AI-based display screen color calibration method, display device and storage medium

Abstract

This application relates to the field of screen display technology, and discloses an AI-based display screen color calibration method, display device, and storage medium. The method includes: calling a content recognition model to identify the content to be displayed and obtaining the content category; constructing a user preference vector based on the content category; obtaining a screen aging compensation factor for the display screen; obtaining the input color value of each pixel in the content to be displayed, wherein the input color value is the color value of the content to be displayed in a standard color space; calibrating the input color value according to the user preference vector and the screen aging compensation factor to obtain a target color value; and displaying the content to be displayed according to the target color value. Since the target color value integrates physical aging compensation and subjective preference adaptation, the display content seen by the user eliminates the color shift caused by screen aging and conforms to their personal subjective preferences under different content categories.

Description

Technical Field

[0001] This application relates to the field of display device technology, and in particular to an AI-based display screen color calibration method, display device, and storage medium. Background Technology

[0002] During the production and use of display screens, due to batch variations in manufacturing processes and the decay of luminous efficiency in various color channels caused by long-term use, the actual displayed colors often deviate from the ideal colors defined in the standard color space. This is especially true in self-emissive display technologies such as Organic Light Emitting Diodes (OLEDs) and Micro-LEDs, where the aging rates of red, green, and blue sub-pixels differ. As usage time accumulates, color shifts and other phenomena may occur on the screen, necessitating color calibration to ensure display quality.

[0003] Existing technologies have proposed various color calibration methods, such as calculating calibration values ​​by obtaining the difference between the actual displayed color and the reference color through color test charts, or calculating the calibration matrix by combining a solid color image with multiple color space conversions. However, these methods all use a fixed standard color as the calibration target, resulting in all users obtaining exactly the same calibration results on the same device. They fail to consider the differences in subjective color preferences among different users, and do not take into account the differentiated color needs of different users, resulting in a lack of personalized adaptability in color calibration results. Summary of the Invention

[0004] The purpose of this application is to provide an AI-based display screen color calibration method, display device, and storage medium, aiming to solve the technical problem that existing technologies cannot adaptively calibrate according to users' differentiated preferences for different content categories, resulting in a lack of personalized adaptability in color calibration results.

[0005] The first aspect of this application provides an AI-based display screen color calibration method, the method comprising:

[0006] The content recognition model is invoked to identify the content to be displayed and obtain the content category; Construct a user preference vector based on the content categories; Obtaining the screen aging compensation factor for the display screen includes: obtaining the cumulative usage time of the display screen and determining the color output capability attenuation ratio based on the cumulative usage time; obtaining the cumulative luminous intensity of each color channel and calculating the channel usage imbalance correction coefficient for each channel based on the cumulative luminous intensity of each color channel; and obtaining the screen aging compensation factor based on the color output capability attenuation ratio and the channel usage imbalance correction coefficient. Obtain the input color value of each pixel in the content to be displayed, wherein the input color value is the color value of the content to be displayed in the standard color space; The input color value is calibrated based on the user preference vector and the screen aging compensation factor to obtain the target color value; The content to be displayed is displayed according to the target color value.

[0007] Optionally, constructing the user preference vector based on the content category includes: Based on the content category, N sets of comparison test images are displayed to the user, and each set of comparison test images includes two image pairs with different color parameters; For each set of comparison test images, the system receives the user's choice of the preferred image from the two image pairs and records the selection result; Construct a user preference vector based on the user's selection results.

[0008] Optionally, constructing a user preference vector based on the user's selection results includes: Based on the selection results, the number of times the user preferred cool colors and warm colors in the color temperature dimension, the number of times they preferred high saturation and low saturation in the saturation dimension, and the number of times they preferred high contrast and low contrast in the contrast dimension are calculated. The color temperature preference coefficient is obtained based on the number of times the user prefers cool colors and the number of times they prefer warm colors in the color temperature dimension, as well as the preset color temperature preference quantification value. Based on the number of times users prefer high saturation and low saturation in the saturation dimension, and the preset saturation preference quantification value, a saturation preference coefficient is obtained. The contrast preference coefficient is obtained based on the number of times the user prefers high contrast and the number of times they prefer low contrast, as well as the preset contrast preference quantification value. A user preference vector is constructed based on the color temperature preference coefficient, the saturation preference coefficient, and the contrast preference coefficient.

[0009] Optionally, the calculation of the channel imbalance correction coefficient based on the cumulative luminance of each color channel includes: The maximum cumulative luminous intensity is determined based on the cumulative luminous intensity of each color channel; Calculate the ratio of the cumulative luminance of each color channel to the maximum cumulative luminance; Based on the ratio and the preset correction intensity coefficient, the channel usage imbalance correction coefficient for each channel is obtained.

[0010] Optionally, the process of obtaining the channel usage imbalance correction coefficient for each channel based on the ratio and a preset correction intensity coefficient includes: Subtract the first product from the unit constant to obtain the imbalance correction coefficient used by the channel; Wherein, the first product is the product of the first difference and the preset correction intensity coefficient; the first difference is the difference between the unit constant and the ratio corresponding to the color channel.

[0011] Optionally, calibrating the input color value based on the user preference vector and the screen aging compensation factor to obtain the target color value includes: Obtain the pixel attributes of each pixel in the content to be displayed; Based on the pixel attributes, the pixels of the content to be displayed are divided into multiple pixel groups; For different pixel groups, determine the contribution ratio of the user preference vector and the screen aging compensation factor in color calibration; The input color value is calibrated according to the contribution ratio of each pixel group to obtain the target color value.

