A method and system for correcting the color gamut of a mobile phone screen

Abstract

The present invention belongs to the field of screen color gamut correction, and discloses a mobile phone screen color gamut correction method and system. The method includes: S1, obtaining ambient light data and screen content data; S2, obtaining color correction parameters based on the ambient light data and the screen content data; S3, obtaining personalized correction parameters based on the color correction parameters; S4, obtaining partition correction parameters based on the personalized correction parameters; S5, obtaining adaptive feedback correction parameters based on the partition correction parameters; S6, obtaining the display color of the screen based on the adaptive feedback correction parameters. The present invention proposes a non-linear color gamut mapping model to ensure that the parameters after color correction can be accurately mapped to the color gamut range that the screen can display. During the mapping process, it is necessary to dynamically adjust the non-linear coefficient of the mapping function according to the actual display characteristics of the screen to ensure the consistency of color display under different brightness and contrast ratios.

Description

Technical Field

[0001] The present invention relates to the field of screen color gamut correction, and particularly to a method and system for correcting the color gamut of a mobile phone screen. Background Art

[0002] With the popularity of mobile devices, especially smart phones, the screen display effect has become one of the important factors affecting the user experience. Color gamut correction is a key link in screen display technology, which directly determines the color accuracy and realism presented by the screen. However, due to the fact that mobile phones need to face variable environmental light and display content in different usage scenarios, static color gamut correction methods often fail to adapt to these changes, resulting in color distortion or unnatural visual experiences. Typical color gamut correction systems usually optimize the display content through preset standard color spaces (such as sRGB, Adobe RGB, etc.), but these methods have several significant drawbacks.

[0003] First of all, most traditional color gamut correction methods are based on static algorithms, and they cannot be adjusted in real time according to the dynamic changes of the external environmental light, display content, and the user's personalized preferences. This means that when users use the mobile phone under different environmental lights (such as strong light, weak light), the color display effect of the screen is not consistent. Especially in high-dynamic content such as videos, games, or complex image scenes, the displayed colors will deviate from the original settings, resulting in inaccurate colors or visual fatigue. Secondly, traditional methods lack support for the personalized needs of users. Different users may have different color preferences in different scenarios. For example, some users prefer high-saturation colors, while some users need to reduce blue light to protect their eyes. However, the existing technologies usually cannot make adaptive adjustments according to these personalized needs and can only set fixed display modes manually, unable to respond dynamically to user needs.

[0004] In addition, existing color gamut correction systems usually ignore the energy-saving requirements. When using the mobile phone for a long time, the screen is the main source of energy consumption, especially in the case of high-brightness and high-contrast display content, the power consumption problem is particularly serious. Currently, in the process of color gamut correction, the technology often globally processes all screen pixels. Even for areas that are insensitive to vision, high-intensity color correction is still performed, which leads to a large amount of unnecessary energy consumption problems and further affects the battery life of the mobile phone.

[0005] Although some color gamut correction systems have started to introduce light sensors in recent years to capture ambient light changes and adjust the screen brightness or color temperature, these technologies are still limited to the adjustment of brightness or simple color temperature and do not perform in-depth optimization on the color itself, making it difficult to ensure accurate color reproduction in complex environments. At the same time, some existing adaptive systems also lack an efficient machine learning mechanism, resulting in color correction adjustments often lagging behind user needs or environmental changes. Due to these drawbacks, the existing technologies have not been able to achieve true dynamic and intelligent adjustment when dealing with color gamut correction and are difficult to meet users' requirements for high color reproduction, low power consumption, and personalization. Summary of the Invention

[0006] The purpose of the present invention is to disclose a method and system for color gamut correction of a mobile phone screen to solve the technical problems raised in the background art.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions:

[0008] On the one hand, the present invention provides a method for color gamut correction of a mobile phone screen, including:

[0009] S1, obtaining ambient light data and screen content data;

[0010] S2, obtaining color correction parameters based on the ambient light data and the screen content data;

[0011] S3, obtaining personalized correction parameters based on the color correction parameters;

[0012] S4, obtaining partition correction parameters based on the personalized correction parameters;

[0013] S5, obtaining adaptive feedback correction parameters based on the partition correction parameters;

[0014] S6, obtaining the display color of the screen based on the adaptive feedback correction parameters;

[0015] Among them, obtaining the ambient light data and the screen content data includes:

[0016] Obtaining the ambient light data through an ambient light sensor, using S env to represent the ambient light data, and the ambient light data includes the intensity I env of the ambient light, the color temperature T env of the ambient light, and the spectral distribution λ env of the ambient light. Then S env can be expressed as:

[0017] S env ={I env , T env , λ env} (1)

[0018] The unit of the intensity of the ambient light is Lux, and the unit of the color temperature of the ambient light is K;

[0019] The screen content data is defined as C disp , and its value is the color information represented by the CIE XYZ space of each pixel. C disp is expressed as:

[0020] C disp = {X, Y, Z} (2)

[0021] Among them, X, Y, and Z respectively represent three standard chromaticity coordinates in the CIE XYZ color space. The RGB values of each pixel are converted to the CIE XYZ space through a standard color conversion matrix:

[0022]

[0023] R, G, and B are respectively the red, green, and blue values of each pixel in the current display content. The value ranges of R, G, and B are from 0 to 255, and the obtained X, Y, and Z are processed in the standard color space.

