Parameter adjustment method and device, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides a parameter adjustment method and device of a pixel compensation system, electronic equipment, a storage medium and a program product. The parameter adjustment method comprises the steps that a first image based on a scene of a test picture displayed on a current display screen is acquired as an input image, and the scene of the test picture displayed on the display screen comprises the steps that in response to the fact that a test pattern is displayed on the display screen and the test pattern is displayed for a first time length, the display screen is switched into a scene of displaying a pure color picture; based on the trained parameter adjustment model, a prediction result corresponding to the input image is obtained, and the prediction result is used for adjusting parameters of the pixel compensation system so as to adjust the display effect of the display screen. According to the parameter adjustment method, the parameters of the pixel compensation system are intelligently adjusted, so that the pixel compensation system can automatically adapt to different use scenes, the need of manual intervention is remarkably reduced, and the labor cost and the time cost are reduced.

Description

Technical Field

[0001] Embodiments of this disclosure relate to a parameter adjustment method and apparatus for a pixel compensation system, an electronic device, a storage medium, and a computer program product. Background Technology

[0002] Micro LED (MLED) technology is an emerging display technology. MLED displays consist of a high-pixel-density two-dimensional MLED array, where each pixel can be addressed, controlled, and independently driven to emit light, offering advantages such as high brightness, high contrast, and low power consumption. However, image retention remains a significant technical challenge. Image retention refers to the temporary "trace" left by an image on the display screen, severely impacting the user's visual experience. This can stem from inconsistent pixel response speeds due to hardware physical characteristics, or from pixel luminous efficiency degradation caused by prolonged display of static images. To mitigate this issue, pixel compensation methods are needed to adjust and balance the brightness performance of each pixel, thereby extending the display's lifespan and optimizing image quality. Summary of the Invention

[0003] At least one embodiment of this disclosure provides a parameter adjustment method for a pixel compensation system, comprising: acquiring a first image of a scene where a test image is displayed on a current display screen as an input image, wherein the scene where the test image is displayed on the display screen includes: a scene where the display screen is switched to display a solid color image in response to displaying a test pattern on the display screen and displaying the test pattern for a first duration; and acquiring a prediction result corresponding to the input image based on a trained parameter adjustment model, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0004] In the parameter adjustment method provided in at least one embodiment of this disclosure, the prediction result includes a list of prediction parameters.

[0005] In the parameter adjustment method provided in at least one embodiment of this disclosure, the first duration is the duration from when the display screen starts displaying the test pattern until the temperature fluctuation value reaches a first range.

[0006] In at least one embodiment of the parameter adjustment method provided in this disclosure, multiple different solid color images correspond to multiple prediction results. The parameter adjustment method further includes: performing a weighted average processing on the multiple prediction results to obtain a weighted prediction result, wherein the weighted prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0007] In the parameter adjustment method provided in at least one embodiment of this disclosure, the trained parameter adjustment model is trained through the following steps: acquiring multiple training sample images and sample labels corresponding to each training sample image; inputting at least one training sample image from the multiple training sample images into the parameter adjustment model to be trained to obtain predicted sample labels; and adjusting the network parameters of the parameter adjustment model to be trained according to the sample labels corresponding to the at least one training sample image and the predicted sample labels.

[0008] In the parameter adjustment method provided in at least one embodiment of this disclosure, the step of obtaining multiple training sample images and sample labels corresponding to each training sample image includes: traversing all combinations of multiple displays and multiple patterns, and for each combination: obtaining training sample images and sample labels corresponding to the training sample images based on the combination of the first display screen among the multiple displays and the first pattern among the multiple patterns.

[0009] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of obtaining a training sample image and a corresponding sample label based on a combination of a first display screen among the multiple display screens and a first pattern among the multiple patterns includes: obtaining a second image based on a scene in which a training image is displayed on the first display screen as the training sample image, wherein the scene in which the training image is displayed on the first display screen includes: a scene in which the first display screen is switched to display a solid color image in response to displaying the first pattern on the first display screen and displaying the first pattern for a first duration; updating the parameter list of the pixel compensation system multiple times, wherein for each update, a third image based on a scene in which the image adjusted by the pixel compensation system is displayed on the first display screen is obtained, and the third image and the corresponding parameter list are saved as a training data pair; and a target parameter list is determined from multiple training data pairs as the sample label corresponding to the training sample image.

[0010] In at least one embodiment of the parameter adjustment method provided in this disclosure, after obtaining a third image of a scene based on the image adjusted by the pixel compensation system displayed on the first display screen, the method further includes: the first display screen being switched to display the first pattern within a predetermined time.

[0011] In at least one embodiment of the parameter adjustment method provided in this disclosure, the multiple updates to the parameter list of the pixel compensation system include: performing a loop process until a preset condition is met, wherein the loop process includes: selecting an unoptimized parameter from the parameter list as a first parameter; optimizing the first parameter; and fixing the first parameter after optimization, wherein the preset condition includes: all parameters in the parameter list have been optimized. In at least one embodiment of the parameter adjustment method provided in this disclosure, the multiple updates to the parameter list of the pixel compensation system include: adjusting the size of the parameter to be optimized according to the third image.

[0012] In the parameter adjustment method provided in at least one embodiment of this disclosure, the step of determining a target parameter list from multiple training data pairs as sample labels corresponding to the training sample images includes: calculating the uniformity of each third image in the multiple training data pairs; and determining the parameter list corresponding to the target third image with the highest uniformity as the target parameter list.

[0013] In the parameter adjustment method provided in at least one embodiment of this disclosure, the uniformity is determined based on the degree of color deviation.

[0014] In the parameter adjustment method provided in at least one embodiment of this disclosure, the multiple patterns include: multiple patterns of different colors, multiple patterns of different areas, multiple patterns of different shapes, and / or multiple patterns of different positions.

[0015] In the parameter adjustment method provided in at least one embodiment of this disclosure, the multiple displays include: multiple displays with different backlight types, multiple displays with different resolutions, and / or displays with different backplane structures. In the parameter adjustment method provided in at least one embodiment of this disclosure, the parameter adjustment model includes: multiple multi-head attention models connected in sequence, wherein each multi-head attention model includes multiple self-attention modules, a normalization module, and a multilayer perceptron module; and a prediction module connected to the last multi-head attention model among the multiple sequentially connected multi-head attention models, configured to obtain a prediction result based on the output of the last multi-head attention model.

[0016] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of obtaining the prediction result corresponding to the input image based on the trained parameter adjustment model includes: segmenting the input image into multiple image blocks; sequentially inputting each image block into a first multi-head attention model among the multiple multi-head attention models, each image block generating multiple encoding vectors through the multiple self-attention modules, concatenating all the encoding vectors to obtain a concatenated encoding vector; processing the concatenated encoding vector through a normalization module and a multilayer perceptron module in the first multi-head attention model to obtain a first output vector; sequentially processing the first output vector through the multi-head attention models other than the first multi-head attention model among the multiple multi-head attention models to obtain a second output vector; and processing the second output vector through a prediction module to generate the prediction result.

[0017] In at least one embodiment of the parameter adjustment method provided in this disclosure, the prediction module includes a normalized exponential function module, which processes the second output vector through the prediction module to generate the prediction result, including: processing the second output vector through the normalized exponential function module to obtain a third output vector; and performing threshold division on the third output vector to obtain the prediction result.

[0018] In at least one embodiment of the parameter adjustment method provided in this disclosure, the method further includes: for each of a plurality of displays: analyzing the color difference between the color of a first camera image and a preset base color, wherein the first camera image is obtained by taking a picture of a scene on the display screen in full display of the preset base color using a camera with preset parameters; establishing a color coordinate transformation matrix based on the color difference to correct the color deviation of the camera image, wherein different types of displays correspond to different color coordinate transformation matrices.

[0019] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of obtaining a first image of the scene displaying the test screen on the current display screen as an input image includes: performing color correction on the target camera image using the color coordinate transformation matrix corresponding to the display screen to obtain the first image, wherein the target camera image is obtained by taking a picture of the scene displaying the test screen on the display screen using a camera with preset parameters.

[0020] At least one embodiment of this disclosure provides a parameter adjustment method for a pixel compensation system, comprising: in response to displaying a test pattern on a current display screen for a first duration, switching the display screen to display a solid color image, and capturing an image of the display screen to generate a first image as an input image; obtaining a prediction result corresponding to the input image based on a trained parameter adjustment model; and adjusting the parameters of the pixel compensation system based on the prediction result to adjust the display effect of the display screen.

[0021] In the parameter adjustment method provided in at least one embodiment of this disclosure, the first duration is the duration from when the display screen starts displaying the test pattern until the temperature fluctuation value reaches a first range.

[0022] In at least one embodiment of the parameter tuning method provided in this disclosure, a parameter tuning model is trained to obtain the trained parameter tuning model; wherein, training the parameter tuning model to obtain the trained parameter tuning model includes: acquiring multiple training sample images and sample labels corresponding to each training sample image; inputting at least one training sample image from the multiple training sample images into the parameter tuning model to be trained to obtain predicted sample labels; wherein, the sample labels corresponding to the at least one training sample image and the predicted sample labels are used to adjust the network parameters of the parameter tuning model to be trained.

[0023] In the parameter adjustment method provided in at least one embodiment of this disclosure, the step of obtaining multiple training sample images and sample labels corresponding to each training sample image includes: traversing all combinations of multiple displays and multiple patterns, and for each combination: obtaining training sample images and sample labels corresponding to the training sample images based on the combination of the first display screen among the multiple displays and the first pattern among the multiple patterns.

[0024] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of obtaining a training sample image and a corresponding sample label based on a combination of a first display screen among the multiple display screens and a first pattern among the multiple patterns includes: in response to displaying the first pattern on the first display screen and displaying the first pattern for a first duration, switching the first display screen to display a solid color image, and taking a picture of the display screen to generate a second image as the training sample image; updating the parameter list of the pixel compensation system multiple times, wherein, for each update, taking a picture of the image displayed on the first display screen after adjustment by the pixel compensation system to generate a third image, wherein the third image and the corresponding parameter list are saved as training data pairs; and determining a target parameter list from multiple training data pairs as the sample label corresponding to the training sample image.

[0025] In at least one embodiment of the parameter adjustment method provided in this disclosure, after capturing the image adjusted by the pixel compensation system displayed on the first display screen, the method further includes: switching the first display screen to display the first pattern within a predetermined time.

