Image comparison method, image comparison device, and image comparison program

Abstract

To provide an image comparison method capable of detecting differences between two images in a short time while suppressing the occurrence of false detection caused by quantization errors or the like.SOLUTION: An image comparison method includes generating a difference image between an original image and a calibration image (S10), and then generating a candidate image representing a part to be a candidate for an edge region (S40). Then, the method includes, based on the original image, the calibration image, and the candidate image, generating an edge image representing the edge region (S50). At that time, pixels that constitute the candidate image are set as processing target pixels, and if the difference between the pixel value of at least one of nine comparison target pixels in the original image and the pixel value of a processing target pixel in the calibration image is equal to or less than a first threshold value, and the difference between the pixel value of at least one of the nine comparison target pixels in the calibration image and the pixel value of a processing target pixel in the original image is equal to or less than the first threshold, it is determined that the processing target pixels are pixels constituting the edge region.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The present invention relates to an image comparison method, an image comparison apparatus, and an image comparison program for comparing two images created for printing (for example, an image before calibration and an image after calibration).

Background Art

[0002] In printing operations, it is often necessary to compare two images created for printing. For example, before actual printing by a printing apparatus, the image to be printed may be corrected by calibration. In such a case, generally, the person who gives the correction instruction (correction instructor) and the person who actually makes the correction to the image to be printed (correction operator) are different. For the correction instructor, in order to confirm whether the correction operator has correctly corrected the image as instructed, a proofreading operation of comparing the image before calibration and the image after calibration is required. In this specification, the term "proofreading" is used in the sense of so-called "digital proofreading" in plate-less printing, rather than in the sense of inspection of printing plates in plate printing.

[0003] Generally, proofreading software is provided with a function of displaying on a computer screen the locations where there are differences between the image before calibration and the image after calibration. By using such proofreading software, the correction instructor can easily confirm whether the correction operator has correctly corrected the image as instructed.

[0004] In connection with the present invention, Japanese Patent Publication No. 2019-211319 discloses an image defect detection device capable of detecting minute dot defects and shortened line defects. The image defect detection device includes a first comparison unit that sets an allowable pattern misalignment range for positional misalignment of a reference image of the object to be inspected included in a reference image and compares the reference image with the object to be inspected, for which the allowable pattern misalignment range has been set; a second comparison unit that sets an allowable pattern misalignment range for positional misalignment of a reference image of the object to be inspected included in the inspection image and compares the inspection image with the reference image with the object to be inspected, for which the allowable pattern misalignment range has been set; and a determination unit that determines whether there is a defect. The determination unit determines that there is an image defect if the comparison result from either the first comparison unit or the second comparison unit indicates that there is a defect in the image of the object to be inspected. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2019-211319 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] When proofreading is performed, it is preferable that only the parts that have actually been corrected through proofreading are detected as differences (displayed on the computer screen as differences). In this regard, it is preferable that even minor corrections be detected as differences if they have been actually corrected according to the instructions of the person who gave the correction instructions. For example, if a line with the smallest line width is added, that added line should be detected as a difference.

[0007] Incidentally, while the comparison between the pre-proofread image and the post-proofread image is performed using raster data obtained through RIP processing, the resolution of the image data contained in the submitted data (data before RIP processing), such as PDF files, may differ from the resolution of the raster data obtained through RIP processing (in other words, the resolution specified when performing RIP processing). Therefore, quantization errors and jagged edges may occur during RIP processing. Due to such quantization errors, the edges of the image may be detected as differences in parts of the image that have not been corrected by the correction worker. For example, if both the pre-proofread image and the post-proofread image contain an image like the one shown in the part labeled 91 in Figure 44, the edges of the image shown in the part labeled 91 in Figure 44 will be detected as differences, as shown in Figure 45. Such phenomena can occur, for example, when a correction worker moves an image and then moves it back to its original position, or when a correction worker shifts the position of the entire image to be printed. Furthermore, such phenomena can occur if there are changes to the software used for image editing or if there are changes to the core part of the RIP processing due to version upgrades, etc. As described above, conventional proofreading software may detect parts that should not be detected as differences as differences (i.e., false detections occur).

[0008] According to the image defect detection device disclosed in Japanese Patent Publication No. 2019-211319, if there is a slight positional misalignment between two images, the misaligned portion is not detected as a difference. However, since an expanded image and a contracted image are created for each of the reference image and the inspection image, and then a comparison is made between the expanded image corresponding to the reference image and the inspection image, between the contracted image corresponding to the reference image and the inspection image, between the expanded image corresponding to the inspection image and the reference image, and between the contracted image corresponding to the inspection image and the reference image (i.e., four combinations are compared), the processing time is long. In addition, when comparing images, calculations are performed for comparison with 9 pixels for each pixel (18 pixels if the sign is positive or negative), which also contributes to the long processing time.

[0009] In view of the above circumstances, the present invention aims to provide an image comparison method that can detect differences between two images in a short time while suppressing the occurrence of false detections caused by quantization errors and the like. [Means for solving the problem]

[0010] The first invention is an image comparison method for comparing a first image, which is a multi-level image, with a second image, which is also a multi-level image. A difference image generation step of generating a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each consisting of one or more pixels, A shrinking step that generates a shrinking image including one or more partial shrinking images by applying a shrinking process to each of the one or more partial difference images, Outline image generation step: Removing the contracted image from the difference image to generate an outline image including one or more partial outline images, each composed of one or more pixels; A candidate image generation step of generating a candidate image that includes one or more partial candidate images, each composed of one or more pixels, which represent candidate portions of an edge region, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation step of generating an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation step of generating a comparison result image by removing the edge image from the difference image, Includes, In the edge image generation step, one or more pixels constituting the one or more partial candidate images included in the candidate image are sequentially designated as processing targets, and the processing target pixel and the eight pixels surrounding the processing target pixel are designated as nine comparison targets. If the difference between the pixel value of at least one of the nine comparison targets in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison targets in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold, then the processing target pixel is determined to be a pixel constituting the edge region.

[0011] The second invention is, in the first invention, In the edge image generation step, in addition to the first condition, if the pixel value of the pixel to be processed in the second image is within the range from the minimum to the maximum pixel value of the nine comparison pixels in the first image, and the second condition is also met, then the pixel to be processed in the first image is determined to be a pixel constituting the edge region.

[0012] The third invention relates to the first or second invention, The aforementioned difference image generation step is: A difference pixel extraction step for extracting pixels with different pixel values ​​between the first image and the second image, In the binarization step to generate the difference image, pixels extracted in the difference pixel extraction step in which the difference between the pixel value in the first image and the pixel value in the second image is equal to or greater than the second threshold are considered to have a difference between the first image and the second image, and pixels extracted in the difference pixel extraction step in which the difference between the pixel value in the first image and the pixel value in the second image is less than the second threshold, and pixels not extracted in the difference pixel extraction step are considered to have no difference between the first image and the second image. It is characterized by including.

[0013] The fourth invention is, in the third invention, The difference image generation step further includes a filtering step that removes partial difference images consisting of a predetermined number of pixels or less from among the one or more partial difference images included in the difference image generated in the binarization step.

[0014] The fifth invention is, in the third invention, The image comparison method is characterized by including a second threshold setting step in which the user sets the second threshold before the binarization step.

[0015] The sixth invention relates to the first or second invention, The outline image generation step is characterized in that the outline image is generated by performing an exclusive OR operation based on the difference image and the contracted image.

