model 300.

[0052] (b3) Output process ST3, based on the ink amount Q1 per unit area corresponding to each of the test images 141 and a plurality of the predicted values ​​(PV1), outputs prediction information 400 indicating whether the ink amount Q1 per unit area of ​​each of the test images 141 is within an appropriate range of the maximum ink amount Qm or outside the appropriate range.

[0053] The above method provides an auxiliary approach for users to determine the maximum ink volume based on their own preferences.

[0054] Method 9 In addition, such as Figure 1 As illustrated, one approach involves a learned model generation apparatus 2 that assists in determining a maximum ink quantity Qm, which is the upper limit of the ink quantity Q1 per unit area on the printing medium ME0, and includes a holding unit (113) and a processing unit 110. The holding unit (113) holds multiple teaching images 121 obtained by reading multiple teaching color marks PA1 with different ink quantities Q1 per unit area. Figures 3 to 5 , Figure 8 As illustrated, the processing unit 110 generates a learned model 300, which enables the computer (100) to function in such a way that, through machine learning based on the relationship between the label LA1 and a plurality of segmented teaching images 122 obtained by segmenting each of the teaching images 121 into a predetermined number (N), a predicted value (PV1) is obtained based on a plurality of segmented test images 142 obtained by segmenting each of the plurality of test images 141 obtained by reading the ink amount Q1 per unit area of ​​the printing medium ME0 into the predetermined number (N), representing the probability that the ink amount Q1 per unit area of ​​each of the segmented test images 142 is appropriate as the maximum ink amount Qm, wherein the label LA1 represents whether the ink amount Q1 per unit area of ​​each of the teaching color marks PA1 is appropriate, exceeds the appropriate range, or is below the appropriate range as the maximum ink amount Qm.

[0055] When the learned model 300 is executed by segmenting multiple test images 142 obtained by reading multiple test color patches PA2 with different ink amounts Q1 per unit area, multiple prediction values ​​(PV1) are obtained. Each prediction value (PV1) represents the probability that the ink amount Q1 per unit area of ​​each segmented test image 142 is appropriate as the maximum ink amount Qm. Thus, for each test color patch PA2, detailed information can be obtained regarding whether the predicted ink amount Q1 per unit area is appropriate or inappropriate as the maximum ink amount Qm. The user can consider the obtained prediction information 400 while incorporating their own preferences into the decision regarding the maximum ink amount Qm. Therefore, the above method provides a learned model generation apparatus that can obtain prediction information that serves as a reference when the user makes a decision regarding the maximum ink amount considering their own preferences.

[0056] Here, the multiple teaching color marks PA1 can be read by a scanner or by a camera, etc. Therefore, the multiple teaching images 121 can be read images obtained by a scanner or captured images obtained by a camera, etc.

[0057] Furthermore, the methods described above can be applied to auxiliary systems including the learned model generation apparatus and the auxiliary apparatus described above, learned model generation methods for generating the learned model described above, learned model generation programs for generating the learned model described above, control programs for the auxiliary apparatus described above, computer-readable non-transitory media recording any of the programs described above, the learned model described above, and computer-readable non-transitory media recording the learned model, etc. Each of the above apparatuses can be composed of multiple dispersed parts.

[0058] (2) Example of the structure of the learned model generation device and auxiliary device: Figure 1 The structure of the auxiliary system 1, including the learned model generation device 2 and the auxiliary device 3, is illustrated schematically. Figure 2 The diagram CH0 on the printed medium ME0 is illustrated schematically. Figure 2 The teaching diagram CH1 and the test diagram CH2 are uniformly represented as diagram CH0. Figure 2 In the diagram, within the double-dotted box, is a schematic diagram illustrating an example of the amount of ink Q1 per unit area.

[0059] Figure 1 The auxiliary system 1 shown includes an information processing device 100 that can be used as a learned model generation device 2 and an auxiliary device 3, and a printer 200 that can generate a printed image IM0 including a chart CH0.

[0060] The information processing device 100 includes a CPU (Central Processing Unit) 111, a ROM (Read Only Memory) 112, a RAM (Random Access Memory) 113, a storage device 114, an input device 115, a display device 116, and a communication I / F (interface) 117. The aforementioned elements (111 to 117) are electrically connected together and configured to input and output information to each other. Furthermore, the ROM 112, RAM 113, and storage device 114 are memories, and at least the ROM 112 and RAM 113 are semiconductor memories. The information processing device 100 includes a processing unit 110, primarily composed of the CPU 111. The RAM 113 is an example of a storage unit. The input device 115 is an example of an operation unit. The display device 116 is an example of a display unit.

[0061] Storage device 114 supports the OS (operating system) not shown, learning program PR1, and maximum ink quantity prediction program PR2. Figure 11 The color conversion LUT (lookup table) 600 shown is stored. The learning program PR1 enables the computer, i.e., the information processing device 100, to function as the learned model generation device 2. In order to execute the learning program PR1, thereby... Figure 2 The multiple teaching images 121 contained in the teaching chart CH1 shown, and the multiple labels LA1 associated with each of the multiple teaching images 121, are stored in RAM 113. After the execution of the learning program PR1, the learned model 300 is stored in the storage device 114. The maximum ink volume prediction program PR2 enables the information processing device 100 to function as an auxiliary device 3. In order to execute the maximum ink volume prediction program PR2, thereby... Figure 2 The multiple test images 141 contained in the test chart CH2 shown are stored in RAM 113. After the execution of the maximum ink volume prediction program PR2, the prediction information 400 is stored in RAM 113. Figure 11 The color conversion LUT 600 shown specifies the correspondence between the coordinate values ​​of R (red), G (green), and B (blue) and the coordinate values ​​of C (blue-green), M (magenta), Y (yellow), and K (black) for multiple grid points GD1. Figure 11 The variable i shown is the variable used to identify each grid point GD1. In addition, since both RAM113 and storage device 114 are memories, storage device 114 can function as an information storage unit, and RAM113 can also store the learned model 300.

[0062] Examples of storage devices 114 include non-volatile semiconductor memories such as flash memory and magnetic storage devices such as hard disks.

[0063] Examples of input devices 115 include indicator devices, physical keys such as keyboards, and touch panels affixed to the surface of the display panel. Input device 115 can also be an external device connected to the main body of the information processing device 100. Examples of display devices 116 include liquid crystal displays and organic EL displays. Display device 116 can also be an external device connected to the main body of the information processing device 100. Communication I / F 117 is connected to communication I / F 220 of printer 200 and inputs and outputs printing data and other information to printer 200.

