Display panel inspection method

Abstract

The inspection method of the display panel includes: training an artificial intelligence model based on life data of a test display panel for aging characteristics; generating virtual lifetime data of the virtual display panel from the aging characteristics using an artificial intelligence model; and predicting a life of the display panel based on the virtual life data.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to and all rights arising from Korean Patent Application No. 10-2023-0089385, filed on Jul. 10, 2023, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to a display panel inspection method and an inspection device for performing the same. More specifically, the present disclosure relates to a display panel inspection method for predicting the life of the display panel and an inspection device for performing the same. Background Art

[0004] As information technology develops, the importance of display devices as a connection medium between users and information has become prominent. Therefore, the use of display devices such as liquid crystal display devices, organic light emitting display devices, etc. has been increasing.

[0005] In order to display an image to the outside, the display panel includes a plurality of sub-pixels that receive an electrical signal to emit light. Each sub-pixel includes a light-emitting element, and the light-emitting element may age as its use increases. In addition, the characteristics of the light-emitting element may change due to aging. In an embodiment, the light-emitting characteristics of the light-emitting element may change over time. Therefore, various studies have been attempted to predict the life of the light-emitting element (i.e., the life of the display panel) based on the aging time of the light-emitting element. Summary of the invention

[0006] A feature of the present disclosure is to provide a method for inspecting a display panel that can predict the life of the display panel through an artificial intelligence model.

[0007] Another feature of the present disclosure is to provide an inspection device for performing the inspection method of a display panel.

[0008] Embodiments of the present disclosure provide a method for inspecting a display panel, including: training an artificial intelligence model based on life data of a test display panel for aging characteristics; generating virtual life data of a virtual display panel from the aging characteristics using the artificial intelligence model; and predicting the life of the display panel based on the virtual life data.

[0009] In an embodiment, the lifespan data of the test display panel may be a brightness maintenance rate according to an aging time of the test display panel.

[0010] In an embodiment, the artificial intelligence model may include a first artificial intelligence model according to a first aging time and a second artificial intelligence model according to a second aging time different from the first aging time.

[0011] In an embodiment, the aging characteristic may include at least one of an aged color, an aged grayscale, an initial aged brightness, an observed color, an observed grayscale, an initial observed brightness, an aged position, and a temperature.

[0012] In an embodiment, training an artificial intelligence model may include: displaying an aging color and an aging grayscale on a pattern of a test display panel; measuring an initial aging brightness of the pattern; displaying an observed color and an observed grayscale on the test display panel; measuring an initial observed brightness of the pattern; measuring a temperature of the test display panel; determining life data of the test display panel; annotating aging characteristics with the life data of the test display panel; and training the artificial intelligence model with the aging characteristics annotated using the life data of the test display panel.

[0013] In an embodiment, the lifespan data of the test display panel may be measured by displaying an observed color and an observed grayscale on the test display panel.

[0014] In an embodiment, the life data of the test display panel may be measured according to an aging time of displaying an aged color and an aged grayscale.

[0015] In an embodiment, the aging characteristic may include at least one of an aged color, an aged grayscale, an initial aged brightness, an observed color, an observed grayscale, an initial observed brightness, an aged position, a temperature, an aged current, an observed current, and a light efficiency.

[0016] In an embodiment, training an artificial intelligence model may include: displaying an aging color and an aging grayscale on a pattern of a test display panel; measuring an initial aging brightness and an aging current of the pattern; displaying an observed color and an observed grayscale on the test display panel; measuring an initial observed brightness and an observed current of the pattern; measuring a temperature of the test display panel; determining life data of the test display panel; annotating aging characteristics with the life data of the test display panel; and training an artificial intelligence model with the aging characteristics annotated using the life data of the test display panel.

[0017] In an embodiment, the artificial intelligence model may be a regression model.

[0018] In an implementation, the artificial intelligence model may be one of a random forest regression model, an extreme tree regression model, and a gradient boosting regression model.

[0019] In an embodiment, the artificial intelligence model may be generated by fusing at least two regression models.

[0020] In an embodiment, predicting the lifetime of the display panel may include: normalizing the virtual lifetime data and generating normal distribution data; and predicting the lifetime of the display panel based on data corresponding to a predetermined ratio of the normal distribution data.

[0021] In an embodiment, predicting the lifetime of the display panel may include: normalizing the virtual lifetime data and the lifetime data of the test display panel and generating normal distribution data; and predicting the lifetime of the display panel based on data corresponding to a predetermined ratio of the normal distribution data.

[0022] Another embodiment of the present disclosure provides an inspection device, comprising: a memory, storing an artificial intelligence model; and a processor, predicting the lifespan of a display panel through the artificial intelligence model stored in the memory, wherein the processor generates virtual lifespan data of a virtual display panel from aging characteristics through the artificial intelligence model and predicts the lifespan of the display panel based on the virtual lifespan data, and the artificial intelligence model is trained based on the lifespan data of a test display panel for aging characteristics.

[0023] In an embodiment, the artificial intelligence model may include a first artificial intelligence model according to a first aging time and a second artificial intelligence model according to a second aging time different from the first aging time.

[0024] In an embodiment, the aging characteristic may include at least one of an aged color, an aged grayscale, an initial aged brightness, an observed color, an observed grayscale, an initial observed brightness, and a temperature.

[0025] In an embodiment, the aging characteristic may include at least one of an aged color, an aged grayscale, an initial aged brightness, an observed color, an observed grayscale, an initial observed brightness, a temperature, an aged current, an observed current, and a light efficiency.

[0026] In an embodiment, the artificial intelligence model may be a regression model.

[0027] In an embodiment, the artificial intelligence model may be generated by fusing at least two regression models.

[0028] According to the display panel inspection method in the embodiment of the present disclosure, by generating virtual life data of a virtual display panel by an artificial intelligence model, the accuracy of life prediction can be improved even when a relatively small number of test display panels are used.

