Gamma adjustment method, computer readable storage medium and computer program product
Adjusting the gamma register value through the target model and error probability density distribution function, the uneven display problem caused by the difference in OLED display model is solved, and the accuracy and debugging efficiency of the gamma register value are improved.
Patent Information
- Application Number
- CN202411295897.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-14
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-09-14
AI Technical Summary
The prior art cannot effectively solve the problem of the difference in gamma register values between different models of OLED display screens, resulting in uneven display effects and low debugging efficiency.
By obtaining the model of the module to be tested, determine its corresponding target model and error probability density distribution function, use these models and functions to adjust the gamma register value until the display screen meets the requirements, ensuring the accuracy and efficiency of the gamma register value.
The deviation of gamma register value between different display models is overcome, the accuracy and debugging efficiency of gamma register value are improved, and the consistency and efficiency of display effects are ensured.
Smart Images

Figure CN118942397A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of display technology, and more specifically, to a gamma adjustment method, a computer-readable storage medium, and a computer program product. Background Art
[0002] OLED (organic light emitting diode) display screens have the characteristics of bright colors, high contrast, faster response and more energy saving, and are widely used in more and more electronic products. After long-term use, OLED display screens may have local uneven brightness or inconsistent color. Gamma correction is an important solution to optimize the display effect of the display screen.
[0003] In the relevant gamma correction scheme, it is necessary to debug the sample display screen first to obtain the gamma data, and then write the gamma data into the gamma register of each display screen. However, due to process errors, different displays will inevitably have a certain degree of deviation. Therefore, the gamma data required by each display screen will also be different. The relevant gamma register value debugging scheme cannot meet the different requirements of different displays and has low efficiency. Summary of the invention
[0004] The purpose of the present disclosure is to provide a gamma adjustment method, a computer-readable storage medium and a computer program product to solve the problem that the gamma register values required by each display screen in the related art are different, improve the accuracy of the gamma register value and the efficiency of debugging the gamma register value.
[0005] In order to achieve the above objectives, the present invention adopts the following technical solutions:
[0006] The present disclosure provides a gamma adjustment method in a first aspect, comprising the following steps:
[0007] Obtain a model of the module to be tested, and determine a target model and an error probability density distribution function corresponding to the model according to the model, wherein the target model is used to represent a mapping relationship between a gamma register value and a target influencing factor, and the error probability density distribution function is used to represent an error distribution between a predicted gamma register value output by the target model and a true gamma register value;
[0008] Obtaining a target impact factor value when the module to be tested displays a target grayscale image, and inputting the target impact factor value into the target model to obtain a first gamma register value;
[0009] Detecting whether a display image of the module to be tested meets the requirement when the module to be tested uses the first gamma register value to perform gamma adjustment;
[0010] When the display screen does not meet the requirements, the first gamma register value is adjusted by the error probability density distribution function, and the adjusted first gamma register value is used to repeatedly perform the step of detecting whether the display screen of the module to be tested meets the requirements when the module to be tested uses the first gamma register value to perform gamma adjustment until the display screen meets the requirements.
[0011] Optionally, the step of adjusting the first gamma register value by using the error probability density distribution function includes:
[0012] Obtaining a first probability value, and obtaining a first error value interval corresponding to the first probability value from the error probability density distribution function;
[0013] A first error value is determined from the first error value interval according to a first method, and the first gamma register value is adjusted with the first error value as a step size.
[0014] Optionally, after the step of adjusting the first gamma register value with the first error value as a step length, the step further includes:
[0015] Detecting whether a display image of the module to be tested meets the requirements when the module to be tested performs gamma adjustment using the adjusted first gamma register value;
[0016] If it does not meet the requirements, repeat the step of adjusting the first gamma register value by using the error probability density distribution function until the display screen of the module to be tested meets the requirements.
[0017] Optionally, the gamma adjustment method further includes:
[0018] Get sample modules of different models;
[0019] For any model, the target model and error probability density distribution function corresponding to the model are determined based on multiple sample modules of the model.
[0020] Optionally, for any model, the step of determining a target model corresponding to the model according to multiple sample modules of the model includes:
[0021] Obtaining an influencing factor when the sample module displays a target grayscale image, wherein the influencing factor includes three spectral stimulation values;
[0022] Performing a correlation analysis on the three spectral stimulation values, and determining a target influencing factor according to the correlation analysis result, wherein the target influencing factor includes at least one of the three spectral stimulation values;
[0023] A target model for representing a mapping relationship between a gamma register value and a target impact factor is established according to the real gamma register value and the target impact factor value of the sample module.
[0024] Optionally, the step of performing a correlation analysis on the spectral stimulus values and determining the target influencing factor according to the correlation analysis result includes:
[0025] detecting whether the three spectral stimulus values are suitable for correlation analysis;
[0026] When the three spectral stimulation values are suitable for correlation analysis, a correlation matrix is calculated, wherein the correlation matrix includes correlation coefficients representing the correlation between any two spectral stimulation values;
[0027] Calculating the eigenvalues of the correlation matrix, and determining the number of spectral stimulation values in the target influencing factor according to the eigenvalues;
[0028] The target impact factor is determined according to the number of spectral stimulation values in the target impact factor and the correlation matrix.
[0029] Optionally, the step of determining the target impact factor according to the number of spectral stimulus values in the target impact factor and the correlation matrix includes:
[0030] When the number of spectral stimulation values in the target impact factor is 1, randomly selecting any spectral stimulation value from the three spectral stimulation values as the target impact factor;
[0031] When the number of spectral stimulation values in the target influencing factor is 2, two spectral stimulation values corresponding to the maximum correlation coefficient values are selected from the correlation matrix as the target influencing factor, and the maximum correlation coefficient value is less than 1.
[0032] Optionally, the step of determining the number of spectral stimulation values in the target influencing factor according to the characteristic value includes:
[0033] Visually displaying a first curve, wherein the first curve is used to represent a mapping relationship between the characteristic value and the number of spectral stimulation values;
[0034] The number of spectral stimulation values corresponding to the inflection points in the first curve is used as the number of spectral stimulation values in the target influencing factor.
[0035] Optionally, the step of establishing a target model for representing a mapping relationship between a gamma register value and a target impact factor according to the real gamma register value and the target impact factor value of the sample module includes:
[0036] Establishing at least two prediction models for representing the mapping relationship between the gamma register value and the target impact factor according to the real gamma register value and the target impact factor value of the sample module;
[0037] For any prediction model, obtaining a predicted gamma register value of the prediction model, and calculating an error between the predicted gamma register value and the true gamma register value;
[0038] The prediction model with the smallest error among the prediction models is used as the target model of the corresponding model of the sample module.
[0039] Optionally, for any prediction model, the step of obtaining a predicted gamma register value of the prediction model and calculating an error between the predicted gamma register value and the actual gamma register value includes:
[0040] For any prediction model, obtain the predicted gamma register value of the prediction model at different grayscales;
[0041] Calculate the error between the predicted gamma register value and the actual gamma register value at different grayscales;
[0042] The error values at different gray levels are summed to obtain the sum of the error values;
[0043] Taking the prediction model with the smallest error among the prediction models as the target model of the model corresponding to the sample module means: taking the prediction model with the smallest sum of error values among the prediction models as the target model of the model corresponding to the sample module.
[0044] Optionally, the color channels of the module to be tested include a first channel, a second channel and a third channel, and the error value includes a first channel error value, a second channel error value and a third channel error value; and the step of summing the error values at different grayscales to obtain the sum of the error values includes: summing the first channel error value, the second channel error value and the third channel error value at different grayscales to obtain the sum of the error values;
[0045] The gamma adjustment method further comprises:
[0046] Determine whether a minimum error value of the target model is greater than or equal to a first error threshold, the minimum error value being a minimum value among the first channel error value, the second channel error value, and the third channel error value;
[0047] If the judgment result is yes, the model parameters of the target model are adjusted until the minimum error value is smaller than the first error threshold.