[0016] The seventh invention relates to the first or second invention, The aforementioned step of generating the comparison result image is: A logical inversion operation step that generates an edge inverted image by performing a logical inversion operation based on the aforementioned edge image, A logical AND operation step to generate the comparison result image by performing a logical AND operation based on the difference image and the edge inversion image. It is characterized by including.

[0017] The eighth invention relates to the first or second invention, In the difference image generation step, a first difference image is generated based on pixels in the first image whose pixel value is higher than the pixel value in the second image, and a second difference image is generated based on pixels in the second image whose pixel value is higher than the pixel value in the first image. In the contraction step, the outline image generation step, the candidate image generation step, the edge image generation step, and the comparison result image generation step, processing based on each of the first difference image and the second difference image is performed. In the comparison result image generation step, as the comparison result image, a first comparison result image obtained by processing based on the first difference image and a second comparison result image obtained by processing based on the second difference image are generated.

[0018] A ninth invention is in the first or second invention, The image comparison method includes a first threshold setting step in which a user sets the first threshold before the edge image generation step.

[0019] A tenth invention is in the first or second invention, The first image and the second image each consist of images of a plurality of ink colors. The processing of the difference image generation step, the contraction step, the outline image generation step, the candidate image generation step, the edge image generation step, and the comparison result image generation step is performed for each ink color.

[0020] An eleventh invention is in the first or second invention, The comparison result image generated in the comparison result image generation step includes a plurality of partial difference result images each composed of one or a plurality of pixels. After the comparison result image generation step, it includes a proximity image combining step of combining two or more partial difference result images that are close to each other by performing dilation processing on each of the plurality of partial difference result images.

[0021] A twelfth invention is in the first or second invention, The image comparison method includes a comparison result display step of displaying the comparison result image on a display unit of a computer. The comparison result display step is characterized by the fact that the parts where a difference is determined between the first image and the second image are displayed in color.

[0022] The 13th invention is, in the 12th invention, In the difference image generation step, a first difference image is generated based on pixels in the first image whose pixel value is higher than the pixel value in the second image, and a second difference image is generated based on pixels in the second image whose pixel value is higher than the pixel value in the first image. The comparison result display step is characterized in that the portion corresponding to the first difference image and the portion corresponding to the second difference image are displayed in different colors.

[0023] The fourteenth invention is an image comparison device for comparing a first image, which is a multi-level image, and a second image, which is also a multi-level image. A difference image generation unit generates a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each consisting of one or more pixels. A shrinking unit that generates a shrunk image including one or more partial shrunk images by applying a shrinking process to each of the one or more partial difference images, An outline image generation unit generates an outline image including one or more partial outline images, each composed of one or more pixels, by removing the contracted image from the difference image. A candidate image generation unit generates a candidate image that includes one or more partial candidate images, each composed of one or more pixels, which represent candidate portions of an edge region, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation unit generates an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation unit generates a comparison result image by removing the edge image from the difference image. Includes, The edge image generation unit sequentially selects one or more pixels constituting the one or more partial candidate images included in the candidate image as processing targets, and uses the processing target pixel and the eight pixels surrounding the processing target pixel as nine comparison targets. The unit determines that the processing target pixel is a pixel constituting the edge region if the first condition is met, which is that the difference between the pixel value of at least one of the nine comparison targets in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison targets in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold.

[0024] The 15th invention is an image comparison program that compares a first image, which is a multi-level image, with a second image, which is also a multi-level image. On the computer, A difference image generation step of generating a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each consisting of one or more pixels, A shrinking step that generates a shrinking image including one or more partial shrinking images by applying a shrinking process to each of the one or more partial difference images, Outline image generation step: Removing the contracted image from the difference image to generate an outline image including one or more partial outline images, each composed of one or more pixels; A candidate image generation step of generating a candidate image that includes one or more partial candidate images, each composed of one or more pixels, which represent candidate portions of an edge region, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation step of generating an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation step of generating a comparison result image by removing the edge image from the difference image, Make it run, In the edge image generation step, one or more pixels constituting the one or more partial candidate images included in the candidate image are sequentially designated as processing targets, and the processing target pixel and the eight pixels surrounding the processing target pixel are designated as nine comparison targets. If the difference between the pixel value of at least one of the nine comparison targets in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison targets in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold, then the processing target pixel is determined to be a pixel constituting the edge region. [Effects of the Invention]

[0025] According to the first invention described above, instead of detecting all areas that are judged to have a difference between the first image and the second image as difference areas, the areas that are judged to have a difference between the first image and the second image, excluding the image edge areas, are detected as difference areas. In other words, the occurrence of false detections, where the edge areas of an image are detected as difference areas even though there is no difference, is suppressed. In the edge image generation step, which generates an edge image representing the edge area, only the pixels that constitute the candidate image representing the candidate part of the edge area are treated as pixels to be processed. Then, a comparison is made between the pixel value of the pixel to be processed in the second image and the pixel values ​​of nine pixels centered on the pixel to be processed in the first image, and between the pixel value of the pixel to be processed in the first image and the pixel values ​​of nine pixels centered on the pixel to be processed in the second image. As a result, it is possible to detect the difference between the two images in a shorter time than before and without causing false detections. As described above, an image comparison method is realized that can detect the difference between two images in a short time while suppressing the occurrence of false detections caused by quantization errors, etc.

[0026] According to the second invention described above, for example, when a thin line with a small difference in gradation from its surroundings is added by modification, the pixels constituting the thin line are not determined to be pixels constituting an edge region, and therefore the thin line is appropriately detected as a difference. In other words, the difference between the two images is detected with high accuracy.

[0027] According to the third invention described above, when generating a difference image, the pixel value of each pixel in one of the first and second images is simply compared with the pixel value of the same pixel in the other of the first and second images. Therefore, a difference image, which is a binary image, is generated in a short time.

[0028] According to the fourth invention described above, the occurrence of false detections caused by various types of noise is suppressed.

[0029] According to the fifth invention described above, the user can set a threshold value for determining whether or not there is a difference between the first image and the second image.

[0030] According to the sixth invention described above, an outline image is generated from a difference image and a scalable image without performing complex calculations.

[0031] According to the seventh invention described above, a comparison result image is generated from the edge image and the difference image without performing complex calculations.

[0032] According to the eighth invention described above, even when the second image is generated by performing both the addition of partial images and the deletion of partial images on the first image, for example, the difference between the first image and the second image can be appropriately detected.

[0033] According to the ninth invention described above, the user can set a threshold value used to determine whether or not a pixel to be processed is a pixel that constitutes an edge region.

[0034] According to the tenth invention described above, it is possible to reduce the amount of memory required for a device that compares two images compared to a configuration that processes multiple ink colors at once.

[0035] According to the 11th invention described above, it is possible to improve the visibility of the differences when they are displayed on the screen.

[0036] According to the 12th invention described above, users can quickly and easily identify the differences between two images.

[0037] According to the 13th invention described above, for example, when a second image is generated by performing both the process of adding and deleting partial images on a first image, the user can distinguish and confirm the added partial image and the deleted partial image.

[0038] According to the 14th invention described above, the same effects as those of the first invention described above can be obtained.

