Print control method, printing apparatus, and program product

Abstract

The present application provides a printing control method, a printing device and a program product. The printing control method includes the following processes: calculating a feature quantity of each of a plurality of object pixels included in a captured image obtained by capturing a medium on which a pattern is formed; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into any one of a pattern cluster and a background cluster according to a predetermined condition; based on the result of being classified into the pattern cluster or the background cluster, obtaining a pattern region for the pattern in the captured image by comparing the captured image with a pattern image related to the pattern formed on the medium; correcting a print image to be printed overlapping the pattern formed on the medium into a corrected print image based on a shape of the pattern region; and printing the corrected print image on the medium.

Description

Technical Field

[0001] This invention relates to a printing control method, printing apparatus, and program. Background Technology

[0002] For example, Patent Document 1 discloses a printing control method that corrects a printed image that overlaps with a pattern by comparing a photographed image obtained by photographing a medium with a pattern and a pattern image related to the pattern.

[0003] Patent Document 1: Japanese Patent Application Publication No. 2021-84359 However, in such printing control methods, it is not easy to accurately identify the patterns formed on the medium. Therefore, it is desirable to improve the correction accuracy of the printed image that overlaps with the pattern. Summary of the Invention

[0004] The printing control method for solving the above-mentioned problem includes the following processing: acquiring a photographic image obtained by photographing a medium with a pattern; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region for the pattern in the photographic image by comparing the photographic image with a pattern image related to the pattern formed on the medium based on the result of whether it is classified into the pattern cluster or the background cluster; correcting a printing image that overlaps with the pattern formed on the medium and prints it into a corrected printing image based on the shape of the pattern region; and printing the corrected printing image on the medium.

[0005] The printing apparatus for solving the above-mentioned problem performs the following processing: acquiring a photographic image obtained by photographing a medium on which a pattern is formed; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region for the pattern in the photographic image by comparing the photographic image with a pattern image related to the pattern formed on the medium based on the result of whether it is classified into the pattern cluster or the background cluster; correcting a printing image that overlaps with the pattern formed on the medium and printing it into a corrected printing image based on the shape of the pattern region; and printing the corrected printing image on the medium.

[0006] The procedure for solving the above-mentioned problem involves at least one computer performing the following processing: acquiring a photographic image obtained by photographing a medium with a pattern; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region in the photographic image by comparing the photographic image with a pattern image related to the pattern formed on the medium based on the result of being classified into the pattern cluster or the background cluster; correcting a printed image that overlaps with the pattern formed on the medium and prints it into a corrected printed image based on the shape of the pattern region; and printing the corrected printed image on the medium. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating the printing apparatus of the first embodiment.

[0008] Figure 2 This is a block diagram illustrating the electrical structure of the printing apparatus according to the first embodiment.

[0009] Figure 3 This is a schematic diagram showing the pattern of the medium in the first embodiment.

[0010] Figure 4 This is a schematic diagram illustrating the pattern image data of the first embodiment.

[0011] Figure 5 This is a schematic diagram illustrating the printed image data of the first embodiment.

[0012] Figure 6 This is a schematic diagram illustrating various pixels in the first embodiment.

[0013] Figure 7 This is a schematic diagram illustrating the clustering of feature quantities of the captured images according to the first embodiment.

[0014] Figure 8 This is a schematic diagram showing the patterned area of ​​the first embodiment.

[0015] Figure 9 This is a flowchart illustrating the pattern learning control process of the first embodiment.

[0016] Figure 10 This is a flowchart illustrating the pattern area learning process in the first embodiment.

[0017] Figure 11 This is a flowchart illustrating the printing control process of the first embodiment.

[0018] Figure 12 This is a flowchart illustrating the process of acquiring the pattern area in the first embodiment.

[0019] Figure 13 This is a flowchart illustrating the pattern area learning process in the second embodiment.

[0020] Figure 14 This is a flowchart illustrating the process of acquiring the pattern area in the second embodiment.

[0021] Figure 15 This is a flowchart illustrating the process of acquiring the pattern area in the third embodiment. Detailed Implementation

[0022] First Implementation Method The following describes one embodiment of the printing control method, printing apparatus, and program. In this description, with the printing apparatus positioned on a horizontal plane, the axis intersecting the horizontal plane is designated as the Z-axis, the axis intersecting the Z-axis is designated as the X-axis, and the axis intersecting both the X-axis and Z-axis is designated as the Y-axis. One direction along the X-axis is designated as the first width direction X1, and the other direction along the X-axis is designated as the second width direction X2. One direction along the Y-axis is designated as the forward direction Y1, and the other direction along the Y-axis is designated as the rearward direction Y2. The upward direction along the Z-axis is designated as the upward direction Z1, and the downward direction along the Z-axis is designated as the downward direction Z2.

[0023] Structure of printing apparatus 11 like Figure 1 As shown, the printing apparatus 11 performs printing on the medium 99. The printing apparatus 11 may also be an inkjet printer that performs printing on the medium 99 by ejecting ink onto the medium 99. The medium 99 has a first side 99A and a second side 99B. The second side 99B is the side opposite to the first side 99A.

[0024] Medium 99 can also be in the form of a strip. Medium 99 can also be unwound from a roll of material. Medium 99 is not limited to strips; it can also be in the form of a single sheet. Medium 99 can also be fabric.

[0025] The printing apparatus 11 includes a printing unit 12. The printing unit 12 may also be disposed above the medium 99 at a position Z1. The printing unit 12 may also be configured to perform printing on the medium 99 by spraying ink onto the medium 99. The printing unit 12 performs printing on the transported medium 99.

[0026] The printing unit 12 may also include a head. Although the printing unit 12 may be a serial head, it may also be a line head. A serial head is a head that scans along the X-axis relative to the medium 99. A line head is a head that records data along the entire X-axis.

[0027] The printing apparatus 11 includes a transport unit 13. The transport unit 13 is configured to transport the medium 99 in the transport direction D. The transport direction D may also be along the Y-axis. The transport direction D may also be directed forward Y1.

[0028] The conveying section 13 may also include a first roller 21 and a second roller 22. The first roller 21 is located upstream of the second roller 22 in the conveying direction D. The first roller 21 is located upstream of the printing section 12 in the conveying direction D. The second roller 22 is located downstream of the printing section 12 in the conveying direction D.

[0029] The conveying unit 13 includes a conveyor belt 23. The conveyor belt 23 is configured to convey the medium 99 while supporting it. The conveyor belt 23 is wound around a first roller 21 and a second roller 22. When the first roller 21 rotates, the conveyor belt 23 moves along the first roller 21 and the second roller 22.