[0012] Optionally, calibrating the input color value according to the contribution ratio of each pixel group to obtain the target color value includes: Based on the user preference vector, the input color value is adjusted to obtain the preference-adjusted color value; The input color value is physically compensated based on the screen aging compensation factor to obtain a physically compensated color value. Based on the contribution ratio, the physical compensation color value and the preference adjustment color value are weighted and summed to obtain the target color value.

[0013] Optionally, the method further includes: The content categories and corresponding user preference vectors are uploaded to the cloud server and stored in association with user accounts. When it is detected that at least one other display device is logged in under the user account, the content category and corresponding user preference vector stored on the cloud server are sent to the other display device, so that the other display device performs color calibration and display according to the received content category and corresponding user preference vector. A second aspect of this application provides a display device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the AI-based display screen color calibration method described above.

[0014] A third aspect of this application provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the AI-based display screen color calibration method.

[0015] This application identifies the content to be displayed by calling a content recognition model to obtain the content category, and then constructs a user preference vector based on the content category. This changes the color calibration objective from simply minimizing physical color difference to optimizing user perception, achieving adaptation to different users' personalized visual preferences. By obtaining the screen aging compensation factor of the display screen, the problem of color shift caused by long-term screen use is solved, providing a physical benchmark for subsequent calibration with user preferences. In the standard color space, after obtaining the input color value of each pixel in the content to be displayed, the input color value is double-calibrated according to the user preference vector and the screen aging compensation factor to obtain the target color value. The content to be displayed is then displayed according to the target color value. Since the target color value integrates physical aging compensation and subjective preference adaptation, the displayed content seen by the user not only eliminates the color shift caused by screen aging, but also conforms to their personal subjective preferences under different content categories, resulting in good color calibration effect. Attached Figure Description

[0016] Figure 1 A flowchart illustrating the AI-based display screen color calibration method provided in this application embodiment; Figure 2 A schematic diagram of the display interface for the comparative test diagram provided in the embodiments of this application; Figure 3 A schematic diagram of the functional modules of the AI-based display screen color calibration device provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of a display device provided in an embodiment of this application.

[0017] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0019] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0020] It should be noted that the execution subject of this application embodiment can be a display device with data processing, network communication and program running functions, such as a tablet computer, personal computer, smartphone, etc., or an electronic device or industrial monitoring system that can realize the above functions.

[0021] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0022] Reference Figure 1 The image shows an AI-based display screen color calibration method provided in an embodiment of this application.

[0023] In this embodiment of the application, the AI-based display screen color calibration method includes the following steps.

[0024] S11, call the content recognition model to recognize the content to be displayed and obtain the content category.

[0025] The content to be displayed refers to the image or video frame data that will be presented on the display screen. Single frames of the content to be displayed are captured at preset time intervals from the operating system interface, application windows, video frames output by the video decoder, screens rendered by the game engine, and content displayed by the web browser. The preset time interval can be a fixed interval (e.g., capturing one frame every 30 frames) or an adaptive interval (e.g., triggering capture only when a significant change in the displayed content is detected). The captured images are preprocessed, for example, by size normalization (e.g., scaling to 224×224 or 227×227 pixels) and pixel value normalization (mapping pixel values ​​to the [0, 1] range).

[0026] Because different types of display content have fundamentally different requirements for color presentation, and traditional color calibration methods use uniform calibration parameters for all display content, they cannot adaptively adjust according to content category. This results in significant differences in performance of the same calibration parameters across different content categories. For example, calibration parameters suitable for watching movies may result in blurry text due to insufficient contrast when used for reading documents; calibration parameters suitable for reading documents may result in a flat and dull image when used for gaming due to insufficient saturation. Therefore, it is necessary to input the pre-processed image into a content recognition model for classification and recognition, enabling the system to possess content awareness capabilities and dynamically adjust the calibration strategy according to the category of content the user is currently viewing, thereby improving the scene adaptability of color calibration and the user experience.

[0027] The content recognition model is a machine learning model trained on a lightweight convolutional neural network, capable of mapping input images to preset content category labels. The lightweight convolutional neural network can include, but is not limited to, network architectures suitable for running on resource-constrained devices such as MobileNet, ShuffleNet, and SqueezeNet. The content category is a classification label indicating the type of content to be displayed. Content categories include, but are not limited to, document / office, movie / video, game, image editing, web browsing, and video call. Different types of content to be displayed have different color rendering requirements.

[0028] In an optional embodiment, to avoid changes in recognition results between adjacent frames, temporal smoothing processing can be performed on the recognition results of multiple consecutive frames. Specifically, a sliding window mechanism can be used to statistically analyze the recognition results of the current frame and the previous M frames (M is a preset value, such as M=5), and the content category with the highest frequency of occurrence can be taken as the final content category.

[0029] S12, Construct a user preference vector based on the content category.

[0030] Because different users have different visual perception physiology and aesthetic habits, their preferences for color parameters such as color temperature, saturation, and contrast of the same display image often differ. If all users' calibration targets are uniformly set to the same standard color, the calibrated display effect, while meeting objective color difference indicators, may not satisfy every user. Therefore, this application actively acquires users' visual preference data based on the content category and quantifies it into a user preference vector. This allows the subsequent color calibration process to be guided by the user's personalized preferences, dynamically generating target color values ​​that meet the user's subjective expectations, thereby improving user satisfaction and personalized adaptability in color calibration.

[0031] User visual preference data refers to a set of information reflecting the subjective preferences of different individual users regarding the way colors are presented on a display. The numerical vector formed by quantifying this user visual preference data is the user preference vector, which represents the user's personalized tendencies in the color dimension.

