Compensation method and device of display panel, driving chip and display device

Abstract

The invention relates to a compensation method and device of a display panel, a driving chip and a display device. The compensation method of the display panel comprises the following steps: acquiring a to-be-displayed picture of the display panel; a pre-stored defect reference picture is obtained, the defect reference picture is a display picture of the display panel under a target test gray scale, and the display picture comprises display defect features of the display panel; inputting the to-be-displayed picture and the defect reference picture into a preset deep learning model to output a compensation picture corresponding to the to-be-displayed picture; displaying a target picture compensated for the display defect according to the picture to be displayed and the compensation picture; and efficient and accurate Demura compensation can be realized.

Description

Technical Field

[0001] This application relates to the field of display technology, and in particular to a compensation method and apparatus for a display panel, a driver chip, and a display device. Background Technology

[0002] With the development of display technology, especially the continuous breakthroughs in OLED (Organic Light-Emitting Diode) display technology and the continuous improvement of related technology industries, the demand for OLED panels is showing a rapid growth trend. OLED panels not only have inherent advantages such as self-illumination, ultra-thinness, energy saving, and vibrant colors, but also enable innovative designs such as full-screen and flexible screens, meeting the diversified needs of modern smart display devices.

[0003] In related technologies, OLED panels may experience display unevenness (mura) during the display process, which affects the display effect of the OLED panel. Summary of the Invention

[0004] Therefore, it is necessary to provide a compensation method and apparatus for a display panel, a driver chip, and a display device to address the aforementioned technical problems, which can achieve efficient and accurate Demura compensation.

[0005] In a first aspect, this application provides a compensation method for a display panel, comprising:

[0006] Obtain the image to be displayed on the display panel;

[0007] Obtain a pre-stored defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel;

[0008] Both the image to be displayed and the defect reference image are input into a preset deep learning model to output a compensation image corresponding to the image to be displayed.

[0009] The target image, after compensating for display defects, is displayed based on the image to be displayed and the compensated image.

[0010] Based on the same inventive concept, in a second aspect, this application provides a compensation device for a display panel, comprising:

[0011] The image storage module is used to store a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel;

[0012] The image acquisition module is used to acquire the image to be displayed on the display panel, and to acquire the pre-stored defect reference image;

[0013] The inference generation module is used to input both the screen to be displayed and the defect benchmark screen into a preset deep learning model to output a compensation screen corresponding to the screen to be displayed.

[0014] The compensation display module is used to display a target image that has been compensated for display defects based on the image to be displayed and the compensation image.

[0015] Based on the same inventive concept, in a third aspect, this application provides a driver chip, comprising:

[0016] A storage unit is used to store at least a defect reference image, wherein the defect reference image is a display image of the display panel under a target test grayscale, and the display image includes display defect features of the display panel;

[0017] The computing unit is equipped with a deep learning model. The computing unit is used to obtain the screen to be displayed on the display panel, obtain the pre-stored defect reference screen, and input both the screen to be displayed and the defect reference screen into the deep learning model so that the deep learning model outputs a compensation screen corresponding to the screen to be displayed, and displays a target screen that has been compensated for the display defect according to the screen to be displayed and the compensation screen.

[0018] Based on the same inventive concept, in a fourth aspect, this application provides a driver chip, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the method described in the first aspect.

[0019] Based on the same inventive concept, in a fifth aspect, this application provides a display device, including a display panel and a driving chip as described in the third or fourth aspect, wherein the driving chip is electrically connected to the display panel.

[0020] Based on the same inventive concept, in a sixth aspect, this application provides a display device, including a display panel, a main control chip, and a driver chip;

[0021] The main control chip includes a storage unit and a processing unit;

[0022] The storage unit is at least used to store a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel;

[0023] The computing unit is pre-configured with a deep learning model. The computing unit is used to obtain the screen to be displayed on the display panel, obtain the pre-stored defect reference screen, and input both the screen to be displayed and the defect reference screen into the deep learning model so that the deep learning model outputs a compensation screen corresponding to the screen to be displayed.

[0024] The driving chip is used to drive the display panel to display a target image that has been compensated for display defects, based on the image to be displayed and the compensated image.