[0039] According to the 15th invention described above, the same effects as those of the first invention described above can be obtained. [Brief explanation of the drawing]

[0040] [Figure 1] This is a block diagram showing the overall configuration of a printing system according to one embodiment of the present invention. [Figure 2] This block diagram shows the hardware configuration of an online calibration server that functions as an image comparison device in the above embodiment. [Figure 3] This figure shows an example of a source image that is subject to image comparison processing in the above embodiment. [Figure 4] This figure shows an example of a calibration image that is subject to image comparison processing in the above embodiment. [Figure 5] This is a block diagram showing the functional configuration for image comparison processing in the above embodiment. [Figure 6] The above embodiment is shown as a flowchart illustrating the procedure for image comparison processing. [Figure 7] This figure illustrates the overall algorithm for image comparison processing in the above embodiment. [Figure 8] This figure illustrates the algorithm for the re-evaluation process, which is performed by the shrinking unit, outline image generation unit, candidate image generation unit, and edge image generation unit in the image comparison process described above. [Figure 9] In the above embodiment, the flowchart shows the detailed procedure for the process in step S10 (the process of generating a difference image) in Figure 6. [Figure 10] This figure shows an example of a grayscale tolerance input screen in the above embodiment. [Figure 11] This figure illustrates the size filter in the above embodiment. [Figure 12] In the above embodiment, this figure shows an example of a positive difference image generated in the process of step S10 in Figure 6. [Figure 13] In the above embodiment, this figure shows an example of a negative difference image generated in the process of step S10 in Figure 6. [Figure 14] This figure shows an example of a difference image in the above embodiment. [Figure 15] In the above embodiment, this figure shows an example of a contracted image generated by the processing in step S20 of Figure 6. [Figure 16] This figure shows an example of an outline image generated in step S30 of Figure 6 in the above embodiment. [Figure 17] This figure illustrates the generation of candidate images in the above embodiment. [Figure 18] In the above embodiment, this figure shows an example of a candidate image generated in step S40 of Figure 6. [Figure 19] This figure illustrates the generation of edge images in the above embodiment. [Figure 20] This figure illustrates the generation of edge images in the above embodiment. [Figure 21] This figure illustrates the generation of edge images in the above embodiment. [Figure 22]This figure illustrates the generation of edge images in the above embodiment. [Figure 23] This figure illustrates the generation of edge images in the above embodiment. [Figure 24] This figure illustrates the generation of edge images in the above embodiment. [Figure 25] In the above embodiment, this figure shows an example of an edge image generated in the process of step S50 in Figure 6. [Figure 26] In the above embodiment, the flowchart shows the detailed procedure for step S60 (the process of generating the comparison result image) in Figure 6. [Figure 27] In the above embodiment, Figure 26 shows an example of an edge-inverted image generated by the processing in step S62. [Figure 28] In the above embodiment, this figure shows an example of a comparison result image generated in step S64 of Figure 26. [Figure 29] This figure illustrates the process of combining two or more adjacent partial difference result images in the above embodiment. [Figure 30] This figure illustrates the process of combining two or more adjacent partial difference result images in the above embodiment. [Figure 31] This figure illustrates the generation of a combined contour image in the above embodiment. [Figure 32] This flowchart illustrates the provision of a first threshold and a second threshold in the above embodiment. [Figure 33] This flowchart illustrates that, in the above embodiment, the processes in steps S10 to S80 in Figure 6 are performed for each ink color. [Figure 34] This diagram illustrates how differences might be missed when an image representing thin, fine lines is added. [Figure 35] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 36] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 37] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 38] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 39] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 40] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 41] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 42] This figure illustrates the concept of adding a second condition to a modified example of the above embodiment. [Figure 43] This figure illustrates that the present invention can also be applied when comparing two multi-level images for purposes other than proofreading. [Figure 44] This is a diagram illustrating a conventional example. [Figure 45] This is a diagram illustrating a conventional example. [Modes for carrying out the invention]

[0041] One embodiment of the present invention will be described below with reference to the attached drawings.

[0042] <1. Printing System Configuration> Figure 1 is a block diagram showing the overall configuration of a printing system according to one embodiment of the present invention. This printing system consists of an online proofing server 10, a print workflow management server 11, an inkjet printer 12, a personal computer 20, and a personal computer 30. The online proofing server 10, the print workflow management server 11, and the inkjet printer 12 are installed at the printing company. The personal computer 20 is installed at the customer company. The personal computer 30 is installed at the production company. The online proofing server 10, the print workflow management server 11, and the inkjet printer 12 are connected by a LAN 41 within the printing company. The online proofing server 10, the personal computer 20, and the personal computer 30 are connected via the Internet 42.

[0043] Computer 20, installed at the client company, is used for ordering prints and issuing correction instructions during the proofreading process. Computer 30, installed at the production company, is used for creating manuscripts according to the order details (creating print-ready data) and for correcting manuscripts during the proofreading process.

[0044] The online proofreading server 10 responds to requests from personal computers (personal computers 20 installed at the customer company or personal computers 30 installed at the production company) via the Internet 42 by displaying proofreading screens on the display units of those personal computers and sending and receiving data between the server and the personal computers. The online proofreading server 10 also performs image comparison processing for proofreading. The images used for proofreading are obtained by applying RIP processing to data sent from personal computers 30 to the online proofreading server 10. In this embodiment, it is assumed that the RIP processing is also performed by the online proofreading server 10. However, the RIP processing may be performed by the print workflow management server 11 or other personal computers (not shown) installed at the printing company.

[0045] Once proofreading is complete and approval has been obtained from the customer company's approver, the data for the print job is provided to the print workflow management server 11. The print workflow management server 11 has a program installed that implements a print workflow system that manages a series of processes for printing using the inkjet printer 12. In other words, the print workflow management server 11 manages the print workflow. For example, it performs processes such as determining the printing order for multiple print jobs to ensure efficient printing.

[0046] The inkjet printing device 12 outputs a printed image (i.e., performs printing) by ejecting ink onto the printing paper, which is the substrate, without using a printing plate, based on the print data generated by the RIP process.

[0047] In this embodiment, the online proofreading server 10 functions as an image comparison device (a device that performs image comparison processing, as described later), but the present invention is not limited to this. A personal computer 20 installed at the customer company or a personal computer 30 installed at the production company may also function as an image comparison device.

[0048] <2. Hardware configuration of the online calibration server> Figure 2 is a block diagram showing the hardware configuration of an online calibration server 10 functioning as an image comparison device in this embodiment. The online calibration server 10 includes a main unit 110, an auxiliary storage device 121, an optical disc drive 122, a display unit 123, a keyboard 124, and a mouse 125. The main unit 110 includes a CPU 111, memory 112, a first disk interface unit 113, a second disk interface unit 114, a display control unit 115, an input interface unit 116, and a network interface unit 117. The CPU 111, memory 112, first disk interface unit 113, second disk interface unit 114, display control unit 115, input interface unit 116, and network interface unit 117 are connected to each other via a system bus. The auxiliary storage device 121 is connected to the first disk interface unit 113. The auxiliary storage device 121 is a magnetic disk drive or the like. The optical disc drive 122 is connected to the second disk interface unit 114. An optical disc 49, such as a CD-ROM or DVD-ROM, which is a computer-readable recording medium, is inserted into the optical disc drive 122. A display unit (display device) 123 is connected to the display control unit 115. The display unit 123 is a liquid crystal display or the like. The display unit 123 is used to display information desired by the operator. A keyboard 124 and a mouse 125 are connected to the input interface unit 116. The keyboard 124 and mouse 125 are used by the operator to input instructions to this online calibration server 10. The network interface unit 117 is a communication interface circuit.