[0029] However, the effects of the present disclosure are not limited to the above-described effects, and various extensions may be made without departing from the spirit and scope of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and other embodiments, advantages and features of the present disclosure will become more apparent by describing in more detail exemplary embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0031] Figure 1 A flow chart showing an embodiment of a method for inspecting a display panel according to the present disclosure is shown.

[0032] Figure 2 Shown according to Figure 1 A block diagram of a display device of a display panel inspection method.

[0033] Figure 3 Shown according to Figure 1 An implementation method of a display panel inspection method for testing a display area of ​​a display panel.

[0034] Figure 4 Shown according to Figure 1 An implementation method of observing color and observing grayscale of a display panel inspection method.

[0035] Figure 5 It is a graph for explaining life span data.

[0036] Figure 6 and Figure 7 Shown according to Figure 1 An implementation method of life data of a display panel inspection method.

[0037] Figure 8 Shown according to Figure 1 A block diagram of an implementation of an artificial intelligence model of a display panel inspection method.

[0038] Fig. 9 An embodiment of life data of a display panel according to a first test of the display panel inspection method in an embodiment of the present disclosure is shown.

[0039] Fig.10 Shown according to Figure 1 A table showing implementation methods of the determination coefficient of the artificial intelligence model of the display panel inspection method.

[0040] Fig.11 Shown according to Figure 1 A graph of normal distribution data of an embodiment of a method for inspecting a display panel.

[0041] Fig.12 A graph illustrating an embodiment of normal distribution data according to a method for inspecting a display panel according to the present disclosure.

[0042] Fig.13 A block diagram showing an inspection device in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0043] Hereinafter, the embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. The following description is intended to provide only sufficient disclosure so that the operation of the present invention can be understood, and any other disclosure is omitted to avoid blurring the scope of the present invention. In addition, the inventive concept can be implemented in different forms and is not limited to the embodiments set forth herein. The embodiments described herein are provided to fully describe the technical concept of the present invention in detail so that those skilled in the art can easily practice the present invention.

[0044] Throughout the specification, when describing an element "connected" to another element, this includes not only "direct connection", but also "indirect connection" and another device in the middle. The terms used herein are for the purpose of describing specific embodiments, and are not intended to limit the scope of the present invention. Throughout the specification, unless explicitly described to the contrary, the word "comprise" and variations such as "comprises", "comprising" will be understood to mean including the stated elements, but do not exclude any other elements. For the purpose of this disclosure, "at least one of X, Y and Z" and "at least one selected from the group consisting of X, Y and Z" can be interpreted as any combination of only X, only Y, only Z or two or more of X, Y and Z, such as, for example, XYZ, XYY, YZ and ZZ. As used herein, the term "and / or" includes any and all combinations of one or more of the related listed items.

[0045] Although the terms first, second, etc. may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another component. Therefore, without departing from the teachings of the present disclosure, the first component discussed below may also be referred to as the second component.

[0046] For the purpose of description, spatial relative terms such as "below", "below", "lower", "above", "upper", etc. may be used herein, and thereby describe the relationship between an element or feature and another (some) element or feature as shown in the drawings. In addition to the orientation depicted in the drawings, spatial relative terms are intended to include different orientations of the device in use, operation and / or manufacture. For example, if the device in the drawings is turned over, the elements described as being "below" or "below" other elements or features will then be oriented "above" other elements or features. Therefore, the term "below" can include both above and below orientations. In addition, the device can be oriented in other ways (e.g., rotated 90 degrees or in other orientations), and therefore, the spatial relative descriptors used herein should be interpreted accordingly.

[0047] Various embodiments are described herein with reference to cross-sectional views that are schematic diagrams of idealized embodiments. Therefore, variations in the shapes of the figures due to, for example, manufacturing techniques and / or tolerances should be expected. Therefore, the embodiments disclosed herein should not be interpreted as limited to the shapes of the specifically shown regions, but should include deviations in shapes due to, for example, manufacturing. Therefore, the regions shown in the drawings are schematic in nature, and their shapes are not intended to illustrate the actual shapes of regions of the device and are not intended to be limiting.

[0048] Figure 1 A flow chart showing an embodiment of a method for inspecting a display panel according to the present disclosure is shown.

[0049] refer to Figure 1 , Figure 1 The inspection method of a display panel in may include training an artificial intelligence model based on life data of a test display panel for aging characteristics (S100), generating virtual life data of a virtual display panel from the aging characteristics using the artificial intelligence model (S200), and predicting the life of the display panel based on the virtual life data (S300).

[0050] In the following, reference will be made to Figures 2 to 13 Describe this in detail.

[0051] Figure 2 Shown according to Figure 1 A block diagram of a display device of a display panel inspection method.

[0052] refer to Figure 2 , the display device may include a display panel 100, a driving controller 200, a gate driver 300, and a data driver 400. In an implementation, the driving controller 200 and the data driver 400 may be integrated on a single chip.

[0053] The display panel 100 may include a display area DA displaying an image and a non-display area NDA disposed adjacent to the display area DA. In an implementation, the gate driver 300 may be disposed (eg, installed) in the non-display area NDA.

[0054] The display panel 100 may include a plurality of gate lines GL, a plurality of data lines DL, and a plurality of sub-pixels SP electrically connected to the gate lines GL and the data lines DL. The gate lines GL may extend in a first direction DR1, and the data lines DL may extend in a second direction DR2 intersecting the first direction DR1.

[0055] The drive controller 200 may receive input image data IMG and an input control signal CONT from a processor (e.g., a graphics processing unit (“GPU”), etc.). In an embodiment, for example, the input image data IMG may include red image data, green image data, and blue image data. In an embodiment, the input image data IMG may also include white image data. In another embodiment, the input image data IMG may include magenta image data, yellow image data, and cyan image data. The input control signal CONT may include a main clock signal and a data enable signal. The input control signal CONT may also include a vertical synchronization signal and a horizontal synchronization signal.