[0032] In an optional implementation, constructing the user preference vector based on the content category includes: Based on the content category, N sets of comparison test images are displayed to the user, and each set of comparison test images includes two image pairs with different color parameters; For each set of comparison test images, the system receives the user's choice of the preferred image from the two image pairs and records the selection result; Construct a user preference vector based on the user's selection results.

[0033] Based on the identified content category, N sets of corresponding comparison test images are selected from a pre-set test image library. The test image library pre-stores multiple sets of test images, and each set of test images has a mapping relationship with at least one content category.

[0034] like Figure 2 As shown, N sets of comparative test images are sequentially displayed to the user, where N is an integer greater than or equal to 1. Preferably, to ensure the reliability of the statistical results, N ≥ 10. Each set of comparative test images includes two image pairs, which have the same image content but different color parameter settings. The color parameters include, but are not limited to, color temperature, saturation, and contrast parameters.

[0035] For each set of comparison test images, the system receives the user's selection instruction for the preferred image between the two images. The user can make the selection via touchscreen, mouse click, or keyboard input. The system records the user's selection result for each set, including the image identifier selected by the user, the corresponding color parameter configuration of the image, and the selection timestamp.

[0036] After selecting N sets of comparative test images, the system performs statistical analysis on the selection results, calculates the user's preference quantification values ​​in each dimension, and combines the preference quantification values ​​in each dimension to form a user preference vector. This user preference vector can be stored in local storage or on a cloud server for later use when dynamically generating target color values.

[0037] In the above-described optional implementation methods, the selection result of a single test image may be subject to chance. By repeatedly testing with multiple sets of test images and recording the selection results, random errors can be effectively reduced, making the preference quantification results more stable and reliable. By allowing users to choose between two images with different color parameters, the user's subjective preference tendencies in dimensions such as color temperature, saturation, and contrast can be intuitively reflected, without the need for complex equipment or professional color knowledge.

[0038] In an optional implementation, constructing a user preference vector based on the user's selection results includes: Based on the selection results, the number of times the user preferred cool colors and warm colors in the color temperature dimension, the number of times they preferred high saturation and low saturation in the saturation dimension, and the number of times they preferred high contrast and low contrast in the contrast dimension are calculated. The color temperature preference coefficient is obtained based on the number of times the user prefers cool colors and the number of times they prefer warm colors in the color temperature dimension, as well as the preset color temperature preference quantification value. Based on the number of times users prefer high saturation and low saturation in the saturation dimension, and the preset saturation preference quantification value, a saturation preference coefficient is obtained. The contrast preference coefficient is obtained based on the number of times the user prefers high contrast and the number of times they prefer low contrast, as well as the preset contrast preference quantification value. A user preference vector is constructed based on the color temperature preference coefficient, the saturation preference coefficient, and the contrast preference coefficient.

[0039] Color temperature characterizes the warmth or coolness of colors in a displayed image. Cool tones correspond to higher color temperature values ​​(e.g., above 6400K), while warm tones correspond to lower color temperature values ​​(e.g., below 4000K). Saturation characterizes the vividness of colors in a displayed image. High saturation makes colors more intense and vibrant, while low saturation makes colors softer and more delicate. Contrast ratio characterizes the degree of difference between light and dark areas in a displayed image. High contrast ratio makes the difference between light and dark areas more pronounced, while low contrast ratio makes the transition between light and dark areas smoother.

[0040] For N sets of comparative test images, calculate the total number of times users selected cool-toned images in the color temperature dimension to obtain the number of times they preferred cool tones; calculate the total number of times users selected warm-toned images to obtain the number of times they preferred warm tones. Similarly, calculate the total number of times users selected corresponding types of images in the saturation dimension to obtain the number of times they preferred high saturation and low saturation; calculate the total number of times users selected corresponding types of images in the contrast dimension to obtain the number of times they preferred high contrast and low contrast.

[0041] The system acquires a preset color temperature preference quantification value and calculates a color temperature preference coefficient based on the frequency of the user's preference for cool and warm tones. The preset color temperature preference quantification value is a constant used to map the user's color temperature preference level to a standardized numerical range. The color temperature preference coefficient can be obtained by normalizing the calculation based on the frequency of cool and warm tone preferences, and then comparing the normalized result with the preset color temperature preference quantification value. In practice, the system first calculates the first difference between the frequency of cool and warm tone preferences, then calculates the first sum between the frequency of cool and warm tone preferences, and finally multiplies the ratio of the first difference to the first sum by the preset color temperature preference quantification value. The color temperature preference coefficient X1 is obtained by multiplying the positive or negative sign of the color temperature preference coefficient X1 by the frequency of the user's preference for cool or warm tones, and its absolute value indicates the intensity of the preference. When X1>0, it indicates that the user prefers cool colors, and the larger the value, the stronger the preference for cool colors. When X1<0, it indicates that the user prefers warm colors, and the larger the absolute value, the stronger the preference for warm colors. When X1=0, it indicates that the user has no significant preference in the color temperature dimension.

[0042] The preset saturation preference quantification value is a pre-defined constant used to map the user's saturation preference level to a standardized numerical range. In practice, the saturation preference coefficient X2 can be calculated following the same process as calculating the color temperature preference coefficient X1. The sign of the saturation preference coefficient X2 indicates the user's preference direction for high or low saturation, and its absolute value indicates the intensity of the preference. When X2 > 0, it indicates that the user prefers high saturation, and the larger the value, the stronger the preference for high saturation; when X2 < 0, it indicates that the user prefers low saturation, and the larger the absolute value, the stronger the preference for low saturation; when X2 = 0, it indicates that the user has no significant preference in the saturation dimension.