[0025] The aforementioned compensation method and apparatus for display panels, driver chip, and display device acquire the display panel's image to be displayed and a pre-stored defect reference image. Both the image to be displayed and the defect reference image are input into a preset deep learning model, causing the deep learning model to output a compensation image corresponding to the image to be displayed. Then, based on the image to be displayed and the compensation image, a target image with compensated display defects is displayed. Thus, the compensation image is directly predicted based on the deep learning model to achieve display defect compensation. Compared to related technologies that achieve full grayscale compensation through pre-stored multiple sets of compensation image data, this method significantly simplifies the compensation method and greatly improves compensation efficiency. Furthermore, it can compensate for different images to be displayed in real time and is applicable to display panels with various defect reference images. The solution is highly flexible and has a low cost. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is one of the flowcharts illustrating a compensation method for a display panel in one embodiment of this application;

[0028] Figure 2 This is a second schematic flowchart of a compensation method for a display panel in one embodiment of this application;

[0029] Figure 3 This is a flowchart illustrating step S103 in one embodiment of this application;

[0030] Figure 4 The compensation method for the display panel in one embodiment of this application also includes a flowchart;

[0031] Figure 5 This is a structural block diagram of a compensation device for a display panel in one embodiment of this application;

[0032] Figure 6 This is a structural block diagram of a driver chip in one embodiment of this application;

[0033] Figure 7 This is one of the structural schematic diagrams of a display device according to an embodiment of this application;

[0034] Figure 8 This is a second schematic diagram of the structure of a display device according to an embodiment of this application. Detailed Implementation

[0035] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0036] It should be noted that when a component is said to be "fixed to" another component, it can be directly attached to the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0037] In this document, spatial terms such as “upper” and “lower” are defined with reference to the accompanying drawings. Therefore, it will be understood that “upper” and “lower” are used interchangeably. It will be understood that when a layer is referred to as being “on” another layer, it can be formed directly on that other layer, or there may be intermediate layers. Therefore, it will be understood that when a layer is referred to as being “directly” on another layer, no intermediate layer is inserted in between.

[0038] In the accompanying drawings, the dimensions of layers and regions may be exaggerated for clarity. It is understood that when a layer or element is referred to as "on" another layer or substrate, the layer or element may be directly on said other layer or substrate, or there may be intermediate layers. Furthermore, it is understood that when a layer is referred to as "between" two layers, the layer may be the only layer between said two layers, or there may be one or more intermediate layers. Additionally, the same reference numerals always denote the same elements.

[0039] In the following text, although terms such as “first” and “second” may be used to describe various components, these components are not necessarily limited to the terms above. The terms above are used only to distinguish one component from another. It will also be understood that expressions used in the singular form include plural expressions unless the singular form has a distinct meaning in the context. Furthermore, in the embodiments below, it will also be understood that the terms “comprising” and / or “having” as used herein indicate the presence of the stated feature or component, but do not exclude the presence or addition of one or more other features or components.

[0040] In the following embodiments, when a layer, region, or element is “connected,” it can be interpreted as the layer, region, or element being connected not only directly but also through other constituent elements placed therebetween. For example, when a layer, region, element, etc., is described as being connected or electrically connected, the layer, region, element, etc., can not only be directly connected or directly electrically connected, but can also be connected or electrically connected through another layer, region, element, etc., placed therebetween.

[0041] In the application documents, the term “and / or” includes any and all combinations of one or more of the related listed items. When a statement such as “at least one of…” follows a list of elements, it modifies the entire list of elements, rather than individual elements within that list.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0043] It should also be understood that the terms “including / comprise” or “have” specify the presence of the stated features, wholes, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0044] Electronic or electrical devices and / or any other related devices or components (e.g., display devices including a display panel and a display panel driver, wherein the display panel driver further includes a drive controller, a gate driver, a gamma reference voltage generator, a data driver, and a transmit driver) according to embodiments of the inventive concepts described herein can be implemented using any suitable hardware, firmware (e.g., application-specific integrated circuits), software, or a combination of software, firmware, and hardware. For example, various components of these devices may be formed on an integrated circuit (IC) chip or on a separate IC chip. Additionally, various components of these devices may be implemented on a flexible printed circuit film, a tape-on-a-package (TCP), a printed circuit board (PCB), or formed on a substrate. Furthermore, various components of these devices may be processes or threads running on one or more processors in one or more computing devices to execute computer program instructions and interact with other system components to perform the various functions described herein. Computer program instructions are stored in memory, which may be implemented in a computing device using standard storage devices such as random access memory (RAM). Computer program instructions may also be stored in other non-transitory computer-readable media such as CD-ROMs, flash drives, etc. Furthermore, those skilled in the art will recognize that the functions of various computing devices may be combined or integrated into a single computing device, or the functions of a particular computing device may be distributed across one or more other computing devices, without departing from the spirit and scope of exemplary embodiments of the present invention.