[0056] The driving controller 200 may generate a first control signal CONT1 , a second control signal CONT2 , and a data signal DATA based on input image data IMG and an input control signal CONT.

[0057] The driving controller 200 may generate a first control signal CONT1 for controlling the operation of the gate driver 300 based on the input control signal CONT to output it to the gate driver 300. The first control signal CONT1 may include a vertical start signal and a gate clock signal.

[0058] The driving controller 200 may generate a second control signal CONT2 for controlling the operation of the data driver 400 based on the input control signal CONT to output it to the data driver 400. The second control signal CONT2 may include a horizontal start signal and a load signal.

[0059] The driving controller 200 may receive input image data IMG and an input control signal CONT to generate a data signal DATA. The driving controller 200 may output the data signal DATA to the data driver 400.

[0060] The gate driver 300 may generate a gate signal for driving the gate line GL in response to the first control signal CONT1 received from the driving controller 200. The gate driver 300 may output the gate signal to the gate line GL. In an embodiment, for example, the gate driver 300 may sequentially output the gate signal to the gate line GL.

[0061] The data driver 400 may receive the second control signal CONT2 and the data signal DATA from the driving controller 200. The data driver 400 may generate a data voltage obtained by converting the data signal DATA into an analog voltage. The data driver 400 may output the data voltage to the data line DL.

[0062] Figure 3 Shown according to Figure 1An embodiment of a display panel inspection method for testing a display area of ​​a display panel, Figure 4 Shown according to Figure 1 An embodiment of a method for inspecting a display panel by observing color and observing grayscale, Figure 5 is a graph used to illustrate life data, and Figure 6 and Figure 7 Shown according to Figure 1 An implementation method of life data of a display panel inspection method.

[0063] For better understanding and ease of description, Figure 5 and Figure 7 The aging time AT and Figure 7 The initial aging brightness AL, initial observed brightness OL and temperature T in represent relative values.

[0064] refer to Figures 3 to 7 , Figure 1 The inspection method of the display panel can display the aging color AC and the aging grayscale AG on the patterns PT1 to PT30 of the test display panels TP1 to TP3, can measure the aging brightness of the patterns PT1 to PT30, can display the observation color OC and the observation grayscale OG on the test display panels TP1 to TP3, can measure the observation brightness of the patterns PT1 to PT30, can measure the temperature T of the test display panels TP1 to TP3, can determine the life data AD of the test display panels TP1 to TP3, can annotate the aging characteristics AF with the life data AD of the test display panels TP1 to TP3, and can train an artificial intelligence model based on the aging characteristics AF annotated with the life data AD of the test display panels TP1 to TP3.

[0065] Figure 1 The inspection method of the display panel may display an image having an aging color AC and an aging grayscale AG (hereinafter, also referred to as an aging image) on the patterns PT1 to PT30 of the test display panels TP1 to TP3 to age the test display panels TP1 to TP3. In addition, Figure 1 The inspection method of the display panel can measure the observed brightness by displaying an image having an observed color OC and an observed grayscale OG (hereinafter, also referred to as an observed image) on the test display panels TP1 to TP3 aged by the aging image. In an embodiment, the aging image may have different colors and grayscales in the respective patterns PT1 to PT30, and the observed image may have the same color and grayscale in the respective patterns PT1 to PT30. That is, the aging brightness of each of the patterns PT1 to PT30 may be the brightness of each of the patterns PT1 to PT30 when the aging image is displayed, and the observed brightness may be the brightness of each of the patterns PT1 to PT30 when the observed image is displayed.

[0066] Here, the aging image may be an image for each of the aging patterns PT1 to PT30, and the observation image may be an image for measuring a characteristic change (i.e., life data AD) according to the aging time AT with a predetermined color (i.e., observation color OC) and a predetermined grayscale (i.e., observation grayscale OG). That is, since the respective patterns PT1 to PT30 are aged with aging images having various aging colors AC and aging grayscales AG, the observed brightness of the observation image showing the same color and grayscale in all the patterns PT1 to PT30 may vary in the respective patterns PT1 to PT30. However, the present disclosure is not limited to the observation image having the same color and grayscale in all the patterns PT1 to PT30.

[0067] The aging time AT may be a time for displaying the aging image. In an embodiment, for example, the life data AD when the aging time AT is 10 may be the life data AD when the display time of the aging image is 10. A detailed description of the life data AD will be described later.

[0068] The initial aging luminance AL and the initial observed luminance OL may be reference luminances for determining the lifetime data AD. A detailed description of this point will be described later.

[0069] In an embodiment, the initial aging luminance AL may be the aging luminance of the patterns PT1 to PT30 measured by displaying an aging image on the patterns PT1 to PT30 of the test display panels TP1 to TP3 when the aging time AT is 0.

[0070] In another embodiment, the initial aging brightness AL may be the aging brightness of the patterns PT1 to PT30 measured by displaying the aging image on the patterns PT1 to PT30 of the test display panels TP1 to TP3, and is the aging brightness of the patterns PT1 to PT30 after the patterns PT1 to PT30 of the test display panels TP1 to TP3 are aged for a predetermined period of time by the aging image. In this case, the aging time AT may be a time obtained by subtracting a predetermined time from a time when the aging image is displayed. That is, when the aging image is displayed for a predetermined time, the aging time AT may be zero.

[0071] In an embodiment, the initial observed brightness OL may be the observed brightness of the patterns PT1 to PT30 measured by displaying the observed images on the test display panels TP1 to TP3 when the aging time AT is 0.