[0043] The preset contrast preference quantification value is a constant used to map the user's contrast preference level to a standardized numerical range. In practice, the contrast preference coefficient X3 can be calculated following the same process as calculating the color temperature preference coefficient X1. The sign of the contrast preference coefficient X3 indicates the user's preference direction for high or low contrast, and its absolute value indicates the intensity of the preference. When X3 > 0, it indicates that the user prefers high contrast, and the larger the value, the stronger the preference for high contrast; when X3 < 0, it indicates that the user prefers low contrast, and the larger the absolute value, the stronger the preference for low contrast; when X3 = 0, it indicates that the user has no significant preference in the contrast dimension.

[0044] The system combines the calculated color temperature preference coefficient X1, saturation preference coefficient X2, and contrast preference coefficient X3 in a preset order to form a user preference vector. That is, the user preference vector is a one-dimensional numerical vector composed of the color temperature preference coefficient X1, saturation preference coefficient X2, and contrast preference coefficient X3, used to represent the user's personalized preference characteristics in the color dimension.

[0045] The above-mentioned optional implementation methods, by statistically analyzing the number of times users express their preferences across various dimensions, transform users' subjective choices into quantifiable statistical data, avoiding the ambiguity and uncertainty of subjective descriptions; by forcing users to make preference choices through bipolar comparisons, eliminating the intermediate ambiguity, achieving clear differentiation of user preferences, and improving the discriminability of preferences; by setting preset preference quantification values, mapping the degree of preference to a standardized numerical range, eliminating the dimensional differences caused by different test group numbers, achieving a standardized representation of user preferences, and making the preference coefficients comparable across users and scenarios; by combining the three coefficients into a user preference vector, the scattered preference data is integrated into a unified vector structure.

[0046] S13, obtain the screen aging compensation factor of the display screen.

[0047] OLED (Organic Light Emitting Diode) and other self-emissive display technologies experience irreversible aging over long-term use. As usage time increases, the quantum efficiency of organic light-emitting materials gradually decreases, leading to a reduction in luminous brightness at the same driving current. Without aging compensation, screen colors will gradually deviate from their initial state, especially when different channels have inconsistent aging rates, resulting in white balance shifts (such as an overall bluish or reddish tint). Obtaining the screen aging compensation factor provides an accurate basis for dynamically generating target color values. The screen aging compensation factor refers to the set of information regarding the degree of display performance degradation that occurs with increasing usage time.

[0048] In an optional implementation, obtaining the screen aging compensation factor of the display screen includes: Obtain the cumulative usage time of the display screen, and determine the color output capability attenuation ratio based on the cumulative usage time; Obtain the cumulative luminance of each color channel, and calculate the channel imbalance correction coefficient for each channel based on the cumulative luminance of each color channel; Based on the color output capability attenuation ratio and the channel usage imbalance correction coefficient, the screen aging compensation factor is obtained.

[0049] An internal timer is set up to monitor the screen's operating status in real time since its initial use. The timer starts whenever the screen switches from off to on and stops when it switches back to off. The total accumulated time is recorded as the cumulative usage duration.

[0050] Based on the cumulative usage time, the color output capability attenuation ratio is calculated according to a preset physical attenuation model. The color output capability attenuation ratio is used to quantify the impact of aging on display performance.

[0051] In one alternative embodiment, the physical decay model can take the form of exponential decay. For example, Y ( t )= e λt ,in, t To accumulate usage time, λ This is the preset physical attenuation coefficient. Physical attenuation is determined by the inherent aging characteristics of the screen's light-emitting materials and can be measured experimentally. Y ( t This indicates the percentage decrease in color output capability due to usage time. Y ( t As usage time increases, the value gradually approaches 0 from 1.

[0052] Due to varying user habits, the luminous intensity of each color channel may differ significantly. For example, displaying a user interface with a blue background for an extended period can cause the blue channel to age faster than the red and green channels. This means that the actual aging level of each color channel is closely related to its luminous intensity, and simply compensating based on usage time alone will result in significant errors. Therefore, this application requires obtaining the cumulative luminous intensity of the red, green, and blue channels separately. Specifically, for each pixel, the contribution of that pixel to the luminous intensity of each channel is calculated based on its RGB pixel value and the current brightness setting. The total luminous intensity of each color channel is obtained by summing the contributions of all pixels.

[0053] Based on the cumulative luminance of each color channel, the channel imbalance correction factor for each color channel is calculated. The channel imbalance correction factor is used to differentiate the aging degree of different channels.

[0054] The screen aging compensation factor for each channel is obtained by multiplying the color output capability attenuation ratio by the channel usage imbalance correction factor. The screen aging compensation factor represents the proportion of color output capability retained by each channel at the current time relative to the initial time. The smaller the value, the more severe the aging attenuation, and the more compensation is required; the larger the value, the milder the aging attenuation, and the less compensation is required.

[0055] The calculated screen aging compensation factors for each channel are associated with the cumulative usage time and the cumulative luminous intensity of each color channel and stored in memory for later use when dynamically generating target color values.

[0056] In the optional embodiments described above, differentiated aging compensation for each color channel is achieved by obtaining the cumulative luminance of each color channel and calculating the channel usage imbalance correction coefficient based on the cumulative luminance of each color channel. Specifically, the higher the cumulative luminance of a certain color channel, the more severe its aging, and the smaller the compensation factor (i.e., more compensation is needed); conversely, the lower the cumulative luminance of a certain color channel, the lighter its aging, and the larger the compensation factor (i.e., less compensation is needed). This differentiated compensation strategy can more accurately restore the color balance of the screen. By simultaneously obtaining the cumulative usage time and the cumulative luminance of each color channel, a more accurate screen aging compensation factor is calculated, providing an accurate basis for aging compensation for the subsequent dynamic generation of target color values, effectively solving the problem of uneven aging caused by differences in usage habits.