[0064] The CPU 111 appropriately reads the information stored in the storage device 114 into the RAM 113, and performs various processes by executing the read program. The CPU 111 executes the learning program PR1 read into the RAM 113, thereby performing processing corresponding to the function of the learned model generation device 2. Furthermore, the CPU 111 executes the maximum ink volume prediction program PR2 read into the RAM 113, thereby performing processing corresponding to the function of the auxiliary device 3. Moreover, the CPU 111 performs color conversion processing, halftone processing, and printing data generation processing by executing a printing control program (not shown). For example, as a color conversion process, the CPU 111 performs processing according to… Figure 11 The color conversion LUT600 will convert each pixel to have R, G, and B 2 8 The process involves converting RGB data with integer values ​​above the tone level into ink volume data. Ink volume data, for example, is represented by 2^32 C, M, Y, and K values ​​in each pixel. 8 Integer values ​​above the tone count. As halftone processing, CPU 111 generates dot data with reduced tone count by performing halftone processing on the ink quantity data. As print data generation processing, CPU 111 performs processing to generate print data by adding command data to the dot data. The computer-readable non-transitory recording medium storing programs (PR1, PR2, etc.) is not limited to the internal storage device of the information processing apparatus 100, and can also be an external recording medium of the information processing apparatus 100.

[0065] Furthermore, the number of CPUs 111 in the processing unit 110 can be one or more. Additionally, part or all of the processing unit 110 can be replaced with hardware such as a DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), or FPGA (Field Programmable Gate Array).

[0066] The information processing device 100 may also include at least a portion of the printer 200. Furthermore, while the information processing device 100 may have all structural elements (111 to 117) within a single housing, it may also be constructed using multiple devices separated in a manner capable of communicating with each other. Therefore, the information processing device 100 may be a single personal computer, a combination of a mobile phone such as a smartphone and one or more personal computers, or a combination of one or more server computers and one or more terminals, etc. Additionally, the learned model generation device 2 and the auxiliary device 3 may also be constructed from separate computers.

[0067] Figure 1 The printer 200 shown is an inkjet printer that ejects C (blue-green) ink, M (magenta) ink, Y (yellow) ink, and K (black) ink, which are inks containing color materials, from the print head 230 to the printing medium ME0. Therefore, Figure 1 The ink 236 shown has four different colors. The printer 200 includes a controller 210, a communication I / F 220 described above, a print head 230, a drive unit 250, and a reading device 260. The reading device 260 can read the graphic CHO on the printed medium ME0. Examples of the reading device 260 include scanners and imaging devices. The reading device 260 can also be the main body of the printer 200 or an external device connected to the information processing unit 100.

[0068] The controller 210 includes a CPU 211, ROM 212, RAM 213, a drive signal transmitter, etc., and controls the operation of the communication I / F 220, printhead 230, drive unit 250, reading device 260, etc. The controller 210 controls the ejection of ink droplets 237 by the printhead 230 based on dot data included in the printing data obtained from the information processing device 100. The controller 210 can also control the relative movement between the printing medium ME0 and the printhead 230 implemented by the drive unit 250. In this way, a printed image IM0 corresponding to the printing data is formed on the printing medium ME0. Furthermore, the controller 210 can also control the transmission of the image read by the reading device 260 from the communication I / F 220 to the information processing device 100. The controller 210 can be configured as a SoC (System on a Chip) or similar device.

[0069] The printhead 230 includes a drive circuit or drive element, and performs printing by ejecting ink droplets 237 from multiple nozzles 234 included in the nozzle array 233 onto the printing medium ME0. Here, a nozzle refers to a small orifice for ejecting ink droplets, and a nozzle array refers to an arrangement of multiple nozzles. Figure 1 The printhead 230 shown includes a C-nozzle array 23C for ejecting C ink droplets 237, an M-nozzle array 23M for ejecting M ink droplets 237, a Y-nozzle array 23Y for ejecting Y ink droplets 237, and a K-nozzle array 23K for ejecting K ink droplets 237. For the drive element, a piezoelectric element that applies pressure to the ink in the pressure chamber connected to the nozzle 234, or a drive element that generates bubbles in the pressure chamber through heat, thereby ejecting the ink droplets 237 from the nozzle 234, can be used. For example, if the binary dot data based on the printing data is "dot formation," the controller 210 outputs a drive signal to the printhead 230 to eject ink droplets for dot formation. Furthermore, if the dot data is three or more values, if the dot data is "large dot formation," the controller 210 outputs a drive signal to eject ink droplets for large dots, and if the dot data is "small dot formation," it outputs a drive signal to eject ink droplets for small dots.

[0070] The printing medium ME0 is not specifically limited and includes paper, cloth, resin, metal, etc. The shape of the printing medium ME0 can be a two-dimensional shape obtained by cutting or a roll.

[0071] like Figure 2 As shown, the chart CH0 on the printing medium ME0 contains multiple pattern columns P0, which include multiple color marks PA0 with different ink ejection amounts 236. Figure 2The multiple pattern columns P0 shown include primary color pattern columns P11, P12, P13, P14, secondary color pattern columns P21, P22, ..., and tertiary color pattern column P31. Primary colors are colors rendered using only one type of ink, secondary colors are colors rendered using two different colors of ink, and tertiary colors are colors rendered using three different colors of ink. Within each pattern column P0, the color marks PA0 are arranged in the order of ink quantity QO1, representing the amount of ink per unit area. Furthermore, color marks PA0 are a collective term for the teaching color marks PA1 included in teaching chart CH1 and the test color marks PA2 included in test chart CH2.

[0072] exist Figure 2 Within the double-dotted box, as a simplified illustrative example, 5×5=25 pixels PX0 are shown as a predetermined number of pixels PX0 corresponding to a unit area. Of course, the predetermined number corresponding to a unit area is not limited to 25; a larger area can also be treated as a unit area. The ink quantity Q1 per unit area refers to the percentage (including percentage) of the number of ink droplets 237 ejected relative to the predetermined number of pixels PX0, and specifically, the percentage converted to the largest ink droplet when ink droplets 237 of different sizes are ejected relative to pixels PX0. Figure 2 The double-dotted box shows the case where the ink amount Q1 per unit area of ​​color swatch PA0 is (20 / 25) × 100 = 80%. However, since multiple types of ink droplets 237 are ejected relative to a single pixel PX0 when forming mixed-color images such as intermediate colors, Q1 may sometimes exceed 100%. For example, the maximum ink amount Q1 per unit area for intermediate colors can reach 200%.

[0073] Each color mark PA0 is a quadrilateral, containing multiple solid areas PA3 and multiple line areas PA4. Figure 2 The diagram illustrates a scenario where four solid areas PA3 exist within each color mark PA0, with line areas PA4 existing between these solid areas PA3. A solid area PA3 refers to an area where the type of ink 236 remains constant and the ink amount Q1 per unit area is uniform. Similarly, a line area PA4, where ink 236 is sprayed, also refers to an area where the type of ink 236 remains constant and the ink amount Q1 per unit area is uniform. For example, the primary color pattern column P11 includes a solid area PA3 for C and a line area PA4 for M; the primary color pattern column P12 includes a solid area PA3 for M and a line area PA4 for Y. Secondary color pattern columns P21, P22, ... include solid areas PA3 for secondary colors, and may also include line areas PA4 for secondary colors. Line areas PA4 may also be areas where ink 236 is not sprayed.