[0024] On the other hand, the present invention provides a mobile phone screen color gamut correction system, including a data acquisition module, a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module, and a fifth acquisition module;

[0025] The data acquisition module is used to acquire ambient light data and screen content data;

[0026] The first acquisition module is used to acquire color correction parameters based on the ambient light data and the screen content data;

[0027] The second acquisition module is used to acquire personalized correction parameters based on the color correction parameters;

[0028] The third acquisition module is used to acquire partition correction parameters based on the personalized correction parameters;

[0029] The fourth acquisition module is used to acquire adaptive feedback correction parameters based on the partition correction parameters;

[0030] The fifth acquisition module is used to acquire the display color of the screen based on the adaptive feedback correction parameters;

[0031] Among them, acquiring the ambient light data and the screen content data includes:

[0032] Acquiring the ambient light data through an ambient light sensor, and using S env to represent the ambient light data. The ambient light data includes the intensity I env of the ambient light, the color temperature T env of the ambient light, and the spectral distribution λenv , then S env can be expressed as:

[0033] S env = {I env , T env , λ env} (1)

[0034] The unit of the intensity of the ambient light is Lux, and the unit of the color temperature of the ambient light is K;

[0035] The screen content data is defined as C disp , and its value is the color information represented by the CIE XYZ space of each pixel. C disp is expressed as:

[0036] C disp = {X, Y, Z} (2)

[0037] Among them, X, Y, and Z respectively represent the three standard chromaticity coordinates in the CIE XYZ color space. The RGB values of each pixel are converted to the CIE XYZ space through a standard color conversion matrix:

[0038]

[0039] R, G, and B are respectively the red, green, and blue values of each pixel in the current display content. The value ranges of R, G, and B are from 0 to 255. The obtained X, Y, and Z after conversion are processed in the standard color space.

[0040] Beneficial effects:

[0041] The spectral conversion neural network is used for high-precision color correction:

[0042] By introducing the spectral conversion neural network, the present invention can accurately map and correct colors based on the current display content of the mobile phone screen and the spectral information of the ambient light. Different from the traditional static correction method, the spectral conversion neural network can dynamically analyze the spectral differences between the screen content and the light conditions, ensuring that the screen colors can be highly consistent with the colors in the real world under different lighting conditions. This not only solves the problem that the traditional static correction cannot adapt to the changes in ambient light, but also greatly improves the color restoration degree and the natural sense of the display, especially in the complex and changeable lighting environment, the effect is particularly obvious.

[0043] User preference adaptive gamut optimization based on meta-learning:

[0044] To solve the problem that the existing technology cannot meet the personalized needs of users, the present invention designs an adaptive optimization mechanism based on meta-learning. The system collects long-term color preference data of users in different scenarios, quickly constructs and updates a meta-learning model, enabling the system to quickly adaptively adjust calibration parameters according to a small amount of new data. This means that no matter whether the user is using the mobile phone for video, gaming, reading or other scenarios, the system can intelligently adjust parameters such as color saturation, brightness and contrast to meet the personalized needs of users, thus overcoming the problems of traditional systems that require manual adjustment of display modes and lack of dynamic response.

[0045] Energy-saving color gamut calibration based on environmental perception:

[0046] To solve the problem of excessive energy consumption caused by global processing of traditional color gamut calibration, the present invention uses an environmental perception module to detect changes in ambient light in real time, and combines a spectral conversion neural network to perform hierarchical processing on different screen areas, dynamically adjusting the intensity and range of color gamut calibration. The system can reduce unnecessary color calibration in non-critical areas according to light changes, and focus on optimizing areas that have a greater impact on the user's visual experience, thus achieving an intelligent energy-saving effect. This not only improves the battery life of the mobile phone, but also ensures that while ensuring color accuracy, unnecessary energy waste is reduced. Brief Description of the Drawings

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0048] Figure 1 It is a schematic diagram of a method for calibrating the color gamut of a mobile phone screen according to the present invention.

[0049] Figure 2 It is a schematic diagram of a system for calibrating the color gamut of a mobile phone screen according to the present invention. Detailed Embodiments

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0051] The present invention proposes a dynamic color gamut correction system based on a spectrum conversion neural network and meta-learning, which can achieve precise correction and adaptive optimization of the color of the mobile phone screen, can not only cope with color changes under different ambient light conditions, but also dynamically adjust according to the user's visual preferences, thereby overcoming the limitations of the prior art.

[0052] As Figure 1 shown in an embodiment, the present invention provides a method for correcting the color gamut of a mobile phone screen, including:

[0053] S1, obtaining ambient light data and screen content data;

[0054] S2, obtaining color correction parameters based on the ambient light data and the screen content data;

[0055] S3, obtaining personalized correction parameters based on the color correction parameters;

[0056] S4, obtaining partition correction parameters based on the personalized correction parameters;

[0057] S5, obtaining adaptive feedback correction parameters based on the partition correction parameters;

[0058] S6, obtaining the display color of the screen based on the adaptive feedback correction parameters;

[0059] Among them, obtaining the ambient light data and the screen content data includes:

[0060] Obtaining the ambient light data through an ambient light sensor, using S env to represent the ambient light data, and the ambient light data includes the intensity I env of the ambient light, the color temperature T env of the ambient light, and the spectral distribution λ env of the ambient light. Then S env can be expressed as:

[0061] S env ={I env , T env , λenv} (1)

[0062] The unit of the intensity of the ambient light is Lux, and the unit of the color temperature of the ambient light is K;

[0063] The ambient light sensor needs to capture the intensity, color temperature, and spectral distribution of the ambient light in real time. The spectral information of the ambient light will be used in the subsequent color correction model to ensure that the content displayed on the screen is consistent with the true color under the ambient light conditions.

[0064] The spectral distribution is usually the distribution of the light intensities of different wavelengths in the visible light range (400nm to 700nm).

