Inspection method, inspection device, and program

Abstract

To provide an inspection method or the like capable of automatically determining a defect mode in a pixel region of a display panel.SOLUTION: An inspection method of a display panel performed by a computer includes: acquisition step (S11) for acquiring an abnormal part image which is an image including an abnormal part of a pixel region obtained by applying image processing using a background difference method to an inspection image of the pixel region of the display panel; generation step (S12) for generating a label image obtained by converting a region showing the abnormal part into a color corresponding to a defect mode of the abnormal part from the abnormal part image using a learned generation model; and determination step (S13) for determining whether a defect mode of the abnormal part has a possibility of non-luminous oozing failure due to deterioration of a functional layer of the pixel region based on the color of the region in the label image. The defect mode includes a break point failure and an oozing failure.SELECTED DRAWING: Figure 14

Description

[Technical Field]

[0001] The present disclosure relates to an inspection method, an inspection device, and a program. [Background technology]

[0002] During the production process of organic EL display panels, various inspections are carried out to maintain product quality.

[0003] Among the various inspection processes, the inspection process for inspecting display defects in pixel areas involves a DS (Dark Spot) inspection, which checks for black stain-like display defects (hereinafter referred to as seepage defects) caused by defects in the moisture barrier layer.

[0004] In the DS inspection, the operator checks the enlarged image of the pixel area, and if an area where moisture has seeped into the light-emitting layer of that pixel area (hereinafter referred to as the seepage area) is visible, the operator determines whether or not there is a seepage defect based on the size of the seepage area.

[0005] However, since display defects in pixel regions include display defects due to dark dots in addition to exudation defects, the operator must also determine the defect mode (hereinafter referred to as defect mode). Both the determination of the defect mode and the measurement of the size of the exudation region are left to the discretion of the operator. This results in differences in the judgment criteria between operators, and even for the same operator, the judgment criteria may fluctuate over time. This results in problems such as overkill, where a good product is judged as defective, and underkill, where a defective product is judged as good.

[0006] In response to this, for example, Patent Document 1 proposes an image classification method for automatically classifying images of defects detected in visual inspection. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2017-054239 A Summary of the Invention [Problem to be solved by the invention]

[0008] However, Patent Document 1 does not consider exudation defects in pixel regions of organic EL display panels as a type of defect. In other words, even if the image classification method of Patent Document 1 is used, the defect type can only be classified into one of foreign matter, defects, and air bubbles, and exudation defects in pixel regions of organic EL display panels cannot be classified.

[0009] The present disclosure has been made in consideration of the above circumstances, and aims to provide an inspection method and the like that can automatically determine defect modes in pixel regions of a display panel. [Means for solving the problem]

[0010] In order to achieve the above object, an inspection method according to one embodiment of the present disclosure is a method for inspecting a display panel performed by a computer, and includes: an acquisition step of acquiring an abnormality image, which is an image including an abnormal portion of the pixel region obtained by performing image processing using a background subtraction method on an inspection image of a pixel region of the display panel; a generation step of generating a label image from the abnormality image by using a trained generation model, in which an area indicating the abnormal portion is converted into a color corresponding to the defect mode of the abnormal portion; and a determination step of determining, based on the color of the area in the label image, whether the defect mode of the abnormal portion is likely to be a seepage defect that does not emit light due to deterioration of the functional layer of the pixel region, wherein the defect modes include a dark spot defect that does not emit light due to an electrical short circuit or electrical open circuit in the pixel region, and a seepage defect.

[0011] This allows automatic determination of the defect mode in the pixel region of the display panel.

[0012] Furthermore, the method may further include, before the determination step, a CNN (Convolutional Neural Network) determination step of obtaining a classification result indicating a defect mode of the abnormal part from the image of the abnormal part using a trained CNN model, and in the determination step, it may be determined whether the defect mode of the abnormal part is likely to be a seepage defect based on the classification result obtained in the CNN determination step and the color of the area in the label image.

[0013] Also, for example, in the determination step, if the defect mode indicated by the classification result acquired in the CNN determination step matches the defect mode indicated by the color of the region in the label image, the computer determines whether the defect mode of the abnormal portion is likely to be a seepage defect; if they do not match, the computer notifies the operator of the mismatch and has the operator determine whether the defect mode of the abnormal portion is likely to be a seepage defect; and the inspection method may further include a measurement step of measuring whether the size of the region in the label image is equal to or greater than a predetermined value if it is determined in the determination step that the defect mode of the abnormal portion is likely to be a seepage defect; and a seepage defect determination step of determining that the abnormal portion is a seepage defect if the size of the region is equal to or greater than the predetermined value in the measurement step.

[0014] Furthermore, for example, if it is determined in the determination step that the defect mode of the abnormal portion is possibly a seepage defect, the method may include a measurement step of measuring whether the size of the area in the label image is equal to or greater than a predetermined value, and a seepage defect determination step of determining that the abnormal portion is a seepage defect if the size of the area is equal to or greater than the predetermined value in the measurement step.

[0015] Furthermore, for example, the trained generative model is trained using a training image of an abnormal part obtained by performing image processing using a background subtraction method on an inspection image of the pixel area of ​​the display panel, which is prepared as training data, and a training label image in which an area indicating an abnormal part shown in the abnormal part image is converted into a color corresponding to the defect mode of the abnormal part, and the defect mode of the abnormal part indicates a dark spot defect, an exudation defect, or normality.

[0016] Here, for example, the trained generative model may be a neural network model based on Generative Adversarial Networks (GAN), or may be a Pix2Pix neural network model.

[0017] Furthermore, for example, the learning image of an abnormal part may be histogram-adjusted so that the background area, excluding the area showing the abnormal part, is uniformly white.

[0018] These general or specific aspects may be realized as an apparatus, a method, an integrated circuit, a computer program, or a computer-readable recording medium such as a CD-ROM, or may be realized as any combination of a system, a method, an integrated circuit, a computer program, and a recording medium. [Effects of the Invention]

[0019] The present disclosure can provide an inspection method and the like that can automatically determine defect modes in pixel regions of a display panel. [Brief explanation of the drawings]