[0072] In another embodiment, the initial observed brightness OL may be the observed brightness of the patterns PT1 to PT30 measured by displaying the observed image on the test display panels TP1 to TP3 after the test display panels TP1 to TP3 are aged for a predetermined period of time by the aging image. In this case, the aging time AT may be a time obtained by subtracting a predetermined time from a time when the aging image is displayed. That is, when the aging image is displayed for a predetermined time, the aging time AT may be zero.

[0073] The lifespan data AD of the test display panels TP1 to TP3 may be measured by displaying the observation color OC and the observation grayscale OG on the test display panels TP1 to TP3.

[0074] The life data AD of the test display panels TP1 to TP3 may be a brightness retention rate according to the aging time AT of the test display panels TP1 to TP3. Here, the brightness retention rate may be a ratio of the brightness when the aging time AT is zero (i.e., the initial observed brightness OL) and the observed brightness after aging for the aging time AT. In an embodiment, for example, Figure 5 As shown in , when the aging time AT is 0, the lifetime data AD may be 1.

[0075] In an embodiment, for example, the aging characteristics AF may include at least one of an aging color AC, an aging grayscale AG, an initial aging brightness AL, an observed color OC, an observed grayscale OG, an initial observed brightness OL, an aging position AP (e.g., the positions of patterns PT1 to PT30), and a temperature T. However, the present disclosure is not limited to these elements included in the aging characteristics AF.

[0076] Here, the aging brightness and the observed brightness are measured by a brightness meter, the temperature T is measured by a thermometer, and the temperature T may be the temperature of the test display panel. That is, even when the aging color AC, the aging gray AG, the observed color OC, the observed gray OG, and the aging position AP are the same, the initial aging brightness AL, the initial observed brightness OL, and the temperature T may be measured differently depending on the test display panel.

[0077] In the illustrated embodiment, thirty patterns PT1 to PT30 are shown, but the present disclosure is not limited to the number of patterns PT1 to PT30.

[0078] refer to Figure 3 , Figure 1 The inspection method of the display panel may display the aged color AC and the aged gray AG in the patterns PT1 to PT30 of the test display panels TP1 to TP3.

[0079] The test display panel may be a display panel used to generate lifetime data AD used to train the artificial intelligence model. In addition, the display panel whose lifetime is predicted does not necessarily have to be different from the test display panel. In an embodiment, for example, the display panel whose lifetime is predicted may be used as a test display panel.

[0080] The display panel may be aged by displaying an aging color AC and an aging grayscale AG for each of the patterns PT1 to PT30. In addition, the aging color AC and the aging grayscale AG of each of the patterns PT1 to PT30 may be different. Therefore, each of the patterns PT1 to PT30 of the display panel may be aged differently. Therefore, the life data AD for testing the display panel described later may be different for each of the patterns PT1 to PT30.

[0081] In an embodiment, for example, 255 grayscale white (W) may be displayed on the first pattern PT1. In an embodiment, for example, 128 grayscale white (W) may be displayed on the second pattern PT2. However, the present disclosure is not limited to the color and the grayscale displayed in each of the patterns PT1 to PT30.

[0082] refer to Figures 3 to 5 , the life data AD of the test display panel can be measured according to the aging time AT showing the aging color AC and the aging grayscale AG.

[0083] In an embodiment, one of red R, green G, blue B, and white W may be displayed on the test display panel. In an embodiment, for example, one of grayscales 0 to 255 may be displayed on the test display panel. However, the present disclosure is not limited to the color and the grayscale displayed on the test display panel.

[0084] The light emitting element of the sub-pixel may age with increasing use. In addition, due to aging, the light emitting characteristics of the light emitting element may change (for example, the brightness may decrease). That is, the life data AD may decrease with increasing aging time AT.

[0085] The aging time AT may be the time for displaying the aging color AC and the aging gray AG on the test display panel. In an embodiment, for example, after displaying the aging color AC and the aging gray AG on the test display panel for the aging time AT, the observation color OC and the observation gray OG may be displayed on the test display panel. In addition, the life data AD may be determined based on the observed brightness of each of the patterns PT1 to PT30 measured when the observation color OC and the observation gray OG are displayed on the test display panel.

[0086] In the illustrated embodiment, it is illustrated that the lifetime data AD is periodically measured every ten aging times AT, however, the lifetime data AD of the present disclosure is not necessarily measured periodically.

[0087] refer to Figure 6 and Figure 7 , the aging characteristics AF are annotated with the lifetime data AD of the test display panels TP1 to TP3, and the artificial intelligence model can be trained using the aging characteristics AF annotated with the lifetime data AD of the test display panels TP1 to TP3.

[0088] In an embodiment, for example, the lifetime data AD may be generated from a plurality of test display panels TP1 to TP3. As the number of test display panels TP1 to TP3 increases, data for training an artificial intelligence model may increase.

[0089] Figure 7 The lifespan data AD of the first test display panel TP1 is shown. The lifespan data AD of the second test display panel TP2 and the third test display panel TP3 may also have the following structure: Figure 7 The form shown in .

[0090] In the illustrated embodiment, three test display panels TP1 to TP3 are shown, however, the present disclosure is not limited to the number of the test display panels TP1 to TP3.

[0091] The aging characteristics AF can be annotated with the life data AD of the test display panel. For better understanding and easy description, Figure 7 The aging characteristics AF include an aged color AC, an aged grayscale AG, an initial aged brightness AL, an observed color OC, an observed grayscale OG, an initial observed brightness OL, an aging position AP (eg, positions of patterns PT1 to PT30 ), and a temperature T. FIG.

[0092] In an embodiment, for example, when the aging time AT is 10, the aging characteristic AF including the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, and the temperature T of 30 can be annotated with the lifetime data AD of 0.94. That is, the artificial intelligence model can be trained to have an output value of 0.94 when the aging characteristic AF including the aging time AT of 10, the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, and the temperature T of 30 is input.