[0049] The auxiliary storage device 121 stores the image comparison program P, other programs, and various data. The CPU 111 reads the image comparison program P stored in the auxiliary storage device 121 into the memory 112 and executes it, thereby realizing various functions for image comparison processing described later. The memory 112 includes RAM and ROM. The memory 112 functions as a work area for the CPU 111 to execute the image comparison program P stored in the auxiliary storage device 121. The image comparison program P is provided stored on the computer-readable recording medium (non-transient recording medium) described above. That is, for example, the user purchases an optical disc 49 as the recording medium for the image comparison program P, inserts it into the optical disc drive 122, reads the image comparison program P from the optical disc 49, and installs it in the auxiliary storage device 121. Alternatively, the image comparison program P transmitted via the network may be received by the network interface unit 117 and installed in the auxiliary storage device 121.

[0050] <3. Image Comparison Processing> <3.1 Assumptions> The image comparison process in this embodiment will now be described. This specification focuses on the case in which the online calibration server 10 compares an image based on data sent from the PC 30 to the online calibration server 10 before the calibration work is performed with an image based on data sent from the PC 30 to the online calibration server 10 after the calibration work is performed. In the following, the image based on data sent from the PC 30 to the online calibration server 10 before the calibration work is performed will be called the "original image," and the image based on data sent from the PC 30 to the online calibration server 10 after the calibration work is performed will be called the "calibrated image." The original image will be denoted by reference numeral 61, and the calibrated image will be denoted by reference numeral 62. In the example described below, it is assumed that the original image 61 is the image shown in Figure 3, and the calibrated image 62 is the image shown in Figure 4. In Figures 3 and 4, one rectangle represents one pixel. The image of the part denoted by reference numeral 502 in Figure 4 is an image obtained by shifting the image of the part denoted by reference numeral 501 in Figure 3 to the left by 0.5 pixels due to the quantization error described above. In Figure 4, the pixel values ​​of pixels with a light shade are smaller than the pixel values ​​of pixels with a dark shade in Figure 4.

[0051] <3.2 Functional Configuration for Image Comparison Processing> Figure 5 is a block diagram showing the functional configuration for image comparison processing. The functional configuration shown in Figure 5 is realized when the image comparison program P is executed on the online calibration server 10. As shown in Figure 5, the online calibration server 10 includes, as functional components for performing image comparison processing, a difference image generation unit 151, a contraction unit 152, an outline image generation unit 153, a candidate image generation unit 154, an edge image generation unit 155, a comparison result image generation unit 156, a nearby image merging unit 157, a merged image contour extraction unit 158, and a comparison result display unit 159. The general operation of each component will be described below.

[0052] The difference image generation unit 151 generates a difference image 64, which is a binary image representing the differences between the original image 61 and the proofread image 62. The difference image 64 generated by the difference image generation unit 151 typically contains one or more partial difference images. Each partial difference image consists of one or more pixels.

[0053] The shrinking unit 152 generates a condensed image 71 by applying a shrinking process to each of the one or more partial difference images included in the difference image 64 generated by the difference image generation unit 151. The condensed image 71 generated by the shrinking unit 152 typically contains one or more partial condensed images. Each partial condensed image consists of one or more pixels.

[0054] The outline image generation unit 153 generates an outline image 72 corresponding to the contour of the difference image 64 by removing the contracted image 71 from the difference image 64. More specifically, the outline image generation unit 153 generates the outline image 72 by removing one or more partial contracted images obtained by the contraction process by the contraction unit 152 from one or more partial difference images contained in the difference image 64. The outline image 72 generated by the outline image generation unit 153 typically contains one or more partial outline images. Each partial outline image consists of one or more pixels.

[0055] The candidate image generation unit 154 generates a candidate image 73 representing a candidate portion of an edge region by removing one or more partial outline images included in the outline image 72 that are adjacent to the partial shrunk image generated by the shrunk unit 152.

[0056] The edge image generation unit 155 generates an edge image representing an actual edge region from among the candidate images 73 based on the original image 61, the calibration image 62, and the candidate image 73. In this regard, the edge image generation unit 155 sequentially selects one or more pixels constituting one or more partial candidate images included in the candidate image 73 as processing targets, and selects the processing target pixel and the eight pixels surrounding it as nine comparison targets. The unit determines that a processing target pixel is a pixel constituting an edge region if the first condition is met, which is that the difference between the pixel value of at least one of the nine comparison targets in the original image 61 and the pixel value of the processing target pixel in the calibration image 62 is less than or equal to a predetermined threshold (hereinafter referred to as the "first threshold"), and the difference between the pixel value of at least one of the nine comparison targets in the calibration image 62 and the pixel value of the processing target pixel in the original image 61 is less than or equal to the first threshold.

[0057] The comparison result image generation unit 156 generates a comparison result image 65 representing the portion where a difference is ultimately determined to exist between the original image 61 and the calibration image 62 by removing the edge image 74 generated by the edge image generation unit 155 from the difference image 64. The comparison result image 65 generated by the comparison result image generation unit 156 typically contains multiple partial difference result images. Each partial difference result image consists of one or more pixels.

[0058] The adjacent image merging unit 157 combines two or more adjacent partial difference result images into a single merged image 75 by performing an expansion process on the comparison result image 65. As a result, for example, images of multiple characters become a single merged image 75.

[0059] The combined image contour extraction unit 158 ​​generates an image (hereinafter referred to as the "combined contour image" for convenience) 76 that represents the region corresponding to the contour of the combined image 75 by removing the image obtained by applying a shrinkage process to the combined image 75 from the combined image 75.

[0060] The comparison result display unit 159 refers to the calibration image 62 and the combined contour image 76 and displays the comparison result image 65 in a manner in which the parts where differences are determined to exist between the original image 61 and the calibration image 62 are highlighted.

[0061] <3.3 Procedure for Image Comparison Processing> Figure 6 is a flowchart showing the steps of the image comparison process. Figure 7 is a diagram illustrating the overall algorithm of the image comparison process. Figure 8 is a diagram illustrating the algorithm of the re-evaluation process, which is performed by the shrinking unit 152, the outline image generation unit 153, the candidate image generation unit 154, and the edge image generation unit 155 within the image comparison process. Referring to these, the steps of the image comparison process (image comparison method) in this embodiment will be described in detail.

[0062] When the image comparison process is started, the difference image generation unit 151 first generates a difference image 64, which is a binary image representing the parts that differ between the original image 61 and the proofread image 62 (step S10).

[0063] Figure 9 is a flowchart showing the detailed procedure for the process in step S10 (the process of generating the difference image 64). In step S10, first, the original image 61 and the proof image 62 are compared pixel by pixel, and pixels with different pixel values ​​between them (difference pixels) are extracted (step S12). In other words, the difference between the original image 61 and the proof image 62 is extracted. The data of the extracted pixels is multi-level data. When a difference is detected for any pixel, the pixel value in the proof image 62 may be larger than the pixel value in the original image 61, or the pixel value in the original image 61 may be larger than the pixel value in the proof image 62. Therefore, in this embodiment, an image based on a set of pixels where the pixel value in the proof image 62 is larger than the pixel value in the original image 61 is treated as a positive difference multi-level image 63a, and an image based on a set of pixels where the pixel value in the original image 61 is larger than the pixel value in the proof image 62 is treated as a negative difference multi-level image 63b (see Figure 7).