[0045] While exemplary embodiments of the display module and the display device including the display module have been specifically described herein, many modifications and variations will be apparent to those skilled in the art. Therefore, it will be understood that the display module and the display device including the display module, constructed according to the principles of the invention, may be implemented in ways other than those specifically described herein. This application is also defined in the claims and their equivalents.

[0046] As mentioned in the background section, in related technologies, OLED panels may experience display unevenness (mura) during the display process, affecting the display effect. The inventors of this application have discovered that methods for improving mura in related technologies mainly include: typically requiring the capture of multiple (5-7) images of the display panel at different gray levels; using mathematical models such as the Gamma function and polynomial fitting to fit a full gray-level brightness curve based on these images, generating a gray-level-brightness response curve for each sub-pixel; then setting a target brightness curve for the entire gray level of the panel (the brightness standard of an ideal uniform image, such as a standard curve conforming to Gamma 2.2); comparing the actual fitted brightness curve of each sub-pixel with the target curve to calculate the brightness compensation value for each gray level; converting the calculated full gray-level compensation matrix into multiple sets of compensation image data (corresponding to compensation parameters for different gray levels) and writing them into the panel's driver chip; during panel display, the driver chip calls the corresponding gray-level compensation parameters based on the current display image's gray level value to correct the driving signal of each sub-pixel in real time. However, this method involves a large amount of computation, resulting in high costs and low efficiency.

[0047] In view of this, embodiments of this application provide a compensation method for display panels, which directly predicts the compensation image based on a deep learning model, enabling efficient and accurate Demura compensation. Furthermore, the compensation method for display panels provided in embodiments of this application can be applied to, for example, but not limited to, OLED display panels.

[0048] In one exemplary embodiment, reference is made to Figure 1 A compensation method for a display panel is provided, which may include the following steps S101 to S104.

[0049] S101, Obtain the screen to be displayed on the display panel.

[0050] The image to be displayed is the content image that the display panel normally outputs, representing the brightness and texture features of the displayed content. As an example, the image to be displayed can be a single frame of an independent image to be displayed, or any single frame from a video stream to be displayed.

[0051] S102, acquire a pre-stored defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel.

[0052] The defect reference image is a photograph taken when the display panel is displayed at the target test grayscale. The defect reference image is used to characterize the display defect features of the display panel, such as display unevenness (mura) defects. These defect features include, for example, the location, degree, and type of the defect, without specific limitations. In other words, the defect reference image can also be understood as a defect reference map of the display panel at a uniform grayscale, characterizing the location, degree, and type of defects. The defect reference image is a pixel-level grayscale image characterizing brightness unevenness and / or color unevenness defects of the display panel.

[0053] As an example, conventional testing equipment for testing and displaying defects can be used to control the display panel to display at a target test grayscale, which may be, but is not limited to, full white, full gray, or full black grayscale. After the display panel has been stably displayed at the target test grayscale for a preset time, a high-precision industrial camera is used to photograph the display panel under standard lighting conditions to obtain a defect reference image of the display panel. As an example, the testing equipment writes the digitized data of the defect reference image (such as a pixel grayscale matrix) into the embedded non-volatile memory (such as OTP (One-Time Programmable) or eFlash (embedded Flash)) of the display panel's driver chip to complete the pre-storage.

[0054] S103: Input both the screen to be displayed and the defect reference screen into the preset deep learning model to output a compensation screen corresponding to the screen to be displayed.