[0074] By observing a printing medium ME0 with multiple color marks PA0 having varying ink amounts Q1 per unit area, the relationship between phenomena such as "interruption" of line areas PA4, "thinning" of line areas PA4, "thickening" of line areas PA4, "adjacency" of line areas PA4, ink "bleeding," ink "agglomeration," and ink "overflow" and the ink amount Q1 per unit area can be determined. Specifically, "interruption" of line areas PA4 refers to the phenomenon of a portion of line area PA4 being missing. "Thinning" of line areas PA4 refers to the phenomenon of line areas PA4 becoming thinner compared to their original width, even if not reaching the level of "interruption." "Thickening" of line areas PA4 refers to the phenomenon of line areas PA4 becoming thicker compared to their original width. "Bleeding" of ink refers to the phenomenon of ink blurring due to ink seeping into surrounding areas. "Agglomeration" of ink refers to the phenomenon of reduced ink dot dispersion due to the aggregation of color materials. Ink "overflow" refers to the phenomenon where the shape of the color mark PA0 is damaged due to ink overflowing from the original area of ​​the color mark PA0. These phenomena are shown in Japanese Patent Application Publication No. 2021-24152.

[0075] The maximum ink quantity Qm, which serves as the upper limit of the ink quantity Q1 per unit area on the printing medium ME0 (refer to...). Figure 9 When there is too much ink, the colors in the darker areas of the printed image IM0 will become saturated, thus reducing image quality. On the other hand, if the maximum ink volume Qm is too small, the color rendering of the printed image IM0 will decrease. After repeated experiments, it was found that the phenomenon described above does not affect the entire color mark PA0, but rather occurs locally on the color mark PA0, and this localized phenomenon affects the image quality of the printed image IM0.

[0076] Therefore, as Figures 3 to 5 As illustrated, the learned model generation device 2 in this specific example determines to generate a learned model 300 for obtaining a predicted value PV1, which is a value representing the probability that the ink amount Q1 per unit area of ​​each segmented test image 142 is appropriate as the maximum ink amount Qm. Here, when the recommended value obtained based on the inference result of the learned model 300 is automatically determined to be the maximum ink amount Qm, a color conversion LUT will be generated according to the inference result even if the user is dissatisfied with the determined maximum ink amount Qm. Therefore, as Figure 6 , Figure 7As illustrated, the auxiliary device 3 in this specific example determines and provides the user with predictive information 400 indicating whether the ink amount Q1 per unit area of ​​each test image 141 is within or outside the appropriate range of the maximum ink amount Qm. The learned model generation device 2 and the auxiliary device 3 are positioned to assist in the determination of the maximum ink amount Qm made by the user.

[0077] Figure 3 The teaching chart image 120 and the test chart image 140 are illustrated schematically. Figure 3 The teaching chart image 120 and the test chart image 140 are shown together.

[0078] The teaching chart image 120 is transmitted via a reading device 260 (see reference). Figure 1 ) to Figure 2 The teaching chart CH1 shown is read. When the reading device 260 reads the teaching chart CH1 on the printed image IM0, it generates multiple teaching images 121 corresponding to the multiple teaching color marks PA1 contained in the teaching chart CH1. Therefore, the multiple teaching images 121 are obtained by reading the multiple teaching color marks PA1 with different ink amounts Q1 per unit area. The multiple teaching images 121 are sent to the information processing device 100 indirectly or directly. When the multiple teaching images 121 are received, the information processing device 100 stores the multiple teaching images 121 in RAM 113. The information processing device 100 may also store the multiple teaching images 121 in the storage device 114. In order to generate Figure 5 The learned model 300 shown is such that each teaching image 121 is segmented along the longitudinal and lateral directions in such a way that it becomes N segmented teaching images 122.

[0079] Test chart image 140 is used to read the data using reading device 260. Figure 2 The test chart CH2 shown is read to obtain the image. When the reading device 260 reads the test chart CH2 on the printed image IM0, it generates multiple test images 141 corresponding to the multiple test color marks PA2 contained in the test chart CH2. Therefore, the multiple test images 141 are obtained by reading the multiple test color marks PA2 with different ink amounts Q1 per unit area. These multiple test images 141 are sent indirectly or directly to the information processing device 100. Upon receiving the multiple test images 141, the information processing device 100 stores the multiple test images 141 in RAM 113. The information processing device 100 may also store the multiple test images 141 in the storage device 114. For utilization... Figure 5The learned model 300 shown has test images 141 segmented in both the longitudinal and lateral directions to become N segmented test images 142. That is, the number of segments of test images 141 is the same as the number of segments N of teaching images 121.

[0080] In addition, the number of segments N is not specifically limited, as long as it is between 50 and 5000 segments that can detect the above phenomenon.

[0081] Figure 4 An example of generating dataset DS1 from multiple teaching images 121 with varying ink amounts Q1 per unit area is illustrated. Figure 4 The “Duty” in the label table TA1 shown refers to the amount of ink Q1 per unit area.

[0082] First, as shown in label table TA1, the operation of establishing a correspondence between the teaching image 121 and the label LA1 according to the amount of each type of ink Q1 per unit area is carried out. Figure 4 In each of the labels LA1 shown, if the ink amount Q1 per unit area of ​​the corresponding teaching color mark PA1 is appropriate as the maximum ink amount Qm, it is "1"; if it exceeds the appropriate range, it is "0"; and if it is below the appropriate range, it is "2". In other words, label LA1 indicates whether the ink amount Q1 per unit area of ​​each teaching color mark PA1 is appropriate, exceeds the appropriate range, or is below the appropriate range as the maximum ink amount Qm. Of course, the value of label LA1 can be appropriately changed. In this specific example, for... Figure 2 , Figure 3 A label table TA1 is generated for each pattern column P0 shown. Additionally, label LA1 is assigned by an observer who has observed the teaching chart CH1. That is, for each pattern column P0, the observer assigns label "1" to the ink quantity Q1 per unit area of ​​the teaching color mark PA1 that is deemed appropriate as the maximum ink quantity Qm, labels "0" to the ink quantity Q1 per unit area of ​​the teaching color mark PA1 that is deemed to exceed the appropriate range as the maximum ink quantity Qm, and labels "2" to the ink quantity Q1 per unit area of ​​the teaching color mark PA1 that is deemed to be below the appropriate range as the maximum ink quantity Qm. In this specific example, there is one ink quantity Q1 per unit area assigned the label "1" indicating "appropriate" for each pattern column P0.

[0083] Next, each teaching image 121 is divided into N segmented teaching images 122, and a process is performed to establish a correspondence between the label LA1 corresponding to the original teaching image 121 and all the segmented teaching images 122. For example, the learned model generation device 2 divides the teaching image "T1_100" with Q1=100% into N segmented teaching images "T1_100_1" to "T1_100_N", and establishes a correspondence between the label "0" with Q1=100% and all the segmented teaching images "T1_100_1" to "T1_100_N". The learned model generation device 2 segments the teaching image "T1_90" with Q1=90% into N segmented teaching images "T1_90_1" to "T1_90_N", and establishes a correspondence between the label "1" with Q1=90% and all segmented teaching images "T1_90_1" to "T1_90_N". The learned model generation device 2 also segments the teaching image "T1_80" with Q1=80% into N segmented teaching images "T1_80_1" to "T1_80_N", and establishes a correspondence between the label "2" with Q1=80% and all segmented teaching images "T1_80_1" to "T1_80_N". This collection of data becomes the dataset DS1, which is input to the neural network that becomes the learned model 300.