[0020] [Figure 1] FIG. 1 is a diagram showing a schematic configuration of an inspection system including an inspection device according to an embodiment. [Figure 2A] FIG. 2A is an example of an enlarged image of an inspection image used in a DS inspection according to an embodiment. [Figure 2B]FIG. 2B is an example of an enlarged image of the test image used in the DS test according to the embodiment. [Figure 3] FIG. 3 is a schematic diagram for explaining the mechanism by which the exudation defect occurs. [Figure 4A] FIG. 4A shows an example of the size of the seeping region that appears in an enlarged image of the test image used in the DS test according to the embodiment. [Figure 4B] FIG. 4B shows an example of the size of the seeping region that appears in an enlarged image of the test image used in the DS test according to the embodiment. [Figure 5] FIG. 5 is a diagram showing an example of the hardware configuration of a computer that realizes the functions of the inspection device according to the embodiment by software. [Figure 6] FIG. 6 is a block diagram illustrating an example of a functional configuration of an inspection device according to an embodiment. [Figure 7] FIG. 7 is another example of an enlarged image of the test image of the organic EL display panel used in the DS test according to the embodiment. [Figure 8] FIG. 8 is a diagram showing an example of an image of an abnormal part obtained from an enlarged image of the inspection image shown in FIG. [Figure 9] FIG. 9 is a diagram illustrating an example of a label image according to the embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of an image pair prepared as training data according to the embodiment. [Figure 11A] FIG. 11A is a diagram showing another example of an image pair prepared as training data according to the embodiment. [Figure 11B] FIG. 11B is a diagram showing another example of an image pair prepared as training data according to the embodiment. [Figure 11C] FIG. 11C is a diagram showing another example of an image pair prepared as training data according to the embodiment. [Figure 11D] FIG. 11D is a diagram showing another example of an image pair prepared as training data according to the embodiment. [Figure 12]FIG. 12 is a diagram conceptually showing a method for training a generative model using image pairs such as those shown in FIGS. 10 to 11D as training data. [Figure 13] FIG. 13 is a diagram for conceptually explaining the size measurement of the color label portion of the label image according to the embodiment. [Figure 14] FIG. 14 is a flowchart showing the operation of the inspection device according to the embodiment. [Figure 15] FIG. 15 is a block diagram showing an example of a functional configuration of an inspection device according to a modified example of the embodiment. [Figure 16] FIG. 16 is a flowchart showing a part of the operation of the inspection device according to the modified example of the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0021] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that each of the embodiments described below represents a specific example of the present disclosure. The numerical values, shapes, materials, specifications, components, component placement and connection configurations, steps, and step order shown in the following embodiments are merely examples and are not intended to limit the present disclosure. Furthermore, among the components in the following embodiments, components that are not recited in the independent claims of the present disclosure will be described as optional components. Furthermore, each drawing is not necessarily an exact illustration. In each drawing, substantially identical components are assigned the same reference numerals, and duplicated descriptions may be omitted or simplified.

[0022] (Embodiment) The inspection device and the like according to this embodiment will be described below.

[0023] [1. Inspection system] The inspection device 10 according to this embodiment will be described below with reference to the drawings.

[0024] 1 is a diagram showing a schematic configuration of an inspection system including an inspection device 10 according to the present embodiment. In this embodiment, an example will be described in which the inspection object of the inspection device 10 is an organic EL display panel 30. The inspection object of the inspection device 10 may also be a display panel using quantum dot light emitting diodes (QLEDs).

[0025] The inspection system shown in FIG. 1 includes an inspection device 10, an imaging device 20, a stage 21, and a stage driving unit 22.

[0026] The inspection device 10 is a device for automatically performing a DS inspection to check for exudation defects in the pixel region of the organic EL display panel 30. As described above, the exudation defect is a display defect in the form of a black stain caused by a defect in the moisture barrier layer. More specifically, the exudation defect according to this embodiment is a display defect in which light is not emitted due to deterioration of a functional layer including an emissive layer in the pixel region, and is one of the defect modes. Typically, the functional layer including the emissive layer deteriorates due to moisture. Furthermore, the exudation defect often appears as a display defect in which moisture has exuded into the emissive layer in the pixel region. In other words, the exudation defect appears as a defect in which an area where moisture has exuded into the emissive layer of the organic EL display panel 30 (exudation area) is visible in an enlarged image of the pixel region, and the size of the exudation area is greater than a predetermined value. The mechanism by which the exudation defect occurs will be described later.

[0027] The imaging device 20 captures an image of an area to be inspected on the organic EL display panel 30, and is configured with, for example, a CCD (Charge Coupled Device) or a CMOS (Complementary Metal-Oxide Semiconductor). More specifically, the imaging device 20 captures an image of a pixel area, which is an area to be inspected on the organic EL display panel 30, to obtain an inspection image of the organic EL display panel 30. The imaging device 20 is controlled by the inspection device 10, but may also be controlled by another computer.

[0028] The stage 21 holds the organic EL display panel 30 .

[0029] The stage driving unit 22 is composed of a ball screw, a guide rail, and a motor, and moves the stage 21 relative to the imaging device 20. The stage driving unit 22 is controlled by the inspection device 10, but may also be controlled by another computer.

[0030] 2A and 2B are examples of enlarged images of the test image used in the DS test according to this embodiment. FIG. 2A shows an example of a case where an exudation region that causes an exudation defect is shown in an enlarged image 91 of a pixel region of the organic EL display panel 30. FIG. 2B shows an example of a case where a dark dot (dark dot region) that causes a dark dot defect is shown in an enlarged image 92 of a pixel region of the organic EL display panel 30. Note that a dark dot is a light-emitting pixel that does not emit light (lights up) due to an electrical short or open circuit in the pixel region, but may also include a light-emitting pixel with low light emission brightness.

[0031] When an operator performs a DS inspection by themselves, they will identify whether or not a bleeding area or a dark spot appears in the enlarged image of the pixel area by its outline shape. A bleeding area is an area where moisture has leaked into the light-emitting layer of the pixel area, and because this area of ​​moisture leakage grows over time, it has the characteristic of having a smooth outline, making it distinguishable from a dark spot area, which does not have a smooth outline. However, the outline is difficult to distinguish, and some operators may mistakenly identify a bleeding area as a dark spot area.

[0032] FIG. 3 is a schematic diagram illustrating the mechanism by which exudation defects occur. FIG. 3 schematically illustrates an example of a cross-sectional view of a pixel region of an organic EL display panel 30. As shown in FIG. 3, the organic EL display panel 30 includes a glass substrate 311, a thin-film transistor layer 312 formed on the glass substrate 311, a light-emitting layer 313 formed on the thin-film transistor layer 312, and a protective film 314 formed on the light-emitting layer 313. The protective film 314 functions as a moisture barrier layer for blocking water. In addition, the organic EL display panel 30 includes an upper substrate formed on the protective film 314 via a filler 315 such as an adhesive. In the example shown in FIG. 3, the upper substrate is formed of a color filter layer 316 and a glass substrate 317. The color filter layer 316 includes a black matrix (BM) that separates the pixel regions and the pixel regions. The upper substrate may be a substrate made of a flexible polarizing plate or the like.