[0065] The screen content data is defined as C disp , and its value is the color information represented by the CIE XYZ space of each pixel. C disp is expressed as:

[0066] C disp ={X, Y, Z} (2)

[0067] Among them, X, Y, and Z respectively represent the three standard chromaticity coordinates in the CIE XYZ color space. The RGB values of each pixel are converted to the CIE XYZ space through a standard color conversion matrix:

[0068]

[0069] R, G, and B are respectively the red, green, and blue values of each pixel in the current displayed content. The value ranges of R, G, and B are from 0 to 255. The obtained X, Y, and Z are processed in the standard color space.

[0070] The color information of the screen display content is also the key to correction. It is necessary to extract the color information from the image currently displayed on the mobile phone screen for matching and correction with the ambient light. To achieve this, the system will extract the RGB value of each pixel from the image frame in real time and convert it to a standardized color space (such as the CIE XYZ space) for more accurate color correction in subsequent processing.

[0071] Through the above technical innovation, the present invention effectively solves the deficiencies of the traditional color gamut correction system in terms of environmental adaptability, response to user personalized needs, and energy efficiency optimization, and realizes the organic combination of high-precision color restoration, intelligent energy saving, and adaptive optimization.

[0072] Preferably, through the time synchronization controller, control S env and C disp to be collected at the same moment.

[0073] To ensure the real-time and accuracy of color correction, the ambient light data S env and the screen display data C disp need to be synchronously collected and processed by the system. In most smartphones, the screen refresh rate is usually 60Hz to 120Hz. Therefore, data collection and processing must be synchronized multiple times per second so that environmental changes can be promptly reflected in the adjustment of the screen color. This can ensure that subsequent correction is based on the ambient light conditions at real time for screen color processing, avoiding data lag.

[0074] During the collection process of ambient light data and screen content data, due to factors such as sensor accuracy and external noise, the collected data may be affected by noise. To improve the stability and accuracy of the data, the system needs to denoise and preprocess the collected data.

[0075] Preferably, for the spectral distribution λ, a denoising method using multi-scale wavelet transform is used to reduce the noise of λ

[0076] env env. λ is wavelet decomposed, divided into multiple frequency bands, and each frequency band is denoised separately.

[0077] env env

[0078] The denoised spectral distribution is represented as λ′:

[0079] env

[0080]

[0081] is the result after denoising processing for the i-th frequency band;

[0082] For C disp , a mean filtering algorithm is used to filter the pixels in RGB, and then the color space conversion is performed. n1 represents the total number of frequency bands.

[0083] Preferably, S2 includes:

[0084] S21, spectral feature extraction:

[0085] F λ = Conv1D(λ env ) (5)

[0086] where F λ is the spectral feature, λ env is the spectral distribution of the ambient light, and Conv1D represents a 1D convolutional neural network; the spectral distribution λ in the ambient light is the intensity change of each wavelength in the light. The present invention

[0087] env

[0088] The spectral features will be extracted by a 1D convolutional neural network. The size and number of convolutional kernels need to match the characteristics of the spectral distribution.

[0089] S22, Screen color feature extraction:

[0090] Use the fully connected layer of the neural network to extract the screen color features. The fully connected layer of this neural network converts the XYZ values of the screen into a high-dimensional feature vector:

[0091] F C = W C ·C disp + b C (6)

[0092] where F C is the screen color feature, and W C and b C are the weights and biases of the fully connected layer respectively;

[0093] To achieve more accurate color correction, the spectral feature F λ and the color feature F C need to be fused. This fusion can be performed by concatenation operation or element-wise product, and the specific choice depends on the experimental results. The fused features will be used to generate correction parameters.

[0094] S23, Feature fusion:

[0095] Feature fusion formula:

[0096] F fusion = F λ ⊕ F C (7)

[0097] where ⊕ represents the feature concatenation operation, and F fusion is the combined feature, and F fusion combines the spectral feature and the screen color feature;

[0098] S24, Generate the spectral non-linear adjustment term R1(λ intensity ):

[0099]

[0100] α is the adjustment coefficient, λ intensity (i) is the spectral intensity of the i-th sampling point measured actually, is the expected spectral intensity; n2 represents the number of sampling points of the spectral intensity data.

[0101] Traditional color correction models often overlook the non - linear impact of light intensity changes in the multi - spectral range on color. To better capture this impact, a spectral non - linear adjustment term is introduced during the correction process. This adjustment term is specifically used to compensate for abnormal bands in the spectrum (such as overly strong or weak specific wavelengths). The role of this term is to suppress overly strong spectral fluctuations and ensure smoother color correction.

[0102] In S25, calculate the color correction parameters using the following formula:

[0103] P corr = W corr ·F fusion + b corr + R1(λ intensity )

[0104] W corr represents the weight matrix of the color correction model, and b corr represents the bias of the color correction model; P corr represents the color correction parameters.

[0105] The fused feature F fusion generates the final color correction parameter P corr through the fully - connected layer. At the same time, a spectral non - linear adjustment term is introduced to adjust the output, making the correction parameters more accurate.

[0106] The core task of the spectral conversion neural network (SCNN) is to establish the mapping relationship between the ambient light S env and the screen content C disp and generate the corrected color parameter P corr .

[0107] Preferably, S3 includes:

[0108] S31, obtain the initial value P user of the personalized correction parameter:

[0109] P user = f θ (P corr , D user )(10)

[0110] f θ is a meta - learning model that learns the adjustment rules of user preferences through the parameter θ, and D user represents the user's historical preference data; D user includes the user's historical adjustment records for saturation, brightness, and contrast;

[0111] D userContains records of the personalized usage habits and visual preferences of the user, such as records of the user's adjustment behaviors of color saturation, brightness, and contrast in different scenarios.

[0112] To quickly adapt to the personalized needs of different users, a meta-learning model f is introduced θ . The core idea of meta-learning is to quickly adjust the parameters of the model through a small amount of new data to make it adapt to new users. The model P corr and D user are used as inputs to output personalized correction parameters P user .