[0048] Optionally, the number of prediction models is three, represented by a first prediction model, a second prediction model and a third prediction model, the first prediction model is a linear model, the second prediction model is a nonlinear model, the third prediction model is a piecewise nonlinear model, the third prediction model includes a first sub-model and a second sub-model, the grayscale of the first sub-model is less than a first grayscale threshold, and the grayscale of the second sub-model is greater than or equal to the first grayscale threshold.
[0049] Optionally, the first probability value is randomly generated, and a value of the first probability value is greater than 0 and less than or equal to 1;
[0050] The step of determining the first error value from the first error value interval according to the first method comprises:
[0051] The mean value, edge value or fixed position value of the first error value interval is determined as the first error value.
[0052] A second aspect of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the gamma adjustment method described above when the program is executed by a processor.
[0053] A third aspect of the present disclosure provides a computer program product, including a computer program, which implements the steps of the gamma adjustment method described above when executed by a processor.
[0054] The beneficial effects of the present disclosure are as follows:
[0055] According to the gamma adjustment method of the embodiment of the present disclosure, on the one hand, different target models correspond to display modules of different models. In this way, when predicting the gamma register value for modules to be tested of different models using their respective target models, the deviation between the gamma register values of display modules of different models can be overcome, thereby improving the accuracy of the predicted gamma register value. On the other hand, there are deviations between the gamma register values of display modules of the same model. By adjusting the predicted first gamma register value through the error probability density distribution function, the deviation between the gamma register values of display modules of the same model can be overcome, thereby further improving the accuracy of the predicted gamma register value. Moreover, since the error between the predicted value and the true value of display modules of the same model is basically similar under a fixed grayscale, by adjusting the first gamma register value through the error probability density distribution function, the accuracy of the adjusted first gamma register value can be guaranteed, and the debugging time of the first gamma register value can be shortened, thereby improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] The specific implementation methods of the present disclosure are further described in detail below with reference to the accompanying drawings.
[0057] Figure 1A flowchart of a gamma adjustment method provided in an embodiment of the present disclosure;
[0058] Figure 2 A flowchart for determining a target model corresponding to any model;
[0059] Figure 3 It is the heat map of the correlation coefficient corresponding to Table 1;
[0060] Figure 4 A schematic diagram of the mapping relationship between the characteristic value and the number of spectral stimulus values;
[0061] Figure 5 is a schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the first prediction model;
[0062] Figure 6 for Figure 5 Data of predicted gamma register values and actual gamma register values corresponding to each target grayscale;
[0063] Figure 7 is a schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the second prediction model;
[0064] Figure 8 for Figure 7 Data of predicted gamma register values and actual gamma register values corresponding to each target grayscale;
[0065] Fig. 9 is a schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the third prediction model;
[0066] Fig.10 for Fig. 9 Data of predicted gamma register values and actual gamma register values corresponding to each target grayscale;
[0067] Fig.11 A flow chart for determining, for any model, an error probability density distribution function corresponding to the model;
[0068] Fig.12 Schematic diagram of the error probability density distribution histogram of sample modules of different models. DETAILED DESCRIPTION
[0069] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.
[0070] Unless otherwise defined, the technical terms or scientific terms used in the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, similar words such as "one", "one" or "the" do not indicate quantity restrictions, but indicate that there is at least one. Similar words such as "include" or "comprise" mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Similar words such as "connect" or "connected" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0071] In related technologies, in order to optimize the display effect of the display module, it is necessary to perform gamma debugging on the display module to obtain the correspondence between each binding point grayscale and the gamma register value, and after the gamma debugging, the gamma register value corresponding to each binding point grayscale needs to be burned into the gamma register.
[0072] There are two main ways to determine the gamma register value of the display module:
[0073] (1) Selecting a number of sample modules from the batch display modules as sample modules, obtaining the gamma register values of the sample modules through debugging, and using the obtained gamma register values as the gamma register values of the batch display modules;
[0074] (2) Selecting several display modules from the batch as sample modules, obtaining the gamma register values of the sample modules through debugging, fitting the gamma register values of the sample modules to obtain a fitting curve, and subsequently determining the gamma register values of the batch display modules using the fitting curve.
[0075] However, due to process errors, different display modules will inevitably have a certain degree of deviation. Therefore, the gamma register values of different display modules may also have certain deviations. On the one hand, there will be certain deviations between display modules of different models. On the other hand, there will be certain deviations between display modules of the same model. Both of the above solutions cannot solve the problem of deviations in the gamma register values of different display modules. In addition, in order to improve the accuracy of the gamma register value, each display module can also be debugged to obtain the gamma register value, but the overall efficiency of the solution is low.
[0076] In order to solve the above technical problems, the embodiments of the present disclosure provide a gamma adjustment method, a computer-readable storage medium, and a computer program product, which are described below in conjunction with various specific embodiments.
[0077] Please refer to Figure 1 , Figure 1 A flowchart of a gamma adjustment method provided in an embodiment of the present disclosure is shown in FIG. Figure 1 As shown, it includes the following steps:
[0078] Step S101, obtain the model of the module to be tested, and determine the target model and error probability density distribution function corresponding to the model according to the model, wherein the target model is used to represent the mapping relationship between the gamma register value and the target influencing factor, and the error probability density distribution function is used to represent the error distribution between the predicted gamma register value output by the target model and the actual gamma register value.
[0079] In the disclosed embodiment, the module to be tested refers to the display module to be tested, that is, the display module for which the gamma register value needs to be determined. Since different types of display modules have certain deviations, it is necessary to predetermine the corresponding target model and error probability density distribution function for different types of display modules.
[0080] Among them, the target influencing factor refers to the main influencing factor among the influencing factors. Specifically, the influencing factor refers to the factor or parameter that will affect the gamma register value. The influencing factor includes three spectral stimulus values, that is, tristimulus values, and the target influencing factor includes at least one spectral stimulus value. The tristimulus value is a numerical value used to approximately describe the three stimulus intensities of the color. It is based on the "three primary colors" theory, that is, any color is the result of the three basic colors of red, green and blue. In the color matching experiment, the number of three primary colors required to achieve color matching with the color to be tested is called the tristimulus value. These values reflect the relative intensity of the color in the three color channels of red, green and blue. In color science, it is usually expressed as X, Y, Z or R, G, B, depending on the color space used. In the display module, the tristimulus value achieves precise control of color and brightness by controlling the intensity of the three color channels of red, green and blue, further affecting the accuracy and consistency of the color.
[0081] In a possible implementation, before step S101, the gamma adjustment method further includes the following steps: obtaining sample modules of different models; for any model, determining the target model and error probability density distribution corresponding to the model based on multiple sample modules of the model. Afterwards, when it is necessary to determine the gamma register value of the module to be tested, first determine the model of the module to be tested, and then obtain the target model and error probability density distribution function that match the model, thereby avoiding the deviation of the gamma register value caused by display modules of different models. In addition, by establishing a special prediction model, i.e., a target model, for each model, it can be ensured that the gamma debugging method can seamlessly support a variety of different models of display modules, thereby improving the compatibility of the gamma debugging method with display modules of different models.
[0082] Step S102, obtaining a target impact factor value when the module to be tested displays a target grayscale image, and inputting the target impact factor value into the target model to obtain a first gamma register value.
[0083] In the disclosed embodiment, the target grayscale picture refers to the picture of the target grayscale, and the target grayscale corresponds to the binding point grayscale that needs to be gamma adjusted. For example: the binding point grayscale of the display module includes grayscale 1, 2, 32, 64, 128, 164, 192, 254, 255, and the target grayscale also includes grayscale 1, 2, 32, 64, 128, 164, 192, 254, 255, etc. In the specific implementation, it is necessary to determine the first gamma register value corresponding to each target grayscale.