[0057] In an optional implementation, the calculation of the channel imbalance correction factor based on the cumulative luminance of each color channel includes: The maximum cumulative luminous intensity is determined based on the cumulative luminous intensity of each color channel; Calculate the ratio of the cumulative luminance of each color channel to the maximum cumulative luminance; Based on the ratio and the preset correction intensity coefficient, the channel usage imbalance correction coefficient for each channel is obtained.

[0058] The preset correction intensity coefficient is an adjustable parameter used to control the degree of influence of color channel imbalance on aging compensation, with a value ranging from 0 to 1. The larger the coefficient, the stronger the correction.

[0059] Among all color channels (R, G, B), the color channel with the highest cumulative luminance is called the maximum cumulative luminance. The maximum cumulative luminance represents the most fatigued or severely degraded color channel. If other color channels are reduced based on the minimum cumulative luminance, the overall brightness will decrease. If the average cumulative luminance is used as the benchmark, some color channels need to be brightened (but aged color channels cannot actually improve brightness and will instead accelerate damage). If the maximum cumulative luminance is used as the benchmark, and relatively less fatigued color channels are reduced to match the most fatigued color channel, maximum and balanced brightness can be maintained.

[0060] The degree of imbalance is quantified by calculating the ratio of the cumulative luminance of each color channel to the maximum cumulative luminance. For a color channel with a ratio <1, it indicates that its cumulative luminance is lower than that of the most overworked color channel, and its aging is relatively mild. For a color channel with a ratio = 1, it indicates that its cumulative luminance is equal to the highest value among all color channels, and it is the color channel with the deepest aging and heaviest workload.

[0061] By actively reducing the output of channels that are used less frequently, the remaining lifespan of each channel is made to decay at a more uniform rate, thus extending the effective lifespan of the screen.

[0062] Based on the ratio and the preset correction intensity coefficient, the channel usage imbalance correction coefficient for each channel can be obtained in the following way: subtract the first product from the unit constant to obtain the channel usage imbalance correction coefficient; wherein, the first product is the product of the first difference and the preset correction intensity coefficient; the first difference is the difference between the unit constant and the ratio corresponding to the color channel.

[0063] The unit constant represents the baseline value in the uncorrected state, that is, the correction system when no adjustments are needed for this color channel. The correction factor can be set to 1 or any positive constant.

[0064] Taking a unit constant of 1 as an example, the difference between the unit constant and the ratio corresponding to a certain color channel is the first difference. The first difference reflects the aging gap between this color channel and the color channel with the largest cumulative luminance. When the cumulative luminance of this color channel is closer to the maximum cumulative luminance, the ratio is closer to 1, and the first difference is closer to 0, indicating that the aging degree of this color channel is closer to the most fatigued channel, and the degree of correction required is lower. Conversely, when the cumulative luminance of this color channel is significantly less than the maximum cumulative luminance, the further the ratio is from 1, the larger the first difference, indicating that this color channel is relatively new and needs to be appropriately reduced to match the most fatigued channel.

[0065] The first difference is multiplied by a preset correction intensity coefficient to obtain the first product. The preset correction intensity coefficient α is a preset adjustable parameter with a value range of [0, 1]. This coefficient controls the degree of correction. When α approaches 0, the correction effect is extremely weak, almost without any compensation, and the overall brightness is preserved to the maximum extent, but the color shift between channels is more noticeable. When α approaches 1, the correction effect is strongest, making the attenuation level of each channel as consistent as possible, improving color accuracy, but the overall brightness will decrease to some extent. When α takes an intermediate value, the system achieves a balance between color accuracy and brightness preservation.

[0066] Finally, subtracting the first product from the unit constant 1 yields the channel imbalance correction coefficient for that color channel. When the color channel has the largest cumulative luminance, the corresponding channel imbalance correction coefficient is 1, and its luminance remains unchanged, requiring no attenuation. When the cumulative luminance of the color channel is less than the maximum cumulative luminance, the channel imbalance correction coefficient is less than 1, indicating that the color channel needs to reduce its luminance according to the calculated ratio, and the degree of reduction is positively correlated with both the first difference and the correction intensity coefficient α.

[0067] The aforementioned optional real-time method, by setting a preset correction intensity coefficient α, can provide a continuously adjustable transition range between complete balancing and no balancing, making the parameter selection space more flexible; the calculation of the channel using the imbalance correction coefficient is small and can be completed in real time, which is highly efficient; it ensures that the channel using the imbalance correction coefficient always falls within a reasonable range, avoiding abnormal situations such as overdrive or negative values.

[0068] S14, obtain the input color value of each pixel in the content to be displayed, wherein the input color value is the color value of the content to be displayed in the standard color space.

[0069] Each pixel of the content to be displayed independently contains color information, and the set of color information from all pixels constitutes the complete content to be displayed. The input color value refers to the original color of the pixel before calibration.

[0070] Standard color spaces can be: sRGB (standard red, green, and blue color space), DCI-P3 (Digital Cinema Color Gamut Standard), or Rec.2020 (Ultra-High Definition Television Color Gamut Standard), etc. A color space defines the digital representation of the three elements of color (hue, saturation, and lightness), ensuring that the same set of color values ​​theoretically corresponds to the same visual perception on different devices. In a standard color space, color values ​​are typically represented by three independent components, such as RGB (red, green, and blue) three-channel values ​​or Lab (lightness, a-channel, and b-channel) color coordinate values.