[0030] The conveying unit 13 includes a roller drive unit 24. The roller drive unit 24 is a drive source for rotating multiple rollers. The roller drive unit 24 may also be an electric motor. Although the roller drive unit 24 is connected to the first roller 21, it may also be connected to the second roller 22, or it may be connected to both the first roller 21 and the second roller 22.

[0031] The transport unit 13 transports the medium 99 such that the surface opposite to the printing unit 12 becomes the printing surface. Thus, when the medium 99 is arranged such that the first surface 99A is opposite to the printing unit 12, the first surface 99A becomes the printing surface.

[0032] The printing apparatus 11 includes an imaging unit 14. The imaging unit 14 is positioned opposite the medium 99. The imaging unit 14 may also be positioned above the medium 99 at a position Z1. The imaging unit 14 is located upstream of the printing unit 12 in the transport direction D. The imaging unit 14 is configured to photograph the medium 99. When transporting the medium 99 with its first surface 99A facing the medium, the imaging unit 14 photographs the first surface 99A of the medium 99.

[0033] The printing apparatus 11 includes a control unit 30. The control unit 30 provides comprehensive control over the printing apparatus 11. The control unit 30 controls various actions performed within the printing apparatus 11. The control unit 30 controls the printing unit 12, the transport unit 13, and the imaging unit 14.

[0034] like Figure 2 As shown, the control unit 30 may also be composed of more than one computer. The control unit 30 may also have more than one processor 31 and more than one memory 32. The processor 31 may also be a CPU (Central Processing Unit). The processor 31 is configured to execute processing based on the program stored in the memory 32.

[0035] The control unit 30 can be configured as a circuit including the following components: α: one or more processors that execute various processes according to a computer program; β: one or more dedicated hardware circuits that execute at least a portion of the various processes; or γ: a combination thereof. The hardware circuit is, for example, an application-specific integrated circuit.

[0036] Memory 32 includes a computer-readable medium accessible by a general-purpose or special-purpose computer. Memory 32 stores program code or instructions configured to cause processor 31 to perform processing. Memory 32 is a non-transitory computer-readable medium for storing programs, but may also include temporary computer-readable media.

[0037] Pattern image data D1 is stored in memory 32. Pattern image data D1 is... Figure 4 The image data of pattern P1 is shown. Pattern image data D1 represents... Figure 3 as well as Figure 4 The shape of pattern P1 is shown. Pattern P1 is set on the first surface 99A. Pattern image data D1 can also be CAD data representing pattern P1 created by software.

[0038] Printed image data D2 is stored in memory 32. Printed image data D2 is... Figure 5 The image data for the printed image P2 shown is shown. Printed image data D2 represents the shape, color, etc., of the printed image P2. The printed image P2 is printed on the first surface 99A. The printed image P2 is printed on the first surface 99A in a manner corresponding to the pattern P1. Printed image data D2 can also be CAD data representing the printed image P2 created by software.

[0039] Image data D3 is stored in memory 32. Image data D3 is image data of the captured image. The captured image is captured by the imaging unit 14. The captured image may also be an image obtained by photographing the medium 99 from the first surface 99A. In particular, the captured image may also be an image obtained by photographing the pattern P1 provided on the first surface 99A. Image data D3 may also include image data for learning and image data for printing.

[0040] Feature data D4 is stored in memory 32. Feature data D4 represents the feature values ​​of each pixel in the captured image. The feature values ​​are used to identify whether each pixel in the captured image is a patterned region corresponding to pattern P1. The feature values ​​are also used to identify whether each pixel in the captured image is a patterned region corresponding to pattern P1 or a background region that is not a patterned region. Feature data D4 may also contain data for learning and data for printing.

[0041] The memory 32 stores pattern region data D5. Pattern region data D5 is data representing whether each pixel in the captured image is a pattern region, based on feature data D4. Pattern region data D5 indicates whether each pixel in the captured image is a pattern region or background region. Pattern region data D5 may also contain learning data and printing data. Learning pattern region data D5 is based on learning feature data D4 calculated from learning captured image data D3. Printing pattern region data D5 is based on printing feature data D4 calculated from printing captured image data D3.

[0042] The categories of media 99 and the printing of media 99 like Figure 3 As shown, medium 99 can also be, for example, jacquard fabric. Jacquard fabric is a fabric in which the pattern P1 is formed by the weaving of fibers such as yarns. The pattern P1 can also be formed by the unevenness produced by the weaving.

[0043] Pattern P1 is provided on the first surface 99A of medium 99. Pattern P1 is formed by graphics, patterns, etc., that are pre-applied to medium 99 to give it an aesthetic design. Pattern P1 is periodically provided on medium 99. Pattern P1 can also be periodically provided on medium 99 in a manner that repeats in the transport direction D.

[0044] like Figure 4 As shown, the pattern image data D1 is the basic data of the pattern P1 set on the first surface 99A. The pattern image data D1 can also be set to be able to identify the pattern area where the pattern P1 is set.

[0045] like Figure 5 As shown, printed image P2 represents the shape, color, etc. of an image printed on the first surface 99A of the medium 99 having pattern P1. Printed image P2 is an image corresponding to the pattern P1 of the first surface 99A.

[0046] like Figure 6As shown, the captured image contains multiple pixels. The captured image contains multiple pixels corresponding to medium 99. The captured image may also not contain multiple pixels that do not correspond to medium 99. Feature data D4 is calculated for each of the multiple pixels in the captured image that correspond to medium 99.

[0047] The pixel that becomes the object of the calculation of feature data D4 is set as object pixel PX0. Although multiple pixels in the pixels constituting the captured image that correspond to medium 99 are each set as object pixel PX0, it is also possible to set all pixels in the pixels constituting the captured image that correspond to medium 99 as object pixel PX0.

[0048] The feature data D4 of the object pixel PX0 can also be calculated based on multiple pixels within a predetermined range centered on the object pixel PX0. The feature data D4 is assigned to each of the multiple object pixels PX0 in the captured image data D3.

[0049] A feature quantity can include at least one type of feature. A feature quantity can also include multiple types of features. A feature quantity can include the grayscale value of the object pixel PX0. A feature quantity can also include the average grayscale value of the reference pixel PX1. A feature quantity can also include the variance of the grayscale value of the reference pixel PX1. A feature quantity can also include the skewness of the grayscale value of the reference pixel PX1. A feature quantity can also include the kurtosis of the grayscale value of the reference pixel PX1.