[0055] In this process, both the image to be displayed and the baseline image with defects are input into a deep learning model. The deep learning model then uses a pre-trained mapping function to infer, generate, and output a compensation image corresponding to the image to be displayed. The compensation image can be understood as a correction image used to eliminate display defects.

[0056] The impact of display defects on display effects is not fixed. For example, the same Mura defect can be perceived very differently by the human eye in displays with different brightness levels (e.g., Mura is more prominent in dark scenes and relatively less noticeable in bright scenes). The essence of deep learning models is adaptive compensation rules: deep learning models dynamically output compensation magnitude based on the characteristics of display defects and the local brightness of the display screen, rather than generating a fixed compensation template.

[0057] S104, Display the target image that has been compensated for the display defects based on the image to be displayed and the compensation image.

[0058] The process involves performing pixel-level calculations on the image to be displayed and the compensation image to obtain a target image that has compensated for display defects, which then drives the display panel to display the target image. Specifically, the input consists of the entire image to be displayed and the defective baseline image; the deep learning model outputs the entire pixel-by-pixel compensation image, which is then processed pixel-by-pixel to obtain the target image.

[0059] Pixel-level operations include at least one of addition, multiplication, or weighted fusion operations. When the pixel-level operation is addition, the pixel value of any pixel in the image to be displayed is added to the compensation value of the corresponding pixel in the compensation image to obtain the pixel value of the corresponding pixel in the target image. When the pixel-level operation is weighted fusion, the weight coefficients of the image to be displayed and the compensation image can be dynamically adjusted according to the defect intensity in the defect reference image.

[0060] The solution in this application directly predicts the compensation image based on a deep learning model. It then displays the target image with the compensation for the display defect based on the image to be displayed and the compensation image, thereby achieving display defect compensation. Compared to related technologies that achieve full grayscale compensation through pre-stored multiple sets of compensation image data, this solution significantly simplifies the compensation method and greatly improves compensation efficiency. Furthermore, it can compensate for different images to be displayed in real time and is applicable to display panels with various defect reference images. The solution is highly flexible and cost-effective. In addition, the solution in this application only requires one captured image of the current display panel (i.e., one defect reference image). Compared to related technologies that require multiple image captures, fitting of full grayscale brightness curves, calculation of compensation values, storage of multiple sets of compensation images, and full grayscale compensation, this solution effectively simplifies the compensation method and reduces compensation costs to a great extent.

[0061] In one exemplary embodiment, reference is made to Figure 2 A compensation method for a display panel is provided, which may include the following steps S201 to S205.

[0062] S201, Obtain the screen to be displayed on the display panel.

[0063] S202, acquire the pre-stored defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel.

[0064] S203, the screen to be displayed and the defect reference screen are partitioned separately to obtain multiple one-to-one corresponding sub-partition screens to be displayed and defect sub-partition screens, wherein the partition boundary of the screen to be displayed coincides with the partition boundary of the defect reference screen.

[0065] The number of sub-partitions to be displayed is positively correlated with the resolution of the display: the number and size of the sub-partitions can be dynamically adjusted according to the resolution of the display; the higher the resolution, the more sub-partitions there are, and the fewer pixels a single sub-partition has.

[0066] As an example, the display screen and the defect reference screen are partitioned separately, including: obtaining the defect intensity difference between adjacent pixels in the defect reference screen, and dividing adjacent pixels with defect intensity differences less than a preset threshold into the same defect sub-partition screen.

[0067] S204, input both the sub-partition image to be displayed and the corresponding defective sub-partition image into the deep learning model to output the partition compensation value that corresponds one-to-one with the sub-partition image to be displayed.

[0068] Among them, the partition compensation value corresponds to the average value of the defect intensity of all pixels in the defect sub-partition, which can be the optimal compensation value output by the deep learning model after extracting the features of the defect sub-partition.

[0069] S205, all pixels within the same sub-partition of the screen to be displayed are calculated at the pixel level using the corresponding partition compensation value to obtain and display the target screen that has been compensated for the display defects.