[0113] Gradient optimization of meta-learning:

[0114] The goal of model training is to minimize the personalized correction error through the gradient descent method. Define a loss function whose value is the deviation between the correction parameter and the user's actual visual needs. Through the meta-learning model, θ is quickly adjusted, and the present invention can enable the model to obtain good optimization effects with a small amount of data:

[0115]

[0116] where η is the learning rate, is the gradient with respect to the model parameter θ, and H user represents the actual visual correction requirements feedback by the user, such as color temperature, brightness, etc. θ * is the result of adjusting θ.

[0117] S32. Obtain the personalized regularization term:

[0118]

[0119] γ is the regularization coefficient, H user (i) is the i-th historical adjustment record in the user's historical preference data, and is the expected value calculated by the system based on the standard setting; n3 represents the total number of historical adjustment records.

[0120] In traditional meta-learning models, the personalized data of users directly affects the optimization of correction parameters. However, in order to improve the stability of personalized correction and prevent overfitting, a personalized regularization term specific to user preferences is designed. This regularization term is used to limit the over-reliance of the model on personalized preferences, thereby making the correction process smoother.

[0121] S33. Obtain the personalized correction parameter:

[0122] P final = P user + R2(D user ) (13)

[0123] P final represents the personalized correction parameter of the final output.

[0124] It can ensure that the user's visual preferences are met, and at the same time, it guarantees the smoothness and stability of the correction.

[0125] Preferably, S4 includes:

[0126] S41, dividing the screen into n4 regions by using a preset partitioning strategy;

[0127] For example, the value of n4 is 2, and the 2 regions include a high-brightness and fast-changing region (such as a dynamic video or game scene) and a low-brightness or static region (such as a text or static picture scene).

[0128] The formula of the partitioning strategy is:

[0129] R zone (i) = Segment(M disp (i))(14)

[0130] where, R zone (i) represents the region number to which the i-th pixel belongs, M disp represents the pixel information matrix of the screen. By analyzing the brightness L disp and the dynamic characteristic D pix in M pix (i), the pixels are divided into corresponding regions; M disp (i) represents the information matrix of the i-th pixel;

[0131] The dynamic characteristic refers to the characteristic of the pixel changing with time, which is used to measure the rate of change of the pixel value (such as brightness, color) between consecutive frames and is used to determine whether the pixel is dynamic or static.

[0132] Segment is a region partitioning function based on brightness and dynamic characteristics, ensuring the separation of the high-brightness and fast-changing region from the low-brightness or static region;

[0133] If the brightness of the pixel exceeds the preset high-brightness threshold and the dynamic characteristic exceeds the dynamic change threshold, the pixel is divided into the high-brightness and fast-changing region; otherwise, the pixel is divided into the low-brightness or static region. Generally, the brightness threshold is set to 200 and the dynamic threshold is set to 50%.

[0134] Traditional full-screen color gamut correction consumes a large amount of computing resources and energy. Through the region characteristic recognition and partitioning correction strategy, the present invention can focus resources on the regions that require high-precision correction, avoid unnecessary computational overhead, and thus achieve a significant energy-saving effect.

[0135] Each region is comprehensively judged based on its brightness, dynamic characteristics, and user preferences to determine its correction intensity and energy efficiency level. All pixels within the region share the same correction strategy.

[0136] S42. For regions with high brightness and rapid changes, the partition correction coefficient is

[0137]

[0138] Where is the energy efficiency regularization term for regions with high brightness and rapid changes (such as regions in videos and game scenes), used to ensure that even in high-energy-consuming scenarios, the upper limit of calculation can be controlled to prevent overcalculation;

[0139] For regions with low brightness or static, the partition correction coefficient is

[0140]

[0141] α low is a regulation parameter, α low <1, is the energy efficiency regularization term for regions with low brightness or static (such as regions in reading mode or static pictures), used to ensure that the visual needs of users are still met in regions with low brightness or static;

[0142] The calculation formula of the energy efficiency regularization term is:

[0143]

[0144] λ1 is the regularization coefficient, D pix (j) represents the dynamic characteristics of the j-th pixel, is the expected dynamic change value of the system, and n5 represents the total number of pixel samples of dynamic characteristics during the calculation of the energy efficiency regularization term, that is, the number of pixels on the screen used to evaluate the dynamic characteristics;

[0145] The present invention uses different color correction strategies to balance the display effect and energy consumption requirements. To make the correction strategy more accurate, the present invention introduces an innovative energy efficiency regularization term, which dynamically adjusts the correction intensity according to the display characteristics of each region to optimize energy consumption.

[0146] Traditional color correction methods often ignore the dynamic characteristics of regions, resulting in energy waste in unnecessary regions. For this reason, the present invention introduces an energy efficiency regularization term based on dynamic changes, which can adaptively adjust the amount of calculation according to the degree of dynamic changes within the region.

[0147] The role of this regularization term is to constrain the regions where the computational load changes dynamically too much, ensuring that the energy consumption does not become too high when performing high-precision correction in these regions.

[0148] S43. Set up a dynamic adjustment mechanism:

[0149] The color correction of the screen not only needs to be optimized in terms of regions, but also needs to adjust the energy efficiency strategy of each partition in real time according to the dynamic changes in the actual usage scenario. For this reason, the present invention designs a dynamic adjustment function It will update the correction strategy of each partition in real time as time t goes by.

[0150] Dynamic adjustment function formula

[0151]

[0152] Among them, η and μ are weight parameters respectively, represents the change rate of the personalized correction parameter over time, represents the change rate of the pixel information matrix over time.

[0153] Through this function, the system can perceive the usage scenario of the screen in real time (such as switching from static text reading to fast-changing video), thereby dynamically adjusting the partition correction strategy to ensure optimal energy efficiency.