[0050] Reference pixel PX1 is the pixel contained in reference region PR1. Reference region PR1 is a region centered on object pixel PX0 with a predetermined number of pixels. Reference pixel PX1 contains object pixel PX0. For example, reference region PR1 can also be a region centered on object pixel PX0 with a vertical dimension of seven pixels and a horizontal dimension of seven pixels. In this case, the number of reference pixels PX1 becomes 49.

[0051] The feature quantity can also be the edge intensity of the object pixel PX0. The feature quantity can also include the average edge intensity of the reference region PR1. The feature quantity can also include the variance of the edge intensity of the reference region PR1. The feature quantity can also include the skewness of the edge intensity of the reference region PR1. The feature quantity can also include the kurtosis of the edge intensity of the reference region PR1.

[0052] The edge intensity of object pixel PX0 can also be calculated based on multiple edge pixels PX2 of the edge region PR2 centered on object pixel PX0. For example, the edge region PR2 can also be a region three pixels vertically and three pixels horizontally centered on object pixel PX0. The edge intensity of object pixel PX0 can also be calculated based on the grayscale difference between object pixel PX0 and its adjacent edge pixels PX2. The edge intensity of object pixel PX0 can also be calculated using the square root of the sum of squares relative to both the vertical and horizontal directions using a Sobel filter.

[0053] Thus, the feature quantities can also include feature quantities related to the object pixel PX0. The feature quantities related to the object pixel PX0 can also include the grayscale value and edge intensity of the object pixel PX0. The feature quantities of the object pixel PX0 can also include feature quantities related to the reference pixel PX1. The feature quantities related to the reference pixel PX1 can also include the grayscale value and edge intensity of the reference pixel PX1.

[0054] Based on feature data D4, the determination of whether a region is a pattern region is performed for each of the multiple object pixels PX0. In this case, based on feature data D4, each of the multiple object pixels PX0 is classified into any one of the multiple feature clusters.

[0055] like Figure 7 As shown, to give a specific example, for each of the multiple object pixels PX0, the pixel is classified into multiple cluster regions corresponding to multiple feature clusters. These multiple cluster regions are displayed on a display device (not shown). For example, the multiple cluster regions may also include a first cluster region R11, a second cluster region R12, a third cluster region R13, and a fourth cluster region R14.

[0056] For each of the multiple clustering regions, a correspondence is established between the cluster and the background cluster. The patterned cluster is the cluster classified as a region with a pattern P1. The background cluster is the cluster classified as a region without a pattern P1.

[0057] For each of the multiple feature clusters classified based on feature data D4, it is possible to identify whether the cluster is a pattern cluster or a background cluster. Therefore, pattern region data D5 is generated based on feature data D4. Pattern region data D5 is assigned to each of the multiple object pixels PX0 in the captured image data D3.

[0058] like Figure 8As shown, for example, the first cluster region R11, the second cluster region R12, and the third cluster region R13 can also be classified as the first region R1 corresponding to the pattern cluster. The first region R1 is equivalent to the pattern region. The fourth cluster region R14 can also be classified as the second region R2 corresponding to the background cluster. The second region R2 is equivalent to the background region.

[0059] During the preparation phase, pattern region data D5 is learned based on feature data D4 generated from the image data D3 used for learning. This preparation phase is earlier than the printing phase, which involves printing on the medium 99.

[0060] During the printing stage, the printed image P2 is corrected based on the feature data D4 generated from the printed image data D3 and the pattern area data D5 learned during the preparation stage, in a manner corresponding to the pattern P1. The corrected printed image P2 is then printed on the medium 99.

[0061] Pattern learning and control processing Reference Figure 9 as well as Figure 10 The pattern learning control process will now be explained. During the preparation phase, the pattern learning control process is executed according to user instructions. During the preparation phase, the medium 99 is precisely positioned relative to the conveyor belt 23. This allows the acquisition of pattern area data D5 for learning from the pattern P1 formed on the medium 99. Subsequent processing will be described as processing executed by the control unit 30, but it can also be processing executed by the processor 31.

[0062] like Figure 9 As shown, in step S11, the control unit 30 performs a pattern image data acquisition process. In this process, the control unit 30 reads the pattern image data D1 from the memory 32. Thus, the control unit 30 acquires the pattern image data D1.

[0063] In step S12, the control unit 30 performs an image capture process. In this process, the control unit 30 controls the transport unit 13 to transport the medium 99. The control unit 30 controls the imaging unit 14 to capture an image of the medium 99 while the transport of the medium 99 is stopped. The control unit 30 stores the image data D3 captured by the imaging unit 14 for learning purposes in the memory 32. Thus, the control unit 30 obtains an image captured by photographing the medium 99 with the pattern P1 as a learning image.

[0064] Pattern area learning and processing In step S13, the control unit 30 performs pattern area learning processing. Specifically, although it will refer to... Figure 10 As will be described later, the control unit 30 learns the pattern area based on the image data D3 used for learning.

[0065] like Figure 10 As shown, in the pattern region learning process, in step S21, the control unit 30 performs edge intensity calculation processing. In this processing, the control unit 30 calculates the edge intensity of each of the multiple object pixels PX0 as an object, based on the image data D3 used for learning. The control unit 30 stores data representing the size of the edge region PR2 used when calculating the edge intensity in the memory 32.

[0066] In step S22, the control unit 30 performs feature calculation processing. In this processing, the control unit 30 calculates the feature quantities of each of the multiple object pixels PX0 contained in the captured image based on the learned image data D3. Specifically, the control unit 30 calculates the feature quantities related to the reference pixel PX1 based on the learned image data D3.

[0067] The control unit 30 stores the feature data D4, which represents the calculated feature values, in the memory 32 in such a way that each of the multiple object pixels PX0 corresponds to one of them. In this way, the control unit 30 assigns feature values ​​to each of the multiple object pixels PX0 in the image data D3 used for learning.

[0068] In step S23, the control unit 30 performs feature normalization processing. In this processing, the control unit 30 performs normalization of multiple types of feature quantities. Specifically, the control unit 30 calculates the normalization parameters required for the normalization of the multiple types of feature quantities. The normalization parameters are parameters that take all of the multiple object pixels PX0 as objects, make the average 0, and make the standard deviation 1. The control unit 30 stores the normalization parameters in the memory 32.

[0069] The control unit 30 takes all of the multiple object pixels PX0 as its object and performs feature normalization based on feature data D4 and normalization parameters for each of the multiple types of feature quantities. Thus, the control unit 30 normalizes the feature quantities for each of the multiple object pixels PX0.