[0093] In an embodiment, for example, when the aging time AT is 10, the aging characteristic AF including the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, and the temperature T of 30 can be annotated with the lifetime data AD of 0.92. That is, the artificial intelligence model can be trained to have an output value of 0.92 when the aging characteristic AF including the aging time AT of 10, the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, and the temperature T of 30 is input.

[0094] In an embodiment, for example, when the aging time AT is 20, the aging characteristic AF including the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, and the temperature T of 30 can be annotated with the lifetime data AD of 0.91. That is, the artificial intelligence model can be trained to have an output value of 0.91 when the aging characteristic AF including the aging time AT of 20, the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, and the temperature T of 30 is input.

[0095] In an embodiment, for example, when the aging time AT is 20, the aging characteristic AF including the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, and the temperature T of 30 can be annotated with the lifetime data AD of 0.88. That is, the artificial intelligence model can be trained to have an output value of 0.88 when the aging characteristic AF including the aging time AT of 20, the aging color AC of white (W), the aging grayscale AG of 255 grayscale, the initial aging brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, and the temperature T of 30 is input.

[0096] The artificial intelligence model may be trained with the annotated aging characteristics. In an embodiment, for example, the trained artificial intelligence model may predict life data by receiving the aging characteristics. In an embodiment, for example, the artificial intelligence model may output virtual life data of a virtual display panel by receiving the aging characteristics. The virtual life data may be life data predicted from any aging characteristics by the artificial intelligence model, and the virtual display panel may be a display panel corresponding to the life data predicted by the artificial intelligence model.

[0097] As described above, when only the aging characteristics of the new test display panel are measured, the artificial intelligence model can generate virtual life data of the virtual display panel. That is, the artificial intelligence model can generate virtual life data of the virtual display panel corresponding to the new test display panel without measuring the life data of the new test display panel.

[0098] Therefore, in Figure 1 In the inspection method of the display panel, the artificial intelligence model generates virtual life data of the virtual display panel, so that the same effect as using a larger number of test display panels can be obtained, and the larger number is the number of test display panels whose life data are measured for actual life prediction.

[0099] Figure 8 Shown according to Figure 1 A block diagram of an implementation of an artificial intelligence model of a display panel inspection method.

[0100] refer to Figure 8 , Figure 1 The inspection method of the display panel can generate virtual life data VAD according to the aging time AT using an artificial intelligence model.

[0101] In an embodiment, the artificial intelligence model may include a first artificial intelligence model according to a first aging time (e.g., 10) and a second artificial intelligence model according to a second aging time (e.g., 20) different from the first aging time.

[0102] In an embodiment, for example, the first artificial intelligence model may generate virtual lifetime data VAD of a first aging time (e.g., 10) by receiving the aging characteristic AF. In an embodiment, for example, the second artificial intelligence model may generate virtual lifetime data VAD of a second aging time (e.g., 20) by receiving the aging characteristic AF.

[0103] In an embodiment, for example, a first artificial intelligence model may be trained using an aging characteristic AF of lifetime data annotated with a first aging time (e.g., 10). In an embodiment, a second artificial intelligence model may be trained using an aging characteristic AF of lifetime data annotated with a second aging time (e.g., 20).

[0104] In the illustrated embodiment, two artificial intelligence models are illustrated, however, the present disclosure is not limited to the number of artificial intelligence models.

[0105] In an embodiment, the aging time AT and the aging characteristics AF can be annotated with the life data, and the artificial intelligence model can be trained with the annotated aging time AT and the aging characteristics AF. In this case, the artificial intelligence model can generate virtual life data VAD by receiving the aging time AT and the aging characteristics AF.

[0106] Fig. 9 An embodiment of life data of a first test display panel according to a method for inspecting a display panel according to the present disclosure is shown.

[0107] Since the inspection method of the display panel in the illustrated embodiment is different from that in the embodiment shown in FIG. Figure 1 The inspection method of the display panel is substantially the same, so the same reference numerals and reference symbols are used for the same or similar constituent elements, and redundant descriptions are omitted.

[0108] For better understanding and ease of description, Fig. 9 The initial aging brightness AL, initial observed brightness OL, temperature T, aging time AT, aging current ACU, observed current OCU and light efficiency LE represent relative values.

[0109] refer to Figure 2 , Figure 3 , Figure 6 as well as Fig. 9 , Fig. 9 The inspection method of the display panel can display the aging color AC and the aging grayscale AG on the patterns PT1 to PT30 of the test display panels TP1 to TP3, can measure the initial aging brightness AL and the aging current ACU of the patterns PT1 to PT30, can display the observation color OC and the observation grayscale OG on the test display panels TP1 to TP3, can measure the initial observation brightness OL and the observation current OCU of the patterns PT1 to PT30, can measure the temperature T of the test display panels TP1 to TP3, can determine the life data AD of the test display panels TP1 to TP3, can annotate the aging characteristics AF with the life data AD of the test display panels TP1 to TP3, and can train an artificial intelligence model based on the aging characteristics AF annotated with the life data AD of the test display panels TP1 to TP3.

[0110] The aging characteristics AF may include at least one of an aging color AC, an aging grayscale AG, an initial aging brightness AL, an observed color OC, an observed grayscale OG, an initial observed brightness OL, an aging position AP (e.g., the positions of patterns PT1 to PT30), a temperature T, an aging current ACU, an observed current OCU, and a light efficiency (also referred to as optical efficiency) LE.

[0111] Here, the aging brightness and the observed brightness are measured with a brightness meter, the temperature T is measured with a thermometer, the aging current ACU and the observed current OCU are measured with an ammeter, and the temperature T may be the temperature of the test display panel. That is, even when the aging color AC, the aging gray AG, the observed color OC, the observed gray OG, and the aging position AP are the same, the initial aging brightness AL, the initial observed brightness OL, the temperature T, the aging current ACU, and the observed current OCU may be measured differently depending on the test display panel.