[0084] Figure 5 A learned model 300 generated by the learned model generation apparatus 2 and used by the auxiliary apparatus 3 is illustrated schematically. In this specific example, the learned model generation apparatus 2 determines to generate a learned model 300 for each type of printing medium ME0, and also generates a learned model 300 for each pattern column P0. When the output resolution of the printer 200 can be changed, the learned model generation apparatus 2 can also generate a learned model 300 for each output resolution.

[0085] The learned model generation device 2 generates a learned model 300 by inputting a dataset DS1, obtained by establishing a correspondence between the label LA1 and all segmentation teaching images 122, into a neural network. The learned model generation device 2 repeatedly performs machine learning on the provisional learned model 300 to increase the probability that the output becomes the label LA1 relative to the input of the segmentation teaching images 122. For example, each time an input is given to the provisional learned model 300, a feature vector for distinguishing the label LA1 is calculated from each segmentation teaching image 122, and the aforementioned machine learning is repeatedly performed based on this feature vector. The neural network can also be described as performing machine learning based on the relationship between the label LA1 and multiple segmentation teaching images 122. The learned model 300, upon inputting the segmented test image 142, is able to output a predicted value PV0 representing the probability that the segmented test image 142 corresponds to label "0", a predicted value PV1 representing the probability that the segmented test image 142 corresponds to label "1", and a predicted value PV2 representing the probability that the segmented test image 142 corresponds to label "2". The learned model 300 enables the information processing device 100 to function in such a way that, based on the segmented test image 142, it obtains a predicted value PV1 representing the probability that the ink amount Q1 per unit area of ​​the segmented test image 142 is appropriate as the maximum ink amount Qm.

[0086] To input multiple segmentation test images 142 into the learned model 300, firstly, as shown in the test image table TA2, a process is performed to establish a correspondence between the test images 141 and the ink amount Q1 per unit area. Next, a process is performed to segment each test image 141 into N segmentation test images 142. For example, the auxiliary device 3 segments the test image "T2_100" with Q=100% into N segmentation test images "T2_100_1" to "T2_100_N". The auxiliary device 3 also segments the test image "T2_90" with Q=90% into N segmentation test images "T2_90_1" to "T2_90_N", and the test image "T2_80" with Q=80% into N segmentation test images "T2_80_1" to "T2_80_N". These segmentation test images 142 are input to the learned model 300, and for each segmentation test image 142, a predicted value PV1 is obtained from the learned model 300.

[0087] However, since there are N predicted values ​​PV1 for each unit area of ​​ink quantity Q1, the auxiliary device 3 decides to perform statistical processing on the N predicted values ​​PV1 according to each ink quantity Q1 per unit area, thereby calculating the suitability index P. When averaging is performed as statistical processing, the auxiliary device 3 calculates the arithmetic mean of the N predicted values ​​PV1 as the suitability index P. Of course, a geometric mean or similar calculation can also be performed instead of the arithmetic mean. Furthermore, the auxiliary device 3 can also arrange the N predicted values ​​PV1 in order (ascending or descending) and calculate the median value of the N predicted values ​​PV1 in that order as the suitability index P. In either case, the suitability index P indicates the probability that the ink quantity Q1 per unit area corresponding to the test image 141 is appropriate.

[0088] Figure 6 An example of predictive information 400 is shown schematically, indicating whether the amount of ink Q1 per unit area of ​​each test image 141 is within or outside the appropriate range of the maximum ink amount Qm.

[0089] like Figure 6 As shown, a correspondence is established between the ink quantity Q1 per unit area and the calculated suitability index P. Using a threshold TH1 as a benchmark, the system outputs whether the ink quantity Q1 per unit area is within the "suitable range," exceeds the suitable range ("excessive"), or falls below the suitable range ("insufficient"). The threshold TH1 is applied to the suitability index P for each ink quantity Q1 per unit area. Figure 6 In the example shown, when the threshold TH1 is 10% and the ink amount Q1 per unit area is 75% to 90%, the external output is "appropriate range" because the suitability indices P(75), P(80), P(85), and P(90) are greater than the threshold TH1. When the ink amount Q1 per unit area is 95% to 100%, the external output is "excessive" because the suitability indices P(95) and P(100) are below the threshold TH1 and Q1 = 95% to 100% exceeds the appropriate range of 75% to 90% ink amount per unit area. When the ink amount Q1 per unit area is below 70%, the external output is "insufficient" because the suitability indices P(70) are below the threshold TH1 and Q1 ≤ 70% is below the appropriate range of 75% to 90% ink amount per unit area. In short, Figure 6 The prediction information 400 shown indicates whether the amount of ink Q1 per unit area of ​​the test image 141 is within the appropriate range of the maximum ink amount Qm, or exceeds the appropriate range, or is below the appropriate range.

[0090] Furthermore, the recommended value 410 for the maximum ink quantity Qm is included within the appropriate range of ink quantity Q1 per unit area. The auxiliary device 3 determines the recommended value 410 for the maximum ink quantity Qm based on the ink quantity Q1 per unit area corresponding to each test image 141 and the suitability index P. The recommended value 410 can also be the ink quantity Q1 per unit area with the highest suitability index P. Alternatively, the recommended value 410 can also be the middle ink quantity per unit area included in the first three ink quantities Q1 per unit area when the ink quantities Q1 per unit area are arranged in ascending or descending order according to the suitability index P. Figure 6 In the example shown, the ink content Q1 of the first three unit areas is within the appropriate range and is 80%, 85%, and 90% respectively. Their midpoint, Q1=85%, is the recommended value 410.

[0091] Figure 7 An example of an auxiliary screen 500 is shown schematically, which includes predictive information 400 of the suitability index P based on the amount of each ink Q1 per unit area.

[0092] Figure 1 The processing unit 110 shown enables the display device 116 to interact with... Figure 2 The multiple test color marks PA2 shown correspond to multiple display color marks 510. For each pattern column P0, the multiple display color marks 510 include multiple appropriate range color marks 511 where the ink amount Q1 per unit area is within an appropriate range, and multiple inappropriate range color marks 512 that are not within the appropriate range color marks 511. Additionally, the multiple inappropriate range color marks 512 include multiple display color marks exceeding the appropriate range and multiple display color marks below the appropriate range. Therefore, for each pattern column P0, the inappropriate range color mark 512 that is higher than the appropriate range color mark 511 indicates that the ink amount Q1 per unit area of ​​the test image 141 exceeds the appropriate range of the maximum ink amount Qm. The inappropriate range color mark 512 that is lower than the appropriate range color mark 511 indicates that the ink amount Q1 per unit area of ​​the test image 141 is below the appropriate range of the maximum ink amount Qm. The processing unit 110 includes, as prediction information 400, display information 515 that makes the multiple appropriate range color marks 511 more prominent by dimming the multiple inappropriate range color marks 512, so that the display device 116 displays the multiple display color marks 510. The display of the multiple inappropriate range color marks 512 can also be a grayed-out display that prohibits operations on each inappropriate range color mark 512. The display information 515 can be said to be information that distinguishes the multiple appropriate range color marks 511 from the multiple inappropriate range color marks 512 as prediction information 400.