[0070] In step S24, the control unit 30 performs feature quantity clustering classification processing. In this processing, the control unit 30 classifies each of the plurality of object pixels PX0 into any one of the plurality of feature quantity clusters based on feature quantities. Specifically, for each of the plurality of object pixels PX0, the control unit 30 classifies the pixel into any one of the plurality of feature quantity clusters based on the normalized feature quantities.

[0071] In detail, the control unit 30 classifies each of the multiple object pixels PX0 into a feature quantity cluster by making the feature quantities, which have been normalized for each of the multiple feature quantity clusters, more similar. The control unit 30 stores the result of classifying the pixel into the feature quantity cluster in a manner corresponding to each of the multiple object pixels PX0 in the memory 32.

[0072] The control unit 30 can also use the K-means method to classify each of the multiple object pixels PX0 into a feature quantity cluster. The value of K used in the K-means method is a predetermined value, for example, it can be any one of 8 to 10. When using the K-means method, the control unit 30 can also store the center value of each feature quantity cluster in the memory 32 as a result of classification into feature quantity clusters.

[0073] In step S25, the control unit 30 performs classification result display processing. In this processing, the control unit 30 causes a display device (not shown) to display the captured image for learning purposes in a manner capable of recognizing the feature clusters obtained by classifying each of the multiple object pixels PX0. The multiple object pixels PX0 can also be displayed individually using colors that can recognize the feature clusters.

[0074] In step S26, the control unit 30 performs a pattern clustering acquisition process. In this process, the control unit 30 acquires data indicating whether the clustering of multiple feature quantities is a pattern clustering or a background clustering, based on instructions from a user input into an input unit (not shown).

[0075] In step S27, the control unit 30 performs pattern clustering setting processing. In this processing, the control unit 30 stores data indicating whether it is pattern clustering or background clustering in the memory 32 in a manner that corresponds to each of the multiple feature quantity clusters.

[0076] Thus, the control unit 30 classifies multiple feature quantity clusters into either the pattern cluster or the background cluster based on predetermined conditions. The predetermined conditions may also be conditions that set at least one of the multiple feature quantity clusters as the pattern cluster according to user instructions.

[0077] In step S28, the control unit 30 performs a learning pattern region setting process. In this process, the control unit 30 stores the object pixels PX0 that have established a correspondence with the feature quantity clusters corresponding to the pattern clusters in the memory 32 as pattern regions.

[0078] Therefore, the control unit 30 assigns each of the multiple pixels contained in the captured image for learning to either a pattern cluster or a background cluster. Thus, the captured image becomes a pattern image after the multiple pixels have been assigned to either a pattern cluster or a background cluster.

[0079] The control unit 30 acquires the pattern region data D5 for learning. Thus, the control unit 30 acquires a learning result indicating whether each of the multiple object pixels PX0 in the pattern region data D5 is a pattern region or a background region.

[0080] The control unit 30 registers the data of the pattern region for representation learning in the memory 32 in a manner corresponding to each of the plurality of object pixels PX0. Thus, the control unit 30 is able to identify whether the region representing the medium 99 is a pattern region or a background region based on the learning results.

[0081] like Figure 9 As shown, in step S14, the control unit 30 performs a learning result display process. In this process, the control unit 30 causes a display unit (not shown) to display the pattern image data D1 and the pattern area data D5.

[0082] In step S15, the control unit 30 performs learning result correction processing. In this processing, the control unit 30 corrects the pattern area data D5 for learning based on instructions input by a user through an input unit (not shown). This allows the pattern area data D5 for learning to be set to pattern area data with less deformation.

[0083] Printing control processing Reference Figure 11 as well as Figure 12 The printing control process is described below. The printing control process is executed according to user instructions during the printing stage. During the printing stage, printing is performed based on the pattern area data D5 prepared during the preparation stage.

[0084] like Figure 11 As shown, in step S41, the control unit 30 communicates with... Figure 9 Step S12 is performed in the same manner for the imaging process. In this process, the control unit 30 controls the transport unit 13 to transport the medium 99. The control unit 30 controls the imaging unit 14 to image the medium 99 while the transport of the medium 99 is stopped. The control unit 30 stores the image data D3 for printing captured by the imaging unit 14 in the memory 32. Thus, the control unit 30 obtains the image obtained by imaging the medium 99 with the pattern P1 as the image for printing.

[0085] Processing of the patterned area during photography In step S42, the control unit 30 performs a process to acquire the patterned area. Specifically, although it will refer to... Figure 12As will be described later, the control unit 30 obtains the pattern area in the printed image based on the printed image data D3 obtained in step S41. The pattern area is the pattern area based on the printed image data D3. In other words, the pattern area is the printed pattern area.

[0086] like Figure 12 As shown, in the process of acquiring the pattern area during imaging, in step S51, the control unit 30... Figure 10 Step S21 is performed in the same manner for edge intensity calculation. In this process, the control unit 30 reads data from the memory 32 representing the size of the edge region PR2 used when calculating the edge intensity. Based on the image data D3 for printing, the control unit 30 calculates the edge intensity of each of the plurality of object pixels PX0 as an object.

[0087] Specifically, the control unit 30 calculates the edge intensity of the object pixels PX0 by treating each object pixel PX0 as an object based on the data representing the size of the edge region PR2 stored in the memory 32 in step S21.

[0088] In step S52, the control unit 30 connects with... Figure 10 Step S22 is performed in the same manner as the feature calculation process. In this process, the control unit 30 calculates the feature quantity for each of the plurality of object pixels PX0 contained in the captured image based on the image data D3 for printing. In particular, the control unit 30 calculates the feature quantity related to the reference pixel PX1 based on the image data D3 for printing.

[0089] The control unit 30 stores the feature data D4, representing the calculated feature values, in the memory 32 in a manner corresponding to each of the plurality of object pixels PX0. Thus, the control unit 30 assigns feature values ​​to each of the plurality of object pixels PX0 in the image data D3 for printing.

[0090] In step S53, the control unit 30 communicates with... Figure 10 Step S23 is performed in the same manner as feature normalization. In this process, the control unit 30 performs normalization of multiple types of feature quantities. Specifically, the control unit 30 reads the normalization parameters required for the normalization of the multiple types of feature quantities from the memory 32. The normalization parameters are... Figure 10 The parameters stored in memory 32 in step S23 are then used by the control unit 30 to normalize the feature values ​​for each of the plurality of object pixels PX0.