[0033] As shown in FIG. 3 , suppose that a foreign substance 320 is introduced into the protective film 314 of the organic EL display panel 30 during its formation, creating gaps in the protective film 314. In other words, suppose that a defect occurs in the moisture barrier layer of the organic EL display panel 30. Then, moisture, indicated as HO in FIG. 3 , descends from the color filter layer 316 and other layers onto the protective film 314. The moisture may include moisture contained in the color filter layer 316 as well as moisture contained in the filler 315. The moisture that descends onto the protective film 314 then penetrates into the gaps in the protective film 314 and is adsorbed onto the light-emitting layer 313. Thus, the moisture that descends onto the protective film 314 penetrates into the protective film 314 via penetration path 321 and is adsorbed onto the light-emitting layer 313. Furthermore, the moisture adsorption region progresses, for example, in the direction of arrow 322, i.e., along the light-emitting layer 313. In this way, the region from which moisture seeps into the light-emitting layer in the pixel region expands. Since heat is the dominant factor in promoting the progression, the seepage area expands with temperature and time. Therefore, the size of the seepage area that will cause a product to be considered defective during DS inspection is determined based on the size of the seepage area that will cause the product to be considered non-defective at the end of the product's lifespan. Here, the size of the seepage area that will cause a product to be defective during DS inspection is, for example, on the order of several tens of microns.

[0034] 4A and 4B show an example of the size of the bleeding area that appears in an enlarged image of the test image used in the DS test according to the present embodiment. 4A and 4B show the size of the bleeding area that appears in an enlarged image of the pixel area of ​​the organic EL display panel 30.

[0035] When an operator performs a DS inspection by himself, he measures the size of the exudation area shown in the enlarged image of the pixel area. However, as shown in Figures 4A and 4B, the exudation area is partially obscured by the boundary between pixels, i.e., the BM, making it difficult for the operator to recognize the edge of the exudation area. As a result, variations in the measurement of the size of the exudation area occur depending on the operator.

[0036] [1-1. Hardware configuration of inspection device 10] Before describing the functional configuration of the inspection device 10 according to this embodiment, an example of the hardware configuration of the inspection device 10 according to this embodiment will be described with reference to FIG.

[0037] FIG. 5 is a diagram showing an example of the hardware configuration of a computer 1000 that realizes the functions of the inspection device 10 according to this embodiment by software.

[0038] 5, the computer 1000 is a computer including an input device 1001, an output device 1002, a CPU 1003, an internal storage 1004, a RAM 1005, a GPU 1006, a reading device 1007, a transmitting / receiving device 1008, and a bus 1009. The input device 1001, the output device 1002, the CPU 1003, the internal storage 1004, the RAM 1005, the GPU 1006, the reading device 1007, and the transmitting / receiving device 1008 are connected by the bus 1009.

[0039] The input device 1001 is a user interface device such as an input button, a touch pad, a touch panel display, etc., and accepts user operations. Note that the input device 1001 may be configured to accept voice operations, remote operations using a remote control, etc., in addition to accepting user touch operations.

[0040] The output device 1002 is also used as the input device 1001, and is configured by a touch pad or a touch panel display, etc., and notifies the user of information that should be made known to the user.

[0041] The internal storage 1004 is a flash memory or the like. The internal storage 1004 may also store in advance at least one of a program for realizing the functions of the inspection device 10 and an application that utilizes the functional configuration of the inspection device 10. The internal storage 1004 may also store a neural network model (such as a generative model), acquired learning data, parameters for an intermediate layer of the model, procedures for performing image processing such as background subtraction, procedures for performing judgments such as a non-dark spot judgment and a DS judgment, which will be described later, and the like.

[0042] The RAM 1005 is a random access memory, and is used to store data and the like when a program or application is executed.

[0043] The GPU 1006 is a graphics processing unit that copies programs, applications, and data stored in the internal storage 1004 to dedicated RAM built into the GPU, and performs image processing according to the instructions contained in the programs and applications.

[0044] The reading device 1007 reads information from a recording medium such as a USB (Universal Serial Bus) memory. The reading device 1007 reads the above-mentioned programs and applications from a recording medium on which the programs and applications are recorded, and stores the programs and applications in the built-in storage 1004.

[0045] The transmitting / receiving device 1008 is a communication circuit for performing wireless or wired communication. The transmitting / receiving device 1008 may communicate with, for example, a server device connected to a network, download the above-mentioned programs and applications from the server device, and store them in the built-in storage 1004.

[0046] The CPU 1003 is a central processing unit that copies programs and applications stored in the internal storage 1004 to the RAM 1005, and sequentially reads and executes instructions contained in the programs and applications from the RAM 1005.

[0047] [1-2. Functional configuration of inspection device 10] Next, each functional configuration of the inspection device 10 according to this embodiment will be described with reference to FIG.

[0048] FIG. 6 is a block diagram showing an example of the functional configuration of the inspection device 10 according to this embodiment.

[0049] 6, the inspection device 10 includes an image acquisition unit 101, a label image generation unit 102, a non-dark spot determination unit 103, a size measurement unit 104, and a DS determination unit 105. Note that the size measurement unit 104 and the DS determination unit 105 are not essential to the inspection device 10, and may be provided externally.

[0050] [1-2-1. Image acquisition unit 101] The image acquisition unit 101 acquires an abnormal part image, which is an image including an abnormal part of a pixel region obtained by performing image processing using a background subtraction method on an inspection image of a pixel region of the organic EL display panel 30.

[0051] In this embodiment, the image acquisition unit 101 acquires an inspection image of a pixel region of the organic EL display panel 30 used for the DS inspection from the imaging device 20. The image acquisition unit 101 also performs image processing using a background subtraction method on the acquired inspection image to generate an abnormal part image, which is a background subtraction image including an abnormal part of the pixel region of the organic EL display panel 30. The image acquisition unit 101 can realize various functions, such as the function of acquiring an inspection image, the function of image processing, and the function of generating an image of an abnormal part, by having a processor execute a control program stored in a memory in a computer that realizes the functions of the inspection device 10.

[0052] Here, background subtraction is image processing that compares an observed image with a background image to extract objects present in the observed image that are not present in the background image. In this embodiment, image processing is performed to extract only the abnormal portion by subtracting normal luminance distribution information from the abnormal pixel image and extracting only the abnormal portion by subtracting normal luminance distribution information from the abnormal pixel image. The abnormal pixel image is an image that includes abnormal pixels such as bleeding regions and dark spots, and is an enlarged image of an inspection image of the pixel region of the organic EL display panel 30. The normal pixel image is an image that includes normal pixels that do not include bleeding regions and dark spots, and is an enlarged image of an inspection image of the pixel region of the organic EL display panel 30.

[0053] The normal pixel image is an image that does not contain an abnormal portion at a position different from the abnormal portion of the test image in the pixel region of the organic EL display panel 30. In other words, the normal pixel image is an image that has the same scale as the image of the region of the test image that contains the abnormal portion. The normal pixel image may be an image obtained from the test image, or may be an image that has the same scale as the test image and is prepared in advance.

[0054] An example of generating an abnormal area image using the background subtraction method will be described below with reference to Figs. 7 and 8. Fig. 7 shows another example of an enlarged image of the test image of the organic EL display panel 30 used in the DS test according to this embodiment. Fig. 7(a) shows an abnormal pixel image 41 including an abnormal area 31a, which is a seepage area, and Fig. 7(b) shows a normal pixel image 42. Fig. 8 shows an example of an abnormal area image 43 obtained from the enlarged image of the test image shown in Fig. 7.