[0154] Through partition correction and the design of an innovative energy efficiency regularization term, the system can intelligently allocate computing resources, ensuring high-precision correction in dynamic and highlighted regions, while reducing the computing intensity in static and low-brightness regions to achieve energy-saving effects. At the same time, through the dynamic adjustment function, the system can adapt to different usage scenarios in real time to further improve energy efficiency.

[0155] Preferably, S5 includes:

[0156] The original calculation formula for the adaptive feedback correction parameter is:

[0157]

[0158] Among them, is the adjusted adaptive feedback correction parameter at time t obtained based on the original calculation formula; P zone represents the partition correction parameter;

[0159] is the correction increment based on the change of ambient light;

[0160] is the correction increment based on user behavior feedback;

[0161] By comparing the current ambient light data and the initial light data Dynamically adjust the correction parameters. For example, when the ambient light intensity increases, the system will automatically increase the screen brightness and adjust the color saturation to adapt to the new light conditions.

[0162] According to the user's operation behavior at time t (such as the user manually adjusting the brightness or turning on the eye protection mode) for adjustment to ensure that the user's operations can be immediately reflected in the display effect.

[0163] Ambient light affects the perceived effect of the screen color, so it is necessary to dynamically adjust the correction parameters according to the light change. The present invention defines an ambient light adjustment item which calculates the correction increment according to the rate of change of the ambient light intensity to calculate the correction increment.

[0164] The calculation formula of

[0165]

[0166] where α env is the adjustment coefficient of the ambient light, is the current ambient light data, is the initial ambient light data;

[0167] The user's operation behavior (such as brightness adjustment, color mode switching) affects their expectations for the screen display, so it is necessary to dynamically adjust the color correction parameters according to these operations. The present invention defines a user behavior feedback adjustment item to calculate the correction increment by parsing the user's operation behavior.

[0168] The calculation formula of

[0169]

[0170] where β user is the adjustment coefficient of the user behavior, is a function representing the result of parsing and quantifying the user's current behavior;

[0171] For example, if the user manually reduces the brightness, this behavior will be fed back and a negative increment will be generated to reduce the brightness part in the correction parameters, so that the screen immediately responds to the user's operation.

[0172] Due to the dynamics of the screen content (for example, the dynamic changes in video content), the present invention needs to design a screen content feedback adjustment item This item is used to cope with the frequent changes in the screen display content, ensuring that the screen can be adaptively adjusted in high-dynamic scenarios (such as rapidly changing video or game scenes).

[0173] Calculate the feedback correction parameters of the screen content

[0174]

[0175] Among them, γ disp is the adjustment coefficient of the screen content feedback, is the analytical function of the dynamic change of the screen content. According to the change situation of the current screen content, a feedback adjustment amount is generated. For example, if there are a large number of rapidly changing dynamic scenes (such as high-frame-rate videos) on the screen, the system will correspondingly increase the color correction frequency and intensity.

[0176] The final calculation formula for the adaptive feedback correction parameter is:

[0177]

[0178] represents the adjusted adaptive feedback correction parameter at time t obtained based on the final calculation formula.

[0179] This formula combines the ambient light, user behavior, and feedback of the screen content, ensuring that the change of each factor can be immediately reflected in the color correction parameter, and ensuring the adaptive adjustment of the display effect.

[0180] Dynamic feedback regularization term

[0181] To avoid waste of unnecessary computing resources caused by overly frequent adjustments, the present invention introduces an innovative dynamic feedback regularization term R5(t), which is used to limit the correction frequency and avoid unnecessary adjustments in a short period of time.

[0182] Regularization term formula:

[0183]

[0184] Among them, λ2 is the regularization strength coefficient, is the estimated value of the desired correction change of the system, is the time change rate of the correction parameter. The role of this regularization term is to suppress overly frequent corrections and ensure that the system maintains a stable adjustment frequency in the actual scenario.

[0185] Preferably, S6 includes:

[0186] S61, set the mapping formula:

[0187]

[0188] is the actual displayed color after mapping. φ is a non - linear gamut mapping function that maps the calibration parameters to the color range that can be actually displayed on the screen;

[0189] The mapping function φ will be dynamically adjusted according to the screen display characteristics and brightness range. Specifically, φ is a non - linear gamut mapping function that adjusts through Gamma correction to ensure that the color parameters are mapped to the gamut range that the screen can actually display. In addition, the φ function also adjusts the color saturation according to the ambient light to ensure enhanced color performance in strong light and avoid oversaturation in weak light.

[0190] The displayed color is the final color value displayed on the screen, which is the RGB value after gamut correction.

[0191]

[0192] P max represents the maximum value of the adaptive feedback correction parameter, which is used to normalize to the range [0, 1]. γ3 represents the Gamma correction coefficient, which is used to control the non - linear adjustment of brightness and contrast, and its value range is [2.2, 2.4]. C max represents the maximum brightness or color value of the screen;

[0193] The actual displayed color of the screen is jointly affected by the color value and brightness. Gamma correction adjusts the non - linear characteristics of brightness and contrast to match the display effect.

[0194] is a non - linear function based on the screen gamut characteristics. It ensures that there is no color distortion on the screen at high brightness, and at the same time, no color details are lost at low brightness. Through different non - linear coefficients, the system can adapt to the color performance under different brightness conditions.

[0195] S62, set the color consistency regularization term C(t):

[0196]

[0197] where γ4 is the weight coefficient for adjusting color consistency, is the actual displayed color value of the i - th region, is the expected displayed color value of this region.