[0071] The original image data of the content to be displayed is read from the display cache, graphics processor, or data stream to be displayed. Then, the original image data is traversed pixel by pixel. For each pixel traversed, its three channel values ​​in the standard color space are extracted. For example, for a scene using the RGB color space, the red channel value, green channel value, and blue channel value of the pixel are read separately.

[0072] S15, calibrate the input color value according to the user preference vector and the screen aging compensation factor to obtain the target color value.

[0073] After obtaining the user preference vector corresponding to the current semantic category and the screen aging compensation factor, the channel values ​​of the input color value are color-calibrated to obtain the target color value. The target color value is used to drive the display screen to emit light from the pixels.

[0074] The parameters of each dimension in the user preference vector can be multiplied and fused with the screen aging compensation factor, and then the fused result can be multiplied with the channel values ​​of the input color value; alternatively, the screen aging compensation factor can be used to physically attenuate the input color value first, and then the user preference vector can be used to subjectively shift the compensated result.

[0075] In an optional implementation, calibrating the input color value according to the user preference vector and the screen aging compensation factor to obtain the target color value includes: Obtain the pixel attributes of each pixel in the content to be displayed; Based on the pixel attributes, the pixels of the content to be displayed are divided into multiple pixel groups; For different pixel groups, determine the contribution ratio of the user preference vector and the screen aging compensation factor in color calibration; The input color value is calibrated according to the contribution ratio of each pixel group to obtain the target color value.

[0076] The content to be displayed is read pixel by pixel, and for each pixel, its pixel attributes are obtained. Pixel attributes include, but are not limited to: brightness level, saturation level, hue region (such as red region, skin region, green region), and spatial frequency characteristics (such as edge detail region, flat region). Pixel attributes are used to determine the sensitivity of the pixel to color distortion and the user's acceptance of color changes in that pixel.

[0077] According to preset grouping rules, each pixel is assigned to a different pixel group. Pixels within the same pixel group have similar or identical pixel attributes. For example, pixels with brightness values ​​higher than a preset high brightness threshold are assigned to the high brightness pixel group; pixels with saturation values ​​lower than a preset low saturation threshold are assigned to the low saturation pixel group; pixels with hue corners in the skin tone range (e.g., hue corners between 0° and 30°) are assigned to the skin tone pixel group; and ordinary pixels that do not belong to any of the above special categories are assigned to the regular pixel group.

[0078] For each pixel group, the contribution ratio of the user preference vector and the screen aging compensation factor is determined separately. The contribution ratio measures the weight or influence of the user preference vector and the screen aging compensation factor in the target color value. The contribution ratio can be a pre-set fixed value or a variable value dynamically calculated based on pixel attributes.

[0079] For example, for the bright pixel group, the aging compensation factor contribution ratio can be set to 0.8 and the user preference vector contribution ratio to 0.2 to prevent over-adjustment from causing loss of highlight details. For the skin tone pixel group, the aging compensation factor contribution ratio can be set to 0.9 and the user preference vector contribution ratio to 0.1 to maximize the accuracy of skin tone reproduction and avoid skin tone distortion caused by user preferences. For the low saturation pixel group, both the aging compensation factor contribution ratio and the user preference vector contribution ratio can be set to 0.5. For the regular pixel group, the aging compensation factor contribution ratio can be set to 0.3 and the user preference vector contribution ratio to 0.7 to enhance visual expressiveness.

[0080] According to the contribution ratio of each pixel group, the input color value of each pixel within the group is calibrated. Specifically, the input color value is adjusted based on the user preference vector to obtain a preference-adjusted color value; the input color value is physically compensated based on the screen aging compensation factor to obtain a physically compensated color value; and the physically compensated color value and the preference-adjusted color value are weighted and summed according to the contribution ratio to obtain the target color value. The target color value integrates physical-level screen aging compensation and subjective-level user preference adaptation.

[0081] S16, Display the content to be displayed according to the target color value.

[0082] After obtaining the target color value through dual calibration using user preference vectors and screen aging compensation factors, the target color value corresponding to each pixel is transmitted to the driving circuit of the display screen. The driving circuit generates corresponding analog voltage or current signals based on the color channel components (e.g., red, green, and blue channel values) in the target color value. These analog voltage or current signals are applied to the light-emitting elements of the corresponding sub-pixels via the pixel driving circuit, controlling the light intensity of each sub-pixel. All pixels emit light synchronously or sequentially row by row according to their target color values. These discrete light-emitting points converge to form a complete image, allowing the user to observe the color-calibrated content to be displayed.

[0083] In an optional implementation, the method further includes: The content categories and corresponding user preference vectors are uploaded to the cloud server and stored in association with user accounts. When it is detected that at least one other display device has been logged in under the user account, the content category and the corresponding user preference vector stored on the cloud server are sent to the other display device, so that the other display device performs color calibration and display according to the received content category and the corresponding user preference vector.

[0084] The display device sends the semantic category and its corresponding user preference vector to the cloud server via the network. Upon receiving the data, the cloud server identifies the currently logged-in user account and stores the semantic category and user preference vector in the storage space corresponding to that user account, establishing a relational index. If an older user preference vector already exists under the same semantic category, it is overwritten and updated. A mapping relationship is established between the uploaded content category and its corresponding user preference vector and the specified user account, ensuring that any device subsequently logged in with that account can access the semantic category and its corresponding user preference vector.