[0055] That is, the image acquisition unit 101 performs image processing using background subtraction to take the difference between the abnormal pixel image 41 shown in Fig. 7(a) and the normal pixel image 42 shown in Fig. 7(b), thereby removing normal brightness distribution information from the abnormal pixel image and extracting only the abnormal portion. As a result, the image acquisition unit 101 can generate an abnormal portion image 43 including an abnormal portion 31b as shown in Fig. 8. The abnormal portion image 43 shown in Fig. 8 is an image in which only the abnormal portion 31a shown in Fig. 7(a) has been extracted.

[0056] In this way, the image acquisition unit 101 can acquire an image of the abnormal area.

[0057] [1-2-2. Label image generation unit 102] The label image generation unit 102 uses a trained generative model to generate a label image from the image of the abnormal portion acquired by the image acquisition unit 101, by converting an area indicating the abnormal portion into a color corresponding to the defect mode of the abnormal portion. The label image generation unit 102 can realize the function of generating a label image using the trained generative model by having a processor execute a control program stored in memory in a computer that realizes the functions of the inspection device 10. The trained generative model is generated by learning from training images of the abnormal portion and training label images prepared as training data. The training images of the abnormal portion are images obtained by performing image processing using a background subtraction method on an inspection image of the pixel region of the organic EL display panel 30. The training label image is an image in which an area indicating the abnormal portion in the image of the abnormal portion is converted into a color corresponding to the defect mode of the abnormal portion. The defect mode of the abnormal portion indicates a dark spot defect, an exudation defect, or normality.

[0058] Fig. 9 is a diagram showing an example of a label image 44 according to the present embodiment. The label image 44 shown in Fig. 9 is an image including a color label portion 31c obtained by converting the region of the abnormal portion 31b in the abnormal portion image 43 shown in Fig. 8 into a color region (color label) corresponding to the defect mode of the abnormal portion 31b.

[0059] In this embodiment, the labeled image generating unit 102 generates labeled image 44 shown in FIG. 9, which includes color labeled portion 31c obtained by converting the region of abnormal portion 31b included in abnormal portion image 43 shown in FIG. 8 into a color label corresponding to the defect mode of abnormal portion 31b. As described above, the defect mode of abnormal portion 31b is, for example, a seepage defect, a dark dot defect, or normal. In addition, in color labeled portion 31c, the defective portion of the region of abnormal portion 31b that is shielded by the boundary between pixels, i.e., BM, is complemented with a color corresponding to the defect mode.

[0060] In this embodiment, a generative model is used, which is a neural network model that complements one of a pair of training images with the other, in order to generate the labeled image 44. Here, the generative model according to this embodiment is, for example, a Pix2Pix neural network model. Pix2Pix is ​​a generative model that automatically extracts relationships hidden between a pair of training images using a neural network and complements one of the pair of images using the extracted relationships. The generative model according to this embodiment may have any configuration as long as it is a neural network model such as a GAN (Generative Adversarial Networks) that performs generative adversarial learning of image pairs. In other words, the generative model according to this embodiment may have any configuration as long as it is a neural network model based on a GAN.

[0061] Next, an example of a method for learning a generative model that generates a labeled image from an abnormal area image will be described.

[0062] FIG. 10 is a diagram illustrating an example of an image pair prepared as training data according to this embodiment. An abnormality image 61 and its label image 62 shown in FIG. 10 are an example of an image pair prepared as training data. (a) of FIG. 10 illustrates an abnormality image 61 including an abnormality 61a, which is a seepage region. (b) of FIG. 10 illustrates a label image 62 including a color label portion 62b in which the region of the abnormality 61a in the abnormality image 61 is converted (replaced) with a color (hatched in the figure) corresponding to the case where the defect mode of the abnormality 61a in the abnormality image 61 is a seepage defect. In the color label portion 62b, the boundary between pixels in the abnormality 61a, i.e., the missing portion shielded by the BM, is converted to a color corresponding to the case where the defect mode of the abnormality 61a is a seepage defect.

[0063] 11A to 11D are diagrams showing other examples of image pairs prepared as training data according to this embodiment.

[0064] Fig. 11A(a) shows an example of a normal image, i.e., an abnormality image without any abnormalities. Fig. 11A(b) shows a label image when the defect mode of the abnormality image is normal. Since the abnormality image shown in Fig. 11A(a) does not have any abnormalities, the label image in Fig. 11A(b) is an image that includes a white color label portion indicating normality.

[0065] Fig. 11B(a) shows an example of an abnormality image including an abnormality that is a seepage region, and Fig. 11B(b) shows a labeled image including color-labeled portions drawn with hatching according to the defect mode of the abnormality image, which is a seepage defect.

[0066] Fig. 11C (a) shows an example of an abnormality image including a dark spot abnormality, and Fig. 11C (b) shows a label image including color-labeled portions drawn with hatching according to the defect mode of the abnormality image, which is a dark spot defect.

[0067] Fig. 11D(a) shows an example of an abnormality image including an abnormality where a seepage region and a dark spot are mixed, and Fig. 11D(b) shows a labeled image including color-labeled portions drawn with hatching according to the defect modes of the abnormality in the abnormality image.

[0068] Fig. 12 is a diagram conceptually showing a method for training a generative model using image pairs such as those shown in Fig. 10 to Fig. 11D as training data. In the example of the input abnormal area image in Fig. 12, in order to conceptually represent the abnormal area image, pixels are indicated by white frames, and the abnormal area is indicated by hatching except for the part that is occluded by the BM.

[0069] When training a generative model that generates a labeled image from an abnormality image, first, multiple image pairs consisting of an abnormality image and a labeled image are prepared, as shown in FIGS. 10 to 11D. Next, as shown in FIG. 12, a GAN-based neural network that constitutes the generative model is trained under supervision so that the input is an abnormality image and the output is a corresponding labeled image. This makes it possible to obtain a generative model that generates a labeled image when an abnormality image is input.

[0070] 12 conceptually illustrates a neural network constituting the generative model according to this embodiment, but as described above, it may be configured as a neural network such as a GAN that performs generative adversarial learning of image pairs such as pix2pix. Furthermore, the neural network constituting the generative model according to this embodiment may have any configuration as long as it is capable of obtaining a generative model that generates a labeled image when an image of an abnormal part is input.

[0071] In this way, the label image generating unit 102 can generate a label image from the abnormal part image acquired by the image acquiring unit 101, including color label portions that are depicted in colors corresponding to the defect modes of the abnormal parts of the abnormal part image.

[0072] The training images of abnormal areas among the image pairs prepared when training a generative model may be histogram-adjusted so that the background area, excluding the area showing the abnormal area, is uniformly blown out, i.e., uniformly converted to a uniform white color represented by a gradation value of 255. This allows the generative model to extract in advance only the information necessary for generating labeled images, thereby training the generative model to become a generative model that can generate labeled images with higher accuracy. Note that when histogram-adjusted training images of abnormal areas are used for training, histogram-adjusted images can also be used for the inspection images used in inspection.