[0198] During the actual display process, the screen may have problems such as uneven display or color calibration imbalance in some areas. To further improve color uniformity and realism, the present invention designs a color consistency regularization term C(t) to balance the color consistency of each display area and ensure that the display effect of the entire screen is uniform under different areas. This regularization term is used to detect and adjust the color differences in each area to ensure the color consistency of the entire screen display. Even under different brightness and ambient light conditions, the color balance can still be guaranteed. Through this regularization, the system can reduce color drift caused by factors such as brightness and contrast in some areas.

[0199] S63, calculate the final display color of the screen:

[0200]

[0201] Wherein, is the final display color after mapping and color consistency adjustment.

[0202] From S5, the present invention has obtained color calibration parameters dynamically adjusted through a real-time feedback mechanism These parameters have been adaptively adjusted according to changes in ambient light, user interaction behavior, and dynamic changes in the screen content. The final gamut calibration of the screen needs to be converted into actual display color values to meet the user's requirements for display quality in the current environment, while maximizing the true color restoration and energy-saving effects of the screen. The range of the screen gamut is usually wider than the output parameters of the color calibration model. To avoid over-saturation or color distortion, the present invention needs to map to the actual displayable color range through appropriate mapping while maintaining the accuracy of color restoration.

[0203] The main function of S6 is to convert the color parameters after a series of calibrations and optimizations into actual displayable color values of the screen. In this process, through a non-linear gamut mapping model and a color consistency adjustment mechanism, it is ensured that the screen display can maintain the accuracy and uniformity of color restoration under different environmental conditions and user operations.

[0204] The present invention proposes a non-linear gamut mapping model to ensure that the parameters after color calibration can be accurately mapped to the displayable gamut range of the screen. The mapping process needs to dynamically adjust the non-linear coefficient of the mapping function according to the actual display characteristics of the screen to ensure the consistency of color display under different brightness and contrast.

[0205] Such as Figure 2In an embodiment shown, the present invention provides a mobile phone screen color gamut correction system, including a data acquisition module, a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module, and a fifth acquisition module;

[0206] The data acquisition module is used to acquire ambient light data and screen content data;

[0207] The first acquisition module is used to acquire color correction parameters based on the ambient light data and the screen content data;

[0208] The second acquisition module is used to acquire personalized correction parameters based on the color correction parameters;

[0209] The third acquisition module is used to acquire partition correction parameters based on the personalized correction parameters;

[0210] The fourth acquisition module is used to acquire adaptive feedback correction parameters based on the partition correction parameters;

[0211] The fifth acquisition module is used to acquire the display color of the screen based on the adaptive feedback correction parameters;

[0212] Among them, acquiring the ambient light data and the screen content data includes:

[0213] Acquiring the ambient light data through an ambient light sensor, and using S env to represent the ambient light data. The ambient light data includes the intensity I env of the ambient light, the color temperature T env of the ambient light, and the spectral distribution λ env of the ambient light. Then S env can be expressed as:

[0214] S env ={I env , T env , λ env} (1)

[0215] The unit of the intensity of the ambient light is Lux, and the unit of the color temperature of the ambient light is K;

[0216] The screen content data is defined as C disp , and its value is the color information represented by the CIE XYZ space of each pixel. C disp is expressed as:

[0217] C disp ={X, Y, Z} (2)

[0218] Among them, X, Y, and Z respectively represent three standard chromaticity coordinates in the CIE XYZ color space. The RGB value of each pixel is converted to the CIE XYZ space through a standard color conversion matrix:

[0219]

[0220] R, G, and B are respectively the red, green, and blue values of each pixel in the current displayed content. The value ranges of R, G, and B are from 0 to 255, and the obtained X, Y, and Z after conversion are processed in the standard color space.