[0091] In step S54, the control unit 30 communicates with... Figure 10 Step S24 performs the feature quantity clustering classification process in the same manner. In this process, the control unit 30 classifies each of the multiple object pixels PX0 into any one of the multiple feature quantity clusters based on the feature quantity. Specifically, for each of the multiple object pixels PX0, the control unit 30 classifies the pixel into any one of the multiple feature quantity clusters based on the normalized feature quantity.

[0092] The control unit 30 reads the classification results into feature quantity clusters from the memory 32. Based on the classification results into feature quantity clusters, the control unit 30 classifies each of the multiple object pixels PX0 into a feature quantity cluster in a manner that makes the normalized feature quantities for each of the multiple feature quantity clusters closer to each other. When using the K-means method, the control unit 30 classifies each of the multiple object pixels PX0 into a feature quantity cluster in a manner that makes the normalized feature quantities for each of the multiple feature quantity clusters closer to the center value of each feature quantity cluster.

[0093] In step S55, the control unit 30 performs pattern clustering assignment processing. In this processing, the control unit 30 retrieves the representation from the memory 32 for each of the multiple feature quantity clusters. Figure 10 The control unit 30 reads the data for either the pattern cluster or the background cluster as set in step S27. Then, it assigns each of the multiple feature quantity clusters to either the pattern cluster or the background cluster.

[0094] Thus, the control unit 30, based on predetermined conditions, categorizes multiple feature quantities into either the pattern cluster or the background cluster. The predetermined conditions can also be conditions corresponding to the pattern region data D5 used for learning.

[0095] In step S56, the control unit 30 performs a pattern region setting process. In this process, the control unit 30 stores the object pixels PX0 that have established a correspondence with the feature quantity clusters corresponding to the pattern clusters in the memory 32 as pattern regions.

[0096] Therefore, the control unit 30 sets whether each of the multiple object pixels PX0 contained in the captured image belongs to the pattern cluster or the background cluster. In this way, the captured image becomes an image after allocating each of the multiple pixels to either the pattern cluster or the background cluster.

[0097] like Figure 11As shown, in step S43, the control unit 30 performs a pattern region acquisition process. In this process, the control unit 30 reads the pattern region data D5 for learning from the memory 32. Thus, the control unit 30 acquires the pattern region for learning.

[0098] In step S44, the control unit 30 performs a pattern area comparison process. In this process, the control unit 30 compares the pattern area data D5 for printing with the pattern area data D5 for learning. Specifically, the control unit 30 calculates the deviation between the pattern area for printing and the pattern area for learning.

[0099] Thus, the control unit 30 compares the captured image with the pattern image related to the pattern P1 formed on the medium 99 based on the classification result of multiple feature clusters, determining whether the cluster is classified into the pattern cluster or the background cluster. Based on the comparison result, the control unit 30 obtains the pattern region in the captured image corresponding to the pattern P1.

[0100] In step S45, the control unit 30 performs a print image data correction process. In this process, the control unit 30 corrects the print image data D2 based on a comparison between the print pattern area data D5 and the learning pattern area data D5. Specifically, the control unit 30 corrects the print image data D2 in a manner that makes the print pattern area data D5 consistent with the learning pattern area data D5. As a result, the control unit 30 generates corrected print image data after correcting the print image data D2.

[0101] Specifically, the control unit 30 corrects the printed image P2, which overlaps with the pattern P1 formed on the medium 99, into a corrected printed image based on a comparison between the shape of the printed pattern area and the shape of the learning pattern area. In other words, the control unit 30 corrects the printed image P2, which overlaps with the pattern P1 formed on the medium 99, into a corrected printed image based on the shape of the printed pattern area.

[0102] In step S46, the control unit 30 performs a printing process. In this process, the control unit 30 controls the printing unit 12 to print the corrected printing image on the medium 99 based on the corrected printing image data obtained in step S45. Thus, the control unit 30 prints the printing image P2 on the medium 99 in a manner that overlaps with the pattern P1 formed on the medium 99.

[0103] In step S47, the control unit 30 determines whether printing has ended. If the control unit 30 determines that printing has not ended, it transfers the process to step S41. If the control unit 30 determines that printing has ended, it terminates the printing control process. Thus, the control unit 30 repeatedly executes steps S41 to S46 until printing ends.

[0104] The function and effects of the first implementation method The function and effects of the first embodiment will be explained.

[0105] (1-1) The control unit 30 classifies the multiple object pixels PX0 contained in the captured image obtained by capturing an image of the medium 99 on which the pattern P1 is formed into any one of multiple feature quantity clusters. The control unit 30 classifies the multiple feature quantity clusters into either the pattern cluster or the background cluster. Based on the result of whether it is classified into the pattern cluster or the background cluster, the control unit 30 obtains the pattern region of the pattern P1 in the captured image by comparing the captured image with the pattern image. The control unit 30 corrects the printed image P2, which is printed overlapping the pattern P1 formed on the medium 99, based on the shape of the pattern region. According to this structure, the pixel can be classified into any one of multiple feature quantity clusters based on the feature quantity of each of the multiple object pixels PX0, and the multiple feature quantity clusters can be classified into either the pattern cluster or the background cluster. Thus, the pattern P1 formed on the medium 99 can be identified with good accuracy. Therefore, the correction accuracy of the printed image P2, which is printed overlapping the pattern P1, can be improved. Thus, by improving the correction accuracy of the printed image P2, the accuracy of printing the printed image P2 corresponding to the pattern P1 can be improved even when there is stretching or deformation on the medium 99.

[0106] In particular, when the medium 99 is a fabric in which the pattern P1 is formed by weaving, there is a tendency to obtain the edges of the fibers such as yarns relative to the weaving direction, but not the edges representing the outline and shape of the pattern P1. In such a medium 99, a further effect can be achieved.

[0107] (1-2) The control unit 30 calculates the feature quantities associated with the reference pixel PX1 contained in the reference region PR1 centered on each of the plurality of object pixels PX0, and uses these as feature quantities for each of the plurality of object pixels PX0. According to this structure, not only based on each of the plurality of object pixels PX0, but also based on the feature quantities associated with the reference pixel PX1 contained in the reference region PR1 centered on each of the plurality of object pixels PX0, the pixel can be classified into any one of the plurality of feature quantity clusters. Therefore, the pattern P1 formed on the medium 99 can be identified with good accuracy. Thus, the correction accuracy of the printed image P2, which is printed overlapping the pattern P1, can be improved.