[0084] In a specific implementation, the module to be tested is controlled to display the target grayscale image, and the color analyzer is used to obtain the tristimulus values when the module to be tested displays the target grayscale image. Since the target influencing factor includes at least one of the tristimulus values, the target influencing factor value can be obtained according to the tristimulus values obtained by the color analyzer. Then, the target influencing factor value is input into the target model to obtain the gamma register value corresponding to the target influencing factor value. In order to distinguish the representation, the gamma register value is recorded as the first gamma register value. After the first gamma register value is obtained, the first gamma register value is written into the gamma register of the module to be tested, and step S103 is executed.
[0085] Step S103 , detecting whether the display image of the module to be tested meets the requirement when the module to be tested uses the first gamma register value to perform gamma adjustment.
[0086] Optionally, after the first gamma register value corresponding to each target grayscale of the module to be tested is determined through the target model and the first gamma register value is written into the gamma register of the module to be tested, it is necessary to further detect whether the display screen of the module to be tested meets the requirements when the module to be tested uses the first gamma register value for gamma adjustment, such as detecting whether the color brightness of the display screen meets the requirements through a color analyzer. If the color brightness of the display screen meets the requirements, it means that the predicted first gamma register value is relatively accurate and can be used directly. If the color brightness of the display screen does not meet the requirements, it means that there is a large deviation between the predicted first gamma register value and the actual gamma register value required by the module to be tested, and the first gamma register value needs to be adjusted to meet the requirements, that is, execute step S104.
[0087] Step S104, when the display screen does not meet the requirements, the first gamma register value is adjusted by the error probability density distribution function, and the adjusted first gamma register value is used to repeatedly perform the step of detecting whether the display screen of the module to be tested meets the requirements when the module to be tested uses the first gamma register value to perform gamma adjustment until the display screen meets the requirements. It is understandable that the adjustment process of the first gamma register value may need to be repeated multiple times until the adjusted first gamma register value makes the display screen of the module to be tested meet the requirements.
[0088] In the disclosed embodiment, since the error probability density distribution function corresponds one-to-one to the model of the display module, that is, the error probability density distribution function is for the same model, and is used to represent the difference distribution between the predicted gamma register value and the actual gamma register value of the display module under the same model, therefore, through the error probability density distribution function, it is possible to roughly predict the error between the predicted value and the actual value at any grayscale of the module to be tested, and use the predicted error value to adjust the first gamma register value, thereby ensuring the reliability of the adjusted first gamma register value so that it meets the requirements.
[0089] Compared with the related art, the gamma adjustment method of the embodiment of the present invention, on the one hand, corresponds to different target models for display modules of different models. In this way, when predicting the gamma register value for modules to be tested of different models using their respective target models, the deviation between the gamma register values of display modules of different models can be overcome, thereby improving the accuracy of the predicted gamma register value. On the other hand, there is a deviation between the gamma register values of display modules of the same model. By adjusting the predicted first gamma register value through the error probability density distribution function, the deviation between the gamma register values of display modules of the same model can be overcome, thereby further improving the accuracy of the predicted gamma register value. Moreover, since the error between the predicted value and the true value of display modules of the same model is basically similar under a fixed grayscale, the first gamma register value can be adjusted through the error probability density distribution function, thereby ensuring the accuracy of the adjusted first gamma register value, shortening the debugging time of the first gamma register value, and improving efficiency.
[0090] The gamma adjustment method of the embodiment of the present disclosure needs to determine the corresponding target model and error probability density distribution function for each model in advance before executing steps S101 to S104.
[0091] In one possible implementation, Figure 2 As shown, for any model, the step of determining a target model corresponding to the model according to multiple sample modules of the model includes steps S201 to S203. Specifically:
[0092] Step S201, obtaining the influencing factors when the sample module displays the target grayscale picture, wherein the influencing factors include three spectral stimulation values. In the specific implementation, the sample module is controlled to display the target grayscale picture, and then the three stimulation values at this time are measured by a color analyzer. Exemplarily, for each target grayscale, the three stimulation values when the sample module displays each target grayscale picture are measured.
[0093] Step S202: performing a correlation analysis on the three spectral stimulation values, and determining a target influencing factor according to the correlation analysis result, wherein the target influencing factor includes at least one of the three spectral stimulation values.
[0094] Optionally, the step of performing a correlation analysis on the three spectral stimulation values and determining the target influencing factor according to the correlation analysis result includes:
[0095] Step S2021, detecting whether the three spectral stimulus values are suitable for correlation analysis;
[0096] Step S2022, when the three spectral stimulation values are suitable for correlation analysis, calculating a correlation matrix, wherein the correlation matrix includes a correlation coefficient representing the correlation between any two spectral stimulation values;
[0097] Step S2023, calculating the eigenvalues of the correlation matrix, and determining the number of spectral stimulation values in the target influencing factor according to the eigenvalues;
[0098] Step S2024: determining the target influencing factor according to the number of spectral stimulation values in the target influencing factor and the correlation matrix.
[0099] In the disclosed embodiment, before performing correlation analysis on the three spectral stimulation values, they are first pre-analyzed to detect whether they are suitable for correlation analysis. When the three spectral stimulation values are suitable for correlation analysis, step S2022 is entered, otherwise step S2022 is not entered, thereby avoiding wasting time and resources on inappropriate data to perform invalid correlation analysis, which helps to ensure the accuracy and effectiveness of the correlation analysis results.
[0100] In a possible implementation manner, whether the tristimulus values are suitable for correlation analysis is determined by using a KMO (Kaiser-Meyer-Olkin) test method and / or a Bartlett sphericity test method.
[0101] Specifically, the KMO test is performed by comparing the relative size of the simple correlation coefficient and the partial correlation coefficient between the original variables. When the sum of the squares of the simple correlation coefficients between all variables is much larger than the sum of the squares of the partial correlation coefficients, the KMO value is close to 1. The closer the KMO value is to 1, the stronger the correlation between the variables and the more suitable the original variables are for factor analysis. When the sum of the squares of the simple correlation coefficients between all variables is close to 0, the KMO value is close to 0. The closer the KMO value is to 0, the weaker the correlation between the variables and the less suitable the original variables are for factor analysis. Usually, when the KMO value is greater than or equal to 0.9, it is very suitable for factor analysis. When the KMO value is greater than 0.6 and less than 0.9, it is very suitable for factor analysis. When the KMO value is less than 0.5, it is not suitable for factor analysis.
[0102] Bartlett's Test of Sphericity is mainly used to test whether the data conforms to the spherical hypothesis, that is, to test whether there is a high degree of correlation between the variables in the data set. The starting point of Bartlett's sphericity test is the correlation coefficient matrix of the variables. Its null hypothesis is that the correlation coefficient matrix is a unit matrix, that is, all elements on the diagonal of the correlation coefficient matrix are 1 (indicating that the correlation coefficient between the variable and itself is 1), and all elements on the non-diagonal are zero (indicating that the variables are uncorrelated). The test statistic is obtained based on the determinant of the correlation coefficient matrix. If the statistic is large and its corresponding accompanying probability value (p value) is less than the significance level set by the user (usually 0.05), the null hypothesis should be rejected, and it is believed that the correlation coefficient matrix cannot be a unit matrix, that is, there is correlation between the original variables, which is suitable for correlation analysis.
[0103] Optionally, the Bartlett sphericity test and the KMO test can be used together to more comprehensively evaluate whether the three stimulus values are suitable for factor analysis. Only when both the KMO test and the Bartlett sphericity test are passed, the three stimulus values are considered suitable for factor analysis.
[0104] In specific implementation, the KMO test method includes the following steps:
[0105] (1) Calculate the simple correlation coefficient matrix between the three spectral stimulus values. The simple correlation coefficient is an indicator to measure the linear relationship between two variables. For example, the simple correlation coefficient is the Pearson correlation coefficient.
[0106] (2) Calculate the partial correlation coefficient, which is the correlation between two variables after considering the influence of other variables. For three spectral stimulus values, it is necessary to calculate the partial correlation coefficient of each pair of spectral stimulus values under the control of the influence of the third spectral stimulus value.