[0026] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of repeatedly updating the parameter list of the pixel compensation system includes: performing a loop process until a preset condition is met, wherein the loop process includes: selecting an unoptimized parameter in the parameter list as a first parameter; optimizing the first parameter; and fixing the first parameter after optimization, wherein the preset condition includes: all parameters in the parameter list have been optimized.

[0027] In the parameter adjustment method provided in at least one embodiment of this disclosure, the multiple updates of the parameter list of the pixel compensation system include: adjusting the size of the parameter to be optimized according to the third image.

[0028] In at least one embodiment of the parameter adjustment method provided in this disclosure, the method further includes: for each of a plurality of displays: using a camera with preset parameters to capture a scene on the display screen in which a preset base color is displayed in full screen, to obtain a first camera image, wherein the first camera image is used to analyze the color difference between the color of the first camera image and the preset base color, and to establish a color coordinate transformation matrix based on the color difference to correct the color deviation of the camera image, wherein different types of displays correspond to different color coordinate transformation matrices.

[0029] In at least one embodiment of the parameter adjustment method provided in this disclosure, the step of taking a picture of the display screen to generate a first image as an input image includes: taking a picture of the display screen using a camera with preset parameters to obtain a target camera image, wherein the first image is obtained by color correction of the target camera image using a color coordinate transformation matrix corresponding to the display screen.

[0030] In at least one embodiment of the parameter adjustment method provided in this disclosure, the method further includes: selecting an environment with stable lighting conditions; adjusting camera parameters, wherein the camera parameters include at least one of focus, exposure, white balance, shooting mode, color mode, resolution, and encoding format.

[0031] At least one embodiment of this disclosure provides a parameter adjustment device for a pixel compensation system, comprising: an acquisition module configured to acquire a first image of a scene where a test image is displayed on a current display screen as an input image, wherein the scene where the test image is displayed on the display screen includes: a scene where the display screen is switched to display a solid color image in response to displaying a test pattern on the display screen for a first duration; and a prediction module configured to acquire a prediction result corresponding to the input image based on a trained parameter adjustment model, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0032] At least one embodiment of this disclosure provides an electronic device, including: at least one processor; at least one memory including one or more computer program modules; wherein the one or more computer program modules are stored in the at least one memory and configured to be executed by the at least one processor, and the one or more computer program modules are used to implement the parameter adjustment method provided in the at least one embodiment above.

[0033] At least one embodiment of this disclosure provides a non-transitory readable storage medium having computer instructions stored thereon, wherein the computer instructions, when executed by at least one processor, implement the parameter adjustment method provided in the at least one embodiment described above. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments will be briefly described below. Obviously, the drawings described below only relate to some embodiments of this disclosure and are not intended to limit this disclosure.

[0035] Figure 1 A flowchart illustrating a pixel compensation method implemented by a pixel compensation system provided in at least one embodiment of this disclosure;

[0036] Figure 2 A flowchart of a parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure;

[0037] Figure 3 A flowchart of a parameter adjustment model training method provided in at least one embodiment of this disclosure;

[0038] Figure 4 A flowchart illustrating a training sample acquisition method provided in at least one embodiment of this disclosure;

[0039] Figures 5A to 5D A schematic diagram illustrating an example of a method for updating a parameter list provided in at least one embodiment of this disclosure;

[0040] Figure 6AA schematic diagram of a parameter adjustment model provided in at least one embodiment of this disclosure;

[0041] Figure 6B A schematic diagram illustrating the generation of prediction results using a parameter adjustment model provided in at least one embodiment of this disclosure;

[0042] Figure 7A A flowchart illustrating a parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure;

[0043] Figure 7B A schematic block diagram of a parameter adjustment device for a pixel compensation system provided in at least one embodiment of this disclosure;

[0044] Figure 8 This is a schematic block diagram of an electronic device provided in at least one embodiment of the present disclosure;

[0045] Figure 9 A schematic block diagram of another electronic device provided for at least one embodiment of this disclosure;

[0046] Figure 10 A schematic block diagram of a non-transitory readable storage medium provided for at least one embodiment of this disclosure. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the described embodiments of this disclosure without creative effort are within the scope of protection of this disclosure.

[0048] Unless otherwise defined, the technical or scientific terms used in this disclosure shall have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms “first,” “second,” and similar terms used in this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as “comprising” or “including” mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as “connected” or “linked” are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as “upper,” “lower,” “left,” and “right” are used only to indicate relative positional relationships, and these relative positional relationships may change accordingly when the absolute position of the described objects changes.

[0049] The present disclosure will now be described through several specific embodiments. To keep the following description of the embodiments of the present disclosure clear and concise, detailed descriptions of known functions and known components may be omitted. When any component of an embodiment of the present disclosure appears in more than one drawing, that component is represented by the same or similar reference numerals in each drawing.

[0050] Micro LED (MLED) technology is an emerging display technology. MLED displays consist of a high-pixel-density two-dimensional MLED array, where each pixel can be addressed, controlled, and independently driven to emit light, offering advantages such as high brightness, high contrast, and low energy consumption. However, image retention remains a significant technical challenge. Image retention refers to the temporary "trace" of an image left on the display screen, a problem that severely impacts the user's visual experience. This can stem from inconsistent pixel response speeds due to the hardware physical characteristics of the MLED display, or from the degradation of pixel luminous efficiency caused by displaying static images for extended periods.

[0051] Specifically, during operation, when different grayscale levels are displayed in different areas of the MLED display, the luminous efficiency of the MLEDs varies due to different heat accumulation in different areas. For example, the luminous efficiency of the MLEDs decreases in areas with high heat accumulation, while it remains high in areas with low heat accumulation. After displaying an image for an extended period, switching the entire display to the same grayscale level can easily result in image retention in areas with high heat accumulation and low MLED luminous efficiency. Ultimately, this leads to uneven image display when the entire display switches to the same grayscale level, reducing the display effect and impacting the user experience.

[0052] To mitigate this problem, pixel compensation can be used to adjust the color of the image in areas of the display screen with high luminous efficiency to match the color of areas with low luminous efficiency, thereby eliminating image retention. This ensures uniform image display even after the entire display has shown an image for an extended period, improving display quality and user experience when switched to the same grayscale level.

[0053] Therefore, a pixel compensation system can be set up to compensate for pixel artifacts on the display screen. In operation, image or video signals are input to the pixel compensation system via an input device. The system processes the input image frames and outputs corresponding compensated image frames. The controller receives these compensated image frames and drives the display screen to display the image after image artifacts have been eliminated.

[0054] Figure 1 A flowchart illustrating a pixel compensation method implemented using a pixel compensation system provided in at least one embodiment of this disclosure. For example, such as... Figure 1 As shown, the pixel compensation method provided in this embodiment includes the following steps S1 to S15: performing pixel compensation on the first image frame among a plurality of image frames to be displayed on the display screen.

[0055] Step S1: Obtain the temperature rise curve of the display screen when the temperature fluctuation value reaches the preset range under the condition of displaying three pure color images of red, blue and green, and obtain the grayscale processing ratio coefficient based on the temperature rise curves under the three conditions.

[0056] Step S2: Convert the first image frame to grayscale according to the scaling factor to obtain the first grayscale image corresponding to the first image frame.

[0057] Step S3: Use time window weights to perform weighted fusion on the first number of historical normalized brightness images accumulated in the cache pool to obtain the heat time accumulation image.

[0058] Step S4: Perform in-screen diffusion filtering on the heat time accumulation image to obtain the in-screen heat diffusion compensation map.

[0059] Step S5: Perform inter-screen diffusion filtering on the heat time accumulation image to obtain the inter-screen heat diffusion compensation map.

[0060] Step S6: Weighted fusion of the intra-screen thermal diffusion compensation map and the inter-screen thermal diffusion compensation map to obtain the first compensation result.

[0061] Step S7: Obtain the maximum compensation value of each pixel in the first grayscale image according to the grayscale-maximum compensation value lookup table to obtain the maximum compensation value map.

[0062] Step S8: Merge the first compensation result with the maximum compensation value map to obtain the composite compensation map.

[0063] Step S9: Subtract the weighted composite compensation image from each of the three channel components in the first image frame to obtain the first compensated image frame. The first compensated image frame is used to drive the display screen to display the first image frame.

[0064] Step S10: Calculate the grayscale mean of the first grayscale image and convert it according to the grayscale-brightness mapping relationship to obtain the second brightness image.

[0065] Step S11: Calculate the mean brightness value of the second brightness image to obtain the mean brightness image.

[0066] Step S12: Perform a boundary search on the mean brightness image to obtain the boundary search map.

[0067] Step S13: Overlay the boundary search map with the second brightness image to obtain an intermediate brightness image affected by the overlay boundary.

[0068] Step S14: Normalize the intermediate brightness image affected by the superimposed boundary to obtain a normalized brightness image, which is used as the first brightness image.

[0069] Step S15: Add the normalized brightness image (first brightness image) to the cache pool, and in response to the fact that the number of historical normalized brightness images in the cache pool has reached a preset threshold before the normalized brightness image corresponding to the first image frame is added, remove the historical normalized brightness image that is furthest from the current time in the cache pool.

[0070] The pixel compensation system integrates dual compensation strategies across temporal and spatial dimensions. Temporally, it predicts current display temperature changes through statistical analysis of historical data. Spatially, it analyzes display characteristics to establish a full-screen temperature model, enabling more precise pixel compensation for different areas of the display. This comprehensive approach, considering both temporal and spatial factors, allows for precise adjustment and optimization of display effects from multiple perspectives, ensuring more accurate compensation results and significantly improving the long-term stability and visual consistency of display content.

[0071] The inventors of this disclosure have noted that the pixel compensation system described above often requires setting a large number of adjustable parameters. This is because different application scenarios have different requirements for MLED displays, and the temperature characteristics exhibited by customized MLED display products under different requirements are also different. Therefore, by providing flexible parameter adjustment options, it is possible to better adapt to various specific application environments and ensure the optimal performance of the display under different conditions.