[0112] In an embodiment, the aging current ACU used to train the artificial intelligence model may be a current applied to the test display panel measured by displaying the aging color AC and the aging gray AG on the test display panel before the test display panel is aged with the aging color AC and the aging gray AG. In another embodiment, the aging current ACU used to train the artificial intelligence model may be a current applied to the test display panel measured by displaying the aging color AC and the aging gray AG on the test display panel after the test display panel is aged with the aging color AC and the aging gray AG for a predetermined time.

[0113] In an embodiment, the observation current OCU may be a current applied to the test display panel measured by displaying an observation image on the test display panel when the aging time AT of the test display panel is zero.

[0114] In another embodiment, the observation current OCU may be a current applied to the test display panel measured by displaying the observation image on the test display panel after the test display panel is aged with the aging image for a predetermined time. In this case, the aging time AT may be a time obtained by subtracting the predetermined time from the time when the aging image is displayed. That is, when the aging image is displayed for a predetermined time, the aging time AT may be zero.

[0115] In an embodiment, the aging current ACU and the observation current OCU may be currents applied to the entire test display panel. In another embodiment, the aging current ACU and the observation current OCU may be currents applied to each of the patterns PT1 to PT30.

[0116] The light efficiency LE may be a ratio of the brightness of each sub-pixel SP to the driving current. In an embodiment, for example, when the current flowing to the light emitting element of each sub-pixel SP is the same, the higher the light efficiency LE, the greater the brightness.

[0117] The light efficiency LE may be different for each sub-pixel SP. The light efficiency LE for training the artificial intelligence model (i.e., the light efficiency LE as the aging characteristic AF) may be measured for each of the patterns PT1 to PT30. In an embodiment, for example, the light efficiency LE for training the artificial intelligence model may be determined based on the light efficiency LE of the sub-pixel SP of each of the patterns PT1 to PT30. In an embodiment, for example, the light efficiency LE for training the artificial intelligence model may be an average value of the light efficiency LE of the sub-pixel SP of each of the patterns PT1 to PT30.

[0118] In an embodiment, for example, when the aging time AT is 10, the lifetime data AD of 0.94 can be used to annotate the aging characteristics AF including the aged color AC of white (W), the aged grayscale AG of 255 grayscale, the initial aged brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, the temperature T of 30, the aged current ACU of 20, the observed current OCU of 10 and the light efficiency LE of the first pattern PT1 of 10. That is, the artificial intelligence model can be trained to have an output value of 0.94 when the input includes an aging time AT of 10, an aging color AC of white (W), an aging grayscale AG of 255 grayscale, an initial aging brightness AL of 120, an observed color OC of blue (B), an observed grayscale OG of 255 grayscale, an initial observed brightness OL of 80, an observation position AP of the first pattern PT1, a temperature T of 30, an aging current ACU of 20, an observed current OCU of 10, and an aging characteristic AF of a light efficiency LE of 10.

[0119] In an embodiment, for example, when the aging time AT is 10, the lifetime data AD of 0.92 can be used to annotate the aging characteristics AF including the aged color AC of white (W), the aged grayscale AG of 255 grayscale, the initial aged brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, the temperature T of 30, the aged current ACU of 20, the observed current OCU of 7 and the light efficiency LE of the first pattern PT1 of 10. That is, the artificial intelligence model can be trained to have an output value of 0.92 when the input includes an aging time AT of 10, an aging color AC of white (W), an aging grayscale AG of 255 grayscales, an initial aging brightness AL of 120, an observed color OC of blue (B), an observed grayscale OG of 128 grayscales, an initial observed brightness OL of 30, an observed position AP of the first pattern PT1, a temperature T of 30, an aging current ACU of 20, an observed current OCU of 7, and an aging characteristic AF of the light efficiency LE of 10.

[0120] In an embodiment, for example, when the aging time AT is 20, the lifetime data AD of 0.91 can be used to annotate the aging characteristics AF including the aged color AC of white (W), the aged grayscale AG of 255 grayscale, the initial aged brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 255 grayscale, the initial observed brightness OL of 80, the observed position AP of the first pattern PT1, the temperature T of 30, the aged current ACU of 20, the observed current OCU of 10, and the light efficiency LE of the first pattern PT1 of 10. That is, the artificial intelligence model can be trained to have an output value of 0.91 when the input includes an aging time AT of 20, an aging color AC of white (W), an aging grayscale AG of 255 grayscale, an initial aging brightness AL of 120, an observed color OC of blue (B), an observed grayscale OG of 255 grayscale, an initial observed brightness OL of 80, an observation position AP of the first pattern PT1, a temperature T of 30, an aging current ACU of 20, an observation current OCU of 10, and an aging characteristic AF of a light efficiency LE of 10.

[0121] In an embodiment, for example, when the aging time AT is 20, the lifetime data AD of 0.88 can be used to annotate the aging characteristics AF including the aged color AC of white (W), the aged grayscale AG of 255 grayscale, the initial aged brightness AL of 120, the observed color OC of blue (B), the observed grayscale OG of 128 grayscale, the initial observed brightness OL of 30, the observed position AP of the first pattern PT1, the temperature T of 30, the aged current ACU of 20, the observed current OCU of 7 and the light efficiency LE of the first pattern PT1 of 10. That is, the artificial intelligence model can be trained to have an output value of 0.88 when the input includes an aging time AT of 20, an aging color AC of white (W), an aging grayscale AG of 255 grayscale, an initial aging brightness AL of 120, an observed color OC of blue (B), an observed grayscale OG of 128 grayscale, an initial observed brightness OL of 30, an observation position AP of the first pattern PT1, a temperature T of 30, an aging current ACU of 20, an observation current OCU of 7, and an aging characteristic AF of the light efficiency LE of 10.

[0122] Fig.10 Shown according to Figure 1 A table showing implementation methods of the determination coefficient of the artificial intelligence model of the display panel inspection method.