[0073] [1-2-3. Non-dark spot determination section 103] The non-dark dot determination unit 103 determines whether the defect mode of the abnormal portion is likely to be a seepage defect, where moisture has seeped into the light-emitting layer of the pixel region, based on the color of the region in the label image. Note that the non-dark dot determination unit 103 can achieve its determination function by, for example, having a processor execute a control program stored in memory in a computer that realizes the functions of the inspection device 10.

[0074] In this embodiment, the non-dark dot determination unit 103 acquires the label image generated by the label image generation unit 102 and determines whether the defect mode of the abnormal portion corresponding to the color label portion is a dark dot defect based on the color of the color label portion included in the acquired label image. For example, assume that the non-dark dot determination unit 103 acquires the label image 44 shown in Fig. 9 generated by the label image generation unit 102. In this case, the non-dark dot determination unit 103 determines that the defect mode of the abnormal portion 31b corresponding to the color label portion 31c is not a dark dot defect based on the color of the color label portion 31c of the acquired label image 44.

[0075] In this way, the non-dead spot determining unit 103 can determine that the defect mode of the abnormal portion 31b corresponding to the color label portion 31c of the acquired label image 44 is likely to be a seepage defect.

[0076] In addition, if the defect mode of the abnormal part 31b corresponding to the color label part 31c is normal, the non-dark spot judgment unit 103 will determine that the inspection result of the DS inspection is OK (good product), and the processing of the abnormal part image will be terminated.

[0077] [1-2-4. Size measurement unit 104] When the non-dark dot determination unit 103 determines that the defect mode of the abnormal portion is likely to be a seepage defect, the size measurement unit 104 measures whether the size of the area of ​​the color label portion in the label image is equal to or greater than a predetermined value. Note that the size measurement unit 104 can realize the measurement function by image processing, for example, by having a processor execute a control program stored in a memory in a computer that realizes the functions of the inspection device 10.

[0078] FIG. 13 is a diagram for conceptually explaining the size measurement of the color label portion 31c of the label image 44 according to the present embodiment.

[0079] In this embodiment, the size measurement unit 104 acquires a label image 44, such as that shown in FIG. 9, for which the non-dark dot determination unit 103 has determined that the defect mode is not a dark dot defect, and measures the size of the area of ​​the color labeled portion 31c in the label image 44 by image processing, for example, as shown in FIG. 13. In the example shown in FIG. 13, the size measurement unit 104 measures X μm and Y μm, i.e., the length (vertical) size and width (horizontal) size, as the size of the area of ​​the color labeled portion 31c in the label image 44. X μm and Y μm shown in FIG. 13 are, for example, 57 μm and 83 μm.

[0080] In this way, the size measurement unit 104 can automatically measure the size of the area of ​​the color label portion 31c of the label image 44, and can measure whether the size of the seeping area is equal to or larger than a predetermined value.

[0081] [1-2-5.DS determination section 105] The DS determination unit 105 determines that the abnormal portion is an exudation defect when the size of the area of ​​the color label portion in the label image is equal to or larger than a predetermined value in the size measurement unit 104. Note that the DS determination unit 105 can realize the above-mentioned determination function by having a processor execute a control program stored in a memory in a computer that realizes the functions of the inspection device 10.

[0082] In this embodiment, for example, if the size of the area of ​​the color label portion 31c in the label image 44 shown in FIG. 13 is equal to or larger than a predetermined value, the DS determination unit 105 determines that the defect mode of the abnormal portion 31a corresponding to the color label portion 31c is a seepage defect.

[0083] In this way, the DS determination unit 105 can automatically determine whether the defect mode of the abnormal part corresponding to the color label part is an exudation defect or not, based on the color of the color label part of the label image and the size of the area.

[0084] [1-3. Operation of Inspection Device 10] An example of the operation of the inspection device 10 configured as above will be described below.

[0085] FIG. 14 is a flowchart showing the operation of the inspection device 10 according to this embodiment.

[0086] First, the inspection device 10 acquires an image of an abnormal portion (S11). More specifically, the image acquisition unit 101 acquires the abnormal portion image, which is a background subtraction image including the abnormal portion of the pixel region obtained by performing image processing using a background subtraction method on the inspection image of the pixel region of the organic EL display panel 30. For example, the image acquisition unit 101 acquires an abnormal portion image 43, which is an image including the abnormal portion 31b, as shown in FIG. 8.

[0087] Next, the inspection device 10 generates a labeled image using the trained generative model (S12). More specifically, the label image generator 102 uses the trained generative model to generate a labeled image by converting a region indicating the abnormal portion from the image of the abnormal portion acquired in step S11 into a color corresponding to the defect mode of the abnormal portion. For example, the label image generator 102 uses the trained generative model to convert the region of the abnormal portion 31b from the image of the abnormal portion 43 shown in FIG. 8 into a color label corresponding to the defect mode of the abnormal portion 31b, thereby generating a labeled image 44 including a color labeled portion 31c as shown in FIG. 9.

[0088] Next, the inspection device 10 determines whether the defect mode of the abnormal portion is likely to be a seepage defect (S13). More specifically, the non-dark dot determination unit 103 determines whether the defect mode of the abnormal portion is likely to be a seepage defect, in which moisture has seeped into the light-emitting layer of the pixel region, based on the color of the region in the label image generated in step S12. For example, the non-dark dot determination unit 103 may determine, based on the color of the color label portion 31c of the label image 44 shown in FIG. 9, that the defect mode of the abnormal portion 31b corresponding to the color label portion 31c is not a dark dot defect.

[0089] If, in step S13, it is determined that the defect mode of the abnormal portion is likely to be a seepage defect (Yes in S13), the inspection device 10 measures the size of the area indicating the abnormal portion in the label image generated in step S12 (S14). More specifically, if it is determined in step S13 that the defect mode of the abnormal portion is likely to be a seepage defect, the size measurement unit 104 measures whether the size of the area of ​​the color labeled portion in the label image is equal to or greater than a predetermined value. For example, if it is determined from the label image 44 shown in FIG. 9 that the defect mode is a dark dot defect and abnormal, the size measurement unit 104 measures the size of the area of ​​the color labeled portion 31c in the label image 44 by image processing, as shown in FIG. 13. Note that, if, in step S13, it is not likely that the defect mode of the abnormal portion is a seepage defect (No in S13), the inspection device 10 ends this process, i.e., the DS inspection.

[0090] Next, the inspection device 10 determines whether the size of the area measured in step S14 is equal to or larger than a predetermined value (S15). More specifically, in step S14, the DS determination unit 105 determines whether the size of the area of ​​the color label portion in the label image measured in step S13 is equal to or larger than a predetermined value. In this embodiment, the predetermined value is a value on the order of several tens of microns.