[0072] Different specifications of MLED displays typically affect their display performance. Here are some of the main specification differences:

[0073] 1. Pixel Pitch (PPI / Pitch) and LED Type. Pixel pitch refers to the distance between the centers of LED pixels, measured in millimeters (mm). A smaller pixel pitch means more pixels per unit area, theoretically allowing for finer images and making it suitable for close-up viewing. Conversely, a larger pixel pitch is better suited for long-distance viewing. Different pixel pitches affect the rate of temperature rise and peak temperature of the entire screen. Furthermore, the different types of LED chips used in MLED displays (e.g., surface mount devices (SMD), on-board COB, etc.) also affect the display's brightness, color performance, viewing angle, and other display effects, while also exhibiting different temperature rise characteristics.

[0074] 2. Resolution / Display Size. Resolution refers to the total number of pixels on a display screen, usually expressed as the number of pixels in width × height. Display screens with different resolutions typically have different sizes, and the size of an MLED display screen directly affects the display effect. For the same image, displays with different resolutions will also exhibit different image retention characteristics.

[0075] 3. Refresh Rate. Refresh rate refers to the frequency at which a display reloads images, measured in Hertz (Hz). A high refresh rate provides smoother image animations, which is especially important for displaying fast-moving scenes. Different refresh rates result in different temperature rise curves for the display.

[0076] 4. Backplate Structure. Different MLED displays may use different backplate structures, such as aluminum backplates or plastic backplates. The backplate structure not only affects heat dissipation performance but also the overall weight of the display and ease of installation.

[0077] 5. Brightness and Contrast Ratio. Brightness is a standard measure of the light intensity of a display screen, measured in nits. Higher brightness makes the screen appear clearer in bright environments. Contrast ratio refers to the difference between the brightest and darkest pixels; a high-contrast screen performs better when displaying dark backgrounds or details in dark areas.

[0078] 6. Color Gamut. Color gamut refers to the range of colors a display screen can show. A wider color gamut provides richer and more accurate color reproduction, suitable for applications with special color requirements.

[0079] For example, such as Figure 1 As shown, the pixel compensation system contains a large number of adjustable parameters, such as the luminance conversion coefficient (3-bit), edge compensation coefficient (10-bit), time accumulation coefficient (3-bit), intra-screen coefficient (3-bit), inter-screen coefficient (3-bit), superposition coefficient (3-bit), compensation coefficient (10-bit), R-channel compensation coefficient (10-bit), G-channel compensation coefficient (10-bit), and B-channel compensation coefficient (10-bit). These parameters are crucial for optimizing display effects. To adapt to different usage scenarios and display characteristics, these parameters need to be adjusted to find the optimal parameter combination. If all parameter values ​​are iterated through, the total number of parameter combinations is 2. (3+10+3+3+3+3+10+10+10+10) =2 65 There are several combinations. Manually adjusting these parameters is not only time-consuming, but also requires engineers to have deep professional knowledge and rich practical experience.

[0080] At least one embodiment of this disclosure provides a parameter adjustment method for a pixel compensation system, comprising: acquiring a first image of a scene where a test image is displayed on a current display screen as an input image, wherein the scene where the test image is displayed on the display screen includes: a scene where the display screen is switched to display a solid color image in response to displaying a test pattern on the display screen for a first duration; acquiring a prediction result corresponding to the input image based on a trained parameter adjustment model, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0081] In the parameter adjustment method of the pixel compensation system provided in at least one embodiment of this disclosure, for example, a deep learning model and intelligent optimization algorithm are used to automatically and intelligently adjust the parameters of the pixel compensation system, so that the pixel compensation system can automatically adapt to different usage scenarios, significantly reduce the need for manual intervention, and reduce labor and time costs.

[0082] Furthermore, in at least one embodiment, by collecting a large amount of display usage data, the parameter tuning model can learn the behavior patterns of a specific display and predict the optimal parameter combinations for different usage scenarios, thereby achieving rapid parameter optimization and adjustment. This not only improves the performance and stability of the display but also helps reduce labor costs during product development and debugging, accelerating the product launch cycle. Moreover, with the continuous advancement of image retention compensation technology, this type of parameter tuning model can continuously learn and evolve to adapt to new display technologies and market demands.

[0083] The parameter adjustment method and apparatus for the pixel compensation system provided in at least one embodiment of this disclosure are applicable to various types of displays that can be adjusted by intervention using pixel compensation methods to adjust the light emission operation of different areas of the display screen, thereby eliminating afterimages. These displays include, but are not limited to, MLED displays. The embodiments of this disclosure do not limit this.

[0084] Figure 2 A flowchart illustrating a parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure. For example, such as Figure 2 As shown, the parameter adjustment method of the pixel compensation system provided in at least one embodiment of this disclosure includes the following steps S101 to S102.

[0085] Step S101: Obtain a first image of the scene where the test screen is displayed on the current display screen as the input image. The scene where the test screen is displayed on the display screen includes: a scene where the display screen is switched to display a solid color screen in response to the display of a test pattern on the display screen and the display of the test pattern reaches a first duration.

[0086] For example, in step S101, the test pattern can be a high-contrast image or other images that can clearly trigger image retention; this embodiment of the present disclosure does not limit this. For example, by displaying the test pattern on the display screen for a first duration and then switching the display screen to display a solid color image, a detectable image retention can be formed on the display screen. For example, a solid color image can be an image filled with a single color, such as white, blue, green, red, etc.; this embodiment of the present disclosure does not limit this. For example, the first image can be an image obtained by capturing the image retention formed on the display screen and preprocessing the captured camera image, such as image color correction. For example, the first image can be used as the input image for a parameter adjustment model.

[0087] For example, in step S101, the first duration is the time from when the display starts showing the test pattern to when the temperature fluctuation value reaches a first range. The first duration can be, for example, 30 minutes, and the corresponding first duration varies for different types of display screens. For example, the temperature fluctuation value reaching the first range may mean that the temperature itself remains constant, or that the temperature changes smoothly over time, exhibiting only a small range of temperature fluctuations; this disclosure does not limit this. For example, this disclosure does not limit the first range of temperature fluctuation values.

[0088] Step S102: Based on the trained parameter adjustment model, obtain the prediction result corresponding to the input image. The prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0089] For example, the prediction result may include a list of prediction parameters. In step S102, the input image is passed to the trained parameter adjustment model, which outputs a list of prediction parameters corresponding to the input image. For example, the prediction parameter list may include the predicted value of at least one of the parameters mentioned above, such as brightness conversion coefficient, edge compensation coefficient, time accumulation coefficient, intra-screen coefficient, inter-screen coefficient, superposition coefficient, compensation coefficient, R-channel compensation coefficient, G-channel compensation coefficient, and B-channel compensation coefficient. The prediction parameter list is used to adjust the parameters of the pixel compensation system to adjust the display effect of the screen and reduce the impact of image retention. For the current display screen, there may be at least one set of parameters for the pixel compensation system that can achieve the optimal compensation effect under the current conditions.

[0090] In at least one embodiment of this disclosure, after a test pattern is displayed on the display screen for a first duration, the display screen can be switched to display solid color images of different colors (e.g., all white, all blue, all green, all red, etc.). For each solid color image, step S101 is executed to obtain multiple first images corresponding to the various solid color images. Furthermore, for each first image, step S102 is executed to obtain the prediction result corresponding to each first image based on the trained parameter adjustment model, thereby obtaining multiple prediction results corresponding to the various solid color images.

[0091] The parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure further includes: performing a weighted average processing on multiple prediction results to obtain a weighted prediction result, wherein the weighted prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen. The weights of the weighted average processing can be designed according to actual needs, and this disclosure does not limit this. This method integrates the influence of multiple colors, which helps to improve the consistency and quality of screen display.

[0092] For example, tests can be conducted based on four pure color images: pure white, pure blue, pure green, and pure red. For each pure color image, steps S101 to S102 are executed to obtain four corresponding prediction results. Next, the four prediction results are weighted and averaged to obtain a weighted prediction result. The weighted prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the screen.

[0093] The parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure further includes performing the following steps S103 to S104 on each of a plurality of display screens. The plurality of display screens includes the current display screen and other display screens required for training the parameter adjustment model. Steps S103 to S104 can be performed before step S101.

[0094] Step S103: Analyze the color difference between the color of the first camera image and the preset base color. The first camera image is obtained by taking a picture of a scene where the preset base color is displayed in full screen on the display screen using a camera with preset parameters.

[0095] For example, in step S103, a camera with preset parameters can be obtained according to the following steps S1031 to S1033.

[0096] Step S1031: Select an environment with stable lighting conditions.

[0097] Step S1032: Camera focus adjustment.

[0098] Step S1033: Camera exposure adjustment.

[0099] For example, in step S1031, the shooting can be carried out in an indoor environment with controllable lighting conditions to avoid significant changes in natural light affecting the shooting results. For example, a blackout curtain can be used to block the influence of natural light, or a consistent artificial light source can be used to provide a more stable lighting effect.

[0100] For example, in step S1032, in order to prevent the focus from changing during continuous shooting, manual focus mode can be selected, and the focus can be adjusted through the camera viewfinder or screen until the image is clear.

[0101] For example, in step S1033, the average brightness of the scene can be measured using the camera's built-in metering function, and the camera's ISO, aperture, and shutter speed can be adjusted based on the measured brightness information so that the captured camera image is neither overexposed nor underexposed.

[0102] Through steps S1031 to S1033 described above, the camera parameters can be adjusted to an ideal state, enabling the acquisition of clear and correctly exposed camera images under consistent lighting conditions. It should be noted that the focusing and exposure in steps S1032 to S1033 are merely examples of camera parameters; various other camera parameters such as white balance, shooting mode, color mode, resolution, and encoding format can also be adjusted. This embodiment of the disclosure does not limit these adjustments. The purpose of camera parameter adjustment is to stabilize shooting conditions and can include all related parameters. Once the camera parameters are determined, they will remain unchanged if the external environment does not change in subsequent steps.

[0103] For example, in step S103, the preset base color is red, green, blue, or other colors that may be needed in the standard color space. For example, when shooting a scene where different preset base colors are displayed in full screen on a display screen, the camera parameters remain unchanged. For example, a color analyzer or image processing software can be used to analyze the color difference between the color of the first camera image and the preset base color. For example, sample color values ​​(RGB values) can be extracted from the first camera image as color vectors in the source color space.

[0104] Step S104: Establish a color coordinate transformation matrix based on color differences to correct color deviations in camera images. Different types of displays correspond to different color coordinate transformation matrices.