[0107] (3) Calculate the KMO value. The calculation formula of the KMO value is as follows:
[0108]
[0109] Among them, r i,j represents the simple correlation coefficient between variable i and variable j, a i,j represents the partial correlation coefficient between variable i and variable j, where i and j represent any two of the three stimulus values.
[0110] (4) According to the calculated KMO value, it is judged whether the three spectral stimulation values are suitable for correlation analysis. In specific implementation, the KMO value between any two spectral stimulation values can be calculated, and according to each KMO value, it is judged whether the three spectral stimulation values are suitable for factor analysis.
[0111] For example, it is assumed that for any two spectral stimulus values among the three stimulus values, such as X and Y, the calculated KMO value is 0.7429, indicating that the correlation between the two spectral stimulus values X and Y is strong and suitable for correlation analysis. Similarly, the KMO values between X and Z, Y and Z can be obtained. When the KMO values between X and Z, Y and Z are also greater than a certain value, such as greater than 0.6, it is determined that the KMO test passes, that is, the three spectral stimulus values are suitable for factor analysis. The next step can directly enter step S2022 or continue to perform the Bartlett sphericity test, and enter step S2022 when the Bartlett sphericity test passes.
[0112] In step S2022, a correlation matrix is calculated, where the correlation matrix includes correlation coefficients representing the correlation between any two spectral stimulus values.
[0113] Optionally, the correlation coefficient may be a Pearson correlation coefficient, a Kendall rank correlation coefficient, a Spearman rank correlation coefficient, etc. In the disclosed embodiment, the Pearson correlation coefficient is used as an example for description. Specifically, the calculation formula of the Pearson correlation coefficient is expressed as:
[0114]
[0115]
[0116] Where X and Y are any two spectral stimulus values, cov represents covariance, σ represents standard deviation, X i Indicates the i-th X, Y i represents the i-th Y, represents the average value of X, represents the average value of Y, and n represents the number of samples. For example, for the sample module of model A, the correlation matrix of the three stimulus values is calculated as shown in Table 1, and the heat map of the correlation coefficient corresponding to Table 1 is shown in Figure 3 shown.
[0117] Table 1 Correlation matrix of model A sample module
[0118]
[0119]
[0120] In Table 1, X, Y, and Z represent three spectral stimulus values. Figure 3 The depth of the medium gray represents the size of the correlation coefficient. The darker the gray, the closer the correlation coefficient is to 1. Figure 3The diagonal part of the coordinate axis represents the correlation between any spectral stimulus value and itself, and its value is 1. For example, the data in the first row and the first column represent the correlation coefficient between the spectral stimulus value X and itself (i.e., the spectral stimulus value X). Figure 3 It can be seen that the correlation coefficients of the three spectral stimulus values are close to 1, and the three spectral stimulus values are almost completely correlated.
[0121] After obtaining the correlation matrix, the eigenvalue of the correlation matrix is further calculated, and the number of spectral stimulus values in the target influencing factor is determined according to the eigenvalue. The eigenvalue of the correlation matrix is a key parameter for measuring the strength of the linear relationship between multiple variables, and its calculation method mainly relies on the general solution method of the matrix eigenvalue. There are many methods for calculating the eigenvalue, such as: definition method, characteristic polynomial method, numerical method, etc. The embodiment of the present disclosure can use the mature eigenvalue calculation method in the relevant technology to calculate the eigenvalue of the correlation matrix.
[0122] Exemplarily, the formula of the matrix eigenvalue is expressed as: λE-A=0, where A represents the z-order matrix, λ represents the eigenvalue of the matrix A, and E represents the unit matrix. For the correlation matrix shown in Table 1, z=3, substituting it into the above formula, three eigenvalues can be obtained after calculation. In order to facilitate the determination of the relationship between the eigenvalue and the number of spectral stimulus values, the eigenvalue and the number of spectral stimulus values are visualized.
[0123] Optionally, the step of determining the number of spectral stimulation values in the target influencing factor according to the characteristic value includes: visually displaying a first curve, wherein the first curve is used to represent the mapping relationship between the characteristic value and the number of spectral stimulation values; and taking the number of spectral stimulation values corresponding to the inflection point in the first curve as the number of spectral stimulation values in the target influencing factor.
[0124] Please refer to Figure 4 , Figure 4 It is a schematic diagram of the mapping relationship between the eigenvalue and the number of spectral stimulation values, wherein the horizontal axis represents the number of spectral stimulation values, and the vertical axis represents the numerical value of the eigenvalue, and then the number of spectral stimulation values corresponding to the inflection point in the first curve is used as the number of spectral stimulation values in the target influencing factor, wherein the inflection point refers to the point where the first curve appears a broken line and the numerical value of the eigenvalue drops sharply, and before the inflection point, increasing the number of spectral stimulation values will cause the numerical value of the eigenvalue to decrease significantly, and after the inflection point, when the number of spectral stimulation values is increased, the numerical change of the eigenvalue is small, and the eigenvalue corresponding to the inflection point is greater than 0. Exemplary, Figure 4In the illustrated embodiment, the horizontal coordinate of the inflection point is 1, that is, the number of spectral stimulation values is 1, that is, when the three stimulation values are almost completely correlated, a spectral stimulation value can be randomly selected from the three stimulation values as the target influencing factor. It can be understood that for different modules, the number of spectral stimulation values corresponding to the inflection point of the first curve calculated may be different, which may be 1, 2 or 3.
[0125] (2) Determine the target impact factor according to the number of spectral stimulus values in the target impact factor and the correlation matrix.
[0126] When the number of spectral stimulation values in the target impact factor is different, the process of determining the target impact factor is also different. Optionally, the step of determining the target impact factor according to the number of spectral stimulation values in the target impact factor and the correlation matrix includes:
[0127] (21) When the number of spectral stimulation values in the target influencing factor is 1, any spectral stimulation value is randomly selected from the three spectral stimulation values as the target influencing factor, such as selecting any one from the spectral stimulation values X, Y, and Z as the target influencing factor.
[0128] (22) When the number of spectral stimulation values in the target influencing factor is 2, two spectral stimulation values corresponding to the maximum correlation coefficient in the correlation matrix are selected as the target influencing factor, and the maximum correlation coefficient is less than 1. Exemplarily, when the number of spectral stimulation values is 2, assuming that the maximum correlation coefficient less than 1 in the correlation matrix is 0.997, and the spectral stimulation values corresponding to the maximum correlation coefficient are X and Y, then X and Y are selected as the target influencing factors.
[0129] (23) When the number of spectral stimulation values in the target influencing factor is 3, the three spectral stimulation values are directly used as the target influencing factor.
[0130] In the disclosed embodiment, since the correlation between the three stimulus values is strong, in order to reduce the number of influencing factors when subsequently establishing the target model and make the current model more accurate and simple, the three stimulus values are firstly subjected to correlation analysis, and then at least one spectral stimulus value is selected from the three stimulus values as the target influencing factor based on the correlation analysis. When the number of determined spectral stimulus values is less than 3, the effect of representing all influencing factors by the main influencing factors is achieved, and data dimensionality reduction can be achieved, that is, the number of variables in the target model is reduced, while retaining the information of the original data as much as possible.
[0131] Step S203 : establishing a target model for representing a mapping relationship between a gamma register value and a target impact factor according to the real gamma register value of the sample module and the target impact factor value.
[0132] In the disclosed embodiment, the gamma register of the sample module stores the gamma register value obtained by debugging, that is, the real gamma register value. After the target impact factor is determined, it is necessary to establish a mapping relationship between the gamma register value and the target impact factor, that is, the target model.
[0133] The number of spectral stimulation values in the target influencing factor will affect the amount of sample data required to establish the target model. The fewer the number of spectral stimulation values in the target influencing factor, the smaller the amount of sample data required to establish the target model. For any target grayscale, the sample data is the target influencing factor value measured by the color analyzer when the sample module displays the target grayscale picture, that is, the spectral stimulation value. Exemplarily, when the target influencing factor includes 1 spectral stimulation value, for any target grayscale, each sample module can be controlled to display the target grayscale picture 3-4 times and measure the corresponding spectral stimulation value; when there are 2 target influencing factors, for any target grayscale, it may be necessary to control each sample module to display the target grayscale picture 10 times or more and measure the corresponding spectral stimulation value.