[0070] In this embodiment, the distribution and severity of display defects within the same local area of ​​a high-resolution display panel are highly similar, and the local brightness characteristics of the image to be displayed are also relatively consistent. Therefore, the image to be displayed and the defect reference image can be partitioned separately, allowing compensation values ​​to be calculated by partition. The compensation values ​​for the same partition are identical, which can also be understood as generating partitioned compensation images. Finally, pixel-level calculations can be completed by partition, significantly reducing the inference time of deep learning models and the computational load on the driver chip, thus improving compensation efficiency.

[0071] In an exemplary embodiment, after driving the display panel to display the target image, the method further includes: detecting display defects in the target image; if the detected defect level is greater than a preset threshold, adjusting the mapping function parameters of the deep learning model and regenerating the compensation image.

[0072] In one exemplary embodiment, the deep learning model may include a two-branch feature fusion network, enabling it to simultaneously capture defect features and content features. The two-branch feature fusion network includes a first feature extraction branch, a second feature extraction branch, and a feature fusion layer.

[0073] refer to Figure 3 Step S103 involves inputting both the screen to be displayed and the defect baseline screen into a preset deep learning model to output a compensation screen corresponding to the screen to be displayed. This may include the following steps S301 to S304.

[0074] S301, extract the brightness and texture features of the image to be displayed through the first feature extraction branch.

[0075] The first feature extraction branch can extract the brightness features (such as the brightness and darkness values ​​of local pixels) and texture features (such as edges and details) of the content to be displayed. The output feature map can distinguish between "smooth areas" and "detail areas". For example, the edges of text in the image to be displayed belong to the detail area, and blurring should be avoided during compensation; a solid color background belongs to the smooth area, and the compensation range can be increased.

[0076] S302, extract the defect distribution features and defect intensity features of the defect reference image through the second feature extraction branch.

[0077] Specifically, convolutional layers can extract the spatial distribution features (characterizing where the defect is located) and intensity features (characterizing the severity of the defect). The output feature map can then be used to mark "high compensation demand areas" and "low compensation demand areas." For example, dark spot areas in a Mura image would be marked as high-weight feature points.

[0078] S303 uses a feature fusion layer to fuse brightness features, texture features, defect distribution features, and defect intensity features to obtain a fused feature map.

[0079] In this process, the features of the two branches are dynamically fused through a feature fusion layer, which enables the deep learning model to prioritize referencing Mura features in defective regions and prioritize protecting content features in detailed regions.

[0080] The feature fusion layer employs, but is not limited to, attention mechanisms. Attention mechanisms can include at least one of spatial attention, channel attention, and hybrid attention mechanisms. Based on the feature fusion layer, the deep learning model can "focus," for example, prioritizing defect features in areas with severe Mura defects and prioritizing content features in detailed areas of the image to be displayed, ultimately generating a fused feature map that provides a basis for generating the compensated image.

[0081] S304, input the fused feature map into a pre-trained mapping function to generate and output a compensated image corresponding to the image to be displayed based on the mapping function.

[0082] The solution in this embodiment achieves the goal of "accurately compensating for defects without damaging the displayed content" through dual-branch feature fusion. Compared with the fixed-parameter compensation algorithm in related technologies, it has stronger adaptability and generalization.

[0083] In one exemplary embodiment, reference is made to Figure 4 The compensation method for the display panel may also include the following steps S401~S402.

[0084] S401, Construct a training dataset, which includes defective baseline sample images, sample images to be displayed, and ideal defect-free sample images.

[0085] The defect reference sample image can be captured by photographing the display panel when it is displayed in the target test grayscale, and is used to characterize the location, intensity, and type of defects in the display panel. The sample images to be displayed cover various typical display content (such as text, images, and video frames) to simulate the actual display scenario of the terminal. The ideal defect-free sample image can be understood as the "standard answer" in training; it can be obtained by using high-precision optical instruments to capture the image of the above-mentioned sample image to be displayed on a defect-free display panel, and use it as the ideal defect-free sample image. Alternatively, the sample image to be displayed and the inverse correction value of the display defect can be superimposed to generate the ideal defect-free sample image, thus artificially synthesizing the ideal defect-free sample image.

[0086] S402, based on the training dataset, uses a combination of pixel loss function and perceptual loss function to train the mapping function of the initial deep learning model until the loss value converges, thus obtaining the deep learning model.