[0123] refer to Fig.10 , the artificial intelligence model may be a regression model. In an embodiment, the artificial intelligence model may be a random forest regression model, an extra tree regression model, a gradient boosting regression model, a decision tree regression model, a K nearest neighbor regression model, a linear regression model, a Bayesian ridge regression model, a ridge regression model, a Huber regression model, or a passive aggressive regression model.

[0124] The coefficient of determination is an index indicating the degree to which the independent variable in the regression model explains the dependent variable, and the higher the coefficient of determination, the higher the correlation between the independent variable and the dependent variable. That is, the higher the coefficient of determination, the higher the prediction accuracy of the artificial intelligence model. Here, the independent variable is the aging characteristics, and the dependent variable is the life span data.

[0125] In an embodiment, for example, the coefficient of determination is a value of the explained sum of squares ("SSE") divided by the total sum of squares ("SST"). The SSE may be the sum of squares of values ​​obtained by subtracting the average value of the life data (or the average value of the virtual life data) from the virtual life data. The SST may be the sum of squares of values ​​obtained by subtracting the average value of the life data (or the average value of the virtual life data) from the life data.

[0126] In an embodiment, for example, a random forest regression model may have a coefficient of determination (R^2) of 0.9976. In an embodiment, for example, an extreme tree regression model may have a coefficient of determination (R^2) of 0.9976. In an embodiment, for example, a gradient boosting regression model may have a coefficient of determination (R^2) of 0.9954. In an embodiment, for example, a decision tree regression model may have a coefficient of determination (R^2) of 0.9953. In an embodiment, for example, a K-nearest neighbor regression model may have a coefficient of determination (R^2) of 0.9921. In an embodiment, for example, a linear regression model may have a coefficient of determination (R^2) of 0.9878. In an embodiment, for example, a Bayesian ridge regression model may have a coefficient of determination (R^2) of 0.9878. In an embodiment, for example, a ridge regression model may have a coefficient of determination (R^2) of 0.9877. In an embodiment, for example, a Huber regression model may have a coefficient of determination (R^2) of 0.9875. In an embodiment, for example, the passive-aggressive regression model may have a coefficient of determination (R^2) of 0.9662.

[0127] In an embodiment, the artificial intelligence model may be generated by fusing at least two regression models. In an embodiment, for example, the artificial intelligence model may be generated by fusing a random forest regression model, an extreme tree regression model, and a gradient boosting regression model. In an embodiment, for example, the artificial intelligence model generated by fusing a random forest regression model, an extreme tree regression model, and a gradient boosting regression model may have a coefficient of determination (R^2) of 0.9980.

[0128] Fig.11 Shown according to Figure 1 A graph of normal distribution data of an embodiment of a method for inspecting a display panel.

[0129] For ease of explanation, Fig.11 It is shown that five virtual lifetime data VAD of the virtual display panel are generated from the aging characteristics in five cases.

[0130] refer to Fig.11 , Figure 1 The inspection method of the display panel may normalize the virtual lifetime data VAD to generate normal distribution data, and may predict the lifetime of the display panel based on data corresponding to a predetermined ratio of the normal distribution data.

[0131] In an embodiment, for example, Fig.11As shown in , normal distribution data can be generated by normalizing five virtual lifetime data VAD of a virtual display panel. In addition, since the worst case must be considered when predicting the lifetime of a display panel, the lifetime of a display panel can be predicted based on data corresponding to the lower 10% of the normal distribution data.

[0132] Virtual lifetime data VAD of a virtual display panel may be generated by inputting aging characteristics measured from display panels other than the test display panel used to measure lifetime data (hereinafter referred to as additional display panels) into the artificial intelligence model. That is, five additional display panels may be required to generate five virtual display panels.

[0133] Figure 1 The inspection method of the display panel can generate virtual life data VAD by measuring only the aging characteristics of the additional display panel without measuring the observed brightness other than the initial observed brightness of the additional display panel. In addition, since the aging characteristics of the additional display panel include values ​​measured without aging the additional display panel (i.e., values ​​measured when the aging time is 0), increasing the number of additional display panels can be less restrictive than increasing the number of test display panels. Therefore, by adding the additional display panel, the effect of increasing the number of display panels for testing can be obtained, and therefore, the accuracy of life prediction can be increased.

[0134] In an embodiment, the lifetime of the display panel may be the aging time required for the brightness retention rate to decrease to 50%. However, the lifetime of the display panel only reflects the change in light emission characteristics due to aging, and the present disclosure does not limit the method of defining the lifetime of the display panel.

[0135] Fig.12 A graph illustrating an embodiment of normal distribution data according to a method for inspecting a display panel according to the present disclosure.

[0136] Since the inspection method of the display panel in the illustrated embodiment is different from that in the embodiment described above, in addition to considering the lifetime data AD when predicting the lifetime of the display panel, Figure 1 The inspection method of the display panel in is basically the same, so the same reference numerals and reference symbols are used for the same or similar constituent elements, and redundant descriptions are omitted.

[0137] For better understanding and ease of description, Fig.12 It is shown that five virtual lifetime data VAD of the virtual display panel are generated from the aging characteristics of five cases, and five lifetime data AD of the test display panel are measured.

[0138] refer to Fig.12 , Fig.12The inspection method of the display panel can normalize the virtual lifetime data VAD and the lifetime data AD of the test display panel to generate normal distribution data, and it can predict the lifetime of the display panel based on data corresponding to a predetermined ratio of the normal distribution data.

[0139] In an embodiment, for example, Fig.12 As shown in , the normal distribution data can be generated by normalizing the five virtual life data VAD of the virtual display panel and the five life data AD of the test display panel. In addition, since the worst case scenario must be considered when predicting the life of the display panel, the life of the display panel can be predicted based on the data corresponding to the lower 10% of the normal distribution data.