[0105] For example, in step S104, a color coordinate transformation matrix is ​​established based on the color differences obtained in step S103. This color coordinate transformation matrix is ​​used for camera image color correction in subsequent steps. For example, the following equation is established:

[0106] Y = A·X

[0107] Where Y is the color vector in the target color space, X is the color vector in the source color space, and A is the color coordinate transformation matrix. Here, the target color space is the standard color space, and the color vector of the default base color is the color vector in the target color space. The color coordinate transformation matrix is ​​obtained by solving a system of linear equations to find the elements in the color coordinate transformation matrix.

[0108] Because the display characteristics of MLED displays are affected by various factors, different types of MLED displays may exhibit some inconsistency when displaying the same color. Therefore, each time a different type of MLED display is replaced, the color coordinate transformation matrix needs to be recalculated according to steps S103-S104 above; that is, different types of displays correspond to different color coordinate transformation matrices. Through steps S103-S104 above, camera image correction can be achieved for the differences between different MLED displays, thereby reducing the impact of color inconsistency on subsequent parameter adjustments.

[0109] For example, one example of step S101 includes the following step S1011.

[0110] Step S1011: Use the color coordinate transformation matrix corresponding to the display screen to perform color correction on the target camera image to obtain the first image, wherein the target camera image is obtained by taking a picture of the scene displayed on the display screen using a camera with preset parameters.

[0111] In step S1011, the camera parameters and environment used when capturing the target camera image are consistent with those used when capturing the first camera image in step S103. For example, the color value of each pixel is extracted from the target camera image as a color vector in the source color space, and the color vector of each pixel is transformed using a color coordinate transformation matrix. The transformed color values ​​are then used to replace the color values ​​of each pixel in the target camera image to generate a color-corrected image, i.e., the first image.

[0112] Figure 3 A flowchart of a parameter adjustment model training method provided in at least one embodiment of this disclosure.

[0113] For example, model training methods include supervised learning, unsupervised learning, and semi-supervised learning. Supervised learning utilizes labeled training samples and adjusts network weights through backpropagation to minimize the difference between the predicted output and the sample labels. Unsupervised learning infers the intrinsic properties and patterns of data from unlabeled training samples, thereby obtaining the relationships between samples. Semi-supervised learning combines the characteristics of supervised and unsupervised learning, using a limited number of labeled training samples and a large number of unlabeled training samples for co-training, such as through transfer learning or pseudo-label learning to improve model performance. It should be noted that the parameter adjustment model training method provided in at least one embodiment of this disclosure is not limited to... Figure 3 The method shown allows you to choose the appropriate model training method based on your actual needs.

[0114] For example, such as Figure 3 As shown, in the parameter adjustment method of the pixel compensation system provided in at least one embodiment of this disclosure, the trained parameter adjustment model is obtained through the following steps S201 to S203.

[0115] Step S201: Obtain multiple training sample images and the sample label corresponding to each training sample image.

[0116] An example of step S201 includes iterating through all combinations of multiple displays and multiple patterns, and performing step S301 for each combination.

[0117] Step S301: Based on the combination of the first display screen among multiple display screens and the first pattern among multiple patterns, obtain the training sample image and the sample label corresponding to the training sample image.

[0118] For example, multiple patterns include: patterns of multiple different colors, patterns of multiple different sizes, patterns of multiple different shapes, and / or patterns of multiple different positions. For example, multiple displays include: displays of multiple different backlight types, displays of multiple different resolutions, and / or displays of different backplane structures. For example, for m types of displays and n types of patterns, there are a total of m×n combinations (where m and n are positive integers).

[0119] For example, in step S301, "first display screen" is used to refer to the display screen that is currently being described, rather than necessarily being the first display screen among multiple display screens; it can be any of the multiple display screens mentioned above. Similarly, "first pattern" can be any of the multiple patterns mentioned above.

[0120] Figure 4 A flowchart illustrating a training sample acquisition method provided in at least one embodiment of this disclosure. For example, such as... Figure 4 As shown, an example of step S301 includes the following steps S3011 to S3013.

[0121] Step S3011: Obtain a second image based on the scene of displaying training images on the first display screen as a training sample image. The scene of displaying training images on the first display screen includes: in response to displaying a first pattern on the first display screen and displaying the first pattern for a first duration, switching the first display screen to display a solid color image.

[0122] For example, in step S3011, by displaying the first pattern on the first display screen for a first duration and then switching the first display screen to display a solid color image, a detectable afterimage can be formed on the first display screen. For example, the solid color image can be an image filled with a single color, such as white, blue, green, or red. This embodiment of the present disclosure does not limit this, and multiple training data pairs can be obtained in subsequent operations based on solid color images of different colors. For example, the second image can be an image obtained by capturing the afterimage formed on the first display screen and preprocessing the captured camera image. Preprocessing, for example, is image color correction, that is, using the color coordinate transformation matrix corresponding to the first display screen to perform color correction on the camera image. The color coordinate transformation matrix can be obtained through the above steps S103 to S104. For example, the second image can be used as a training sample image.

[0123] Step S3012: Update the parameter list of the pixel compensation system multiple times. For each update, obtain a third image of the scene displayed on the first display screen after adjustment by the pixel compensation system, and save the third image and the corresponding parameter list as training data pairs.

[0124] For example, in step S3012, a parameter update strategy can be set for updating the parameter list of the pixel compensation system, such as gradually adjusting the parameter values ​​or using a search algorithm.

[0125] For example, in step S3012, after each update of the parameter list, the image adjusted by the pixel compensation system after the parameter update is displayed in real time on the first display screen, at which time the color of the afterimage on the first display screen changes. For example, the third image can be obtained by taking a picture of the changed afterimage on the first display screen and preprocessing the captured camera image. The preprocessing operation can be referred to the above example, and will not be repeated here.

[0126] In step S3012, after acquiring a third image of a scene based on the image adjusted by the pixel compensation system displayed on the first display screen, the first display screen is switched to display the first pattern within a predetermined time.

[0127] For example, to avoid the afterimage effect on the display being affected by a solid color image, the display needs to be switched back to showing the first image immediately after shooting. The preset time can be sufficiently short, such as less than one second.

[0128] When updating the parameter list of the pixel compensation system, attempting to iterate through all parameter values ​​would result in a massive space of parameter combinations, making it virtually impossible to iterate through all combinations within a finite amount of time. Furthermore, the vast majority of these parameter combinations are invalid. To reduce the traversal of invalid parameter combinations, the following example illustrates a method that can be used.

[0129] For example, one example of updating the parameter list of the pixel compensation system multiple times in step S3012 includes the following step S3012a.

[0130] Step S3012a: Execute the loop process until the preset condition is met. The loop process includes: selecting an unoptimized parameter from the parameter list as the first parameter; optimizing the first parameter, and fixing the first parameter after optimization. The preset condition includes that all parameters in the parameter list have been optimized.

[0131] For example, in step S3012a, some parameters of the pixel compensation system can be decoupled, which means that a certain parameter can be adjusted to the optimal value first, fixed, and then other parameters can be adjusted for optimization.

[0132] For example, another example of updating the parameter list of the pixel compensation system multiple times in step S3012 includes step S3012b.

[0133] Step S3012b: Adjust the size of the parameter to be optimized based on the third image.

[0134] For example, in step S3012b, after each parameter update, the effect of real-time compensation is recorded in the third image. The direction of parameter adjustment can be determined based on the color differences in different regions of the third image, i.e., whether the parameter needs to be increased or decreased. For example, for the compensation coefficient parameter, a value that is too small or too large will directly affect the effect of real-time compensation. Therefore, by analyzing the effect of real-time compensation recorded in the third image, it can be determined whether the parameter needs to be increased or decreased.

[0135] Figures 5A-5D This is a schematic diagram illustrating an example of a method for updating a parameter list provided in at least one embodiment of the present disclosure.

[0136] For example, in step S3011, by displaying on the first display screen, as shown in the image... Figure 5AWhen the first pattern shown reaches a first duration and the first display screen is switched to display a solid color image, a detectable afterimage can be formed on the first display screen. After capturing and preprocessing the afterimage formed on the first display screen, a result such as... Figure 5B The second image shown. The higher the grayscale value or brightness value of the pixels in the first pattern, the darker the color of the corresponding pixels in the second image.

[0137] For example, in step S3012, the "compensation coefficient" from the parameter list is selected as the parameter to be adjusted. This updates the parameter list of the pixel compensation system for the first time, and the image adjusted by the pixel compensation system after parameter updates is displayed in real-time on the first display screen. At this time, the color of the afterimage on the first display screen changes. After capturing and preprocessing the changed afterimage on the first display screen, the following can be obtained: Figure 5C The third image is shown, and the third image and its corresponding parameter list are saved as a training data pair. The third image is analyzed, as follows: Figure 5C As shown, the color of region 2 is lighter than that of region 1, indicating that the compensation value in region 2 is too small. Ideally, the compensation value in region 2 should be consistent with that in region 1. Therefore, we can conclude that the current compensation coefficient is too small, and determine that the direction of parameter adjustment in the next update should be to increase the compensation coefficient.

[0138] Next, the compensation coefficient is increased, the parameter list of the pixel compensation system is updated again, and the image adjusted by the updated pixel compensation system is displayed in real time on the first display screen. At this time, the color of the afterimage on the first display screen changes again. After taking a picture of the changed afterimage on the first display screen and preprocessing it, the following can be obtained: Figure 5D The third image is shown, and the third image and its corresponding parameter list are saved as a training data pair. The third image is analyzed, as follows: Figure 5D As shown, the color of region 2 is darker than that of region 1, indicating that the compensation value in region 2 is too large. Therefore, we can conclude that the current compensation coefficient is too large, and determine that the direction of parameter adjustment in the next update is to reduce the compensation coefficient.

[0139] Repeat the above steps until the compensation coefficient is adjusted to the optimal value, at which point the color in region 2 is closest to the color in region 1. Fix the compensation coefficient and stop adjusting it. Then select the next parameter to be adjusted in the parameter list and continue adjusting until all parameters in the parameter list have been optimized.

[0140] Step S3013: Determine a list of target parameters from multiple training data pairs, which will serve as the sample labels corresponding to the training sample images.

[0141] For example, one example of step S3013 includes steps S3013a to S3013b.

[0142] Step S3013a: Calculate the uniformity of the third image in each of the multiple training data pairs.

[0143] Step S3013b: Determine the parameter list corresponding to the target third image with the highest uniformity as the target parameter list.