[0093] In addition to displaying the prediction information 400, the processing unit 110 also enables the display device 116 to display the prediction information. Figure 6 The recommended color 520 for the recommended value 410 is displayed. Figure 7 The auxiliary screen 500 shown has recommended color codes 520 according to each pattern column P0. Figure 7 Each recommended color bar 520 shown is surrounded by a thick line to make it more conspicuous. Recommended color bars 520 are examples of recommendation information output externally.

[0094] The input device 115 is capable of accepting the operation of selecting any one of a plurality of appropriate range color marks 511 for each pattern column P0. Alternatively, the input device 115 can also accept the operation of selecting any one of a plurality of display color marks 510 for each pattern column P0. The input device 115 can be described as an operation unit for accepting operations on the plurality of display color marks 510 displayed on the display device 116.

[0095] (3) Specific examples of learned model generation processing: Figure 8 The learned model generation process implemented by the learned model generation device 2 is illustrated schematically. Referring also to the following... Figures 1 to 5 The learning model generation process in steps S102 to S110 will be explained below. Additionally, the term "step" may sometimes be omitted, and the symbol for the step may be shown in parentheses.

[0096] The main body of the learned model generation process is the processing unit 110. The learned model generation process begins when the input device 115 receives an operation to generate the learned model 300.

[0097] When the learned model generation process begins, the processing unit 110 performs the formation on the printing medium ME0. Figure 2 Control of the shown teach chart CH1 (S102). As described above, the teach chart CH1 includes multiple teach color marks PA1 with different ink amounts Q1 per unit area. For example, the storage device 114 stores teach chart printing data that causes the printer 200 to print the teach chart CH1, and the processing unit 110 forms the teach chart CH1 on the printing medium ME0 by sending the teach chart printing data to the printer 200. Alternatively, if the teach chart CH1 has been prepared, the processing in S102 can be omitted.

[0098] Next, the processing unit 110 causes the reading device 260 to read the teaching chart CH1 on the printed medium ME0 and obtain the generated teaching chart image 120 (see reference). Figure 3The teaching chart image 120 is stored in RAM 113 (S104). The teaching chart image 120 includes multiple teaching images 121 corresponding to multiple teaching color marks PA1, each with a different ink amount Q1 per unit area.

[0099] Next, the processing unit 110 performs the process of assigning the label LA1 to each teaching color mark PA1, and as follows: Figure 4 As shown in the label table TA1, a correspondence is established between the label LA1 and each teaching image 121 (S106). As described above, the label LA1 indicates whether the ink amount Q1 per unit area of ​​each teaching color mark PA1 is appropriate, exceeds the appropriate range, or is below the appropriate range as the maximum ink amount Qm. The processing of assigning the label LA1 can also be a process of accepting the input of the label LA1 value for each teaching color mark PA1 via the input device 115. In this case, the observer of the teaching chart CH1 only needs to input "1" when the ink amount Q1 per unit area of ​​the teaching color mark PA1 is appropriate as the maximum ink amount Qm, input "0" when it exceeds the appropriate range, and input "2" when it is below the appropriate range. When the value of the label LA1 is input, the processing unit 110 establishes a correspondence between the teaching image 121 and the value of the label LA1 for each pattern column P0 by establishing a correspondence between each ink amount Q1 per unit area and the value of the label LA1, thereby generating the label table TA1.

[0100] Next, the processing unit 110 generates... Figure 4 The dataset DS1 (S108) is shown. At this time, the processing unit 110 divides each teaching image 121 into N segmented teaching images 122, and establishes a correspondence between the label LA1 corresponding to the original teaching image 121 and all segmented teaching images 122. Thus, a dataset DS1 is generated for each pattern column P0, with the label LA1 corresponding to each segmented teaching image 122.

[0101] Finally, the processing unit 110 performs machine learning with the dataset DS1 as input and generates a learned model 300 (S110). Figure 5 As shown, the learned model 300 enables the information processing device 100 to function in such a way that, by taking the input of the segmented test image 142, it obtains predicted values ​​PV0, PV1, and PV2 representing the probability that the segmented test image 142 corresponds to the label LA1. In other words, the processing unit 110 generates the learned model 300 described above through machine learning based on the relationship between the label LA1 and multiple segmented teaching images 122.

[0102] (4) Specific examples of auxiliary processing: Figure 9The auxiliary processing performed by the auxiliary device 3 is illustrated schematically. Here, S206 corresponds to the segmentation process (a1) and the segmentation step ST1. S208 corresponds to the prediction value acquisition process (a2) and the prediction value acquisition step ST2. S210 to S214 correspond to the output process (a3) ​​and the output step ST3. S216 corresponds to the maximum ink quantity setting process (a4 or a5) and the maximum ink quantity determination step ST4. Figure 10 The illustration shows that in Figure 9 The classification process implemented in S212. See also below. Figures 1 to 7 The auxiliary processing steps from S202 to S216 will be explained.

[0103] The main body of the auxiliary processing is the processing unit 110. The auxiliary processing begins when the input device 115 receives an operation to determine the maximum ink volume Qm.

[0104] When auxiliary processing begins, the processing unit 110 performs the forming of such a process on the printing medium ME0. Figure 2 The control of the test chart CH2 shown is described in section S202. As mentioned above, the test chart CH2 includes multiple test color marks PA2 with different ink amounts Q1 per unit area. For example, the storage device 114 stores test chart printing data that causes the printer 200 to print the test chart CH2, and the processing unit 110 forms the test chart CH2 on the printing medium ME0 by sending the test chart printing data to the printer 200. Alternatively, if the test chart CH2 has been prepared, the processing in section S202 can be omitted.

[0105] Next, the processing unit 110 causes the reading device 260 to read the test chart CH2 on the printed medium ME0 and obtain the generated test chart image 140 (see reference). Figure 3 The test chart image 140 is stored in RAM 113 (S204). The test chart image 140 includes multiple test images 141 corresponding to multiple test color marks PA2, each with a different ink amount Q1 per unit area.

[0106] Next, the processing unit 110 divides each test image 141 into N segments to obtain N segmented test images 142 (S206).