[0064] After a positive difference multi-level image 63a and a negative difference multi-level image 63b are obtained, a binarization process is performed on each of the positive difference multi-level image 63a and the negative difference multi-level image 63b (step S14). In the binarization process, the absolute value of the difference of each pixel is compared with a predetermined threshold (hereinafter referred to as the "second threshold") TH2. Pixels whose absolute difference is greater than or equal to the second threshold TH2 are assigned a value of 1 after binarization, while pixels whose absolute difference is less than the second threshold TH2 are assigned a value of 0 after binarization. Thus, among the difference pixels extracted in step S12, pixels whose difference between the pixel value in the original image 61 and the pixel value in the proof image 62 is greater than or equal to the second threshold TH2 are considered to have a difference between the original image 61 and the proof image 62, while pixels whose difference between the pixel value in the original image 61 and the pixel value in the proof image 62 is less than the second threshold TH2 are considered to have no difference between the original image 61 and the proof image 62.

[0065] Incidentally, the second threshold TH2 can be set in advance by the user. However, in this embodiment, the user does not directly specify the second threshold TH2, but rather specifies the percentage (%) corresponding to the second threshold TH2. Hereinafter, this percentage specified by the user will be referred to as the "tone tolerance". Figure 10 shows an example of a tone tolerance input screen 80 for the user to input the tone tolerance. This tone tolerance input screen 80 includes a text box 801 for inputting the tone tolerance value, an OK button 802, and a Cancel button 803. When the OK button 802 is pressed with a value entered in the text box 801, the second threshold TH2 is calculated. If the Cancel button 803 is pressed, the second threshold TH2 is not calculated, and the tone tolerance input screen 80 is hidden. If the multi-level data is 8-bit data, and the value (percentage) entered in the text box 801 is represented by V, the second threshold TH2 is calculated by the following equation (1). TH2 = (V / 100) × 255 ... (1)

[0066] After the binarization process is completed, a process called "size filtering" is performed (step S16) to remove partial difference images from the binary image generated by the binarization process that consist of pixels less than or equal to a predetermined number of isolated errors, NI. For example, suppose the number of isolated errors NI is set to "2" and the binary image to be processed contains four partial difference images 81 to 84, as shown in part A of Figure 11. In this case, partial difference image 81 consists of one pixel and partial difference image 82 consists of two pixels, so partial difference images 81 and 82 are removed by the size filtering. As a result, a difference image like the one shown in part B of Figure 11 is obtained. Note that the process in step S16 is not mandatory, and a procedure that omits the process in step S16 can be adopted. However, performing the process in step S16 suppresses the occurrence of false detections caused by various types of noise.

[0067] The processing in step S10 is completed when the size filter is finished. As described above, if the original image 61 is the image shown in Figure 3, the calibration image 62 is the image shown in Figure 4, and the number of isolated errors NI is set to "1", then the difference image (positive difference image 64a) corresponding to the positive difference multi-level image 63a is obtained as the image shown in Figure 12, and the difference image (negative difference image 64b) corresponding to the negative difference multi-level image 63b is obtained as the image shown in Figure 13. Note that the part of the image labeled 621 in Figure 4 is removed by the size filter.

[0068] As can be seen from Figure 7, the re-evaluation process (the process from steps S20 to S60 in Figure 6) is also divided into a process based on the positive difference image 64a and a process based on the negative difference image 64b. However, below we will focus on one of these processes and explain the case in which the re-evaluation process is performed based on the difference image 64 as shown in Figure 14.

[0069] In step S20, the shrinking unit 152 performs a shrinking process on each of the one or more partial difference images included in the difference image 64 generated in step S10. In this example, the difference image 64 (see Figure 14) contains five partial difference images 641 to 645, and a shrinking process with a shrinking width of 1 pixel is performed on each of these five partial difference images 641 to 645. This generates a shrinking image 71 as shown in Figure 15. The shrinking image 71 shown in Figure 15 contains one partial shrinking image 711. This partial shrinking image 711 is the image obtained by the shrinking process on the partial difference image 644 in Figure 14.

[0070] After the condensed image 71 is generated, the outline image generation unit 153 generates an outline image 72 by removing one or more partial condensed images generated in step S20 from one or more partial difference images included in the difference image 64 (step S30). More specifically, the outline image 72 is generated by performing an exclusive OR (XOR) operation between the difference image 64 (see Figure 14) and the condensed image 71 (see Figure 15). In this example, step S30 generates an outline image 72 as shown in Figure 16. The outline image 72 contains five partial outline images 721 to 725. The partial outline image 724 is an image obtained by removing the partial condensed image 711 shown in Figure 15 from the partial difference image 644 in Figure 14.

[0071] After the outline image 72 is generated, the candidate image generation unit 154 generates a candidate image 73 by removing one or more partial outline images included in the outline image 72 that are adjacent to the partial contraction image generated in step S20 (step S40). In this regard, if the partial contraction image 711 shown in Figure 15 is represented by a diagonal line, the relationship between the partial contraction image 711 and the partial outline image 724 in Figure 16 is as shown in Figure 17. As can be seen from Figure 17, the partial outline image 724 is adjacent to the partial contraction image 711. Therefore, in step S40, the partial outline image 724 is removed from the outline image 72. As a result, a candidate image 73 as shown in Figure 18 is generated. In this example, the candidate image 73 includes four partial candidate images 731 to 734.

[0072] After the candidate image 73 is generated, the edge image generation unit 155 generates an edge image 74 representing the edge region (step S50). In step S50, one or more pixels that constitute one or more partial candidate images included in the candidate image 73 generated in step S40 are sequentially selected as processing pixels, and if the first condition described above is met, the processing pixel is determined to be a pixel that constitutes an edge region. Step S50 will be explained in detail with reference to Figures 19 to 24.

[0073] As described above, the image of the portion labeled 502 in Figure 4 is obtained by shifting the image of the portion labeled 501 in Figure 3 to the left by 0.5 pixels due to quantization error. Therefore, the relationship between the image of the portion labeled 502 in Figure 4 and the image of the portion labeled 501 in Figure 3 is as shown in Figure 19. The candidate image 73 corresponding to the portion shown in Figure 19 is as shown in Figure 20. In step S50, one or more pixels constituting the partial candidate image included in the candidate image 73 are sequentially designated as processing targets, but here we focus on the operation when the pixel labeled 541 in Figure 20 is the processing target pixel. At this time, the processing target pixel 541 and the eight pixels surrounding the processing target pixel 541 become comparison targets. That is, nine pixels centered on the processing target pixel 541 become comparison targets.

[0074] In this case, first, the pixel value of the processing target pixel 541 (see Figure 21) in the calibration image 62 is sequentially compared with the pixel values ​​of the nine comparison target pixels in the original image 61 (the nine pixels located within the thick frame labeled with reference numeral 54 in Figure 22). Then, it is determined whether the difference between the pixel value of at least one of the nine comparison target pixels in the original image 61 and the pixel value of the processing target pixel 541 in the calibration image 62 is less than or equal to the first threshold TH1. In this example, it is assumed that the difference between the pixel values ​​of the three processing target pixels 541 to 543 in the original image 61 and the pixel value of the processing target pixel 541 in the calibration image 62 is less than or equal to the first threshold TH1. Furthermore, the pixel value of the processing target pixel 541 (see Figure 22) in the original image 61 is sequentially compared with the pixel values ​​of the nine comparison target pixels in the calibration image 62 (the nine pixels located within the thick frame labeled with reference numeral 54 in Figure 21). Next, it is determined whether the difference between the pixel value of at least one of the nine comparison pixels in the calibration image 62 and the pixel value of the processing target pixel 541 in the original image 61 is less than or equal to the first threshold TH1. In this example, it is determined that the difference between the pixel values ​​of the three processing target pixels 541 to 543 in the calibration image 62 and the pixel value of the processing target pixel 541 in the original image 61 is less than or equal to the first threshold TH1. As a result, this case satisfies the first condition, and the processing target pixel 541 is determined to be a pixel that constitutes an edge region.