[0144] For example, in step S3013a, all the third images acquired in step S3012 need to be judged for uniformity. Uniformity is an important indicator for measuring the compensation effect of the pixel compensation system. The higher the uniformity, the better the display consistency of the entire display screen, that is, the better the compensation effect of the pixel compensation system.

[0145] For example, uniformity is determined based on the degree of color deviation.

[0146] Color cast is typically described using color deviation values ​​(Delta uv, Duv). The color deviation value represents the uniformity of color across an object's surface; it reflects the intensity of color across different areas. If an object's surface has multiple colors, or different hues and brightness levels, it indicates uneven color distribution, and the color deviation value will be correspondingly higher. Therefore, the lower the color deviation value of an image, the higher its uniformity.

[0147] For example, in steps S3013a to S3013b, the color deviation value of each third image is calculated, and the parameter list corresponding to the third image with the smallest color deviation value is the parameter list of the pixel compensation system with the best compensation effect at present, and it is used as the sample label corresponding to the training sample image.

[0148] Step S202: Input at least one training sample image from multiple training sample images into the training parameter adjustment model to obtain the predicted sample label.

[0149] Figure 6A This is a schematic diagram of a parameter adjustment model provided in at least one embodiment of the present disclosure.

[0150] For example, such as Figure 6A As shown, the parameter adjustment model 600 provided in at least one embodiment of this disclosure includes a deep learning model, which includes a plurality of multi-head attention models 610 and a prediction module 620 connected in sequence. Each multi-head attention model 610 includes a plurality of self-attention modules 611, a normalization module 612 and a multilayer perceptron module 613; the prediction module 620 is connected to the last multi-head attention model 610 among the plurality of multi-head attention models 610 connected in sequence, and is configured to obtain a prediction result based on the output of the last multi-head attention model 610.

[0151] In at least one embodiment of this disclosure, the multi-head attention model refers to the Transformer model, which captures the relationships between image patches through a self-attention mechanism. This mechanism allows the model to focus on the local and global relationships of the image during image processing and ignore irrelevant background information.

[0152] In at least one embodiment of this disclosure, the display screen can be a spliced ​​screen comprising multiple sub-display screens, which together display an image frame. For example, the multiple sub-display screens of the spliced ​​screen can be spliced ​​in an M row × N column arrangement, where M and N are positive integers, and at least one of them is greater than 1. This disclosure does not limit the splicing method. Since the temperature characteristics of the spliced ​​screen are related to each sub-display screen that makes up the spliced ​​screen, the multi-head attention model is suitable for parameter adjustment model construction in this scenario.

[0153] For example, one example of step S202 includes steps S2021 to S2025.

[0154] Step S2021: Divide the input image into multiple image blocks.

[0155] For example, in step S2021, the input image is a training sample image, which can be segmented according to the layout of the splicing screen. For example, when multiple sub-displays of the splicing screen are spliced ​​in an M-row × N-column arrangement, the training sample image can be segmented into M × N image blocks, where M is the number of rows and N is the number of columns.

[0156] Step S2022: Input each image block into the first multi-head attention model of multiple multi-head attention models in sequence. Each image block generates multiple encoding vectors through multiple self-attention modules. Concatenate all the encoding vectors to obtain concatenated encoding vectors.

[0157] For example, step S2022 is the process of encoding image patches using a self-attention mechanism. In step S2022, assuming the parameter adjustment model includes P multi-head attention models connected in sequence, each image patch is sequentially input into the first multi-head attention model. In each multi-head attention model, each image patch generates multiple encoding vectors through multiple self-attention modules. The encoding vectors capture the features within the image patch and the relationships between them. The encoding vectors corresponding to all image patches are concatenated to form a concatenated encoding vector, which contains the feature information of all image patches.

[0158] Step S2023: Process the concatenated encoded vector through the normalization module and the multilayer perceptron (MLP) module in the first multi-head attention model to obtain the first output vector.

[0159] For example, in step S2023, the concatenated encoded vector is passed through a normalization module to adjust the data distribution, making it have zero mean and unit variance, thereby improving the model's learning efficiency and stability. The normalized concatenated encoded vector is then passed through a multilayer perceptron module. A multilayer perceptron module typically includes at least one fully connected layer for performing nonlinear transformations on the normalized features to extract higher-level feature representations. After the above processing, the first output vector is obtained.

[0160] Step S2024: Process the first output vector sequentially through multiple multi-head attention models other than the first multi-head attention model to obtain the second output vector.

[0161] For example, in step S2024, assuming the parameter adjustment model includes P multi-head attention models connected in sequence, the first output vector is input into the second multi-head attention model, and then sequentially through the third, fourth, and so on, up to the Pth multi-head attention model. After processing by the second to Pth multi-head attention models, a second output vector is obtained, which contains the feature representation after multiple self-attention mechanisms.

[0162] Step S2025: Process the second output vector through the prediction module to generate the prediction result.

[0163] For example, in step S2025, the prediction module extracts useful information from the second output vector and transforms it into a specific prediction result. The prediction module typically includes a normalized exponential function (Softmax) module or a fully connected layer, etc. The prediction result generated in step S2025 can be used as the predicted sample label corresponding to the training sample image.

[0164] For example, in the parameter adjustment model provided in at least one embodiment of this disclosure, the prediction module includes a normalized exponential function module, and an example of step S2025 includes the following steps S2025a to S2025b.

[0165] Step S2025a: Process the second output vector through the normalized exponential function module to obtain the third output vector.

[0166] Step S2025b: Threshold the third output vector to obtain the prediction result.

[0167] Figure 6B This is a schematic diagram illustrating the generation of prediction results using a parameter adjustment model provided in at least one embodiment of this disclosure. Figure 6B This is an example of steps S2021 to S2025 above.

[0168] For example, such as Figure 6BAs shown, the parameter adjustment model includes P multi-head attention models connected in sequence. Step S2021 involves segmenting the input image into multiple image patches. Step S2022 involves sequentially inputting each image patch into the first multi-head attention model. Each image patch generates multiple encoding vectors through multiple self-attention modules within the first multi-head attention model. All encoding vectors are concatenated to obtain a concatenated encoding vector. Step S2023 involves processing the concatenated encoding vector through the normalization module and multilayer perceptron module in the first multi-head attention model to obtain a first output vector. Step S2024 involves sequentially processing the first output vector through the second to the Pth multi-head attention models to obtain a second output vector. Step S2025 involves processing the second output vector through a prediction module to generate a prediction result.

[0169] Step S203: Adjust the network parameters of the model by adjusting the parameters to be trained based on the sample labels and predicted sample labels corresponding to at least one training sample image.

[0170] For example, in step S203, a loss function is selected to calculate the difference between the predicted sample label and the sample label output by the model. In each round of iterative training, the model adjusts the model parameters through backpropagation based on the loss value to gradually reduce the loss and improve the prediction accuracy. For example, the loss function can be the mean squared error loss function (MSE Loss), the mean absolute error loss function (MAELoss), etc., or a combination of different loss functions can be selected. This embodiment of the disclosure does not limit this.

[0171] For example, similar to step S202 above, one example of step S102 includes the following steps S1021 to S1025.

[0172] Step S1021: Divide the input image into multiple image blocks.

[0173] Step S1022: Input each image block into the first multi-head attention model of multiple multi-head attention models in sequence. Each image block generates multiple encoding vectors through multiple self-attention modules. Concatenate all the encoding vectors to obtain concatenated encoding vectors.

[0174] Step S1023: The concatenated encoded vector is processed by the normalization module and the multilayer perceptron module in the first multi-head attention model to obtain the first output vector.

[0175] Step S1024: Process the first output vector sequentially through multiple multi-head attention models other than the first multi-head attention model to obtain the second output vector.

[0176] Step S1025: Process the second output vector through the prediction module to generate the prediction result.

[0177] For a detailed description of steps S1021 to S1025, please refer to the description of steps S2021 to S2025 above, which will not be repeated here.

[0178] The parameter adjustment method for a pixel compensation system provided in at least one embodiment of this disclosure uses a deep learning model and intelligent optimization algorithm to automatically and intelligently adjust the parameters of the pixel compensation system, enabling the pixel compensation system to automatically adapt to different usage scenarios, significantly reducing the need for manual intervention, and lowering labor and time costs.

[0179] The following is an example of a parameter adjustment model training method provided by at least one embodiment of the present disclosure, including the following steps S401 to S409.

[0180] Step S401: For each of the multiple displays, analyze the color difference between the color of the first camera image and the preset base color. The first camera image is obtained by taking a picture of the scene on the display screen where the preset base color is displayed in full screen using a camera with preset parameters.

[0181] Step S402: For each of the multiple displays, establish a color coordinate transformation matrix based on color differences to correct color deviations in camera images. Different types of displays correspond to different color coordinate transformation matrices.

[0182] Step S403: Traverse all combinations of multiple displays and multiple patterns. For each combination: based on the combination of the first display among multiple displays and the first pattern among multiple patterns, perform the following steps S404 to S407.

[0183] Step S404: Perform color correction on the second target camera image using the color coordinate transformation matrix corresponding to the first display screen to obtain the second image. The second target camera image is obtained by capturing the scene on the first display screen displaying the training image using a camera with preset parameters. The scene on the first display screen displaying the training image includes: a scene where, in response to displaying a first pattern on the first display screen for a first duration, the first display screen is switched to display a solid color image.

[0184] Step S405: Update the parameter list of the pixel compensation system multiple times. For each update, use the color coordinate transformation matrix corresponding to the first display screen to perform color correction on the third camera image to obtain the third image. The third target camera image is obtained by taking a picture of the scene displayed on the first display screen after adjustment by the pixel compensation system using a camera with preset parameters, and the third image and the corresponding parameter list are saved as training data pairs.

[0185] Step S406: Calculate the uniformity of the third image in each of the multiple training data pairs.

[0186] Step S407: Determine the parameter list corresponding to the target third image with the highest uniformity as the sample label.

[0187] Step S408: Input at least one training sample image from multiple training sample images into the training parameter adjustment model to obtain the predicted sample label.

[0188] Step S409: Adjust the network parameters of the model by adjusting the parameters to be trained based on the sample labels and predicted sample labels corresponding to at least one training sample image.

[0189] The following is an example of a parameter adjustment method for a pixel compensation system provided in at least one embodiment of the present disclosure, including the following steps S501 to S504.