[0134] In a possible implementation, the step of establishing a target model for representing a mapping relationship between a gamma register value and a target impact factor according to a real gamma register value of a sample module and a target impact factor value includes:
[0135] Step S2031: establishing at least two prediction models for representing a mapping relationship between a gamma register value and a target impact factor according to the real gamma register value and the target impact factor value of the sample module.
[0136] Step S2032: for any prediction model, obtain the predicted gamma register value of the prediction model, and calculate the error between the predicted gamma register value and the true gamma register value.
[0137] Step S2033: The prediction model with the smallest error among the prediction models is used as the target model of the model corresponding to the sample module.
[0138] In the embodiment of the present disclosure, in order to improve the prediction accuracy of the target model, at least two prediction models are established, and then the prediction model with the smallest error is selected from the at least two prediction models as the target model. Optionally, the number of the prediction models is three, which are respectively represented as the first prediction model, the second prediction model and the third prediction model.
[0139] In a possible implementation, the first prediction model, the second prediction model, and the third prediction model are of different types. Optionally, the first prediction model is a linear model, the second prediction model is a nonlinear model, the third prediction model is a piecewise nonlinear model, the third prediction model includes a first sub-model and a second sub-model, the grayscale of the first sub-model is less than a first grayscale threshold, and the grayscale of the second sub-model is greater than or equal to the first grayscale threshold.
[0140] With such a setting, on the one hand, different types of prediction models may learn different feature representations and rules from the data. By establishing multiple different types of prediction models and selecting the prediction model with the best performance, a higher prediction accuracy than that of a single prediction model can be obtained. On the other hand, different types of prediction models have different assumptions and preferences for data. Therefore, they may have different adaptability to different types of data. By selecting the prediction model that best suits the sample module, the generalization ability of the prediction model can be enhanced. In addition, establishing multiple prediction models and selecting the one with the best performance as the final target model can reduce the risk of failure of a single type of prediction model. In some cases, the performance of a single type of prediction model may fluctuate greatly due to changes in data or the influence of noise. By combining multiple different types of prediction models, such fluctuations can be reduced and stability can be improved.
[0141] Optionally, the first prediction model is a linear model, which can be expressed as:
[0142] When S=1, R=a1*X1+b1, G=a2*X1+b2, B=a3*X1+b3;
[0143] When S=2, R=a1*X1+b1*X2+c1, G=a2*X1+b2*X2+c2, B=a3*X1+b3*X2+c3;
[0144] When S=3, R=a1*X1+b1*X2+c1*X3+d1, G=a2*X1+b2*X2+c2*X3+d2, B=a3*X1+b3*X2+c3*X3+d3;
[0145] Among them, R, G, B represent the predicted gamma register values corresponding to the red, green, and blue color channels respectively, X1, X2, and X3 represent the target influencing factors, S represents the number of spectral stimulus values in the target influencing factors, a1, b1, c1, d1, a2, b2, c2, d2, a3, b3, c3, and d3 represent the weight coefficients obtained by fitting, and when S=1, X1 is one of the spectral stimulus values X, Y, and Z; when S=2, X1 and X2 are any one of X, Y, and Z respectively, and X1 and X2 are different; when S=3, X1, X2, and X3 are one of X, Y, and Z respectively, and X1, X2, and X3 are different.
[0146] In a specific implementation, the first prediction model shown above is obtained by fitting the real gamma register value of the sample module with the measured target impact factor value of the sample module (ie, the spectral stimulation value).
[0147] The fitting process is described by taking the first prediction model of a certain color channel (for example, the red channel) when S=1 as an example. For example: the display module has a total of m binding point grayscales, that is, corresponding to m target grayscales, the m target grayscales are respectively recorded as G1, G2.......Gm, the real gamma register values of the red channel of the sample module are respectively recorded as GAM11, GAM12.......GAM1m, and the target influencing factor values (such as spectral stimulus values X or Y or Z) under the target grayscales G1, G2.......Gm measured by the color analyzer are respectively recorded as X11, X12.......X1m. When fitting, multiple coordinate points (X11, GAM11), (X12, GAM12)...(X1m, GAM1m) are used for fitting to obtain the first prediction model R=a1*X1+b1 of the red channel. Similarly, the first prediction models of the green channel and the blue channel can be fitted. When S is greater than 1, the principle of fitting to obtain the first prediction model is the same, which will not be repeated here.
[0148] For example, please refer to Figure 5 and Figure 6 , Figure 5 Schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the first prediction model, Figure 6 for Figure 5 The data of the predicted gamma register value and the real gamma register value corresponding to each target grayscale in, Figure 5 The horizontal axis represents the grayscale, the vertical axis represents the gamma register value, the blue curve represents the prediction curve formed by the predicted gamma register values at different grayscales obtained by the first prediction model, and the orange curve represents the real gamma register value of the sample module. It can be understood that the real gamma register value of the sample module is discrete data, and each real gamma register value corresponds to a target grayscale or a binding point grayscale, but for the convenience of description, Figure 5 The continuous orange curves represent multiple real gamma register values. Figure 6 The first column of data represents the target grayscale, the R column of data represents the real gamma register value of the red channel corresponding to each target grayscale, the R_pre column of data represents the predicted gamma register value of the red channel corresponding to each target grayscale obtained by the first prediction model, and the R_diff column of data represents the difference between the R_pre column of data and the R column of data. Similarly, the G column of data represents the real gamma register value of the green channel corresponding to each target grayscale, the G_pre column of data represents the predicted gamma register value of the green channel corresponding to each target grayscale obtained by the first prediction model, and the G_diff column of data represents the difference between the G_pre column of data and the G column of data; the B column of data represents the real gamma register value of the blue channel corresponding to each target grayscale, the B_pre column of data represents the predicted gamma register value of the blue channel corresponding to each target grayscale obtained by the first prediction model, and the B_diff column of data represents the difference between the B_pre column of data and the B column of data. The acquisition process of the R_pre column data is as follows: the target impact factor value of the sample module at each target grayscale is measured by a color analyzer, and the target impact factor value is substituted into the first prediction model of the red channel to obtain the predicted gamma register value of the red channel corresponding to the target grayscale. Similarly, the G_pre column data and the B_pre column data can be obtained.
[0149] refer to Figure 5 and Figure 6 The smaller the difference between the predicted gamma register value and the actual gamma register value, that is, the smaller the error, the higher the prediction accuracy of the first prediction model. Conversely, the larger the difference between the predicted gamma register value and the actual gamma register value, the lower the prediction accuracy of the first prediction model. Figure 5 and Figure 6 Represents the first prediction model obtained based on a sample module of a certain model, Figure 5 and Figure 6 It can be seen that for the sample module of this model, the difference between the predicted gamma register value and the actual gamma register value is large. In order to ensure the prediction accuracy of the final target model, if the error of the first prediction model is greater than the first error threshold, it means that the prediction accuracy is very low at this time, and the first prediction model cannot be used as the target model. It is understandable that for sample modules of other models, the prediction accuracy of the first prediction model can also be higher. Figure 5 and Figure 6 Just a set of sample data is provided.
[0150] Optionally, the second prediction model is a nonlinear model, and the maximum is three times, and the default target impact factor includes only one spectral stimulus value, such as any one of X, Y and Z. According to the previous analysis, it can be known that the three spectral stimulus values usually have a high correlation, so the default target impact factor includes only one spectral stimulus value. Exemplarily, the second prediction model can be expressed as:
[0151] R=a1*X1 3 +b1*X1 2 +c1*X1+d1;
[0152] G=a2*X1 3 +b2*X1 2 +c2*X1+d2;
[0153] B=a3*X1 3 +b3*X1 2 +c3*X1+d3;
[0154] R, G, B represent the predicted gamma register values corresponding to the red, green, and blue color channels respectively. X1 represents the target influencing factor, which can be any one of X, Y, and Z. a1, b1, c1, d1, a2, b2, c2, d2, a3, b3, c3, and d3 represent the weight coefficients obtained by fitting.