[0091] In step S15, if the size of the area is equal to or larger than a predetermined value (Yes in S15), the inspection device 10 determines that the abnormal portion of the abnormal portion image acquired in step S11 is a seepage defect (S16). More specifically, if the size of the area of ​​the color labeled portion in the label image is determined to be equal to or larger than a predetermined value in step S15, the DS determination unit 105 determines that the abnormal portion is a seepage defect. For example, if the size of the area of ​​color labeled portion 31c in label image 44 shown in FIG. 13 is equal to or larger than a predetermined value, the DS determination unit 105 determines that the defect mode of abnormal portion 31a corresponding to color labeled portion 31c is a seepage defect.

[0092] On the other hand, if the size of the area is not equal to or larger than the predetermined value in step S15 (No in S15), the inspection device 10 determines that the abnormal portion in the abnormal portion image acquired in step S11 is not a seepage defect (S17). More specifically, if the size of the color-labeled portion in the label image is determined to be smaller than the predetermined value in step S15, the DS determination unit 105 determines that the abnormal portion is not a seepage defect. For example, if the size of the color-labeled portion 31c in the label image 44 shown in FIG. 13 is smaller than the predetermined value, the DS determination unit 105 determines that the defect mode of the abnormal portion 31a corresponding to the color-labeled portion 31c is not a seepage defect. Then, the DS determination unit 105 determines that the pixel area of ​​the organic EL display panel 30 having the abnormal portion image including the abnormal portion 31a is a non-defective product.

[0093] [1-4. Effects, etc.] The inspection device 10 etc. of this embodiment acquires an abnormal part image, which is an image including an abnormal part of the pixel region, by performing image processing using background subtraction on an inspection image of a pixel region of the organic EL display panel 30. Furthermore, the inspection device 10 etc. of this embodiment uses a trained generative model to generate a label image in which an area indicating an abnormal part from the acquired abnormal part image is converted into a color corresponding to the defect mode of the abnormal part. Then, based on the color of the area in the generated label image, it is determined whether the defect mode of the abnormal part is likely to be a seepage defect.

[0094] In this way, the inspection device 10 etc. of the present embodiment can use a trained generative model to generate a labeled image in which the defective part partially occluded by the BM in the abnormal part is complemented from the abnormal part image, which is a background subtraction image including the abnormal part in the pixel region, and which is color-coded for each defect mode. In other words, the inspection device 10 etc. of the present embodiment can automatically determine the defect mode in the pixel region of the organic EL display panel 30.

[0095] Here, when the inspection device 10 etc. of this embodiment determines that the defect mode of the abnormal part may be a seepage defect, it measures whether the size of the area of ​​the color label part in the label image is equal to or larger than a predetermined value.

[0096] Since such a size, i.e., the size of the color label indicating the exudation region where the missing portion has been filled and which may be a seepage defect, can be easily measured, it is possible to automatically and accurately determine whether the abnormal portion in the pixel region is a seepage defect or not. In other words, the inspection device 10 etc. of the present embodiment can automatically determine the defect mode in the pixel region of the organic EL display panel 30.

[0097] Therefore, since it is possible to automate the defect mode judgment and the measurement of the size of the seepage area, which were previously left to the judgment of the operator in DS inspection, it is possible to solve the problems of differences in judgment criteria between operators and fluctuations in judgment criteria over time even for the same operator. As a result, it becomes possible to perform stable DS inspections using the same criteria, which not only significantly improves inspection efficiency but also solves the problems of overkill and underkill.

[0098] According to the inspection device 10 etc. of the present embodiment, a method of generating a labeled image uses a trained generative model such as a generative model configured by a neural network.

[0099] That is, the trained generative model is trained using training images of abnormal areas and training label images prepared as training data. The training images of abnormal areas are background subtraction images obtained by performing image processing using background subtraction on an inspection image of the pixel region of the organic EL display panel 30. The training label images are images in which the areas showing the abnormal areas shown in the abnormal area images are converted into colors corresponding to the defect modes of the abnormal areas. The defect modes of the abnormal areas indicate dark dot defects, exudation defects, or normality. Here, the trained generative model is a neural network model based on GAN, and may be, for example, a Pix2Pix neural network model.

[0100] This allows the generative model to be trained in advance by preparing paired images of learning background subtraction images (learning abnormality images) showing various defect modes such as exudation defects and dark dot defects and learning label images (in which missing portions are filled in and color-coded for each defect mode). Therefore, using a generative model that is a neural network model good at recognition and completion, it is possible to automatically generate, with high accuracy, labeled images in which missing portions are filled in and color-coded for each defect mode from abnormality images obtained by performing image processing using the background subtraction method. Therefore, using the labeled images, it is possible to automatically and accurately determine the defect mode and measure the size of the exudation region, thereby enabling automatic and accurate determination of whether the abnormal portion of a pixel region is an exudation defect.

[0101] The training images of abnormal areas may be histogram-adjusted so that the background area, excluding the area showing the abnormal area, is uniformly overexposed, i.e., converted to a uniform white color indicated by a gradation value of 255. This allows the generative model to extract in advance only the information necessary for generating labeled images, and therefore allows training to result in a generative model that can generate labeled images with higher accuracy.

[0102] (Variation) In the above embodiment, a trained generative model is used to generate a label image from an image of an abnormal part in a background subtraction image, and the generated label image is used to determine whether the defect mode of the abnormal part is a seepage defect. In order to further improve the accuracy of determining whether the defect mode of the abnormal part is a seepage defect, a model different from the generative model, i.e., a CNN (Convolutional Neural Network) model, may be used to determine the defect mode of the abnormal part and double-check the determination result. The following describes this case, focusing on the differences from the above embodiment.

[0103] [2-1. Inspection device 10A] 15 is a block diagram showing an example of the functional configuration of an inspection device 10A according to a modification of the present embodiment. The inspection device 10A according to this modification is different from the inspection device 10 shown in FIG. 6 in that a CNN determination unit 106A is added and the function of the non-dark dot determination unit 103A is different.

[0104] [2-1-1.CNN determination unit 106A] The CNN determination unit 106A can determine the defect mode of the abnormal part using a model different from the generative model. More specifically, the CNN determination unit 106A acquires a classification result indicating the defect mode of the abnormal part from the image of the abnormal part using a trained CNN model. Note that the CNN determination unit 106A can realize the above-mentioned determination function by having a processor execute a control program stored in a memory in a computer that realizes the functions of the inspection apparatus 10A.

[0105] The trained CNN model is trained as follows. First, multiple training images of abnormal areas, which are background subtraction images, are prepared as training data for each class number corresponding to a defect mode. The training images of abnormal areas are the training images of abnormal areas in the above-described embodiment, and are background subtraction images obtained by performing image processing using a background subtraction method on an inspection image of the pixel region of the organic EL display panel 30. The class number can be determined, for example, as 0 for normal, 1 for exudation defects, 2 for dark dot defects, or 4 for a mixture of exudation defects and dark dot defects. The combination of exudation defects and dark dot defects may be assigned the numbers 1 and 2. The CNN model is trained so that the training images of abnormal areas are used as input and the class numbers are used as output. Using the trained CNN model obtained through this training, the CNN determination unit 106A can obtain a classification result indicating the defect mode of the abnormal area from the abnormal area image.