[0190] Step S501: For the current display screen, analyze the color difference between the color of the first camera image and the preset base color. The first camera image is obtained by taking a picture of the scene on the display screen where the preset base color is displayed in full screen using a camera with preset parameters.

[0191] Step S502: For the current display screen, establish a color coordinate transformation matrix based on color differences to correct color deviations in the camera image.

[0192] Step S503: Perform color correction on the target camera image using the color coordinate transformation matrix corresponding to the current display screen to obtain a first image. The target camera image is obtained by capturing the scene on the display screen where the test image is displayed using a camera with preset parameters. The scene on the display screen where the test image is displayed includes: a scene where the display screen is switched to display a solid color image in response to displaying a test pattern on the display screen for a first duration.

[0193] Step S504: Based on the trained parameter adjustment model, obtain the prediction result corresponding to the input image, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0194] Figure 7AThis is a flowchart illustrating a parameter adjustment method for a pixel compensation system provided in at least one embodiment of the present disclosure. Figure 7A As shown, the parameter adjustment method includes steps S710 to S730, which can be performed by a user using the parameter adjustment device.

[0195] Step S710: In response to displaying a test pattern on the current display screen for a first duration, the display screen is switched to display a solid color image, and the display screen image is captured to obtain a first image as the input image.

[0196] For example, the first duration is the time from when the display starts showing the test pattern until the temperature fluctuation value reaches a first range.

[0197] Step S720: Based on the trained parameter adjustment model, obtain the prediction result corresponding to the input image.

[0198] Step S730: Based on the prediction results, adjust the parameters of the pixel compensation system to adjust the display effect of the screen.

[0199] For a detailed description of steps S710 to S730, please refer to the description of steps S101 to S102, which will not be repeated here.

[0200] In the parameter adjustment method of the pixel compensation system provided in at least one embodiment of this disclosure, in order to avoid the impact of inconsistencies that may occur when different types of MLED displays display the same color, it is necessary to perform the following step S740 on each of the multiple displays. The multiple displays include the current display and other displays used to train the parameter adjustment model. Step S740 can be performed before step S710.

[0201] Step S740: Use a camera with preset parameters to take a picture of the scene on the display screen where the preset basic color is displayed in full screen, and obtain the first camera image.

[0202] Step S740 describes the operations performed by the user during the preparation phase of camera image color correction. In step S740, a first camera image is used to analyze the color difference between the first camera image and a preset base color, and a color coordinate transformation matrix is ​​established based on the color difference to correct the color deviation of the camera image. Different types of displays correspond to different color coordinate transformation matrices.

[0203] An example of capturing the display screen in step S710 to obtain a first image as the input image is as follows: using a camera with preset parameters to capture the display screen to obtain a target camera image, wherein the first image is obtained by color correction of the target camera image using the color coordinate transformation matrix corresponding to the display screen.

[0204] For example, in step S710, the user can use a camera with preset parameters to capture the display screen to obtain a target camera image. After preprocessing such as color correction, the target camera image can be used to obtain a first image, which can be used as the input image for the trained parameter adjustment model.

[0205] In order to obtain the trained parameter tuning model, step S750 is included before steps S710 to S730.

[0206] Step S750: Train the parameter tuning model to obtain the trained parameter tuning model.

[0207] An example of step S750 includes steps S751 and S752.

[0208] Step S751: Obtain multiple training sample images and the sample label corresponding to each training sample image.

[0209] Step S752: Input at least one training sample image from the multiple training sample images into the parameter adjustment model to obtain the predicted sample label. The sample label corresponding to at least one training sample image and the predicted sample label are used to adjust the network parameters of the parameter adjustment model to be trained.

[0210] An example of step S752 above includes iterating through all combinations of multiple displays and multiple patterns, and performing step S7520 for each combination.

[0211] Step S7520: Based on the combination of the first display screen among multiple display screens and the first pattern among multiple patterns, obtain the training sample image and the sample label corresponding to the training sample image.

[0212] An example of step S7520 includes the following steps S7521 to S7523.

[0213] Step S7521: In response to displaying a first pattern on the first display screen for a first duration, the first display screen is switched to display a solid color image, and the display screen of the first display screen is captured to generate a second image as a training sample image.

[0214] For example, in step S7521, the user can use a camera with preset parameters to capture the display screen of the first display screen to obtain the corresponding camera image. After preprocessing such as color correction, the camera image can be used to obtain a second image as a training sample image.

[0215] Step S7522: Update the parameter list of the pixel compensation system multiple times. For each update, take a picture of the image adjusted by the pixel compensation system displayed on the first display screen to generate a third image. The third image and the corresponding parameter list are saved as training data pairs.

[0216] For example, in step S7522, the user can use a camera with preset parameters to capture the display screen of the first display screen to obtain the corresponding camera image. After the camera image undergoes preprocessing such as color correction, a third image can be obtained.

[0217] In step S7522, after capturing the image adjusted by the pixel compensation system displayed on the first display screen, the method further includes: switching the first display screen to display the first pattern within a predetermined time.

[0218] For example, one example of updating the parameter list of the pixel compensation system multiple times in step S7522 includes the following step S7522a. This can be performed by the user or implemented directly by software.

[0219] Step S7522a: Execute the loop process until the preset condition is met. The loop process includes: selecting an unoptimized parameter from the parameter list as the first parameter; optimizing the first parameter, and fixing the first parameter after optimization. The preset condition includes that all parameters in the parameter list have been optimized.

[0220] For example, another example of updating the parameter list of the pixel compensation system multiple times in step S7522 includes step S7522b.

[0221] Step S7522b: Adjust the size of the parameter to be optimized based on the third image.

[0222] Step S7523: Determine a list of target parameters from multiple training data pairs, which will serve as the sample labels corresponding to the training sample images.

[0223] For example, users can also set camera parameters through the following steps S760-S770, and use the camera with preset parameters to take pictures in each of the above steps.

[0224] Step S760: Select an environment with stable lighting conditions;

[0225] Step S770: Adjust camera parameters, wherein the camera parameters include at least one of focus, exposure, white balance, shooting mode, color mode, resolution, and encoding format.

[0226] The specific descriptions of each step performed by the user can be found in the descriptions in the above embodiments, and will not be repeated here.

[0227] Figure 7B This is a schematic block diagram of a parameter adjustment device for a pixel compensation system provided in at least one embodiment of the present disclosure.

[0228] For example, such as Figure 7B As shown, the parameter adjustment device 700 includes an acquisition module 701 and a prediction module 702.

[0229] For example, in at least one embodiment of this disclosure, the acquisition module 701 is configured to acquire a first image of a scene where a test image is displayed on the current display screen as an input image. The scene where the test image is displayed on the display screen includes a scene where, in response to displaying a test pattern on the display screen for a first duration, the display screen is switched to display a solid color image. For example, the acquisition module 701 can implement step S101, the specific implementation of which can be found in the description of step S101, and will not be repeated here.

[0230] For example, in at least one embodiment of this disclosure, the prediction module 702 is configured to obtain a prediction result corresponding to the input image based on a trained parameter adjustment model, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen. For example, the prediction module 702 can implement step S102, and its specific implementation method can be referred to the relevant description of step S102, which will not be repeated here.

[0231] For example, in at least one embodiment of this disclosure, the prediction result includes a list of prediction parameters.

[0232] For example, in at least one embodiment of this disclosure, the first duration is the duration from when the display screen starts displaying the test pattern until the temperature fluctuation value reaches a first range.

[0233] For example, in at least one embodiment of this disclosure, multiple different solid color images correspond to multiple prediction results. The parameter adjustment device 700 further includes a weighting module, which is configured to perform weighted averaging on the multiple prediction results to obtain a weighted prediction result. The weighted prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

[0234] For example, in at least one embodiment of this disclosure, the trained parameter tuning model is trained through a model training module, which includes a sample acquisition module, a sample prediction module, and a network parameter adjustment module. The sample acquisition module is configured to acquire multiple training sample images and sample labels corresponding to each training sample image; the sample prediction module is configured to input at least one training sample image from the multiple training sample images into the parameter tuning model to be trained to obtain predicted sample labels; and the network parameter adjustment module is configured to adjust the network parameters of the parameter tuning model to be trained based on the sample labels corresponding to at least one training sample image and the predicted sample labels.

[0235] For example, in at least one embodiment of this disclosure, the sample acquisition module is further configured to: traverse all combinations of multiple displays and multiple patterns, and for each combination: based on the combination of the first display among the multiple displays and the first pattern among the multiple patterns, acquire training sample images and sample labels corresponding to the training sample images.

[0236] For example, in at least one embodiment of this disclosure, the sample acquisition module includes an acquisition unit, an update unit, and a determination unit. The acquisition unit is configured to acquire a second image based on a scene where a training image is displayed on a first display screen as a training sample image, wherein the scene where the training image is displayed on the first display screen includes: a scene where, in response to displaying a first pattern on the first display screen for a first duration, the first display screen is switched to display a solid color image; the update unit is configured to update the parameter list of the pixel compensation system multiple times, wherein, for each update, a third image based on a scene where the image adjusted by the pixel compensation system is displayed on the first display screen is acquired, and the third image and the corresponding parameter list are saved as training data pairs; the determination unit is configured to determine a target parameter list from multiple training data pairs as sample labels corresponding to the training sample images.

[0237] For example, in at least one embodiment of this disclosure, after acquiring a third image of a scene based on a pixel-compensated image displayed on a first display screen, the first display screen is switched to display a first pattern within a predetermined time.

[0238] For example, in at least one embodiment of this disclosure, the update unit is further configured to: perform a loop process until a preset condition is met, the loop process including: selecting an unoptimized parameter from the parameter list as a first parameter; optimizing the first parameter; and fixing the first parameter after optimization, wherein the preset condition includes: all parameters in the parameter list have been optimized.

[0239] For example, in at least one embodiment of this disclosure, the updating unit is further configured to adjust the size of the parameter to be optimized based on the third image.

[0240] For example, in at least one embodiment of this disclosure, the determining unit is further configured to: calculate the uniformity of each third image in a plurality of training data pairs; and determine the parameter list corresponding to the target third image with the highest uniformity as the target parameter list.

[0241] For example, in at least one embodiment of this disclosure, uniformity is determined based on the degree of color deviation.

[0242] For example, in at least one embodiment of this disclosure, the multiple patterns include: multiple patterns of different colors, multiple patterns of different areas, multiple patterns of different shapes, and / or multiple patterns of different positions.