[0155] In the specific implementation, the weight coefficients shown above are obtained by fitting the real gamma register of the sample module and the target impact factor value (i.e., the spectral stimulation value) of the sample module measured. The specific fitting process can refer to the specific example of the first prediction model, which will not be described in detail here. For example, in one embodiment, the weight coefficients obtained by polynomial fitting are as follows:
[0156] a1=0.00000887, b1=-0.01484406, c1=7.90668233, d1=1555.03343137;
[0157] a2=0.00001253, b2=-0.02174869, c2=11.60321901, d2=1131.31947988;
[0158] a3=0.00000581, b3=-0.01088559, c3=6.54118545, d3=1874.70832545;
[0159] Please refer to Figures 7 and 8 , Figure 7 is a schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the second prediction model, Figure 8 for Figure 7 The data of the predicted gamma register value and the real gamma register value corresponding to each target grayscale in, Figure 7 The horizontal axis in the middle represents the grayscale, the vertical axis represents the gamma register value, the continuous curve represents the prediction curve formed by the predicted gamma register values at different grayscales obtained by the second prediction model, the blue scattered value represents the true gamma register value of the sample module, and the red curve on the left represents the prediction curve corresponding to the red channel, the green curve in the middle represents the prediction curve corresponding to the green channel, and the blue curve on the right represents the prediction curve corresponding to the blue channel. Figure 8 The data definitions of each column can be referred to Figure 6 The difference is that Figure 8 The data in the R_pre column represent the predicted gamma register values of the red channel corresponding to each target grayscale obtained by the second prediction model, the data in the G_pre column represent the predicted gamma register values of the green channel corresponding to each target grayscale obtained by the second prediction model, and the data in the B_pre column represent the predicted gamma register values of the blue channel corresponding to each target grayscale obtained by the second prediction model. Figure 8 It can be seen that in the second prediction model, the error degrees of low grayscale and high grayscale are different. At low grayscale, the difference between the predicted gamma register value and the actual gamma register value is small, and at high grayscale, the difference between the predicted gamma register value and the actual gamma register value is large.
[0160] In order to improve the prediction accuracy, the third prediction model in the embodiment of the present disclosure adopts a piecewise nonlinear model, and different sub-models are used for different grayscale intervals. Optionally, the grayscale interval of the display module is divided into a first grayscale interval and a second grayscale interval with the first grayscale threshold as the dividing point, wherein the first grayscale interval is a low grayscale interval with a grayscale less than the first grayscale threshold, and the second grayscale interval is a high grayscale interval with a grayscale greater than or equal to the first grayscale threshold. Correspondingly, the third prediction model includes a first sub-model and a second sub-model, and for each target grayscale in the first grayscale interval, the first sub-model is used to predict its gamma register value, and for each target grayscale in the second grayscale interval, the second sub-model is used to predict its gamma register value.
[0161] Optionally, the third prediction model is expressed as:
[0162]
[0163] Where y represents the predicted gamma register value, X1 represents the target influencing factor, Gray represents the target grayscale, G0 represents the first grayscale threshold, and a1, b1, c1, d1, a2, b2, c2, and d2 represent the weight coefficients obtained by fitting. For example:
[0164] For the red channel, the weight coefficient of the third prediction model is:
[0165] a1=0.00000211, b1=-0.00420886, c1=3.39095058, d1=1978.0759;
[0166] a2=0.08608976, b2=-4.04854718, c2=69.20743046, d2=1453.7703.
[0167] For the green channel, the weight coefficient of the third prediction model is:
[0168] a1=0.00000224, b1=-0.00472095, c1=4.01479007, d1=1869.6743;
[0169] a2=0.75912124, b2=-34.10615732, c2=448.48592615, d2=73.7117.
[0170] For the blue channel, the weight coefficient of the third prediction model is:
[0171] a1=0.00000162, b1=-0.00333009, c1=2.87800404, d1=2264.0376;
[0172] a2=0.05921602, b2=-3.13091602, c2=57.32745871, d2=1828.841.
[0173] Please refer to Figures 9 and 10 , Fig. 9 Schematic diagram of the error comparison between the predicted gamma register value and the actual gamma register value of the third prediction model. Fig.10 for Fig. 9 The data of the predicted gamma register value and the real gamma register value corresponding to each target grayscale in, Fig. 9 The horizontal axis represents the grayscale, the vertical axis represents the gamma register value, the continuous curve represents the prediction curve formed by the predicted gamma register values at different grayscales obtained by the third prediction model, and the scattered value represents the actual gamma register value of the sample module. Specifically, Fig. 9 The three continuous curves in the first column on the left correspond to the prediction curves of the red channel, green channel, and blue channel respectively. The three continuous curves in the second column in the middle correspond to the prediction curves formed by the second sub-models of the red channel, green channel, and blue channel respectively. The three curves in the third column on the right correspond to the prediction curves formed by the first sub-models of the red channel, green channel, and blue channel respectively. Fig.10 The definition of each column data can refer to Figure 6 The difference is that Fig.10 The data in the R_pre column represent the predicted gamma register values of the red channel corresponding to each target grayscale obtained by the third prediction model, the data in the G_pre column represent the predicted gamma register values of the green channel corresponding to each target grayscale obtained by the third prediction model, and the data in the B_pre column represent the predicted gamma register values of the blue channel corresponding to each target grayscale obtained by the third prediction model. Fig.10 and Figure 8 By comparison, it can be seen that Fig.10 The error of each target grayscale is compared Figure 8 Reduced.
[0174] In the present disclosure, Figures 5 to 10 The first prediction model, the second prediction model, and the third prediction model shown are all obtained from the same sample module. It can be seen that different prediction models have different corresponding prediction gamma register values, and the differences are relatively large. In order to improve the accuracy of the predicted gamma register value, the embodiment of the present disclosure discloses a solution of using the prediction model with the smallest error among the prediction models as the target model of the corresponding model of the sample module, that is, selecting the one with the best prediction accuracy from the first prediction model, the second prediction model, and the third prediction model as the final target model.
[0175] In a possible implementation, for any prediction model, the step of obtaining a predicted gamma register value of the prediction model and calculating an error between the predicted gamma register value and the actual gamma register value includes:
[0176] (1) For any prediction model, obtain the predicted gamma register value of the prediction model at different grayscales; (2) calculate the error value between the predicted gamma register value and the actual gamma register value at different grayscales, and sum the error values at different grayscales to obtain the sum of the error values, which is used to represent the error between the predicted gamma register value of the prediction model and the actual gamma register value.
[0177] The different grayscales in the above steps (1) and (2) are the target grayscales, i.e. the binding point grayscales. For any grayscale of any prediction model, the target impact factor value measured by the color analyzer at the grayscale is input into the prediction model to obtain the predicted gamma register value of the prediction model at the grayscale.
[0178] Assume that the red channel, green channel and blue channel of the display module are recorded as the first channel, the second channel and the third channel respectively, and correspondingly, the color channels of the module to be tested and the sample module include the first channel, the second channel and the third channel. At this time, the error value calculated in step (2) includes the first channel error value, the second channel error value and the third channel error value, and the step of summing the error values under different grayscales to obtain the sum of the error values includes: summing the first channel error value, the second channel error value and the third channel error value under different grayscales to obtain the sum of the error values, and the specific calculation formula is as follows:
[0179]
[0180] Among them, Diff represents the sum of error values, R_diff i Represents the red channel error value of the i-th grayscale, G_diff i Indicates the green channel error value of the i-th grayscale, B_diff i represents the blue channel error value of the i-th grayscale, n1 represents the number of grayscales, if the display module includes 27 target grayscales, then n1 is equal to 27.