[0221] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for color gamut correction of a mobile phone screen, characterized in that: include: S1, obtaining ambient light data and screen content data; S2, obtaining color correction parameters based on ambient light data and screen content data; S3, obtaining personalized correction parameters based on the color correction parameters; S4, obtaining a partition correction parameter based on the personalized correction parameter; S5, obtaining an adaptive feedback correction parameter based on the partition correction parameter; S6, obtaining the display color of the screen based on the adaptive feedback correction parameter; Among them, obtaining ambient light data and screen content data includes: The ambient light sensor is used to obtain ambient light data. env Represents ambient light data, which includes the intensity of ambient light I env , color temperature of ambient light T env and the spectral distribution of ambient light λ env , then S env Can be expressed as: S env ={I env ,T env ,λ env } (1) The unit of ambient light intensity is Lux, and the unit of ambient light color temperature is K; The screen content data is defined as C disp , whose value is the color information of each pixel in CIE XYZ space, C disp It is expressed as: C disp ={X,Y,Z} (2) Among them, X, Y, and Z represent the three standard chromaticity coordinates in the CIE XYZ color space. The RGB value of each pixel is converted to the CIE XYZ space through the standard color conversion matrix: R, G, B are the red, green, and blue values ​​of each pixel in the current display content, respectively. The value range of R, G, and B is 0 to 255. The converted X, Y, and Z are processed in the standard color space; For the spectral distribution λ env , the denoising method of multi-scale wavelet transform is used to env To reduce noise, env Perform wavelet decomposition and transform λ env Divide into multiple frequency bands, and perform noise reduction on each frequency band. The spectral distribution after noise reduction is expressed as λ′ env : is the result of noise reduction processing on the i-th frequency band; For C disp , the mean filter algorithm is used to filter the pixels in RGB, and then the color space is converted after filtering. n1 represents the total number of frequency bands; S2 includes: S21, spectral feature extraction: F λ =Conv1D(λ env ) (5) Among them, F λ is the spectral characteristic, λ env is the spectral distribution of ambient light, Conv1D represents a 1D convolutional neural network; S22, screen color feature extraction: Use the fully connected layer of the neural network to extract the screen color features. The fully connected layer of the neural network converts the XYZ values ​​of the screen into a high-dimensional feature vector: F C =W C ·C disp +b C (6) Among them, F C is the screen color characteristic, W C and b C are the weights and biases of the fully connected layer respectively; S23, feature fusion: Feature fusion formula: in, represents the feature concatenation operation, F fusion is the joint feature, F fusion Combines spectral characteristics and screen color characteristics; S24, generate the spectral nonlinear adjustment term R1(λ intensity ): α is the adjustment coefficient, λ intensity (i) is the actual measured spectral intensity of the i-th sampling point, is the expected spectral intensity; n2 represents the number of sampling points of spectral intensity data; S25, calculate the color correction parameters using the following formula: P corr =W corr ·F fusion +b corr +R1(λ intensity ) W corr represents the weight matrix of the color correction model, b corr represents the bias of the color correction model; P corr Represents color correction parameters; S3 includes: S31, obtaining the initial value P of the personalized correction parameter user : P user =f θ (P corr ,D user ) (10) f θ It is a meta-learning model that learns the adjustment rules of user preferences through parameters θ. user Represents the user's historical preference data; D user Includes historical records of user adjustments to saturation, brightness, and contrast; S32, obtain personalized regularization items: γ is the regularization coefficient, H user (i) is the i-th historical adjustment record in the user's historical preference data, and It is the expected value calculated by the system based on the standard setting; n3 represents the total number of historical adjustment records; S33, obtaining personalized correction parameters: P final =P user +R2(D user ) (13) P final The personalized correction parameters representing the final output; S4 includes: S41, using a preset partitioning strategy to divide the screen into n4 areas; The formula for the partitioning strategy is: R zone (i)=Segment(M disp (i)) (14) Among them, R zone (i) indicates the region number to which the i-th pixel belongs, M disp Represents the pixel information matrix of the screen, through M disp The brightness L in (i) pix and dynamic characteristics D pix Analyze and divide the pixels into corresponding areas; M disp (i) represents the information matrix of the i-th pixel; Segment is a region division function based on brightness and dynamic characteristics, ensuring that high-brightness and rapidly changing areas are separated from low-brightness or static areas; S42, for areas with high brightness and rapid changes, the partition correction coefficient is : in, It is an energy efficiency regularization term for areas with high brightness and rapid changes. It is used to ensure that the upper limit of calculation can be controlled even in high energy consumption scenarios to prevent over-calculation. For low-light or static areas, the partition correction factor is : α low is the adjustment parameter, It is an energy efficiency regularization term for low-brightness or static areas, which is used to ensure that the user's visual needs are still met in low-brightness or static areas; The calculation formula of energy efficiency regularization term is: λ1 is the regularization coefficient, D pix (j) represents the dynamic characteristics of the jth pixel, is the dynamic change value expected by the system; n5 represents the number of pixels on the screen used to evaluate dynamic characteristics when calculating the energy efficiency regularization term; S43, set up dynamic adjustment mechanism: Dynamic adjustment function formula : Among them, η and μ are weight parameters, represents the rate of change of personalized correction parameters over time, Represents the rate of change of the pixel information matrix over time; S5 includes: The original calculation formula of the adaptive feedback correction parameter is: in, is the adaptive feedback correction parameter adjusted at time t based on the original calculation formula; P zone represents the partition correction parameter; is the correction increment based on ambient light changes; It is a correction increment based on user behavior feedback; The calculation formula is: Among them, α env is the adjustment factor of ambient light, is the current ambient light data, is the initial ambient light data; The calculation formula is: Among them, β user is the adjustment coefficient of user behavior, It is a function that represents the analysis and quantification results of the user's current behavior; Calculate screen content feedback correction parameters : Among them, γ disp is the adjustment coefficient of screen content feedback, It is a parsing function for the dynamic changes of screen content; The final calculation formula of the adaptive feedback correction parameter is: represents the adjusted adaptive feedback correction parameter at time t obtained based on the final calculation formula.

2. A method for color gamut correction of a mobile phone screen according to claim 1, characterized in that: Through the time synchronization controller, control S env and C disp The collection was performed at the same time.

3. A method for color gamut correction of a mobile phone screen according to claim 1, characterized in that S6 include: S61, set the mapping formula: is the actual displayed color after mapping, φ is the nonlinear color gamut mapping function, and the correction parameter Map to the color range that the screen can actually display; P max Represents the maximum value of the adaptive feedback correction parameter, which is used to Normalized to the range of [0,1], γ3 represents the Gamma correction coefficient, which is used to control the nonlinear adjustment of brightness and contrast, and its value range is [2.2,2.4]. max Indicates the maximum brightness or color value of the screen; S62, set the color consistency regularization term C(t): Among them, γ4 is the weight coefficient for adjusting color consistency, is the actual display color value of the i-th region, is the desired display color value of the area; S63, calculate the final display color of the screen: in, It is the final display color after mapping and color consistency adjustment.