[0140] Fig.13 A block diagram of an embodiment of an inspection device according to the present disclosure is shown.

[0141] refer to Fig.13 , Figure 1 The inspection method of the display panel may be implemented by an inspection device 1000. The inspection device 1000 may include a communicator 1100, a nonvolatile memory 1200, a processor 1300, and a system memory 1400. In an implementation, for example, the inspection device 1000 may be a computer device.

[0142] The nonvolatile memory 1200 may store an artificial intelligence model. The processor 1300 may use the artificial intelligence model stored in the nonvolatile memory 1200 (or the artificial intelligence model loaded in the system memory 1400) to predict the life of the display panel.

[0143] The communicator 1100 may access a network through wired communication and / or wireless communication in response to the control of the processor 1300 .

[0144] The nonvolatile memory 1200 may include at least one of storage media such as a flash memory, a hard disk, and a multimedia card that retains stored data even when power is turned off. The nonvolatile memory 1200 may write and read data in response to the control of the processor 1300.

[0145] The processor 1300 may control the operation of the communicator 1100, the nonvolatile memory 1200, and the system memory 1400. The processor 1300 may perform various processes such as calculations and comparisons by including a plurality of processing units (or a plurality of cores). In an embodiment, for example, the processor 1300 may include a central processing unit ("CPU"). In an embodiment, for example, the processor 1300 may also include a graphics processing unit.

[0146] The processor 1300 may load the execution code and / or instruction words into the system memory 1400 and execute the loaded execution code and / or instruction words. In an embodiment, for example, the processor 1300 may perform an operation for measuring the life of the display panel by loading an artificial intelligence model into the system memory 1400 and executing the loaded artificial intelligence model.

[0147] In addition, the processor 1300 may load an operating system into the system memory 1400 and execute the loaded operating system to control global operations of the inspection apparatus 1000 .

[0148] The system memory 1400 may include at least one of a random access memory ("RAM"), a read only memory ("ROM"), and other types of computer-readable storage media. Fig.13 , the system memory 1400 is illustrated as a constituent element separate from the processor 1300, but this is merely an example, and at least a portion of the system memory 1400 may be included in the processor 1300. The system memory 1400 may be divided into a plurality of constituent elements as needed.

[0149] In this embodiment, the inspection device may include at least one of a brightness meter for measuring brightness, a thermometer for measuring temperature, and an ammeter for measuring current. In another embodiment, the inspection device may receive brightness data, temperature data, and current data from the outside.

[0150] The present disclosure may be applicable to display devices and electronic devices including the same. In an embodiment, for example, the present disclosure may be applicable to digital televisions ("TV"), three-dimensional ("3D") TVs, mobile phones, smart phones, tablet computers, VR devices, personal computers ("PC"), home electronic devices, laptop computers, personal digital assistants ("PDAs"), portable media players ("PMPs"), digital cameras, music players, portable game consoles, navigation, etc.

[0151] While the disclosure has been described in connection with what are presently considered to be practical embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but on the contrary is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for inspecting a display panel, the method comprising: Training AI models based on lifetime data of test display panels for aging characteristics; generating virtual life data of a virtual display panel from the aging characteristics using the artificial intelligence model; as well as The lifespan of the display panel is predicted based on the virtual lifespan data.

2. The display panel inspection method according to claim 1, wherein: The life data of the test display panel is a brightness retention rate according to an aging time of the test display panel.

3. The display panel inspection method according to claim 1, wherein: The artificial intelligence model includes a first artificial intelligence model according to a first aging time and a second artificial intelligence model according to a second aging time different from the first aging time.

4. The display panel inspection method according to claim 1, wherein: The aging characteristics include at least one of aging color, aging grayscale, initial aging brightness, observed color, observed grayscale, initial observed brightness, aging position, and temperature, and Training the artificial intelligence model includes: displaying the aging color and the aging grayscale on the pattern of the test display panel; measuring the initial aged brightness of the pattern; Displaying the observed color and the observed grayscale on the test display panel; measuring the initial observed brightness of the pattern; measuring the temperature of the test display panel; determining the life data of the test display panel; annotating the aging characteristics with the lifetime data of the test display panel; and The artificial intelligence model is trained with the aging characteristics annotated using the lifetime data of the test display panel.

5. The method for inspecting a display panel according to claim 4, wherein: The life data of the test display panel is measured by displaying the observed color and the observed grayscale on the test display panel.

6. The method for inspecting a display panel according to claim 4, wherein: The life data of the test display panel is measured according to an aging time of displaying the aged color and the aged grayscale.

7. The display panel inspection method according to claim 1, wherein: The aging characteristics include at least one of aging color, aging grayscale, initial aging brightness, observed color, observed grayscale, initial observed brightness, aging position, temperature, aging current, observed current, and light efficiency, and Training the artificial intelligence model includes: displaying the aging color and the aging grayscale on the pattern of the test display panel; measuring the initial aging brightness and the aging current of the pattern; Displaying the observed color and the observed grayscale on the test display panel; measuring the initial observed brightness and the observed current of the pattern; measuring the temperature of the test display panel; determining the life data of the test display panel; annotating the aging characteristics with the lifetime data of the test display panel; and The artificial intelligence model is trained with the aging characteristics annotated using the lifetime data of the test display panel.

8. The method for inspecting a display panel according to claim 1, wherein: The artificial intelligence model is a regression model.

9. The display panel inspection method according to claim 1, wherein: Predicting the life of the display panel includes: normalizing the virtual life data and generating normal distribution data; and The lifespan of the display panel is predicted based on data corresponding to a predetermined ratio of the normal distribution data.

10. The method for inspecting a display panel according to claim 1, wherein: Predicting the life of the display panel includes: normalizing the virtual lifetime data and the lifetime data of the test display panel and generating normal distribution data; and The lifespan of the display panel is predicted based on data corresponding to a predetermined ratio of the normal distribution data.

Images