[0075] Next, we focus on the operation when the pixel labeled 551, one of the four pixels constituting the partial candidate image 734 shown in Figure 18, is the pixel to be processed. Note that the image of the vicinity of the pixel to be processed 551 in the original image 61 is as shown in Figure 23, and the image of the vicinity of the pixel to be processed 551 in the calibration image 62 is as shown in Figure 24.

[0076] In this case as well, first, it is determined whether the difference between the pixel value of at least one of the nine comparison pixels in the original image 61 (the nine pixels located within the thick frame labeled with reference numeral 55 in Figure 23) and the pixel value of the processing target pixel 551 in the calibration image 62 is less than or equal to the first threshold TH1. In this example, it is determined that the difference between the pixel value of any of the nine comparison pixels in the original image 61 and the pixel value of the processing target pixel 551 in the calibration image 62 is greater than the first threshold TH1. Furthermore, it is determined whether the difference between the pixel value of at least one of the nine comparison pixels in the calibration image 62 (the nine pixels located within the thick frame labeled with reference numeral 55 in Figure 24) and the pixel value of the processing target pixel 551 in the original image 61 is less than or equal to the first threshold TH1. In this example, it is determined that the difference between the pixel values ​​of the seven comparison pixels 552 to 558 in the calibration image 62 and the pixel value of the processing target pixel 551 in the original image 61 is less than or equal to the first threshold TH1. Based on the above, this case does not satisfy the first condition, and therefore, it is determined that the pixel 551 to be processed is not a pixel that constitutes an edge region.

[0077] As described above, in step S50, an image like the one shown in Figure 25 is generated as an edge image 74 representing the edge region.

[0078] Subsequently, the comparison result image generation unit 156 generates a comparison result image 65 by removing the edge image 74 generated in step S50 from the difference image 64 (step S60).

[0079] Figure 26 is a flowchart detailing the steps of the process in step S60 (the process of generating the comparison result image 65). In step S60, first, a logical inversion operation is performed based on the edge image 74 (step S62). This generates the edge inversion image 741 as shown in Figure 27. Next, a logical AND operation is performed based on the difference image 64 (see Figure 14) and the edge inversion image 741 (step S64). This generates the comparison result image 65 as shown in Figure 28. The comparison result image 65 shown in Figure 28 contains two partial difference result images 651 and 652. The process in step S60 is completed when the comparison result image 65 is generated by the logical AND operation described above.

[0080] After the comparison result image 65 is generated, the proximity image merging unit 157 merges two or more adjacent partial difference result images into one (step S70). In step S70, each of the multiple partial difference result images is subjected to an expansion process, for example, with an expansion width of 10 to 15 pixels. As a result, two or more adjacent partial difference result images are merged. Regarding this step S70, let us assume, for example, that the comparison result image 65 shown in Figure 29 is generated in step S60. At this time, the partial difference result image of the portion labeled 655 in Figure 29 is subjected to the expansion process to generate the merged image labeled 75(1) in Figure 30, and the partial difference result image of the portion labeled 656 in Figure 29 is subjected to the expansion process to generate the merged image labeled 75(2) in Figure 30.

[0081] Next, the combined image contour extraction unit 158 ​​removes the image obtained by applying a shrinkage process to the combined image 75 generated in step S70, thereby generating a combined contour image 76 (step S80). More specifically, first, the combined image 75 is subjected to a shrinkage process, for example, with a shrinkage width of 1 pixel. This generates a shrunk combined image in the same manner as in step S20 described above. Next, an exclusive OR (XOR) operation is performed between the combined image 75 and the shrunk combined image. This generates the combined contour image 76. For example, the combined contour image labeled 76(1) in Figure 31 is generated based on the combined image labeled 75(1) in Figure 30, and the combined contour image labeled 76(2) in Figure 31 is generated based on the combined image labeled 75(2) in Figure 30.

[0082] Finally, the comparison result display unit 159 displays the comparison result image 65 on the display unit 123 in a manner in which the parts where differences have been determined between the original image 61 and the proof image 62 are colored (step S90). In this regard, the comparison result image 65 may be displayed in a manner such as "parts with no difference: displayed in gray, parts of the difference corresponding to the positive difference image 64a: colored with a first specified color specified by the user, and parts of the difference corresponding to the negative difference image 64b: colored with a second specified color specified by the user," or the comparison result image 65 may be displayed in a manner such as "parts with no difference: displayed in gray, parts of the difference: colored with the color of the proof image 62." The image comparison process is completed when the comparison result image 65 is displayed on the display unit 123 in this manner.

[0083] In this embodiment, the image comparison process described above is performed by the printing company's online proofing server 10 (see Figure 1). It is preferable that a step for the user to set a first threshold TH1 is provided before step S50. It is also preferable that a step for the user to set a second threshold TH2 is provided before step S14. In this regard, for example, as shown in Figure 32, the step S6 for setting the first threshold TH1 and the step S8 for setting the second threshold TH2 may be provided before step S10.

[0084] Incidentally, the image subject to the image comparison process consists of images of each of several ink colors. The processes in steps S10 to S80 in Figure 6 are performed for each ink color. For example, an image composed of four colors, C (cyan), M (magenta), Y (yellow), and K (black), is subject to the image comparison process, and the processes in steps S10 to S80 in Figure 6 are performed for each of these four colors. In this case, the overall flow of the image comparison process is as shown in Figure 33, for example. Alternatively, for example, an image composed of (Z+4) colors, which are C, M, Y, and K plus Z (Z is a natural number) feature colors, is subject to the image comparison process, and the processes in steps S10 to S80 in Figure 6 are performed for each of these (Z+4) colors. Furthermore, for example, in the case of print data with only two colors, K and M, the image composed of K and M is subject to the image comparison process, and the processes in steps S10 to S80 in Figure 6 are performed for each of the two colors. By adopting a configuration that processes each ink color separately, it is possible to reduce the amount of memory required for the device that functions as an image comparison device (in this embodiment, the online calibration server 10) compared to a configuration that processes multiple ink colors together.

[0085] In this embodiment, step S10 is the difference image generation step, step S12 is the difference pixel extraction step, step S14 is the binarization step, step S16 is the filtering step, step S20 is the shrinking step, step S30 is the outline image generation step, step S40 is the candidate image generation step, step S50 is the edge image generation step, step S60 is the comparison result image generation step, step S62 is the logical inversion operation step, step S64 is the logical AND operation step, step S70 is the adjacent image merging step, and step S90 is the comparison result display step. The original image 61 corresponds to the first image, and the calibration image 62 corresponds to the second image.