[0087] The mapping function is determined by the weight parameters of the deep learning network. Its structure can be a two-branch feature fusion network, adaptable to dual-input scenarios of "defect + content". The loss function is the core supervision metric during training, used to measure the difference between the model's predicted target image and the "standard answer".

[0088] The pixel loss function calculates the pixel value error between the "sample image to be displayed + compensated sample image" and the "ideal defect-free sample image," ensuring that the brightness and chromaticity of the compensated image are consistent with the target, thus eliminating the visual deviation caused by display defects. In other words, the role of the pixel loss function is to ensure that the final compensated target image accurately compensates for display defects in terms of brightness and chromaticity values. As an example, the pixel loss function may include at least one of Mean Absolute Error (MAE) and Mean Square Error (MSE).

[0089] The perceptual loss function uses a pre-trained image feature model to extract high-level semantic features of the image, avoiding visual distortions such as blurred text and color aberrations in the compensated output. Optionally, adversarial loss can be introduced, i.e., a discriminator network is introduced to make the generated target image closer to the visual effect of a real, flawless image, improving the subjective experience for the human eye. In other words, the essence of the perceptual loss function is different from the pixel loss function, which directly compares pixel values. Instead, it inputs the target image and the ideal image into a pre-trained visual feature model (such as a VGG network), extracts high-level visual features of the image (such as text edges, image texture, and color distribution), and then compares the differences in these features to calculate the "visual error value." The role of the perceptual loss function is to ensure that the final compensated target image is consistent with the ideal image in human visual perception.

[0090] The network weights are iteratively optimized using the backpropagation algorithm, allowing the model to gradually approach the optimal mapping function. In this embodiment, a batch of defective baseline sample images and a batch of sample images to be displayed are input into the network to predict compensation sample images. The loss value between "sample image to be displayed + compensation sample image" and "ideal defect-free sample image" is calculated. The loss error is backpropagated to update the network weight parameters, allowing the loss value to gradually decrease. The above process is repeated until the loss converges, meaning the model's compensation effect stabilizes and reaches the expected level.

[0091] As an example, pixel loss and perceptual loss can be added together with certain weights to obtain the combined loss; for example, combined loss = a × pixel loss + b × perceptual loss, where a and b are weighting coefficients. This ensures that display defects are accurately eliminated (pixel loss) while maintaining visual distortion-free display content (perceptual loss), thus compensating for display defects on the display panel and improving display quality. Furthermore, during training, the combined loss gradually decreases as the model learns (the error becomes smaller). When the combined loss reaches a stable value and no longer decreases significantly after multiple training rounds, it indicates that the model has learned to accurately compensate for display defects without disrupting the image, and training can be stopped.

[0092] After training, the model can be tested using an independent validation dataset (a dataset trained without parameters) to verify its compensation effect under new display defects and new content to be displayed, thus avoiding overfitting. If the validation results are satisfactory, the trained network weights can be solidified into the final mapping function for subsequent inference.

[0093] In this embodiment, pixel loss is aimed at "numerical accuracy" to ensure that display defects are largely offset, and perceptual loss is aimed at "visual naturalness" to ensure that the displayed content is not distorted. The combined loss is a weighted sum of pixel loss and perceptual loss, which takes into account both eliminating display defects and improving user experience, so that a clear picture is finally displayed to the user. The convergence of loss values ​​is a signal that the model training is complete, indicating that the model's compensation ability has reached a stable standard.

[0094] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the accompanying drawings may include multiple steps or stages, which are not necessarily completed at the same time, but may be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but may be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0095] Based on the same inventive concept, and referring to Figure 5 A compensation device for a display panel is provided, which may include a screen storage module 510, a screen acquisition module 520, an inference generation module 530, and a compensation display module 540.

[0096] The image storage module 510 is used to store a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel.

[0097] The image acquisition module 520 is used to acquire the image to be displayed on the display panel, as well as to acquire the pre-stored defect reference image.

[0098] The inference generation module 530 is used to input both the screen to be displayed and the defect benchmark screen into a preset deep learning model to output a compensation screen corresponding to the screen to be displayed.

[0099] The compensation display module 540 is used to display a target image that has been compensated for display defects based on the image to be displayed and the compensation image.