[0107] Next, the processing unit 110 sets each segmented test image 142 as input to enable the learned model 300 to execute, thereby obtaining the predicted value PV1 (refer to) representing the label "1" within the appropriate range. Figure 5(S208). In S208, N predicted values ​​PV1 are obtained for each segmented test image 142. In S208, the processing unit 110 may also obtain predicted values ​​PV0 representing the label "0" which is above the appropriate range, and may also obtain predicted values ​​PV2 representing the label "2" which is below the appropriate range.

[0108] Next, the processing unit 110 performs statistical processing on the N predicted values ​​PV1 obtained by executing the learned model 300 for each test image 141, thereby calculating, as shown in the figure. Figure 6 The appropriateness index P is shown (S210). For example, the processing unit 110 calculates the appropriateness index P as the arithmetic mean of N predicted values ​​PV1 for each test image 141. As described above, the appropriateness index P indicates the probability that the amount of ink Q1 per unit area corresponding to the test image 141 is appropriate.

[0109] Alternatively, the N predicted values ​​PV0 can be considered together in the calculation of the suitability index P, and the N predicted values ​​PV2 can also be considered together in the calculation of the suitability index P.

[0110] Next, the processing unit 110 performs a classification process (S212) to determine the ink quantity Q1 per unit area. This will be referred to below. Figure 10 Let's illustrate the classification process with an example.

[0111] When the hierarchical processing begins, the processing unit 110 causes the display device 116 to... Figure 10 The appropriate range selection screen 530 is displayed (S302). The appropriate range selection screen 530 has multiple options 531 for substantially selecting the threshold TH1 in the appropriateness index P for each ink amount Q1 applied per unit area, and an "OK" button 532. The multiple options 531 include "narrow" for judging the appropriate range of the maximum ink amount Qm under strict conditions, "normal" for judging the appropriate range under recommended conditions, and "wide" for judging the appropriate range under lenient conditions. When "narrow" is selected, the threshold TH1 is set larger than when "normal" is selected, and when "wide" is selected, the threshold TH1 is set smaller than when "normal" is selected. The input device 115 can accept the operation of selecting any one of the multiple options 531. After accepting the operation of option 531, when the input device 115 accepts the operation of the "OK" button 532, the processing unit 110 sets the threshold TH1 according to the selected option 531. Therefore, the input device 115 can be described as an operation unit for accepting operations that change the threshold TH1.

[0112] Next, the processing unit 110 classifies the amount of ink Q1 per unit area that exceeds the threshold TH1 as "appropriate range" (S304). Figure 6 In the example shown, since the suitability indices P(75), P(80), P(85), and P(90) are greater than the threshold TH1, Q1 = 75% to 90% is classified as "suitable range". It can be said that the processing unit 110 determines that the amount of ink per unit area Q1, for which the calculated suitability index P exceeds the threshold TH1, is within the suitable range.

[0113] Next, the processing unit 110 determines a recommended value 410 for the maximum ink quantity Qm from an appropriate range of ink quantity Q1 per unit area (refer to...). Figure 6 (S306). For example, the processing unit 110 determines the ink amount per unit area with the largest suitability index P among the ink amounts per unit area Q1 within the appropriate range as the recommended value 410. Alternatively, the processing unit 110 may also determine the ink amount per unit area of ​​the middle value included in the ink amounts per unit area of ​​the top three suitability index P rankings among the ink amounts per unit area Q1 within the appropriate range as the recommended value 410.

[0114] In the above manner, the processing unit 110 determines the recommended value 410 based on the ink amount Q1 per unit area corresponding to each test image 141 and multiple predicted values ​​PV1.

[0115] Next, the processing unit 110 classifies the amount of ink Q1 per unit area that does not exceed the threshold TH1, based on whether it is "excessive" or "insufficient" (S308). Figure 6 In the example shown, since the suitability indices P(95) and P(100) of Q1 > 90% are below the threshold TH1, Q1 = 95 to 100% is classified as "excessive". Furthermore, since the suitability indices P(70) of Q1 < 75% are below the threshold TH1, Q1 ≤ 70% is classified as "insufficient". It can be said that the processing unit 110 determines that the amount of ink Q1 per unit area for which the calculated suitability index P does not exceed the threshold TH1 is outside the appropriate range.

[0116] Using the above method, for each type of printing medium ME0, the amount of ink Q1 per unit area is further classified into three categories: "appropriate range," "excessive," and "insufficient" for each pattern column P0. This classification can also be implemented for each output resolution. In addition, if the color mark 510 is displayed in the same way in "excessive" and "insufficient," the processing unit 110 can also uniformly classify "excessive" and "insufficient" as "outside the appropriate range."

[0117] Finally, the processing unit 110 generates prediction information 400 for display based on the classification of ink quantity Q1 per unit area (see reference). Figure 7 (S310). Figure 7 The prediction information 400 shown includes display information 515 for distinguishing multiple appropriate range color marks 511 from multiple inappropriate range color marks 512, and is supplemented with information for displaying the recommended color mark 520 representing the recommended value 410.

[0118] As described above, the processing unit 110 generates prediction information 400 based on the suitability index P of the ink amount Q1 per unit area.

[0119] After the classification process is completed, the processing unit 110 causes the display device 116 to display the auxiliary screen 500, which includes multiple classified display color marks 510 containing ink amount Q1 per unit area. Figure 9 (S214 shown). Figure 7 The auxiliary screen 500 shown has predictive information 400 indicating whether the ink amount Q1 per unit area of ​​each test image 141 is within or outside the appropriate range of the maximum ink amount Qm. In addition to the predictive information 400, the auxiliary screen 500 also has a recommended color scale 520 indicating a recommended value 410.

[0120] Based on the above method, it can be said that the processing unit 110 performs the following processing, that is, it outputs prediction information 400 based on the amount of ink per unit area Q1 corresponding to each test image 141 and multiple prediction values ​​PV1, and in addition to outputting prediction information 400, it also outputs recommended color mark 520.

[0121] Finally, the processing unit 110 accepts the operation for any one of the plurality of appropriate range color marks 511 via the input device 115, and sets the ink amount Q1 per unit area corresponding to the appropriate range color mark 511 that has been operated to the maximum ink amount Qm (S216). For example, in Figure 7 In the auxiliary screen 500 shown, when the appropriate range color mark 511 (Q1=80%) in the pattern column P0 of "C / M" is operated, the processing unit 110 sets the maximum ink amount Qm of C to 80%. In this case, the maximum ink amount Qm is set to a value different from the recommended value of 85% shown in the recommended color mark 520. Of course, when the recommended color mark 520 in the pattern column P0 of "C / M" is operated, the processing unit 110 sets the maximum ink amount Qm of C to the recommended value of 85%.

[0122] The maximum ink quantity Qm is not limited to being set for each pattern column P0; it can be set for either the primary colors or the secondary colors. In this case, when the appropriate range of color stops 511 with Q1=80% is operated on in any of the primary color pattern columns P0, the processing unit 110 sets the maximum ink quantity Qm for the primary colors to 80%. When the maximum ink quantity Qm obtained by summarizing the primary colors is set, the learned model 300 for the primary colors can also be generated through machine learning based on the dataset DS1 obtained by summarizing the primary colors. When the maximum ink quantity Qm obtained by summarizing the secondary colors is set, the learned model 300 for the secondary colors can also be generated through machine learning based on the dataset DS1 obtained by summarizing the secondary colors.