[0243] For example, in at least one embodiment of this disclosure, the multiple displays include: multiple displays with different backlight types, multiple displays with different resolutions, and / or displays with different backplane structures.

[0244] For example, in at least one embodiment of this disclosure, the parameter tuning model includes: a plurality of multi-head attention models connected in sequence, wherein each multi-head attention model includes a plurality of self-attention modules, a normalization module and a multilayer perceptron module; and a prediction module connected to the last multi-head attention model among the plurality of multi-head attention models connected in sequence, configured to obtain a prediction result based on the output of the last multi-head attention model.

[0245] For example, in at least one embodiment of this disclosure, the prediction module is further configured to: segment the input image into multiple image blocks; sequentially input each image block into a first multi-head attention model among multiple multi-head attention models, each image block generates multiple encoding vectors through multiple self-attention modules, concatenate all the encoding vectors to obtain a concatenated encoding vector; process the concatenated encoding vector through a normalization module and a multilayer perceptron module in the first multi-head attention model to obtain a first output vector; sequentially process the first output vector through multiple multi-head attention models other than the first multi-head attention model among multiple multi-head attention models to obtain a second output vector; and process the second output vector through the prediction module to generate a prediction result.

[0246] For example, in at least one embodiment of this disclosure, the prediction module includes a normalized exponential function module. In the prediction module, the second output vector is processed by the prediction module to generate a prediction result, including: processing the second output vector through the normalized exponential function module to obtain a third output vector; and performing threshold division on the third output vector to obtain the prediction result.

[0247] For example, in at least one embodiment of this disclosure, the parameter adjustment device 700 further includes a color correction module configured to: for each of a plurality of displays: analyze the color difference between the color of a first camera image and a preset base color, wherein the first camera image is obtained by taking a picture of a scene on the display screen displaying the preset base color in full screen using a camera with preset parameters; establish a color coordinate transformation matrix based on the color difference to correct the color deviation of the camera image, wherein different types of displays correspond to different color coordinate transformation matrices.

[0248] For example, in at least one embodiment of this disclosure, the acquisition module 701 is further configured to: perform color correction on the target camera image using the color coordinate transformation matrix corresponding to the display screen to obtain a first image, wherein the target camera image is obtained by taking a picture of the scene displaying the test screen on the display screen using a camera with preset parameters.

[0249] It should be noted that the above-mentioned acquisition module 701, prediction module 702, acquisition unit, update unit and determination unit can be implemented by software, hardware, firmware or any combination thereof. For example, the acquisition module 701 and prediction module 702 can be implemented as acquisition circuit and prediction circuit respectively. The embodiments of this disclosure do not limit their specific implementation methods.

[0250] It should be understood that the parameter adjustment device 700 provided in the embodiments of this disclosure can be used to implement the aforementioned parameter adjustment method, and can also achieve similar technical effects as the aforementioned parameter adjustment method, which will not be elaborated here.

[0251] It should be noted that in the embodiments of this disclosure, the parameter adjustment device 700 may include more or fewer circuits or units, and the connection relationship between the various circuits or units is not limited and can be determined according to actual needs. The specific configuration of each circuit is not limited and can be constructed from analog devices, digital chips, or other suitable methods according to circuit principles.

[0252] Figure 8 This is a schematic block diagram of an electronic device provided in at least one embodiment of the present disclosure.

[0253] For example, such as Figure 8As shown, the electronic device 800 includes at least one processor 801 and at least one memory 802. The at least one memory 802 includes one or more computer program modules. These computer program modules are stored in the at least one memory 802 and configured to be executed by the at least one processor 801. The one or more computer program modules include instructions for performing a parameter adjustment method provided in at least one embodiment of this disclosure. When executed by the at least one processor 801, these instructions can perform one or more steps in the parameter adjustment method provided in at least one embodiment of this disclosure. The memory 802 and the processor 801 can be interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0254] For example, processor 801 can be a central processing unit (CPU), a digital signal processor (DSP), or other processing units with data processing and / or program execution capabilities, such as a field-programmable gate array (FPGA); for example, the central processing unit (CPU) can be an x86 or ARM architecture. Processor 801 can be a general-purpose processor or a special-purpose processor, and can control other components in electronic device 800 to perform desired functions.

[0255] For example, memory 802 may include any combination of one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), portable compact disc read-only memory (CD-ROM), USB memory, flash memory, etc. One or more computer program modules may be stored on the computer-readable storage medium, and processor 801 may run one or more computer program modules to implement various functions of electronic device 800. Various application programs and various data, as well as various data used and / or generated by the application programs, may also be stored in the computer-readable storage medium. The specific functions and technical effects of electronic device 800 can be referred to the description of parameter adjustment methods above, and will not be repeated here.

[0256] Figure 9 This is a schematic block diagram of another electronic device provided for at least one embodiment of the present disclosure.

[0257] The electronic device in at least one embodiment of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, personal digital assistants (PDAs), tablet computers (PADs), portable multimedia players (PMPs), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 9 The electronic device 900 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0258] For example, such as Figure 9 As shown, in some examples, electronic device 900 includes a processing unit (e.g., central processing unit, graphics processor, etc.) 901, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 902 or a program loaded from storage device 908 into random access memory (RAM) 903. The RAM 903 also stores various programs and data required for the operation of the computer system. The processing unit 901, ROM 902, and RAM 903 are connected via bus 904. An input / output (I / O) interface 905 is also connected to bus 904.

[0259] For example, the following components can be connected to I / O interface 905: input devices 906 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 907 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 908 including, for example, magnetic tapes, hard disks, etc.; and communication devices 909 including, for example, network interface cards such as LAN cards, modems, etc. Communication device 909 allows electronic device 900 to communicate wirelessly or wiredly with other devices to exchange data and perform communication processing via networks such as the Internet. Drive 910 is also connected to I / O interface 905 as needed. Removable media 911, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on drive 910 as needed so that computer programs read from them can be installed into storage device 908 as needed. Although Figure 9 An electronic device 900 including various devices is shown; however, it should be understood that implementation or inclusion of all shown devices is not required. More or fewer devices may be implemented or included alternatively.

[0260] For example, the electronic device 900 may further include a peripheral interface (not shown in the figure). This peripheral interface can be various types of interfaces, such as a USB interface, a Lightning interface, etc. The communication device 909 can communicate wirelessly with a network and other devices, such as the Internet, an intranet, and / or a wireless network such as a cellular telephone network, a wireless local area network (LAN), and / or a metropolitan area network (MAN). Wireless communication can use any of a variety of communication standards, protocols, and technologies, including but not limited to Global System for Mobile Communications (GSM), Enhanced Data GSM Environment (EDGE), Wideband Code Division Multiple Access (W-CDMA), Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Bluetooth, Wi-Fi (e.g., based on IEEE 802.11a, IEEE 802.11b, IEEE 802.11g, and / or IEEE 802.11n standards), Voice over Internet Protocol (VoIP), Wi-MAX, protocols for email, instant messaging, and / or Short Message Service (SMS), or any other suitable communication protocol.

[0261] For example, the electronic device 900 can be any device such as a mobile phone, tablet computer, laptop computer, e-book, game console, television, digital photo frame, or navigator, or any combination of data processing devices and hardware. The embodiments disclosed herein do not limit this.

[0262] For example, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For instance, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 909, or installed from a storage device 908, or installed from a ROM 902. When the computer program is executed by a processing device 901, the parameter adjustment methods disclosed in embodiments of this disclosure are performed.

[0263] It should be noted that the computer-readable medium described above in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In embodiments of this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In embodiments of this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0264] The aforementioned computer-readable medium may be included in the aforementioned electronic device 900; or it may exist independently and not assembled into the electronic device 900.

[0265] Figure 10 A schematic block diagram of a non-transitory readable storage medium provided for at least one embodiment of this disclosure.

[0266] For example, such as Figure 10 As shown, a non-transiently readable storage medium 1000 stores computer instructions 1001, which, when executed by a processor, perform one or more steps of the parameter adjustment method described above.

[0267] For example, the non-transiently readable storage medium 1000 can be any combination of one or more computer-readable storage media. For instance, one computer-readable storage medium may contain computer-readable program code for acquiring a first image of a scene based on a test image displayed on the current screen as an input image, and another computer-readable storage medium may contain computer-readable program code for acquiring a prediction result corresponding to the input image based on a trained parameter adjustment model. Of course, the above-mentioned program codes can also be stored in the same computer-readable medium, and the embodiments of this disclosure do not limit this.

[0268] For example, when the program code is read by a computer, the computer can execute the program code stored in the computer's storage medium to perform, for example, the parameter adjustment method provided in any embodiment of this disclosure.

[0269] For example, the storage medium may include a memory card for a smartphone, a storage component for a tablet computer, a hard drive for a personal computer, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), portable compact disc read-only memory (CD-ROM), flash memory, or any combination of the above storage media, or other suitable storage media. For example, the readable storage medium may also be... Figure 8 The memory 802 in the memory is described in the foregoing content and will not be repeated here.

[0270] At least one embodiment of this disclosure provides a computer program product, which includes a computer program / instructions, wherein when the computer program / instructions are executed by at least one processor, they perform the parameter adjustment method provided in at least one embodiment of this disclosure.

[0271] Although the present disclosure has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to the embodiments of the present disclosure, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present disclosure are within the scope of protection claimed by the present disclosure.

[0272] The following points should be noted regarding this disclosure:

[0273] (1) The accompanying drawings of the embodiments of this disclosure only involve the structures involved in the embodiments of this disclosure. Other structures can be referred to the general design.

[0274] (2) For clarity, the thickness of layers or regions in the drawings used to describe embodiments of the present disclosure is enlarged or reduced, i.e., these drawings are not drawn to actual scale.

[0275] (3) Where there is no conflict, the embodiments of this disclosure and the features in the embodiments can be combined with each other to obtain new embodiments.

[0276] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. The scope of protection of this disclosure should be determined by the scope of protection of the claims.

Claims

1. A parameter adjustment method for a pixel compensation system, comprising: The first image of the scene where the test screen is displayed on the current display screen is obtained as the input image, wherein the scene where the test screen is displayed on the display screen includes: a scene where the display screen is switched to display a solid color screen in response to the display of a test pattern on the display screen and the display of the test pattern reaches a first duration; Based on the trained parameter adjustment model, the prediction result corresponding to the input image is obtained, wherein the prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

2. The parameter adjustment method according to claim 1, wherein, The prediction results include a list of prediction parameters.