[0085] A user account is a unique identifier registered by a user on a display device or related application system. When it is detected that at least one other display device is logged in under the same user account, the cloud server will send the content categories and corresponding user preference vectors associated with that user account to the other display devices via the network. Other display devices are one or more other display devices that are associated with the same user account, distinct from the currently performing upload operation, such as the user's mobile phone, tablet, TV, monitor, or laptop. When a user logs in to another display device, that device actively requests the associated content categories and corresponding user preference vectors from the cloud server.

[0086] After receiving the above data, other display devices store the content category and the corresponding user preference vector locally. Subsequently, when other display devices perform color calibration operations, they can directly call the received user preference vector and perform semantic recognition, preference matching, aging compensation, and color calibration on the content to be displayed according to the aforementioned method, and finally output the display.

[0087] In the above optional implementations, users do not need to repeat preference tests or manually set them on each display device, which greatly simplifies the configuration process for multi-device users and improves the user experience. When different display devices display content of the same semantic category, they can present a unified subjective color style because they use the same user preference vector, avoiding the visual disjointedness caused by differences in default calibration between different devices.

[0088] This application identifies the content to be displayed by calling a content recognition model to obtain the content category, and then constructs a user preference vector based on the content category. This changes the color calibration objective from simply minimizing physical color difference to optimizing user perception, achieving adaptation to different users' personalized visual preferences. By obtaining the screen aging compensation factor of the display screen, the problem of color shift caused by long-term screen use is solved, providing a physical benchmark for subsequent calibration with user preferences. In the standard color space, after obtaining the input color value of each pixel in the content to be displayed, the input color value is double-calibrated according to the user preference vector and the screen aging compensation factor to obtain the target color value. The content to be displayed is then displayed according to the target color value. Since the target color value integrates physical aging compensation and subjective preference adaptation, the displayed content seen by the user not only eliminates the color shift caused by screen aging, but also conforms to their personal subjective preferences under different content categories, resulting in good color calibration effect.

[0089] Figure 3 This is a functional block diagram of the AI-based display screen color calibration device provided in the embodiments of this application.

[0090] In some embodiments, the AI-based display screen color calibration device 30 may include multiple functional modules composed of program code segments. The program code of each program segment in the AI-based display screen color calibration device 30 may be stored in the memory of a smart terminal and executed by at least one processor to perform (see details). Figure 1 (Description) AI-based display screen color calibration function.

[0091] In this embodiment, the AI-based display screen color calibration device 30 can be divided into multiple functional modules according to its functions. These functional modules may include: a category recognition module 301, a preference construction module 302, a factor acquisition module 303, a color acquisition module 304, a color calibration module 305, and a content display module 306. The term "module" in this application refers to a series of computer-readable instruction segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be detailed in subsequent embodiments.

[0092] The category recognition module 301 is used to call the content recognition model to recognize the content to be displayed and obtain the content category; The preference construction module 302 is used to construct a user preference vector based on the content category; The factor acquisition module 303 is used to acquire the screen aging compensation factor of the display screen. The color acquisition module 304 is used to acquire the input color value of each pixel in the content to be displayed, wherein the input color value is the color value of the content to be displayed in the standard color space. The color calibration module 305 is used to calibrate the input color value according to the user preference vector and the screen aging compensation factor to obtain the target color value; The content display module 306 is used to display the content to be displayed according to the target color value.

[0093] It should be understood that the various variations and specific embodiments of the AI-based display screen color calibration method provided in the above embodiments are also applicable to the AI-based display screen color calibration device in this embodiment. Through the detailed description of the AI-based display screen color calibration method described above, those skilled in the art can clearly understand the implementation process of the AI-based display screen color calibration device in this embodiment. For the sake of brevity, it will not be described in detail here.

[0094] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the embodiment of the AI-based display screen color calibration method.

[0095] See Figure 4 The diagram shown is a structural schematic of a smart terminal provided in an embodiment of this application. In a preferred embodiment of this application, the display device 4 includes: a memory 401, at least one processor 402, at least one communication bus 403, and a display screen 404.

[0096] Those skilled in the art should understand that Figure 4 The structure of the smart terminal shown does not constitute a limitation of the embodiments of this application. The display device 4 may also include more or fewer other hardware or software, or different component arrangements than shown.

[0097] In some embodiments, the memory 401 stores a computer program and an operating system. When executed by the at least one processor 402, the computer program implements all or part of the steps in the AI-based display screen color calibration method described above. The memory 401 includes read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data. Further, the computer-readable storage medium may primarily include a program storage area and a data storage area, wherein the program storage area may store the operating system, at least one application program required for a function, etc.

[0098] In some embodiments, the at least one processor 402 is the control unit of the display device 4, connecting various components of the display device 4 via various interfaces and lines. It executes programs or modules stored in the memory 401 and calls data stored in the memory 401 to perform various functions and process data of the display device 4. For example, when the at least one processor 402 executes a computer program stored in the memory, it implements all or part of the steps of the AI-based display screen color calibration method described in this application embodiment; or it implements all or part of the functions of the AI-based display screen color calibration device. The at least one processor 402 may be composed of integrated circuits, such as a single-packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips.

[0099] In some embodiments, the at least one communication bus 403 is configured to enable communication between the memory 401 and the at least one processor 402, etc. Although not shown, the display device 4 may also include a power supply (e.g., a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 402 via a power management device, thereby enabling functions such as charging, discharging, and power consumption management. The power supply may also include one or more DC or AC power supplies, a rechargeable power fault detection circuit, a power converter or inverter, a power status indicator, or any other components.