[0181] Correspondingly, the step of using the prediction model with the smallest error among the prediction models as the target model of the model corresponding to the sample module is: using the prediction model with the smallest sum of error values among the prediction models as the target model of the model corresponding to the sample module.
[0182] In a possible implementation, the gamma adjustment method further includes the following steps:
[0183] (1) Determine whether the minimum error value of the target model is greater than or equal to a first error threshold, wherein the minimum error value is the minimum value among the first channel error value, the second channel error value, and the third channel error value; (2) If the determination result is yes, adjust the model parameters of the target model until the minimum error value is less than the first error threshold. Exemplarily, the first channel error value is represented by R_diff i , the second channel error value is expressed as G_diff i , the third channel error value is expressed as B_diff i , the minimum error value is R_diff i 、G_diff i 、B_diff i The minimum value in .
[0184] Among them, adjusting the model parameters of the target model is the process of adjusting the weight coefficients in the target model. When adjusting, more sample data needs to be obtained for refitting. For example, when the minimum error value of the target model is greater than or equal to the first error threshold, it is necessary to control the sample module to display the target grayscale image multiple times, and use the color analyzer to obtain the target impact factor value each time the sample module displays the target grayscale image, and then use more sample data for fitting to obtain a new weight coefficient, that is, to obtain a new target model. Through repeated adjustments, the minimum error value of the target model is less than the first error threshold.
[0185] In one possible implementation, see Fig.11 For any model, the step of determining the error probability density distribution function corresponding to the model according to multiple sample modules of the model includes:
[0186] Step S301, according to the target model of the model, obtain the predicted gamma register value of the target model at different grayscales, and calculate the error value between the predicted gamma register value and the actual gamma register value at different grayscales.
[0187] In step S302, the obtained error values are grouped according to their sizes, and the frequency of each group of error values is calculated. Then, an error probability density distribution histogram is drawn with the error value as the horizontal coordinate and the frequency as the vertical coordinate. The histogram can intuitively display the error distribution.
[0188] Step S303, fitting the error probability density distribution histogram to obtain an error probability density distribution function. In specific implementation, a suitable probability density distribution function can be selected to fit the distribution of error values according to the shape and characteristics of the histogram.
[0189] The error probability density distribution function can be used to calculate the probability that the error value falls within a certain interval. For example, the calculation formula is as follows:
[0190]
[0191] Where X represents the error value, and P(a≤X≤b) represents the probability that the error value is greater than or equal to a and less than or equal to b. Similarly, based on the error probability density distribution function, when a certain probability value is known, the corresponding error value interval can be obtained. Optionally, the error value interval corresponding to a certain probability value can be determined by solving an equation or looking up a table.
[0192] For example, please refer to Fig.12 , Fig.12It is the error probability density distribution histogram of sample modules of different models. The horizontal axis represents the error value, and the vertical axis represents the frequency. The positive and negative values represent the deviation direction of the error value, that is, the deviation direction of the predicted value relative to the true value. A positive sign indicates that the predicted value is greater than the true value, and a negative sign indicates that the predicted value is less than the true value. Fig.12 In the error probability density distribution histogram on the left, the error value mainly ranges from -150 to 150, and is mainly distributed between -50 and 40. In the error probability density distribution histogram on the right, the error value mainly ranges from -150 to 150, but is mainly distributed between -50 and 70. It can be seen that the error probability density distribution of sample modules of different models is different, and the error probability density distribution functions obtained by subsequent fitting are also different.
[0193] For each model of sample module, after determining the target model and error probability density distribution function corresponding to each model through the above description, step S101 to step S104 can be executed.
[0194] In a possible implementation manner, the step of adjusting the first gamma register value by using an error probability density distribution function includes:
[0195] Step S401, obtaining a first probability value, and obtaining a first error value interval corresponding to the first probability value from the error probability density distribution function.
[0196] Optionally, the first probability value is randomly generated, and the value of the first probability value is greater than 0 and less than or equal to 1. Exemplarily, in the error probability density distribution, the error value has an 80% probability of falling within the error value interval of 0-10. If the randomly generated first probability value α is between 0 and 0.8, the first error value interval is determined to be 0-10.
[0197] Step S402: determine a first error value from the first error value interval according to a first method, and adjust the first gamma register value with the first error value as a step size.
[0198] The first error value interval may include a group of error values, and when adjusting the first gamma register value, a specific error value needs to be selected from the first error value interval as a step size for adjustment. Optionally, the method of selecting the specific error value is recorded as the first method.
[0199] Exemplarily, the step of determining the first error value from the first error value interval according to the first method includes: determining the mean value, edge value or fixed position value of the first error value interval as the first error value, which can also be expressed as, the first method is to select the mean value, edge value (which may be the maximum value, the minimum value or one of the two) or 3 / 4 value (i.e., fixed position value) of the first error value interval as the first error value. Afterwards, the first gamma register value is adjusted with the first error value as the step length, and the specific adjustment process is: the sum of the first error value and the first gamma register value is used as the adjusted first gamma register value, and the adjusted first gamma register value is written into the gamma register of the module to be tested.
[0200] Step S403, when the module to be tested uses the adjusted first gamma register value to perform gamma adjustment, whether the display screen of the module to be tested meets the requirements, if it meets the requirements, then the adjustment is terminated, if it does not meet the requirements, the step of adjusting the first gamma register value by the error probability density distribution function is repeated until the display screen of the module to be tested meets the requirements. Specifically, when it does not meet the requirements, return to step S401, that is, randomly generate a probability value as the first probability value again, and repeat steps S402 to S403.
[0201] It is understandable that the first probability value randomly generated again may be the same as or different from the first probability value randomly generated for the first time. In other embodiments, the first probability value may not be randomly generated, but adjusted in a certain order, such as the first probability values obtained are 0.8, 0.5, 0.4, 0.2, etc.
[0202] In the disclosed embodiment, for display modules of the same model, the error between the predicted value and the true value of each display module is basically similar under a fixed grayscale. Therefore, when adjusting the first gamma register value of the module to be tested, a probabilistic value can be obtained from the error probability density distribution of the fixed grayscale, such as randomly determining a first error value according to the error probability density distribution function and adjusting the first gamma register value with the first error value as the step size. The adjustment process can take into account the randomness and uncertainty of the error, thereby more comprehensively reflecting the actual situation. At the same time, since the error probability density distribution function is derived based on sample data of multiple sample modules, the scheme also has certain statistical reliability and scientificity. For any module to be tested, the accuracy of the adjusted first gamma register value can be guaranteed, and the debugging time of the first gamma register value can be shortened. In addition, the error probability density distribution function of the embodiment of the present disclosure not only includes the size of the error, but also includes the direction of the error (reflected by the positive or negative error value). By quantifying the size and direction of the error, the deviation between the predicted value and the true value can be more accurately represented. When the error probability density distribution function is subsequently used to determine the first error value, the obtained first error value also includes both size and direction. When the first error value is used to adjust the first gamma register value, the reliability of the adjusted first gamma register value can be improved and its stability can be increased.
[0203] Optionally, the gamma adjustment method in the embodiment of the present disclosure can be applied to a driving unit of a display module, which includes a gamma register for storing gamma register values. The gamma adjustment method in the embodiment of the present disclosure is used to generate a first gamma register value that meets the requirements and store it in the gamma register.
[0204] Based on the same inventive concept, the second aspect of the present disclosure provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the gamma adjustment method described above are implemented. In the specific implementation process, the computer storage medium may include: a universal serial bus flash drive (USB, Universal Serial Bus Flash Drive), a mobile hard disk, a read-only memory (ROM, Read Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and other storage media that can store program codes.
[0205] Based on the same inventive concept, the third aspect of the present disclosure provides a computer program product, including a computer program, which implements the steps of the gamma adjustment method described above when executed by a processor. Since the principle of solving the problem by the above computer program is similar to that of the display driving method, the implementation of the above computer program can refer to the implementation of the display driving method, and the repeated parts will not be repeated.