[0108] (1-3) The feature quantity associated with the reference pixel PX1 includes at least one of the gray value of the reference pixel PX1 and the edge intensity of the reference pixel PX1. According to this structure, the pixel can be classified into any one of multiple feature quantity clusters based on at least one of the gray value and the edge intensity of the reference pixel PX1. As a result, the pattern P1 formed on the medium 99 can be identified with good accuracy. Therefore, the correction accuracy of the printed image P2, which is printed overlapping the pattern P1, can be improved.

[0109] (1-4) The control unit 30 normalizes the feature values ​​for each of the multiple object pixels PX0. Based on the normalized feature values, the control unit 30 classifies the pixel into any one of the multiple feature value clusters. According to this structure, the feature values ​​for each of the multiple object pixels PX0 can be normalized, and the pixel can be classified into any one of the multiple feature value clusters based on the normalized feature values. By normalizing the feature values, the uneven influence of each feature value can be mitigated. As a result, the pattern P1 formed on the medium 99 can be identified with good accuracy. Therefore, the correction accuracy of the printed image P2, which is printed overlapping the pattern P1, can be improved.

[0110] (1-5) The pattern image is an image in which each of the multiple pixels contained in the pattern image is assigned to either a pattern cluster or a background cluster. Based on this structure, by comparing the pattern image and the captured image, the pattern region for the pattern P1 in the captured image can be obtained, where the pattern image is an image in which each of the multiple pixels contained in the pattern image is assigned to either a pattern cluster or a background cluster. Therefore, the pattern P1 formed on the medium 99 can be identified with good accuracy. Consequently, the correction accuracy of the printed image P2, which is printed overlapping with the pattern P1, can be improved.

[0111] Second Implementation Method Next, the second embodiment will be described. In the following description, descriptions of structures that are the same as those in the previously described embodiments will be omitted or simplified, and descriptions of structures that are different from those in the previously described embodiments will be provided.

[0112] In the second embodiment, the control unit 30 may also classify each of the plurality of object pixels PX0 into any one of the plurality of feature quantity clusters based on a plurality of feature quantities of the same available range. The plurality of feature quantities of the same available range may also be grayscale values ​​for each reference pixel PX1.

[0113] like Figure 13 As shown, in the pattern region learning process, the control unit 30 does not execute step S21. Therefore, the control unit 30 may not need to obtain the edge intensity of the object pixel PX0 as the feature quantity for each of the multiple object pixels PX0. In the pattern region learning process, the control unit 30 does not execute step S23. Therefore, the control unit 30 may not need to perform feature quantity normalization.

[0114] like Figure 14 As shown, in the process of acquiring the pattern area, the control unit 30 does not execute step S51. Therefore, the control unit 30 may not acquire the edge intensity of the target pixel PX0 as a feature quantity for each of the multiple target pixels PX0. In the process of acquiring the pattern area, the control unit 30 does not execute step S53. Therefore, the control unit 30 may not perform feature quantity normalization.

[0115] The function and effects of the second embodiment The function and effects of the second embodiment will be explained.

[0116] (2-1) In addition, the control unit 30 does not perform characteristic quantity normalization. Therefore, the control load can be reduced.

[0117] (2-2) The control unit 30 does not acquire the edge intensity of the object pixel PX0 as a feature quantity for each of the multiple object pixels PX0. Therefore, the control load can be reduced.

[0118] Third Implementation Method Next, the third embodiment will be described.

[0119] In the third embodiment, instead of calculating the feature values ​​for multiple non-object pixels in the captured image, the pixel can be classified into multiple feature value clusters based on the feature values ​​for multiple object pixels PX0.

[0120] The captured image contains multiple object pixels PX0 and multiple non-object pixels. The multiple non-object pixels differ from the multiple object pixels PX0. When the captured image contains multiple pixels corresponding to medium 99, the multiple pixels corresponding to medium 99 include multiple object pixels PX0 and multiple non-object pixels. The multiple object pixels PX0 are pixels extracted from the multiple pixels corresponding to medium 99 at a predetermined ratio. The multiple object pixels PX0 may also be fewer than the multiple non-object pixels.

[0121] Processing of the patterned area during photography like Figure 15 As shown, in the process of capturing the patterned area, in steps S52 to S56, the control unit 30 performs control over each of the plurality of object pixels PX0 contained in the captured image. In steps S52 to S56, the control unit 30 does not perform control over each of the plurality of non-object pixels contained in the captured image. When step S56 ends, the control unit 30 transfers the process to step S57.

[0122] In step S57, the control unit 30 performs pattern region interpolation processing. In this processing, the control unit 30 interpolates the classification of each of the multiple non-object pixels (PX0) into either the pattern cluster or the background cluster based on the result of whether each of the multiple object pixels PX0 is classified into the pattern cluster or the background cluster.

[0123] The control unit 30 may also perform nearest-neighbor interpolation on each of the multiple object pixels PX0, based on whether it is a pattern region or a background region, to interpolate whether it is a pattern cluster or a background cluster for each of the multiple non-object pixels. The control unit 30 may also correct the interpolation result by smoothing the interpolation result after nearest-neighbor interpolation.

[0124] In step S58, the control unit 30 performs interpolation pattern region registration processing. In this processing, the control unit 30 assigns either an interpolated pattern cluster or a background cluster to each of the plurality of non-object pixels among the plurality of pixels contained in the pattern image. Thus, the captured image becomes an image in which each of the plurality of object pixels PX0 and the plurality of non-object pixels has been assigned either a pattern cluster or a background cluster.

[0125] The role and effects of the third implementation method The function and effects of the third embodiment will be explained.

[0126] (3-1) The control unit 30 interpolates the classification of each of the multiple object pixels PX0 into either the pattern cluster or the background cluster based on the result of whether each of the multiple object pixels PX0 is classified into the pattern cluster or the background cluster. According to this structure, even without calculating the feature values ​​for each of the multiple non-object pixels, interpolation can be performed based on the feature values ​​for each of the multiple object pixels PX0 to determine whether the classification is pattern cluster or background cluster. Therefore, the control load can be reduced.

[0127] Change Example This embodiment can be modified and implemented as follows. This embodiment and the following modifications can be combined and implemented with each other within the scope of technical inconsistency.

[0128] • As a feature, any feature can be used. For example, a feature using a gray-level co-occurrence matrix (GLCM) can also be used. • The control unit 30 can also classify multiple object pixels PX0 into any one of the multiple feature clusters based on a single feature, rather than based on multiple types of feature values. The single feature value can also be the gray value for each of the multiple object pixels PX0.