[0106] [2-1-2. Non-dark spot determination section 103A] In addition to the functions described in the above embodiment, the non-dark dot determination unit 103A has the function of double-checking the determination results. In other words, the non-dark dot determination unit 103A further determines whether the defect mode of the abnormal portion is likely to be an exudation defect based on the classification result acquired by the CNN determination unit 106A and the color of the area in the label image. In this way, the automatic determination results can be double-checked, thereby further improving the accuracy of the exudation defect determination. Note that the non-dark dot determination unit 103A can realize its determination function by, for example, having a processor execute a control program stored in memory in a computer that realizes the functions of the inspection device 10A.

[0107] More specifically, if the defect mode indicated by the classification result acquired by the CNN determination unit 106A matches the defect mode indicated by the color of the region in the label image, the non-dark dot determination unit 103A determines whether the defect mode of the abnormal portion is likely to be a seepage defect. On the other hand, if the defect mode indicated by the classification result acquired by the CNN determination unit 106A does not match the defect mode indicated by the color of the region in the label image, the non-dark dot determination unit 103A notifies the operator of this mismatch. This allows the non-dark dot determination unit 103A to allow the operator to determine whether the defect mode of the abnormal portion is likely to be a seepage defect. The operator simply determines whether the defect mode of the abnormal portion is likely to be a seepage defect, i.e., whether a seepage region exists in the image of the abnormal portion. If the defect mode of the abnormal portion is likely to be a seepage defect, i.e., if a seepage region exists in the image of the abnormal portion, the operator measures the size of the seepage region and determines whether it is greater than or equal to a predetermined value.

[0108] [2-2. Operation of Inspection Device 10A] An example of the operation of the inspection device 10A configured as above will be described below.

[0109] Fig. 16 is a flowchart showing part of the operation of the inspection device 10A according to a modified example of the present embodiment. The operation of the inspection device 10A according to this modified example differs from the operation of the inspection device 10 shown in Fig. 14 in the processing content of step S12. More specifically, in this modified example, the processing of step S12A shown in Fig. 16 is performed instead of the processing of step S12 shown in Fig. 14.

[0110] First, the inspection device 10A acquires an image of an abnormal part (S11). More specifically, the image acquisition unit 101 acquires the image of an abnormal part, which is a background subtraction image including an abnormal part of the pixel region, obtained by performing image processing using a background subtraction method on the inspection image of the pixel region of the organic EL display panel 30.

[0111] Next, in step S12A, the inspection device 10A generates a labeled image using the trained generative model (S121). More specifically, the label image generator 102 uses the trained generative model to generate a labeled image from the image of the abnormal part acquired in step S11, by converting the area indicating the abnormal part into a color corresponding to the defect mode of the abnormal part.

[0112] Furthermore, in step S12A, the inspection device 10A uses the trained CNN model to acquire a classification result indicating the defect mode of the abnormal portion from the image of the abnormal portion acquired in step S11 (S122). More specifically, the CNN determination unit 106A uses the trained CNN model to acquire a classification result indicating the defect mode of the abnormal portion from the image of the abnormal portion acquired in step S11.

[0113] Next, in step S12A, the inspection device 10A determines whether the defect mode indicated by the classification result obtained in step S122 matches the defect mode indicated by the color indicating the abnormal part in the label image generated in step S121 (S123).

[0114] In step S123, if the defect modes match (Yes in S123), the process proceeds to step S13 shown in Fig. 14. In this way, by double-checking the defect mode determination of the abnormal part, the accuracy of the exudation defect determination can be further improved.

[0115] On the other hand, if the defect modes do not match in step S123 (No in S123), the operator is notified that the defect modes do not match (S124). In this way, by double-checking the determination of the defect mode of the abnormal part and allowing the operator to make a decision if they do not match, it is possible to more accurately determine the exudation defect.

[0116] [2-3. Effects, etc.] In this modification, in order to further improve the determination accuracy of the inspection apparatus 10 of the embodiment, a CNN model is prepared in addition to the generative model, and is trained using training images of abnormal parts, which are background subtraction images classified in advance for each defect mode. As a result, when an abnormal part image, which is a background subtraction image, is input, the trained CNN can output a class number corresponding to the defect mode.

[0117] Furthermore, the inspection device 10A, etc. of this modified example uses a trained generative model to generate a label image from an image of an abnormal part, and uses a trained CNN to acquire a classification number corresponding to the defect mode of the abnormal part in the image of the abnormal part. The inspection device 10A, etc. of this modified example can double-check the automatic determination result by comparing the defect mode determined using the label image generated using the trained generative model with the classification result indicating the defect mode obtained using the trained CNN model. This can further improve the accuracy of determining exudation defects. Note that, if the double-check results show that the defect mode determined using the generative model does not match the classification result indicating the defect mode, an operator can make a judgment. Having an operator make a judgment in the case of a mismatch allows for more accurate determination of exudation defects.

[0118] (Other embodiments) While the inspection device and inspection method according to the present disclosure have been described above based on the embodiments and modifications, the present disclosure is not limited to these embodiments and modifications. As long as they do not deviate from the gist of the present disclosure, various modifications that a person skilled in the art could conceive of to the embodiments and modifications, and other forms constructed by combining some of the components in the embodiments and modifications, are also included within the scope of the present disclosure.

[0119] The following embodiments may also be included within the scope of one or more aspects of the present disclosure.

[0120] (1) Some of the components constituting the above-mentioned inspection device may be a computer system consisting of a microprocessor, ROM, RAM, GPU, hard disk unit, display unit, keyboard, mouse, etc. A computer program is stored in the RAM or hard disk unit. The microprocessor achieves its function by operating in accordance with the computer program. Here, the computer program is composed of a combination of multiple instruction codes that indicate commands to a computer to achieve a predetermined function.

[0121] (2) Some of the components constituting the above-mentioned inspection device may be configured as a single system LSI (Large Scale Integration). A system LSI is an ultra-multifunctional LSI manufactured by integrating multiple components on a single chip, and specifically, is a computer system configured including a microprocessor, ROM, RAM, GPU, etc. A computer program is stored in the RAM. The system LSI achieves its functions by the microprocessor or the GPU operating in accordance with the computer program.

[0122] (3) Some of the components constituting the above-mentioned inspection device may be composed of an IC card or a standalone module that can be attached to each device. The IC card or the module is a computer system composed of a microprocessor, ROM, RAM, GPU, etc. The IC card or the module may include the above-mentioned ultra-multifunctional LSI. The IC card or the module achieves its functions by the microprocessor or the GPU operating in accordance with a computer program. The IC card or the module may be tamper-resistant.