4. A mobile phone screen color gamut correction system, characterized in that: It includes a data acquisition module, a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module and a fifth acquisition module; The data acquisition module is used to acquire ambient light data and screen content data; The first acquisition module is used to acquire color correction parameters based on ambient light data and screen content data; The second acquisition module is used to acquire personalized correction parameters based on the color correction parameters; The third acquisition module is used to acquire the partition correction parameter based on the personalized correction parameter; The fourth acquisition module is used to acquire the adaptive feedback correction parameter based on the partition correction parameter; The fifth acquisition module is used to acquire the display color of the screen based on the adaptive feedback correction parameter; Among them, obtaining ambient light data and screen content data includes: The ambient light sensor is used to obtain ambient light data. env Represents ambient light data, which includes the intensity of ambient light I env , color temperature of ambient light T env and the spectral distribution of ambient light λ env , then S env Can be expressed as: S env ={I env ,T env ,λ env } (1) The unit of ambient light intensity is Lux, and the unit of ambient light color temperature is K; The screen content data is defined as C disp , whose value is the color information of each pixel in CIE XYZ space, C disp It is expressed as: C disp ={X,Y,Z} (2) Among them, X, Y, and Z represent the three standard chromaticity coordinates in the CIE XYZ color space. The RGB value of each pixel is converted to the CIE XYZ space through the standard color conversion matrix: R, G, B are the red, green, and blue values ​​of each pixel in the current display content, respectively. The value range of R, G, and B is 0 to 255. The converted X, Y, and Z are processed in the standard color space; For the spectral distribution λ env , the denoising method of multi-scale wavelet transform is used to env To reduce noise, env Perform wavelet decomposition and transform λ env Divide into multiple frequency bands, and perform noise reduction on each frequency band. The spectral distribution after noise reduction is expressed as λ′ env : is the result after noise reduction processing of the i-th frequency band; For C disp , the mean filter algorithm is used to filter the pixels in RGB, and then the color space is converted after filtering. n1 represents the total number of frequency bands; Obtain color correction parameters based on ambient light data and screen content data, including: S21, spectral feature extraction: F λ =Conv1D(λ env ) (5) Among them, F λ is the spectral characteristic, λ env is the spectral distribution of ambient light, Conv1D represents a 1D convolutional neural network; S22, screen color feature extraction: Use the fully connected layer of the neural network to extract the screen color features. The fully connected layer of the neural network converts the XYZ values ​​of the screen into a high-dimensional feature vector: F C =W C ·C disp +b C (6) Among them, F C is the screen color characteristic, W C and b C are the weights and biases of the fully connected layer respectively; S23, feature fusion: Feature fusion formula: in, represents the feature concatenation operation, F fusion is the joint feature, F fusion Combines spectral characteristics and screen color characteristics; S24, generate the spectral nonlinear adjustment term R1(λ intensity ): α is the adjustment coefficient, λ intensity (i) is the actual measured spectral intensity of the i-th sampling point, is the expected spectral intensity; n2 represents the number of sampling points of spectral intensity data; S25, calculate the color correction parameters using the following formula: P corr =W corr ·F fusion +b corr +R1(λ intensity ) W corr represents the weight matrix of the color correction model, b corr represents the bias of the color correction model; P corr Represents color correction parameters; Obtain personalized correction parameters based on color correction parameters, including: S31, obtaining the initial value P of the personalized correction parameter user : P user =f θ (P corr ,D user ) (10) f θ It is a meta-learning model that learns the adjustment rules of user preferences through parameters θ. user Represents the user's historical preference data; D user Includes historical records of user adjustments to saturation, brightness, and contrast; S32, obtain personalized regularization items: γ is the regularization coefficient, H user (i) is the i-th historical adjustment record in the user's historical preference data, and It is the expected value calculated by the system based on the standard setting; n3 represents the total number of historical adjustment records; S33, obtaining personalized correction parameters: P final =P user +R2(D user ) (13) P final The personalized correction parameters representing the final output; Obtaining partition correction parameters based on personalized correction parameters, including: S41, using a preset partitioning strategy to divide the screen into n4 areas; The formula for the partitioning strategy is: R zone (i)=Segment(M disp (i)) (14) Among them, R zone (i) indicates the region number to which the i-th pixel belongs, M disp Represents the pixel information matrix of the screen, through M disp The brightness L in (i) pix and dynamic characteristics D pix Analyze and divide the pixels into corresponding areas; M disp (i) represents the information matrix of the i-th pixel; Segment is a region division function based on brightness and dynamic characteristics, ensuring that high-brightness and rapidly changing areas are separated from low-brightness or static areas; S42, for areas with high brightness and rapid changes, the partition correction coefficient is : in, It is an energy efficiency regularization term for areas with high brightness and rapid changes. It is used to ensure that the upper limit of calculation can be controlled even in high energy consumption scenarios to prevent over-calculation. For low-light or static areas, the partition correction factor is : α low is the adjustment parameter, It is an energy efficiency regularization term for low-brightness or static areas, which is used to ensure that the user's visual needs are still met in low-brightness or static areas; The calculation formula of energy efficiency regularization term is: λ1 is the regularization coefficient, D pix (j) represents the dynamic characteristics of the jth pixel, is the dynamic change value expected by the system; n5 represents the number of pixels on the screen used to evaluate dynamic characteristics when calculating the energy efficiency regularization term; S43, set up dynamic adjustment mechanism: Dynamic adjustment function formula : Among them, η and μ are weight parameters, represents the rate of change of personalized correction parameters over time, Represents the rate of change of the pixel information matrix over time; Acquiring adaptive feedback correction parameters based on the partition correction parameters includes: The original calculation formula of the adaptive feedback correction parameter is: in, is the adaptive feedback correction parameter adjusted at time t based on the original calculation formula; P zone represents the partition correction parameter; is the correction increment based on ambient light changes; It is a correction increment based on user behavior feedback; The calculation formula is: Among them, α env is the adjustment factor of ambient light, is the current ambient light data, is the initial ambient light data; The calculation formula is: Among them, β user is the adjustment coefficient of user behavior, It is a function that represents the analysis and quantification results of the user's current behavior; Calculate screen content feedback correction parameters : Among them, γ disp is the adjustment coefficient of screen content feedback, It is a parsing function for the dynamic changes of screen content; The final calculation formula of the adaptive feedback correction parameter is: P t adj represents the adjusted adaptive feedback correction parameter at time t obtained based on the final calculation formula.

Images