[0129] The edge intensity of object pixel PX0 can also be calculated using multiple pixels contained in a region smaller than the reference region PR1. Alternatively, the edge intensity of object pixel PX0 can be calculated using multiple pixels contained in a region wider than the reference region PR1. Finally, the edge intensity of object pixel PX0 can be calculated using multiple reference pixels PX1 within the reference region PR1.

[0130] While the control unit 30 uses the K-means method when classifying each of the multiple object pixels PX0 into a feature cluster, it is not limited to this. The control unit 30 may also use the X-means method when classifying each of the multiple object pixels PX0 into a feature cluster.

[0131] The control unit 30 can also display supplementary information (not shown) on a display device other than color differentiation, in a manner capable of identifying feature quantity clusters that classify multiple object pixels PX0 individually. Alternatively, the control unit 30 can also display supplementary information on a display device (not shown) without color differentiation, in a manner capable of identifying feature quantity clusters that classify multiple object pixels PX0 individually. As supplementary information, the range of grayscale values ​​of the object pixel PX0 may be included for each of the multiple feature quantity clusters.

[0132] • The control unit 30 may assign each of the multiple feature quantity clusters to the pattern cluster, but not to the background cluster. The control unit 30 may also assign each of the multiple feature quantity clusters to the background cluster, but not to the pattern cluster. The control unit 30 may also assign each of the multiple feature quantity clusters to either the pattern cluster or the background cluster.

[0133] It is also possible to set multiple types of pattern clusters. These multiple types of pattern clusters may include a first pattern cluster and a second pattern cluster. The first pattern cluster and the second pattern cluster are different pattern clusters. Alternatively, the first pattern cluster may be a cluster containing the first pattern, and the second pattern cluster may be a cluster containing the second pattern. The control unit 30 may also assign each of the multiple feature quantity clusters to either the first pattern cluster or the second pattern cluster.

[0134] The control unit 30 may also assign a cluster other than the pattern cluster and the background cluster for each of the multiple feature quantity clusters. The control unit 30 may also assign a retained cluster other than the pattern cluster and the background cluster for each of the multiple feature quantity clusters. The retained cluster may also be a cluster that retains both the pattern cluster and the background cluster. The retained cluster may also be assigned as a non-object pixel, as in the third embodiment.

[0135] • The control unit 30 may also not acquire the pattern image data D1 during the pattern learning control processing. Alternatively, during the pattern learning control processing, the control unit 30 may assign each of the multiple object pixels PX0 to either a pattern region or a background region by comparing the pattern image data D1 with the acquired image data D3 used for learning. In this case, the control unit 30 may also assign regions in the acquired image data D3 that have a higher similarity to the pattern image data D1 as pattern regions. Thus, the predetermined condition may include whether the similarity to the pattern image data D1 is higher.

[0136] The pattern area data D5 for learning can also be generated without executing the pattern learning control process, and instead be pre-stored in the memory 32. The pattern area data D5 for learning can also be pre-assigned to the pattern image data D1 and stored in the memory 32. That is, the pattern image associated with the pattern P1 formed on the medium 99 can be either the learning image data D3 acquired during the preparation stage, or the pattern image data D1. The control unit 30 can also execute printing control processing instead of pattern learning control processing.

[0137] • The medium 99 is not limited to a roll. The medium 99 can also be paper, resin film or sheet, resin and metal composite film, laminated film, fabric, non-woven fabric, metal foil, metal film, ceramic sheet, and clothing, etc.

[0138] • The liquid can be arbitrarily selected as long as it is a substance that can be recorded on the medium 99 by adhering to it. For example, ink may include substances in which functional material particles composed of solid matter such as color materials or metal particles are dissolved, dispersed or mixed in a solvent, and may include various compositions such as water-based ink, oil-based ink, colloidal ink, and hot-melt ink.

[0139] As used in this specification, the expression "at least one" means one or more of the desired options. As an example, if there are two options, "at least one" as used in this specification means only one option or both of the two options. As another example, if there are three or more options, "at least one" as used in this specification means only one option or any combination of two or more options.

[0140] Postscript The following text describes the technical concepts and their effects as understood from the implementation methods and modifications described above. These technical concepts and their effects can be combined with each other within a technically compatible framework.

[0141] [1] The printing control method includes the following processing: acquiring a photographed image obtained by photographing a medium with a pattern; calculating a feature quantity for each of a plurality of object pixels contained in the photographed image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region for the pattern in the photographed image by comparing the photographed image with a pattern image related to the pattern formed on the medium based on the result of whether it is classified into the pattern cluster or the background cluster; correcting a printing image that overlaps with the pattern formed on the medium and prints it into a corrected printing image based on the shape of the pattern region; and printing the corrected printing image on the medium.

[0142] According to this structure, a pixel can be classified into any one of multiple feature quantity clusters based on the feature quantity for each of multiple object pixels, and these multiple feature quantity clusters can be classified into either a pattern cluster or a background cluster. Therefore, patterns formed on a medium can be identified with good accuracy. Consequently, the correction accuracy of printed images printed while overlapping with patterns can be improved.

[0143] [2] In the above printing control method, it can also be set that the calculation of the feature quantity for each of the plurality of object pixels includes the following process: calculating the feature quantity related to the reference pixel contained in the reference region centered on each of the plurality of object pixels.

[0144] According to this structure, a pixel can be classified into any of several feature quantity clusters not only based on each of the multiple object pixels, but also based on the feature quantities associated with reference pixels contained in a reference region centered on each of the multiple object pixels. This allows for the accurate identification of patterns formed on a medium. Therefore, the correction accuracy of printed images that overlap with patterns can be improved.

[0145] [3] In the above printing control method, it can also be set that the feature quantity associated with the reference pixel includes at least one of the gray value of the reference pixel and the edge intensity of the reference pixel.

[0146] According to this structure, a pixel can be classified into any one of multiple feature value clusters based on at least one of the grayscale value and edge intensity of the reference pixel. This allows for the accurate identification of patterns formed on a medium. Therefore, the correction accuracy of printed images that overlap with patterns can be improved.

[0147] [4] In the above printing control method, it can also be set to include the following processing: normalizing the feature quantity for each of the plurality of object pixels; classifying to any one of the plurality of feature quantity clusters includes the following processing: classifying the pixel to any one of the plurality of feature quantity clusters based on the normalized feature quantity.

[0148] According to this structure, feature values ​​for each of multiple object pixels can be normalized, and the pixel can be classified into any one of the multiple feature value clusters based on the normalized feature values. Therefore, patterns formed on a medium can be identified with good accuracy. Consequently, the correction accuracy of printed images printed overlapping with patterns can be improved.