[0123] Furthermore, when the processing unit 110 prohibits the input device 115 from accepting operations on each inappropriate range color mark 512, even if an inappropriate range color mark 512 is operated, the corresponding ink amount Q1 per unit area will not be set to the maximum ink amount Qm. On the other hand, the processing unit 110 may also allow the input device 115 to accept operations on each inappropriate range color mark 512 in addition to operations on each appropriate range color mark 511. For example, in Figure 7 In the auxiliary screen 500 shown, when the inappropriate range color mark 512 of Q1=95% in the pattern column P0 of "C / M" is operated, the processing unit 110 sets the maximum ink amount Qm of C or the primary color to 95%.

[0124] The maximum ink volume Qm is determined to be used in the production of the color conversion LUT 600 (reference) during the color conversion process. Figure 11 It is used for purposes such as ) and others. Figure 11 As shown, a correspondence is established between the coordinate values ​​(C, M, Y, K) = (Ci, Mi, Yi, Ki) of the ink volume data and the grid point GD1 with the coordinate values ​​(R, G, B) of the RGB data as (Ri, Gi, Bi). In this case, the processing unit 110 generates a color conversion LUT 600 such that the ink volume obtained by adding the ink volume corresponding to coordinate value Ci, the ink volume corresponding to coordinate value Mi, the ink volume corresponding to coordinate value Yi, and the ink volume corresponding to coordinate value Ki is less than or equal to the maximum ink volume Qm. When color conversion processing is performed according to the color conversion LUT 600 generated in this way, the ink volume per unit area on the printed image IM0 is limited to less than or equal to the maximum ink volume Qm.

[0125] Of course, the color conversion LUT is not limited to the color conversion LUT 600 described above. The input coordinate values ​​of the color conversion LUT can also be C, M, and Y coordinate values, C, M, Y, and K coordinate values, etc. The output coordinate values ​​of the color conversion LUT can also be C, M, Y, K, and the coordinate values ​​of a spot color, etc. Examples of spot colors include Or (orange), Gr (green), Lc (light blue-green) with lower concentration compared to C, Lm (light magenta) with lower concentration compared to M, Dy (deep yellow) with higher concentration compared to Y, and Lk (light black) with lower concentration compared to K, etc. Furthermore, the processing unit 110 can convert RGB data, etc., into ink volume data according to the color conversion LUT that may exceed the maximum ink volume Qm, and generate printing data after converting the ink volume of each pixel of the ink volume data to below the maximum ink volume Qm.

[0126] like Figure 5 As shown, through Figure 8 The learned model 300 generated by the learned model generation process shown enables the information processing device 100 to function in such a way that, based on multiple test images 141 obtained by reading multiple test color marks PA2 with different ink amounts Q1 per unit area on the printing medium ME0, each is divided into N segmented test images 142, and a predicted value PV1 representing the probability that the ink amount Q1 per unit area of ​​each segmented test image 142 is appropriate as the maximum ink amount Qm is obtained. When the learned model 300 is executed with N segmented test images 142 as input, N predicted values ​​PV1 representing the probability that the ink amount Q1 per unit area of ​​each segmented test image 142 is appropriate as the maximum ink amount Qm are obtained, wherein the N segmented test images 142 are segmented test images obtained by segmenting each of the multiple test images 141 obtained by reading multiple test color marks PA2 with different ink amounts Q1 per unit area. Therefore, for each test color mark PA2, we can obtain detailed information on whether the predicted ink amount Q1 per unit area as the maximum ink amount Qm is appropriate or inappropriate.

[0127] pass Figure 9The prediction information 400 output by the auxiliary processing shown is based on the ink amount Q1 per unit area corresponding to each test image 141 and N predicted values ​​PV1. The prediction information 400 is not limited to a single recommended value, but rather indicates whether the ink amount Q1 per unit area of ​​each test image 141 is within or outside the appropriate range of the maximum ink amount Qm. Thus, the user can incorporate their own preferences into the determination of the maximum ink amount Qm while referring to the appropriate range predicted with a margin of error. Therefore, in this specific example, there is an auxiliary device that allows the user to consider their preferences when determining the maximum ink amount.

[0128] (5) Variation example: Various variations of this invention can be listed.

[0129] For example, features other than the label LA1 and the segmentation teaching image 122 can be added to the dataset DS1 used for machine learning. When the learned model generation device 2 generates a learned model 300 that summarizes primary or secondary colors, the color information of the solid region PA3 and the line region PA4 can be appended to the dataset DS1. In this case, the auxiliary device 3 executes the learned model 300 by taking the segmentation test image 142, the color information of the solid region PA3, and the color information of the line region PA4 as input, thereby obtaining the predicted value PV1 and outputting the predicted information 400. When the learned model generation device 2 generates a learned model 300 that summarizes multiple types of printing media ME0, the type information of printing media ME0 can also be appended to the dataset DS1. In this case, the auxiliary device 3 executes the learned model 300 by taking the segmentation test image 142 and the type information of printing media ME0 as input, thereby obtaining the predicted value PV1 and outputting the predicted information 400. Furthermore, features such as output resolution can also be added to the dataset DS1.

[0130] Color mark PA0, which includes teaching color mark PA1 and test color mark PA2, can also be a solid color mark where the type of ink 236 remains unchanged and the amount of ink Q1 per unit area is uniform, provided that there is no line area PA4. In this case, since phenomena such as ink "bleeding," ink "agglomeration," and ink "overflow" may also occur, the learned model 300 can be used to obtain the predicted value PV1, and the predicted information 400 can be output.

[0131] pass Figure 9The auxiliary processing shown indicates that the amount of ink for the test color mark PA2 selected by the user is appropriate for the user as the maximum ink amount Qm. Therefore, the test chart CH2 used in determining the maximum ink amount Qm can also be set as an additional teaching chart CH1, thereby enabling the learned model generation device 2 to perform additional machine learning. By using the maximum ink amount Qm as appropriate and based on the relationship between the label and the segmentation test image 142 as an additional segmentation teaching image 122, the learned model 300 is updated, wherein the label indicates whether the ink amount Q1 per unit area of ​​each additional teaching color mark PA1 is appropriate, exceeds the appropriate range, or is below the appropriate range as the maximum ink amount Qm.

[0132] In the above processing, for example, a judgment on whether it is "more than" can be replaced by a judgment on whether it is "above", and a judgment on whether it is "below" can be replaced by a judgment on whether it is "smaller". Such cases where the judgments are replaced in these ways are also included in the approach of this application.

[0133] (6) Summary: As explained above, according to the present invention, it is possible to provide structures in various ways that allow users to determine the maximum ink amount according to their own wishes. Of course, the above-mentioned basic functions and effects can also be obtained in a configuration consisting only of structural elements involved in independent technical solutions.