[0086] <4. Effects> In the image comparison process of this embodiment, not all areas where a difference is determined between the original image 61 and the proofread image 62 are detected as differences. Rather, the areas where a difference is determined between the original image 61 and the proofread image 62, excluding the image edge areas, are detected as differences. In other words, even if quantization errors occur, the image edge areas are not detected as differences.

[0087] Incidentally, according to the method disclosed in Japanese Patent Publication No. 2019-211319, as described above, it is necessary to compare four combinations (comparison of the dilated image corresponding to the reference image with the inspection image, comparison of the contracted image corresponding to the reference image with the inspection image, comparison of the dilated image corresponding to the inspection image with the reference image, and comparison of the contracted image corresponding to the inspection image with the reference image), and when comparing each combination, it is necessary to perform calculations for comparison with 9 pixels for each pixel (18 pixels in total, considering the sign of the pixels). In contrast, according to this embodiment, when generating an edge image 74 representing an edge region, only the pixels constituting the candidate image 73 representing the candidate portion of the edge region are treated as the pixels to be processed. Then, a comparison is made between the pixel value of the pixels to be processed in the calibration image 62 and the pixel values ​​of 9 pixels centered on the pixels to be processed in the original image 61, and between the pixel value of the pixels to be processed in the original image 61 and the pixel values ​​of 9 pixels centered on the pixels to be processed in the calibration image 62. In this way, according to this embodiment, it is possible to detect the difference between two images in a shorter time than before and without causing false detections. Furthermore, the reduced processing time leads to lower power consumption, which can contribute to achieving the SDGs (Sustainable Development Goals).

[0088] As described above, this embodiment provides an image comparison method that can detect differences between two images in a short time while suppressing the occurrence of false detections caused by quantization errors and the like.

[0089] <5. Variation> In the above embodiment, in the case where the image near the processing target pixel 551 in the original image 61 is as shown in Figure 23 and the image near the processing target pixel 551 in the proof image 62 is as shown in Figure 24, the processing target pixel 551 is determined not to be a pixel constituting an edge region. A similar case is assumed where the image near the processing target pixel 551 in the original image 61 is as shown in Figure 23 and the image near the processing target pixel 551 in the proof image 62 is as shown in Figure 34 (hereinafter referred to as the "thin line case" for convenience). It is assumed that the image of the part labeled with reference numeral 56 in Figure 34 is a light-colored image obtained by shifting an image representing a thin line with a width of 1 pixel to the left by 0.5 pixels due to quantization error.

[0090] In the case of thin lines as described above, in step S50 of Figure 6, it is determined that the difference between the pixel values ​​of the nine comparison pixels in the original image 61 (the nine pixels located within the thick frame labeled with reference numeral 55 in Figure 23) and the pixel value of the processing target pixel 551 in the calibration image 62 (see Figure 34) is less than or equal to the first threshold TH1, and that the difference between the pixel values ​​of the nine comparison pixels in the calibration image 62 (the nine pixels located within the thick frame labeled with reference numeral 55 in Figure 34) and the pixel value of the processing target pixel 551 in the original image 61 (see Figure 23) is less than or equal to the first threshold TH1. In other words, since the first condition is met, it is determined that the processing target pixel 541 is a pixel that constitutes an edge region. In that case, the part labeled with reference numeral 56 in Figure 34 is not detected as a difference (a detection omission occurs). Therefore, a method for appropriately detecting thin lines with small gradation differences from the surroundings as a difference when such thin lines are added by correction will be described below as a modified example of the above embodiment.

[0091] First, let's explain the concept behind this modified example. For the image of the portion labeled 501 in Figure 3, the graph representing the pixel values ​​of the portion indicated by the line A-A' in Figure 35 is as shown in Figure 36. Similarly, for the image of the portion labeled 502 in Figure 4, the graph representing the pixel values ​​of the portion indicated by the line B-B' in Figure 37 is as shown in Figure 38. Here, the pixel labeled 81 in Figures 35 and 37 is referred to as the "pixel of interest." As can be seen from Figures 35 to 38, among the nine pixels centered around the pixel of interest 81 in the original image, there are pixels with pixel values ​​less than or equal to the pixel value of the pixel of interest 81 in the calibration image, and pixels with pixel values ​​greater than or equal to the pixel value of the pixel of interest 81 in the calibration image. In other words, the pixel value of the pixel of interest 81 in the calibration image is within the range from the minimum to the maximum pixel value of the nine pixels centered around the pixel of interest 81 in the original image. Furthermore, the pixel value of the target pixel 81 in the original image is within the range of the minimum and maximum pixel values ​​of the nine pixels centered around the target pixel 81 in the calibrated image.

[0092] Regarding the image shown in Figure 34, the graph representing the pixel values ​​of the portion indicated by the line C-C' in Figure 39 is as shown in Figure 40. Similarly, regarding the corresponding portion of the original image, the graph representing the pixel values ​​of the portion indicated by the line D-D' in Figure 41 is as shown in Figure 42. Here, the pixel labeled 83 in Figures 39 and 41 is referred to as the "pixel of interest." As can be seen from Figures 39 to 42, the pixel value of the pixel of interest 83 in the calibration image is greater than the pixel values ​​of any of the nine pixels centered around the pixel of interest 83 in the original image. Thus, for portions that do not constitute the edges of the image, the pixel value of the pixel of interest in at least one of the original image and the calibration image is outside the range of the minimum and maximum pixel values ​​of the nine pixels centered around the pixel of interest in the other of the original image and the calibration image.

[0093] In view of the above, in this modified example, in step S50 of Figure 6, the edge image generation unit 155 determines that a pixel to be processed is a pixel constituting an edge region if, in addition to the first condition described above, the pixel value of the pixel to be processed in the calibration image is within the range from the minimum to the maximum pixel value of the nine comparison pixels in the original image, and the pixel value of the pixel to be processed in the original image is within the range from the minimum to the maximum pixel value of the nine comparison pixels in the calibration image.

[0094] According to this modified version, when a thin line with a small difference in gradation from its surroundings is added by the correction, the pixels constituting the thin line are not determined to be pixels constituting an edge region. In other words, the thin line is appropriately detected as a difference. Thus, according to this modified version, an image comparison method is realized that can accurately detect differences between two images in a short time while suppressing the occurrence of false detections caused by quantization errors, etc.

[0095] <6. Others> The present invention is not limited to the embodiments described above, and can be implemented in various modified forms without departing from the spirit of the invention. For example, although the above embodiments described an example of comparing two images for the purpose of proofreading, the invention is not limited to this. The present invention can also be applied when comparing two multi-level images (first image 51 and second image 52) for purposes other than proofreading and displaying the comparison result image obtained from the comparison (see Figure 43). Furthermore, for example, the comparison result image 65 generated in step S60 may be displayed directly on the display unit 123 without performing the processing in steps S70 and S80 in Figure 6. [Explanation of symbols]

[0096] 10… Online calibration server (image comparison device) 11…Print Workflow Management Server 12… Inkjet printing equipment 20…(Customer company's) personal computer 30…(The production company's) computer 61...Original image 62…Proofreading image 64...Difference image 65...Comparison result image 71...Collapsed image 72…Outline image 73… Candidate image 74…Edge image 80... Tonal tolerance input screen 151...Difference image generation unit 152... Contraction part 153... Outline image generation unit 154... Candidate image generation unit 155...Edge image generation unit 156...Comparison result image generation unit 157... Close-up image merging section 158... Combined image contour extraction section 159…Comparison result display section P...Image comparison program

Claims

1. An image comparison method that compares a first image, which is a multi-level image, with a second image, which is also a multi-level image. A difference image generation step of generating a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each composed of one or more pixels; A shrinking step that generates a shrinking image including one or more partial shrinking images by applying a shrinking process to each of the one or more partial difference images, Outline image generation step: Removing the contracted image from the difference image to generate an outline image including one or more partial outline images, each composed of one or more pixels, A candidate image generation step of generating a candidate image that includes one or more partial candidate images, each composed of one or more pixels, which represent candidate portions of an edge region, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation step of generating an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation step of generating a comparison result image by removing the edge image from the difference image, Includes, In the edge image generation step, one or more pixels constituting the one or more partial candidate images included in the candidate image are sequentially designated as processing targets, and the processing target pixel and the eight pixels surrounding the processing target pixel are designated as nine comparison targets, and if the difference between the pixel value of at least one of the nine comparison targets in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison targets in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold, then the processing target pixel is determined to be a pixel constituting the edge region, characterized in that the image comparison method satisfies the first condition.