[0149] [5] In the above printing control method, the pattern image can also be set as an image in which each of the multiple pixels contained in the pattern image is assigned to either the pattern cluster or the background cluster.

[0150] According to this structure, by comparing a pattern image and a captured image, the pattern region corresponding to the pattern in the captured image can be obtained. The pattern image is an image in which each of the multiple pixels contained in the pattern image is assigned whether it is a pattern cluster or a background cluster. Therefore, patterns formed on a medium can be identified with good accuracy. Consequently, the correction accuracy of printed images that overlap with patterns can be improved.

[0151] [6] In the above printing control method, the captured image may also be configured to include the plurality of object pixels and the plurality of non-object pixels that are different from the plurality of object pixels, and include the following processing: based on the result of whether each of the plurality of object pixels is classified into the pattern cluster or the background cluster, interpolation is performed on each of the plurality of non-object pixels to determine whether it belongs to the pattern cluster or the background cluster.

[0152] According to this structure, even without calculating the feature values ​​for each of the multiple non-object pixels, it is possible to interpolate whether the clustering is for patterns or for the background based on the feature values ​​for each of the multiple object pixels. Therefore, the control load can be reduced.

[0153] [7] The printing apparatus performs the following processing: acquiring a photographic image obtained by photographing a medium on which a pattern is formed; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region for the pattern in the photographic image by comparing the photographic image with a pattern image related to the pattern formed on the medium based on the result of being classified into the pattern cluster or the background cluster; correcting a printing image that overlaps with the pattern formed on the medium and printing it into a corrected printing image based on the shape of the pattern region; and printing the corrected printing image on the medium. According to this structure, the same effect as [1] is achieved.

[0154] [8] The program causes at least one computer to perform the following processing: acquiring a photographic image obtained by photographing a medium with a pattern; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity; classifying the plurality of feature quantity clusters into either a pattern cluster or a background cluster according to predetermined conditions; obtaining a pattern region in the photographic image for the pattern by comparing the photographic image with a pattern image related to the pattern formed on the medium based on the result of being classified into the pattern cluster or the background cluster; correcting a printed image that overlaps with the pattern formed on the medium to a corrected printed image based on the shape of the pattern region; and printing the corrected printed image on the medium. According to this structure, the same effect as [1] is achieved.

[0155] Symbol Explanation 11…Printing apparatus; 12…Printing section; 13…Conveying section; 14…Picking section; 21…First roller; 22…Second roller; 23…Conveyor belt; 24…Roller drive section; 30…Control section; 31…Processor; 32…Memory; 99…Media; 99A…First surface; 99B…Second surface; D…Conveying direction; D1…Pattern image data; D2…Printed image data; D3…Picking image data; D4…Feature data; D5…Pattern area data; P1… Pattern; P2… Printed image; PR1… Reference area; PR2… Edge area; PX0… Object pixel; PX1… Reference pixel; PX2… Edge pixel; R1… First region; R11… First cluster region; R12… Second cluster region; R13… Third cluster region; R14… Fourth cluster region; R2… Second region; X1… First width direction; X2… Second width direction; Y1… Front; Y2… Back; Z1… Above; Z2… Below.

Claims

1. A printing control method, characterized in that, The process includes the following steps: obtaining a photographic image of a medium with a pattern; calculating a feature value for each of the plurality of object pixels contained in the photographic image; and classifying each of the plurality of object pixels into any one of the plurality of feature value clusters based on the feature value. Based on predetermined conditions, the multiple feature quantities are clustered and classified into either the pattern cluster or the background cluster. Based on whether the image is classified into the pattern cluster or the background cluster, the pattern region in the captured image is obtained by comparing the captured image with a pattern image related to the pattern formed on the medium; based on the shape of the pattern region, the printed image that overlaps with the pattern formed on the medium is corrected into a corrected printed image; and the corrected printed image is printed on the medium.

2. The printing control method as described in claim 1, characterized in that, The calculation of the feature quantity for each of the plurality of object pixels includes the following process: calculating the feature quantity associated with the reference pixel contained in the reference region centered on each of the plurality of object pixels.

3. The printing control method as described in claim 2, characterized in that, The feature quantity associated with the reference pixel includes at least one of the grayscale value of the reference pixel and the edge intensity of the reference pixel.

4. The printing control method according to any one of claims 1 to 3, characterized in that, The process includes the following steps: normalizing the feature values ​​for each of the plurality of object pixels; classifying the pixel into any one of the plurality of feature value clusters includes the following steps: classifying the pixel into any one of the plurality of feature value clusters based on the normalized feature values.

5. The printing control method according to any one of claims 1 to 3, characterized in that, The pattern image is an image in which each of the plurality of pixels contained in the pattern image is assigned to either the pattern cluster or the background cluster.

6. The printing control method according to any one of claims 1 to 3, characterized in that, The captured image includes the plurality of object pixels and a plurality of non-object pixels that are different from the plurality of object pixels. The processing includes the following: based on the result of whether each of the plurality of object pixels is classified into the pattern cluster or the background cluster, interpolation is performed on each of the plurality of non-object pixels to determine whether it belongs to the pattern cluster or the background cluster.

7. A printing apparatus, characterized in that, The process includes the following steps: obtaining a photographic image of a medium with a pattern; calculating a feature value for each of the plurality of object pixels contained in the photographic image; and classifying each of the plurality of object pixels into any one of the plurality of feature value clusters based on the feature value. According to predetermined conditions, the multiple feature quantities are clustered and classified into either the pattern cluster or the background cluster; Based on whether the image is classified into the pattern cluster or the background cluster, the pattern region in the captured image is obtained by comparing the captured image with a pattern image related to the pattern formed on the medium; based on the shape of the pattern region, the printed image that overlaps with the pattern formed on the medium is corrected into a corrected printed image; and the corrected printed image is printed on the medium.

8. A program product, characterized in that, It includes a program that causes at least one computer to perform the following processes: acquiring a photographic image obtained by photographing a medium with a pattern; calculating a feature quantity for each of a plurality of object pixels contained in the photographic image; and classifying each of the plurality of object pixels into any one of a plurality of feature quantity clusters based on the feature quantity. Based on predetermined conditions, the multiple feature quantities are clustered and classified into either the pattern cluster or the background cluster. Based on whether the image is classified into the pattern cluster or the background cluster, the pattern region in the captured image is obtained by comparing the captured image with a pattern image related to the pattern formed on the medium; based on the shape of the pattern region, the printed image that overlaps with the pattern formed on the medium is corrected into a corrected printed image; and the corrected printed image is printed on the medium.

Images