[0100] The solution in this embodiment, the compensation device for the display panel, and the driving method for the display panel belong to the same inventive concept, can solve the same technical problem, and thus achieve the same technical effect. Repeated content will not be described again here.

[0101] Based on the same inventive concept, and referring to Figure 6A driver chip is provided, comprising a storage unit 610 and a processing unit 620. The storage unit 610 is used to store at least a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes display defect features of the display panel. The processing unit 620 is preset with a deep learning model, and the processing unit 620 is used to obtain the image to be displayed of the display panel, obtain the pre-stored defect reference image, and input both the image to be displayed and the defect reference image into the deep learning model so that the deep learning model outputs a compensation image corresponding to the image to be displayed, and displays the target image with the display defect compensated according to the image to be displayed and the compensation image.

[0102] In this embodiment, the driver chip is a hardware unit that directly outputs pixel driving voltage to the display panel. The driver chip executes the compensation method for the display panel, significantly reducing the latency from "compensation calculation" to "pixel driving," achieving real-time compensation, and avoiding screen stuttering and ghosting. This approach is more suitable for mid-to-low-end display panels such as mobile phone screens and tablet screens. The model algorithm is lightweight and requires less computing power. Furthermore, the driver chip and the display panel driving method belong to the same inventive concept, solve the same technical problem, and achieve the same technical effect; therefore, repeated details will not be elaborated here.

[0103] Based on the same inventive concept, a driver chip is provided, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the compensation method for the display panel provided in any of the above embodiments.

[0104] The solution in this embodiment, the compensation method for the driver chip and the display panel, belong to the same inventive concept, can solve the same technical problem, and thus achieve the same technical effect. Repeated content will not be repeated here.

[0105] Based on the same inventive concept, and referring to Figure 7 A display device is provided, comprising a display panel and a driver chip as provided in any of the above embodiments, wherein the driver chip is electrically connected to the display panel. The solution of this embodiment, the compensation method for the display device and the display panel, belong to the same inventive concept, can solve the same technical problem, and thus achieve the same technical effect; therefore, repeated details are not described here.

[0106] Based on the same inventive concept, and referring to Figure 8A display device is provided, comprising a display panel, a main control chip, and a driver chip. The main control chip includes a storage unit 810 and a processing unit 820. The storage unit 810 is used to store at least a defect reference image, wherein the defect reference image is a display image of the display panel under a target test grayscale, and the display image includes display defect features of the display panel. The processing unit 820 is pre-loaded with a deep learning model, and is used to acquire the image to be displayed on the display panel, acquire the pre-stored defect reference image, and input both the image to be displayed and the defect reference image into the deep learning model so that the deep learning model outputs a compensated image corresponding to the image to be displayed. The driver chip is used to drive the display panel to display the target image with compensated display defects based on the image to be displayed and the compensated image.

[0107] In this embodiment, the main control chip executes deep learning model inference and generates compensated images, while the driver chip receives compensation data, performs pixel-level calculations, and drives the display panel. This approach is more suitable for high-end display panels such as 8K TV screens and professional industrial control screens, and can support more complex deep learning models and greater computing power. Furthermore, the driving methods for the display device and display panel belong to the same inventive concept, solve the same technical problems, and achieve the same technical effects; therefore, repeated details will not be elaborated here.

[0108] It is understood that the display device in the embodiments of this application can be any product or component with display function, such as OLED display device, QLED display device, electronic paper, mobile phone, tablet computer, television, monitor, laptop computer, digital photo frame, navigator, wearable device, Internet of Things device, etc., and the embodiments disclosed in this application do not limit this.

[0109] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification. The above embodiments only illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A compensation method for a display panel, characterized in that, include: Obtain the image to be displayed on the display panel; Obtain a pre-stored defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel; Both the image to be displayed and the defect reference image are input into a preset deep learning model to output a compensation image corresponding to the image to be displayed. The target image, after compensating for display defects, is displayed based on the image to be displayed and the compensated image.