3. The parameter adjustment method according to claim 1, wherein, The first duration is the time from when the display screen starts displaying the test pattern until the temperature fluctuation value reaches a first range.

4. The parameter adjustment method according to claim 1, wherein, Multiple different solid color images correspond to multiple prediction results, and the parameter adjustment method further includes: The multiple prediction results are weighted and averaged to obtain a weighted prediction result, which is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

5. The parameter adjustment method according to claim 1, wherein, The trained parameter tuning model is obtained through the following steps: Obtain multiple training sample images and the sample label corresponding to each training sample image; Input at least one of the multiple training sample images into the training parameter adjustment model to obtain the predicted sample label; The network parameters of the training parameter adjustment model are adjusted based on the sample labels corresponding to the at least one training sample image and the predicted sample labels.

6. The parameter adjustment method according to claim 5, wherein, The step of obtaining multiple training sample images and the sample label corresponding to each training sample image includes: Iterate through all combinations of multiple displays and multiple patterns, for each of the following combinations: Based on the combination of the first display screen among the multiple display screens and the first pattern among the multiple patterns, training sample images and corresponding sample labels are obtained.

7. The parameter adjustment method according to claim 6, wherein, The step of obtaining training sample images and corresponding sample labels based on the combination of a first display screen among the multiple display screens and a first pattern among the multiple patterns includes: A second image based on the scene of displaying training images on the first display screen is obtained as the training sample image, wherein the scene of displaying training images on the first display screen includes: a scene in which the first display screen is switched to display a solid color image in response to displaying the first pattern on the first display screen and displaying the first pattern for the first duration; The parameter list of the pixel compensation system is updated multiple times. For each update, a third image of the scene on the first display screen that is displayed after being adjusted by the pixel compensation system is obtained, and the third image and the corresponding parameter list are saved as a training data pair. A list of target parameters is determined from multiple training data pairs and used as sample labels corresponding to the training sample images.

8. The parameter adjustment method according to claim 7, wherein, After acquiring a third image of the scene based on the image adjusted by the pixel compensation system displayed on the first display screen, The first display screen is switched to display the first pattern within a predetermined time.

9. The parameter adjustment method according to claim 7, wherein, The multiple updates to the parameter list of the pixel compensation system include: The loop process is executed until a preset condition is met, and the loop process includes: Select the unoptimized parameter from the parameter list as the first parameter; The first parameter is optimized, and after optimization, the first parameter is fixed. The preset condition includes: all parameters in the parameter list have been optimized.

10. The parameter adjustment method according to claim 7, wherein, The multiple updates to the parameter list of the pixel compensation system include: Adjust the size of the parameter to be optimized based on the third image.

11. The parameter adjustment method according to claim 7, wherein, The step of determining a list of target parameters from multiple training data pairs as sample labels corresponding to the training sample images includes: Calculate the uniformity of each third image in the plurality of training data pairs; The parameter list corresponding to the target third image with the highest uniformity is determined as the target parameter list.

12. The parameter adjustment method according to claim 11, wherein, The uniformity is determined based on the degree of color deviation.

13. The parameter adjustment method according to claim 6, wherein, The various patterns include: Patterns of various colors, patterns of various sizes, patterns of various shapes, and / or patterns in various positions.

14. The parameter adjustment method according to claim 6, wherein, The plurality of displays includes: Displays with various backlight types, various resolutions, and / or different back panel structures.

15. The parameter adjustment method according to any one of claims 1-14, wherein, The parameter adjustment model includes: Multiple multi-head attention models are connected in sequence, wherein each multi-head attention model includes multiple self-attention modules, normalization modules and multilayer perceptron modules; The prediction module, connected to the last of the multiple multi-head attention models in sequence, is configured to obtain the prediction result based on the output of the last multi-head attention model.

16. The parameter adjustment method according to claim 15, wherein, The method of obtaining the prediction result corresponding to the input image based on the trained parameter adjustment model includes: The input image is divided into multiple image blocks; Each image block is sequentially input into the first multi-head attention model among the multiple multi-head attention models. Each image block generates multiple encoding vectors through the multiple self-attention modules. All the encoding vectors are concatenated to obtain a concatenated encoding vector. The concatenated encoded vector is processed by the normalization module and the multilayer perceptron module in the first multi-head attention model to obtain the first output vector. The first output vector is processed sequentially through the multi-head attention models other than the first multi-head attention model to obtain the second output vector; The second output vector is processed by the prediction module to generate the prediction result.

17. The parameter adjustment method according to claim 16, wherein, The prediction module includes a normalized exponential function module. The second output vector is processed by the prediction module to generate the prediction result, including: The second output vector is processed by the normalized exponential function module to obtain the third output vector; The prediction result is obtained by thresholding the third output vector.

18. The parameter adjustment method according to any one of claims 1-14, further comprising: For each of the multiple displays: Analyze the color difference between the color of the first camera image and the preset base color, wherein the first camera image is obtained by taking a picture of a scene on the display screen in full display using a camera with preset parameters; A color coordinate transformation matrix is ​​established based on the color differences to correct color deviations in camera images. Different types of displays correspond to different color coordinate transformation matrices.

19. The parameter adjustment method according to claim 18, wherein, The step of acquiring a first image of the scene displaying the test screen on the current display screen as the input image includes: The target camera image is color-corrected using the color coordinate transformation matrix corresponding to the display screen to obtain the first image, wherein the target camera image is obtained by taking a picture of the scene displaying the test screen using a camera with preset parameters.

20. A method for adjusting parameters of a pixel compensation system, comprising: In response to displaying a test pattern on the current display screen for a first duration, the display screen is switched to display a solid color image, and the display screen image is captured to generate a first image as an input image. Based on the trained parameter adjustment model, the prediction result corresponding to the input image is obtained; Based on the prediction results, the parameters of the pixel compensation system are adjusted to adjust the display effect of the display screen.

21. The parameter adjustment method according to claim 20, wherein, The first duration is the time from when the display screen starts displaying the test pattern until the temperature fluctuation value reaches a first range.

22. The parameter adjustment method according to claim 20, further comprising: The parameter tuning model is trained to obtain the trained parameter tuning model; The step of training the parameter tuning model to obtain the trained parameter tuning model includes: Obtain multiple training sample images and the sample label corresponding to each training sample image; Input at least one of the multiple training sample images into the training parameter adjustment model to obtain the predicted sample label; The sample labels corresponding to the at least one training sample image and the predicted sample labels are used to adjust the network parameters of the model to be trained.

23. The parameter adjustment method according to claim 22, wherein, The step of obtaining multiple training sample images and the sample label corresponding to each training sample image includes: Iterate through all combinations of multiple displays and multiple patterns, for each of the following combinations: Based on the combination of the first display screen among the multiple display screens and the first pattern among the multiple patterns, training sample images and corresponding sample labels are obtained.

24. The parameter adjustment method according to claim 23, wherein, The step of obtaining training sample images and corresponding sample labels based on the combination of a first display screen among the multiple display screens and a first pattern among the multiple patterns includes: In response to displaying the first pattern on the first display screen and displaying the first pattern for the first duration, the first display screen is switched to display a solid color image, and the display screen of the first display screen is captured to generate a second image as the training sample image. The parameter list of the pixel compensation system is updated multiple times. For each update, the image displayed on the first display screen after adjustment by the pixel compensation system is captured to generate a third image. The third image and the corresponding parameter list are saved as training data pairs. A list of target parameters is determined from multiple training data pairs and used as sample labels corresponding to the training sample images.

25. The parameter adjustment method according to claim 24, wherein, After capturing the image displayed on the first display screen, adjusted by the pixel compensation system, the process further includes: The first display screen is switched to display the first pattern within a predetermined time.

26. The parameter adjustment method according to claim 24, wherein, The multiple updates to the parameter list of the pixel compensation system include: The loop process is executed until a preset condition is met, and the loop process includes: Select the unoptimized parameter from the parameter list as the first parameter; The first parameter is optimized, and after optimization, the first parameter is fixed. The preset condition includes: all parameters in the parameter list have been optimized.

27. The parameter adjustment method according to claim 24, wherein, The multiple updates to the parameter list of the pixel compensation system include: Adjust the size of the parameter to be optimized based on the third image.

28. The parameter adjustment method according to any one of claims 27, further comprising: For each of the multiple displays: The first camera image is obtained by taking a picture of a scene where the preset base colors are displayed in full screen on the display screen using a camera with preset parameters. The first camera image is used to analyze the color difference between the first camera image and the preset base color, and to establish a color coordinate transformation matrix based on the color difference to correct the color deviation of the camera image. Different types of displays correspond to different color coordinate transformation matrices.

29. The parameter adjustment method according to claim 20, wherein, The step of capturing the display screen to generate a first image as an input image includes: The target camera image is obtained by capturing the displayed screen image using a camera with preset parameters. The first image is obtained by color correction of the target camera image using the color coordinate transformation matrix corresponding to the display screen.

30. The parameter adjustment method according to claim 20, further comprising: Choose an environment with stable lighting conditions; Adjust camera parameters, wherein the camera parameters include at least one of focus, exposure, white balance, shooting mode, color mode, resolution, and encoding format.

31. A parameter adjustment device for a pixel compensation system, comprising: The acquisition module is configured to acquire a first image based on the scene of the test screen being displayed on the current display screen as an input image, wherein the scene of the test screen being displayed on the display screen includes: a scene in which the display screen is switched to display a solid color screen in response to the display of a test pattern on the display screen and the display of the test pattern reaches a first duration; The prediction module obtains the prediction result corresponding to the input image based on the trained parameter adjustment model. The prediction result is used to adjust the parameters of the pixel compensation system to adjust the display effect of the display screen.

32. An electronic device, comprising: At least one processor; At least one memory, including one or more computer program modules; The one or more computer program modules are stored in the at least one memory and configured to be executed by the at least one processor, and the one or more computer program modules are used to implement the parameter adjustment method according to any one of claims 1-19.

33. A non-transitory readable storage medium having computer instructions stored thereon, wherein, When the computer instructions are executed by at least one processor, they implement the parameter adjustment method according to any one of claims 1-19.

34. A computer program product comprising computer instructions, wherein, When the computer instructions are executed by at least one processor, the parameter adjustment method according to any one of claims 1-19 is performed.

Images