[0123] (4) Furthermore, some of the components constituting the above-mentioned inspection device may be the computer program or the digital signal recorded on a computer-readable recording medium, such as a flexible disk, hard disk, CD-ROM, MO, DVD, DVD-ROM, DVD-RAM, BD (Blu-ray (registered trademark) Disc), semiconductor memory, etc. Alternatively, the components may be the digital signal recorded on such a recording medium.

[0124] Furthermore, some of the components constituting the above-mentioned determination device may transmit the computer program or the digital signal via a telecommunications line, a wireless or wired communication line, a network such as the Internet, data broadcasting, etc.

[0125] (5) The present disclosure may be embodied as the methods described above, a computer program for implementing these methods on a computer, or a digital signal comprising the computer program.

[0126] (6) The present disclosure may also be directed to a computer system having a microprocessor, a GPU, and a memory, wherein the memory stores the computer program, and the microprocessor or the GPU operates in accordance with the computer program.

[0127] (7) The program or the digital signal may also be implemented by another independent computer system by recording it on the recording medium and transferring it, or by transferring the program or the digital signal via the network, etc.

[0128] (8) In addition, some of the components that make up the above-mentioned inspection device may be performed by a cloud or server device.

[0129] (9) The above-described embodiments and modifications may be combined with each other. [Industrial Applicability]

[0130] The present disclosure can be used in an inspection method, an inspection device, a program, etc. that can automatically determine whether there are any black stain-like display defects caused by defects in the moisture barrier layer in an inspection process that inspects display defects in the pixel area of ​​an organic EL display panel or a display panel that uses quantum dot light-emitting elements. [Explanation of symbols]

[0131] 10, 10A Inspection equipment 20 Imaging device 21 Stages 22 Stage drive unit 30 Organic EL display panel 31a, 31b, 61a Abnormal part 31c, 62b color label part 41 Abnormal pixel image 42 Normal pixel image 43, 61 Abnormal area images 44, 62 Label images 91, 92 Enlarged images 101 Image acquisition unit 102 Label image generation unit 103, 103A Non-dark spot determination section 104 Size Measurement Section 105 DS judgment section 106A CNN judgment section 311, 317 Glass substrate 312 Thin Film Transistor Layer 313 Light-emitting layer 314 Protective film 315 Filler 316 Color filter layer 320 Foreign matter 321 Invasion Pass 322 Arrow 1000 computers 1001 Input Device 1002 Output device 1003 CPU 1004 Internal Storage 1005 RAM 1006 GPU 1007 Reading device 1008 Transmitting and receiving device 1009 Bus

Claims

1. A computer-implemented display panel inspection method, comprising: an acquisition step of acquiring an abnormal part image, which is an image including an abnormal part of the pixel region, by performing image processing using a background subtraction method on an inspection image of the pixel region of the display panel; a generation step of generating, from the image of the abnormal part, a labeled image in which a region indicating the abnormal part is converted into a color corresponding to a defect mode of the abnormal part, using the trained generation model; a determining step of determining whether the defect mode of the abnormal portion is likely to be a seepage defect in which light is not emitted due to deterioration of a functional layer in the pixel region, based on the color of the region in the label image; The defect modes include a dark spot defect in which no light is emitted due to an electrical short circuit or an electrical open circuit in a pixel area, and a seepage defect. Testing method.

2. Furthermore, before the judgment step, a trained CNN (Convolutional Neural Network) a CNN determination step of obtaining a classification result indicating a defect mode of the abnormal portion from the image of the abnormal portion using a model; In the determination step, the classification result obtained in the CNN determination step and the label and determining whether the defect mode of the abnormal portion is likely to be an exudation defect based on the color of the region in the image. The inspection method according to claim 1 .

3. In the determination step, the defect model indicated by the classification result obtained in the CNN determination step is If the code matches the defect mode indicated by the color of the area in the label image, the computer determines whether the defect mode of the abnormal portion is likely to be an exudation defect; If there is no match, notify the operator of the fact that there is no match and have the operator determine whether the defect mode of the abnormal portion is likely to be an exudation defect; The inspection method further comprises: a measuring step of measuring whether or not a size of the region in the label image is equal to or larger than a predetermined value when it is determined in the determining step that the defect mode of the abnormal portion may be an exudation defect; and a seepage defect determination step of determining that the abnormal portion is a seepage defect when the size of the region is equal to or larger than the predetermined value in the measurement step. The inspection method according to claim 2 .

4. a measuring step of measuring whether or not a size of the region in the label image is equal to or larger than a predetermined value when it is determined in the determining step that the defect mode of the abnormal portion may be an exudation defect; and a seepage defect determination step of determining that the abnormal portion is a seepage defect when the size of the region is equal to or larger than the predetermined value in the measurement step. The inspection method according to claim 1 .

5. the trained generative model is trained using a training image of an abnormal part obtained by performing image processing using a background subtraction method on an inspection image of a pixel region of the display panel, which is prepared as training data, and a training label image obtained by converting an area showing an abnormal part shown in the training image of an abnormal part into a color corresponding to a defect mode of the abnormal part; The defect mode of the abnormal portion indicates a dark spot defect, a seepage defect, or a normal state. The inspection method according to any one of claims 1 to 4.

6. The trained generative model is a neural network model based on GAN (Generative Adversarial Networks). The inspection method according to claim 5.

7. The trained generative model is a Pix2Pix neural network model. The inspection method according to claim 5.

8. The learning image of the abnormal part is histogram-adjusted so that the background area excluding the area showing the abnormal part is uniformly white. The inspection method according to claim 5.

9. A computer-based display panel inspection device, an image acquisition unit that acquires an abnormal part image, which is an image including an abnormal part of the pixel region, obtained by performing image processing using a background subtraction method on an inspection image of the pixel region of the display panel; a label image generating unit that generates a label image from the image of the abnormal part by converting a region indicating the abnormal part into a color corresponding to a defect mode of the abnormal part using a trained generation model; a non-dark dot determination unit that determines whether the defect mode of the abnormal portion is likely to be a seepage defect in which light is not emitted due to deterioration of a functional layer in the pixel region, based on the color of the region in the label image; The defect modes include a dark spot defect in which no light is emitted due to an electrical short circuit or an electrical open circuit in a pixel area, and a seepage defect. Inspection equipment.

10. A program for causing a computer to execute a display panel inspection method, an acquisition step of acquiring an abnormal part image, which is an image including an abnormal part of the pixel region, by performing image processing using a background subtraction method on an inspection image of the pixel region of the display panel; a generation step of generating, from the image of the abnormal part, a labeled image in which a region indicating the abnormal part is converted into a color corresponding to a defect mode of the abnormal part, using the trained generation model; a determining step of determining whether the defect mode of the abnormal portion is likely to be a seepage defect in which light is not emitted due to deterioration of a functional layer in the pixel region, based on the color of the region in the label image; The defect modes include a dark spot defect in which no light is emitted due to an electrical short circuit or an electrical open circuit in a pixel area, and a seepage defect. program.

Images