2. The image comparison method according to claim 1, characterized in that, in the edge image generation step, in addition to the first condition, if the pixel value of the pixel to be processed in the second image is within the range from the minimum to the maximum pixel value of the nine comparison target pixels in the first image, and the pixel value of the pixel to be processed in the first image is within the range from the minimum to the maximum pixel value of the nine comparison target pixels in the second image, then the pixel to be processed is determined to be a pixel constituting the edge region.

3. The aforementioned difference image generation step is: A difference pixel extraction step for extracting pixels with different pixel values ​​between the first image and the second image, In the binarization step to generate the difference image, pixels extracted in the difference pixel extraction step in which the difference between the pixel value in the first image and the pixel value in the second image is equal to or greater than the second threshold are considered to have a difference between the first image and the second image, and pixels extracted in the difference pixel extraction step in which the difference between the pixel value in the first image and the pixel value in the second image is less than the second threshold, and pixels not extracted in the difference pixel extraction step are considered to have no difference between the first image and the second image. The image comparison method according to claim 1 or 2, characterized by including the following:

4. The image comparison method according to claim 3, characterized in that the difference image generation step further includes a filtering step to remove partial difference images consisting of a predetermined number of pixels or less from among the one or more partial difference images included in the difference image generated in the binarization step.

5. The image comparison method according to claim 3, characterized in that it includes a second threshold setting step in which the user sets the second threshold before the binarization step.

6. The image comparison method according to claim 1 or 2, characterized in that the outline image is generated by performing an exclusive OR operation based on the difference image and the contracted image in the outline image generation step.

7. The aforementioned step of generating the comparison result image is: A logical inversion operation step that generates an edge inverted image by performing a logical inversion operation based on the aforementioned edge image, A logical AND operation step to generate the comparison result image by performing a logical AND operation based on the difference image and the edge inversion image. The image comparison method according to claim 1 or 2, characterized by including the following:

8. In the difference image generation step, a first difference image is generated based on pixels in the first image whose pixel value is higher than the pixel value in the second image, and a second difference image is generated based on pixels in the second image whose pixel value is higher than the pixel value in the first image. In the shrinking step, the outline image generation step, the candidate image generation step, the edge image generation step, and the comparison result image generation step, processing is performed based on the first difference image and the second difference image, respectively. The image comparison method according to claim 1 or 2, characterized in that the comparison result image generation step generates a first comparison result image obtained by processing based on the first difference image and a second comparison result image obtained by processing based on the second difference image as the comparison result image.

9. The image comparison method according to claim 1 or 2, characterized in that it includes a first threshold setting step in which the user sets the first threshold before the edge image generation step.

10. The first and second images consist of images of multiple ink colors. The image comparison method according to claim 1 or 2, characterized in that the processing of the difference image generation step, the shrinking step, the outline image generation step, the candidate image generation step, the edge image generation step, and the comparison result image generation step is performed for each ink color.

11. The comparison result image generated in the comparison result image generation step includes multiple partial difference result images, each consisting of one or more pixels. The image comparison method according to claim 1 or 2, characterized in that, after the comparison result image generation step, it includes a proximity image merging step of merging two or more adjacent partial difference result images by applying an expansion process to each of the plurality of partial difference result images.

12. The comparison results display step includes displaying the comparison result image on the computer's display unit, The image comparison method according to claim 1 or 2, characterized in that in the comparison result display step, portions where differences are determined to exist between the first image and the second image are displayed in color.

13. In the difference image generation step, a first difference image is generated based on pixels in the first image whose pixel value is higher than the pixel value in the second image, and a second difference image is generated based on pixels in the second image whose pixel value is higher than the pixel value in the first image. The image comparison method according to claim 12, characterized in that in the comparison result display step, the portion corresponding to the first difference image and the portion corresponding to the second difference image are displayed in different colors.

14. An image comparison device that compares a first image, which is a multi-level image, with a second image, which is also a multi-level image. A difference image generation unit generates a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each composed of one or more pixels. A shrinking unit that generates a shrinking image including one or more partial shrinking images by applying a shrinking process to each of the one or more partial difference images, An outline image generation unit generates an outline image including one or more partial outline images, each composed of one or more pixels, by removing the contracted image from the difference image. A candidate image generation unit generates a candidate image that represents a candidate portion of an edge region, which includes one or more partial candidate images, each composed of one or more pixels, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation unit generates an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation unit generates a comparison result image by removing the edge image from the difference image. Includes, Image comparison device characterized in that the edge image generation unit sequentially sets one or more pixels constituting the one or more partial candidate images included in the candidate image as processing target pixels, and sets the processing target pixels and eight pixels surrounding the processing target pixels as nine comparison target pixels, and determines that the processing target pixel is a pixel constituting the edge region if the first condition is met such that the difference between the pixel value of at least one of the nine comparison target pixels in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison target pixels in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold.

15. An image comparison program that compares a first image, which is a multi-level image, with a second image, which is also a multi-level image. On the computer, A difference image generation step of generating a difference image based on the first image and the second image, which is a binary image representing the portion where there is a difference between the first image and the second image, and which includes one or more partial difference images, each composed of one or more pixels; A shrinking step that generates a shrinking image including one or more partial shrinking images by applying a shrinking process to each of the one or more partial difference images, Outline image generation step: Removing the contracted image from the difference image to generate an outline image including one or more partial outline images, each composed of one or more pixels, A candidate image generation step of generating a candidate image that includes one or more partial candidate images, each composed of one or more pixels, which represent candidate portions of an edge region, by removing a partial outline image adjacent to the partial contraction image from the one or more partial outline images mentioned above. An edge image generation step of generating an edge image representing an edge region based on the first image, the second image, and the candidate image, A comparison result image generation step of generating a comparison result image by removing the edge image from the difference image, Make it run, An image comparison program characterized in that, in the edge image generation step, one or more pixels constituting the one or more partial candidate images included in the candidate image are sequentially designated as processing targets, and the processing target pixel and the eight pixels surrounding the processing target pixel are designated as nine comparison targets, and if the difference between the pixel value of at least one of the nine comparison targets in the first image and the pixel value of the processing target pixel in the second image is less than or equal to a first threshold, and the difference between the pixel value of at least one of the nine comparison targets in the second image and the pixel value of the processing target pixel in the first image is less than or equal to the first threshold, then the processing target pixel is determined to be a pixel constituting the edge region.

Images