2. The compensation method for the display panel according to claim 1, characterized in that, The step of inputting both the image to be displayed and the defect reference image into a preset deep learning model to output a compensated image corresponding to the image to be displayed includes: The screen to be displayed and the defect reference screen are partitioned to obtain multiple one-to-one corresponding sub-partition screens to be displayed and defect sub-partition screens, wherein the partition boundary of the screen to be displayed coincides with the partition boundary of the defect reference screen. The sub-partition image to be displayed and the corresponding defective sub-partition image are both input into the deep learning model to output a partition compensation value that corresponds one-to-one with the sub-partition image to be displayed. The step of displaying a target image that has been compensated for display defects based on the image to be displayed and the compensated image includes: All pixels within the same sub-partition of the image to be displayed are subjected to pixel-level calculations using the corresponding partition compensation value to obtain and display the target image.

3. The compensation method for the display panel according to claim 2, characterized in that, The step of partitioning the screen to be displayed and the defect reference screen includes: Obtain the defect intensity difference between adjacent pixels in the defect reference image, and classify adjacent pixels whose defect intensity difference is less than a preset threshold into the same defect sub-partition image.

4. The compensation method for the display panel according to claim 1, characterized in that, Also includes: Construct a training dataset, wherein the training dataset includes defective baseline sample images, sample images to be displayed, and ideal defect-free sample images; Based on the training dataset, the mapping function of the initial deep learning model is trained using a combination of pixel loss function and perceptual loss function until the loss value converges, thus obtaining the deep learning model.

5. The compensation method for the display panel according to claim 1, characterized in that, The deep learning model includes a first feature extraction branch, a second feature extraction branch, and a feature fusion layer; The step of inputting both the image to be displayed and the defect reference image into a preset deep learning model to output a compensated image corresponding to the image to be displayed includes: The brightness and texture features of the image to be displayed are extracted through the first feature extraction branch; The defect distribution features and defect intensity features of the defect reference image are extracted through the second feature extraction branch; The brightness feature, texture feature, defect distribution feature, and defect intensity feature are fused through the feature fusion layer to obtain a fused feature map; The fused feature map is input into a pre-trained mapping function to generate and output a compensated image corresponding to the image to be displayed.

6. The compensation method for the display panel according to claim 1, characterized in that, After driving the display panel to display the target image, the method further includes: The target image is subjected to display defect detection. If the detected defect level is greater than a preset threshold, the mapping function parameters of the deep learning model are adjusted and a compensation image is regenerated.

7. A compensation device for a display panel, characterized in that, include: The image storage module is used to store a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel; The image acquisition module is used to acquire the image to be displayed on the display panel, and to acquire the pre-stored defect reference image; The inference generation module is used to input both the screen to be displayed and the defect benchmark screen into a preset deep learning model to output a compensation screen corresponding to the screen to be displayed. The compensation display module is used to display a target image that has been compensated for display defects based on the image to be displayed and the compensation image.

8. A driver chip, characterized in that, include: A storage unit is used to store at least a defect reference image, wherein the defect reference image is a display image of the display panel under a target test grayscale, and the display image includes display defect features of the display panel; The computing unit is equipped with a deep learning model. The computing unit is used to acquire the screen to be displayed on the display panel, acquire the pre-stored defect reference screen, and input both the screen to be displayed and the defect reference screen into the deep learning model so that the deep learning model outputs a compensation screen corresponding to the screen to be displayed, and displays a target screen that has been compensated for the display defect according to the screen to be displayed and the compensation screen.

9. A driver chip, characterized in that, The method includes a memory and a processor, the memory storing a computer program, characterized in that the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

10. A display device, characterized in that, It includes a display panel and a driver chip as described in claim 8, wherein the driver chip is electrically connected to the display panel.

11. A display device, characterized in that, Includes display panel, main control chip, and driver chip; The main control chip includes a storage unit and a processing unit; The storage unit is at least used to store a defect reference image, wherein the defect reference image is a display image of the display panel under the target test grayscale, and the display image includes the display defect features of the display panel; The computing unit is pre-configured with a deep learning model. The computing unit is used to obtain the screen to be displayed on the display panel, obtain the pre-stored defect reference screen, and input both the screen to be displayed and the defect reference screen into the deep learning model so that the deep learning model outputs a compensation screen corresponding to the screen to be displayed. The driving chip is used to drive the display panel to display a target image that has been compensated for display defects, based on the image to be displayed and the compensated image.

Images