[0134] Furthermore, it is also possible to implement structures that involve replacing or altering the structures disclosed in the examples described above, or structures that involve replacing or altering the structures disclosed in the examples described above with known technologies. This invention also includes these structures.

[0135] Symbol Explanation 1…Auxiliary system; 2…Learned model generation device; 3…Auxiliary device; 100…Information processing device; 110…Processing unit; 113…RAM; 115…Input device; 116…Display device; 120…Teaching chart image; 121…Teaching image; 122…Segmented teaching image; 140…Test chart image; 141…Test image; 142…Segmented test image; 200…Printer; 210…Controller; 230…Print head; 236…Ink; 237…Ink droplet; 250…Driver; 260…Reading device; 300…Learned model; 400…Prediction information; 410…Recommended value; 500…Auxiliary screen; 510…Display color mark; 511…Appropriate range color mark; 512…Inappropriate range color mark; 515…Display information; 520…Recommended color mark; 530… Appropriate range selection screen; 600… Color conversion lookup table; CH0… Chart; CH1… Teaching chart; CH2… Test chart; IM0… Printed image; DS1… Data set; LA1… Label; ME0… Printing medium; P… Appropriateness index; P0, P11 to P14, P21 to P22, P31… Pattern column; PA0… Color mark; PA1… Teaching color mark; PA2… Test color mark; PA3… Solid area; PA4… Line area; PR1… Learning procedure; PR2… Maximum ink volume prediction procedure; PV0, PV1, PV2… Predicted value; Q1… Ink volume per unit area; Qm… Maximum ink volume; QO1… Ink volume order; ST1… Segmentation process; ST2… Predicted value acquisition process; ST3… Output process; ST4… Maximum ink volume determination process; TH1… Threshold.

Claims

1. An auxiliary device for assisting in determining a maximum ink quantity as an upper limit of ink quantity per unit area on a printing medium, and comprising: The holding unit holds multiple test images obtained by reading multiple test color marks with different amounts of ink per unit area on the printing medium. The processing unit is capable of executing a learned model that enables the computer to function in such a way that, based on a plurality of segmented test images obtained by segmenting each of the test images into a predetermined number, it obtains a predicted value representing the probability that the amount of ink per unit area of ​​each of the segmented test images is appropriate as the maximum amount of ink. The processing unit performs the following processing: The segmentation process involves dividing each of the test images into the predetermined number of segments to obtain multiple segmented test images. The predicted value acquisition process involves executing the learned model by taking each of the segmented test images as input, thereby obtaining the predicted value. The output processing, based on the amount of ink per unit area corresponding to each of the test images and a plurality of the predicted values, outputs predictive information indicating whether the amount of ink per unit area of ​​each of the test images is within or outside the appropriate range of the maximum ink amount.

2. The auxiliary device as described in claim 1, wherein, The processing unit performs the following processing in the output processing: For each of the test images, an appropriateness index is calculated by performing statistical processing on the multiple predicted values ​​obtained by executing the learned model, the appropriateness index representing the probability that the amount of ink per unit area corresponding to the test image is appropriate; The prediction information is generated based on the suitability index regarding the amount of ink per unit area; and Output the prediction information.

3. The auxiliary device as described in claim 2, wherein, In the output processing, the processing unit compares the calculated suitability index with a threshold, and determines that the amount of ink per unit area where the suitability index exceeds the threshold is within the appropriate range. The auxiliary device also includes an operation unit for accepting operations to change the threshold.

4. The auxiliary device as described in any one of claims 1 to 3, wherein, The processing unit enables the display unit to display multiple display color marks that correspond to the multiple test color marks respectively. The plurality of display color marks includes a plurality of appropriate range color marks whose corresponding ink amount per unit area is within the appropriate range, and a plurality of inappropriate range color marks that are not within the appropriate range. In the output processing, the processing unit causes the display unit to display the plurality of display color marks in such a way that it includes display information that distinguishes the plurality of appropriate range color marks from the plurality of inappropriate range color marks as the prediction information.

5. The auxiliary device as described in claim 4, wherein, It also includes an operation unit for handling operations on the plurality of display color bars displayed on the display unit. The processing unit also performs a maximum ink quantity setting process, which is to set the ink quantity per unit area corresponding to the operated display color mark to the maximum ink quantity when the operation unit receives an operation on any one of the plurality of display color marks.

6. The auxiliary device as described in claim 4, wherein, It also includes an operation unit for handling operations on the plurality of display color bars displayed on the display unit. The processing unit also performs a maximum ink quantity setting process, which is to set the ink quantity per unit area corresponding to the operated appropriate range color mark as the maximum ink quantity when the operation unit is prohibited from accepting operations on the plurality of inappropriate range color marks and the operation unit accepts an operation on any one of the plurality of appropriate range color marks.

7. The auxiliary device as described in any one of claims 1 to 3, wherein, In the output processing, the processing unit determines a recommended value for the maximum ink amount based on the ink amount per unit area corresponding to each of the test images and a plurality of the predicted values, and outputs recommendation information representing the recommended value in addition to outputting the predicted information.

8. An auxiliary method comprising using a computer to assist in determining a maximum ink quantity as an upper limit of ink quantity per unit area on a printing medium, wherein, The computer is capable of executing a learned model that enables it to function in such a way that it obtains a predicted probability, based on a plurality of segmented test images, that the amount of ink per unit area of ​​each of the segmented test images is appropriate as the maximum amount of ink. The plurality of test images are images obtained by segmenting each of a plurality of test images obtained by reading multiple test color marks with different amounts of ink per unit area on the printing medium into a predetermined number of segments. The auxiliary method includes: The segmentation process involves dividing each of the test images into the predetermined number of segments, thereby obtaining multiple segmented test images. The prediction value acquisition process involves executing the learned model by taking each of the segmented test images as input, thereby obtaining the prediction value. The output process, based on the amount of ink per unit area corresponding to each of the test images and a plurality of the predicted values, outputs predictive information indicating whether the amount of ink per unit area of ​​each of the test images is within an appropriate range or outside the appropriate range of the maximum ink amount.

9. A learned model generation apparatus for assisting in determining a maximum ink amount as an upper limit of ink amount per unit area on a printing medium, and comprising: The holding unit holds multiple teaching images obtained by reading multiple test color marks with different amounts of ink per unit area on the printing medium; The processing unit generates a learned model that enables the computer to function in such a way that, through machine learning based on the relationship between labels and a plurality of segmented teaching images obtained by segmenting each of the teaching images into a predetermined number of segmented teaching images, a predicted value representing the probability that the ink amount per unit area of ​​each of the segmented test images is appropriate as the maximum ink amount is obtained by segmenting each of a plurality of test images obtained by reading a plurality of test color marks with different ink amounts per unit area on the printing medium into the plurality of segmented test images, wherein... The label indicates whether the amount of ink per unit area of ​​each of the teaching color marks is appropriate, exceeds an appropriate range, or is below an appropriate range as the maximum amount of ink.

Images