[0100] In some embodiments, the display screen 404 includes a plurality of pixel units arranged in an array, each pixel unit comprising at least one red sub-pixel, one green sub-pixel, and one blue sub-pixel; a pixel driving circuit layer for driving each sub-pixel to emit light; and a light-emitting functional layer covering the pixel driving circuit layer. The light-emitting functional layer emits light of a corresponding color under electrical signal excitation. The display screen 404 further includes a row scanning driving circuit and a column data driving circuit electrically connected to the pixel driving circuit layer. The row scanning driving circuit is used to select pixel units row by row, and the column data driving circuit is used to write a driving voltage or driving current corresponding to a target color value to the selected pixel units. Optionally, the display screen 404 may further include a touch sensing layer, which is integrated above the light-emitting functional layer or embedded in the pixel driving circuit layer, for detecting user touch operations on the display screen. Optionally, the display screen 404 is any one of an organic light-emitting diode display screen, a micro light-emitting diode display screen, or a quantum dot light-emitting diode display screen.

[0101] The display device 4 may also include a Bluetooth module, a Wi-Fi module, internal memory, a network interface, an input location, and a display screen, etc., which will not be described in detail here.

[0102] The integrated unit, implemented as a software functional module, can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a smart terminal to execute portions of the methods described in the various embodiments of this application.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0104] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

Claims

1. A display screen color calibration method based on AI, characterized in that, The method includes: The content recognition model is invoked to identify the content to be displayed and obtain the content category; Construct a user preference vector based on the content categories; Obtaining the screen aging compensation factor for the display screen includes: obtaining the cumulative usage time of the display screen and determining the color output capability attenuation ratio based on the cumulative usage time; obtaining the cumulative luminous intensity of each color channel and calculating the channel usage imbalance correction coefficient for each channel based on the cumulative luminous intensity of each color channel; and obtaining the screen aging compensation factor based on the color output capability attenuation ratio and the channel usage imbalance correction coefficient. Obtain the input color value of each pixel in the content to be displayed, wherein the input color value is the color value of the content to be displayed in the standard color space; The input color value is calibrated based on the user preference vector and the screen aging compensation factor to obtain the target color value; The content to be displayed is displayed according to the target color value.

2. The AI-based display screen color calibration method according to claim 1, characterized in that, The construction of the user preference vector based on the content category includes: Based on the content category, N sets of comparison test images are displayed to the user, and each set of comparison test images includes two image pairs with different color parameters; For each set of comparison test images, the system receives the user's choice of the preferred image from the two image pairs and records the selection result; Construct a user preference vector based on the user's selection results.

3. The AI-based display screen color calibration method according to claim 2, characterized in that, The process of constructing a user preference vector based on user selection results includes: Based on the selection results, the number of times the user preferred cool colors and warm colors in the color temperature dimension, the number of times they preferred high saturation and low saturation in the saturation dimension, and the number of times they preferred high contrast and low contrast in the contrast dimension are calculated. The color temperature preference coefficient is obtained based on the number of times the user prefers cool colors and the number of times they prefer warm colors in the color temperature dimension, as well as the preset color temperature preference quantification value. Based on the number of times users prefer high saturation and low saturation in the saturation dimension, and the preset saturation preference quantification value, a saturation preference coefficient is obtained. The contrast preference coefficient is obtained based on the number of times the user prefers high contrast and the number of times they prefer low contrast, as well as the preset contrast preference quantification value. A user preference vector is constructed based on the color temperature preference coefficient, the saturation preference coefficient, and the contrast preference coefficient.

4. The AI-based display screen color calibration method according to claim 1, characterized in that, The calculation of the channel imbalance correction coefficient based on the cumulative luminance of each color channel includes: The maximum cumulative luminous intensity is determined based on the cumulative luminous intensity of each color channel; Calculate the ratio of the cumulative luminance of each color channel to the maximum cumulative luminance; Based on the ratio and the preset correction intensity coefficient, the channel usage imbalance correction coefficient for each channel is obtained.

5. The AI-based display screen color calibration method according to claim 4, characterized in that, The method for obtaining the channel usage imbalance correction coefficient for each channel based on the ratio and a preset correction intensity coefficient includes: Subtract the first product from the unit constant to obtain the imbalance correction coefficient used by the channel; Wherein, the first product is the product of the first difference and the preset correction intensity coefficient; the first difference is the difference between the unit constant and the ratio corresponding to the color channel.

6. The AI-based display screen color calibration method according to claim 1, characterized in that, The step of calibrating the input color value according to the user preference vector and the screen aging compensation factor to obtain the target color value includes: Obtain the pixel attributes of each pixel in the content to be displayed; Based on the pixel attributes, the pixels of the content to be displayed are divided into multiple pixel groups; For different pixel groups, determine the contribution ratio of the user preference vector and the screen aging compensation factor in color calibration; The input color value is calibrated according to the contribution ratio of each pixel group to obtain the target color value.

7. The AI-based display screen color calibration method according to claim 6, characterized in that, The step of calibrating the input color value according to the contribution ratio of each pixel group to obtain the target color value includes: Based on the user preference vector, the input color value is adjusted to obtain the preference-adjusted color value; The input color value is physically compensated based on the screen aging compensation factor to obtain a physically compensated color value. Based on the contribution ratio, the physical compensation color value and the preference adjustment color value are weighted and summed to obtain the target color value.

8. The AI-based display screen color calibration method according to claim 1, characterized in that, The method further includes: The content categories and corresponding user preference vectors are uploaded to the cloud server and stored in association with user accounts. When it is detected that at least one other display device has been logged in under the user account, the content category and the corresponding user preference vector stored on the cloud server are sent to the other display device, so that the other display device performs color calibration and display according to the received content category and the corresponding user preference vector.

9. A display device, characterized in that, The display device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the AI-based display screen color calibration method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the AI-based display screen color calibration method as described in any one of claims 1 to 8.

Images