[0206] The computer program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0207] Among them, the display module in the embodiment of the present disclosure can be an organic light-emitting diode (OLED) display module. It is understandable that the display module can also be set to other types according to actual needs. For example, the display module can also be a quantum dot light emitting diode (QLED) display module or a micro light emitting diode (MicroLED) display module.
[0208] Obviously, the above embodiments of the present disclosure are merely examples for clearly illustrating the present disclosure, and are not limitations on the implementation methods of the present disclosure. For ordinary technicians in this field, other different forms of changes or modifications can be made on the basis of the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present disclosure are still within the protection scope of the present disclosure.
Claims
1. A gamma adjustment method, characterized in that: The following steps are involved: Obtain a model of the module to be tested, and determine a target model and an error probability density distribution function corresponding to the model according to the model, wherein the target model is used to represent a mapping relationship between a gamma register value and a target influencing factor, and the error probability density distribution function is used to represent an error distribution between a predicted gamma register value output by the target model and a true gamma register value; Obtaining a target impact factor value when the module to be tested displays a target grayscale image, and inputting the target impact factor value into the target model to obtain a first gamma register value; Detecting whether a display image of the module to be tested meets the requirement when the module to be tested uses the first gamma register value to perform gamma adjustment; When the display screen does not meet the requirements, the first gamma register value is adjusted by the error probability density distribution function, and the adjusted first gamma register value is used to repeatedly perform the step of detecting whether the display screen of the module to be tested meets the requirements when the module to be tested uses the first gamma register value to perform gamma adjustment until the display screen meets the requirements.
2. The gamma adjustment method according to claim 1, characterized in that: The step of adjusting the first gamma register value by using the error probability density distribution function comprises: Obtaining a first probability value, and obtaining a first error value interval corresponding to the first probability value from the error probability density distribution function; A first error value is determined from the first error value interval according to a first method, and the first gamma register value is adjusted with the first error value as a step size.
3. The gamma adjustment method according to claim 2, characterized in that: After the step of adjusting the first gamma register value with the first error value as a step length, the following step further includes: Detecting whether a display image of the module to be tested meets the requirements when the module to be tested performs gamma adjustment using the adjusted first gamma register value; If it does not meet the requirements, repeat the step of adjusting the first gamma register value by using the error probability density distribution function until the display screen of the module to be tested meets the requirements.
4. The gamma adjustment method according to claim 1, characterized in that: The gamma adjustment method further comprises: Get sample modules of different models; For any model, the target model and error probability density distribution function corresponding to the model are determined based on multiple sample modules of the model.
5. The gamma adjustment method according to claim 4, characterized in that: For any model, the step of determining a target model corresponding to the model according to multiple sample modules of the model includes: Obtaining an influencing factor when the sample module displays a target grayscale image, wherein the influencing factor includes three spectral stimulation values; Performing a correlation analysis on the three spectral stimulation values, and determining a target influencing factor according to the correlation analysis result, wherein the target influencing factor includes at least one of the three spectral stimulation values; A target model for representing a mapping relationship between a gamma register value and a target impact factor is established according to the real gamma register value and the target impact factor value of the sample module.
6. The gamma adjustment method according to claim 5, characterized in that: The steps of performing correlation analysis on the spectral stimulus values and determining the target influencing factor according to the correlation analysis results include: detecting whether the three spectral stimulus values are suitable for correlation analysis; When the three spectral stimulation values are suitable for correlation analysis, a correlation matrix is calculated, wherein the correlation matrix includes correlation coefficients representing the correlation between any two spectral stimulation values; Calculating the eigenvalues of the correlation matrix, and determining the number of spectral stimulation values in the target influencing factor according to the eigenvalues; The target impact factor is determined according to the number of spectral stimulation values in the target impact factor and the correlation matrix.
7. The gamma adjustment method according to claim 6, characterized in that: The step of determining the target impact factor according to the number of spectral stimulus values in the target impact factor and the correlation matrix comprises: When the number of spectral stimulation values in the target impact factor is 1, randomly selecting any spectral stimulation value from the three spectral stimulation values as the target impact factor; When the number of spectral stimulation values in the target influencing factor is 2, two spectral stimulation values corresponding to the maximum correlation coefficient values are selected from the correlation matrix as the target influencing factor, and the maximum correlation coefficient value is less than 1.
8. The gamma adjustment method according to claim 6, characterized in that: The step of determining the number of spectral stimulation values in the target influencing factor according to the characteristic value comprises: Visually displaying a first curve, wherein the first curve is used to represent a mapping relationship between the characteristic value and the number of spectral stimulation values; The number of spectral stimulation values corresponding to the inflection points in the first curve is used as the number of spectral stimulation values in the target influencing factor.
9. The gamma adjustment method according to claim 5, characterized in that: The step of establishing a target model for representing the mapping relationship between the gamma register value and the target impact factor according to the real gamma register value of the sample module and the target impact factor value comprises: Establishing at least two prediction models for representing the mapping relationship between the gamma register value and the target impact factor according to the real gamma register value and the target impact factor value of the sample module; For any prediction model, obtaining a predicted gamma register value of the prediction model, and calculating an error between the predicted gamma register value and the true gamma register value; The prediction model with the smallest error among the prediction models is used as the target model of the corresponding model of the sample module.
10. The gamma adjustment method according to claim 9, characterized in that: For any prediction model, the steps of obtaining a predicted gamma register value of the prediction model and calculating an error between the predicted gamma register value and the actual gamma register value include: For any prediction model, obtain the predicted gamma register value of the prediction model at different grayscales; Calculate the error value between the predicted gamma register value and the actual gamma register value at different grayscales; The error values at different gray levels are summed to obtain the sum of the error values; Taking the prediction model with the smallest error among the prediction models as the target model of the model corresponding to the sample module means: taking the prediction model with the smallest sum of error values among the prediction models as the target model of the model corresponding to the sample module.
11. The gamma adjustment method according to claim 10, characterized in that: The color channels of the module to be tested include a first channel, a second channel and a third channel, and the error values include a first channel error value, a second channel error value and a third channel error value; The step of summing the error values at different grayscales to obtain the sum of the error values includes: summing the first channel error value, the second channel error value and the third channel error value at different grayscales to obtain the sum of the error values; The gamma adjustment method further comprises: Determine whether a minimum error value of the target model is greater than or equal to a first error threshold, the minimum error value being a minimum value among the first channel error value, the second channel error value, and the third channel error value; If the judgment result is yes, the model parameters of the target model are adjusted until the minimum error value is smaller than the first error threshold.
12. The gamma adjustment method according to claim 9, characterized in that: The number of the prediction models is three, which are respectively represented as a first prediction model, a second prediction model and a third prediction model. The first prediction model is a linear model, the second prediction model is a nonlinear model, and the third prediction model is a piecewise nonlinear model. The third prediction model includes a first sub-model and a second sub-model. The grayscale of the first sub-model is less than a first grayscale threshold, and the grayscale of the second sub-model is greater than or equal to the first grayscale threshold.
13. The gamma adjustment method according to claim 3, characterized in that: The first probability value is randomly generated, and a value of the first probability value is greater than 0 and less than or equal to 1; The step of determining the first error value from the first error value interval according to the first method comprises: The mean value, edge value or fixed position value of the first error value interval is determined as the first error value.
14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the gamma adjustment method according to any one of claims 1 to 13 are implemented.
15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the gamma adjustment method according to any one of claims 1 to 13 are implemented.
Citation Information
Patent Citations
Gamma adjusting and recalibrating method and system of OLED module based on multi-objective relational mathematic model
CN108597442A
Adjusting device, method for adjusting gamma voltage and white balance of display screen and medium
CN111243550A
Test adjusting system based on AMOLED Gamma method
CN115188328A
Display panel and debugging method and debugging device thereof
CN116631318A
Gamma adjustment method and device of display panel, electronic equipment and storage medium
CN118038790A
Cited By
Gamma debugging method, computer readable storage medium and computer program product
CN120014975A
Sensor-based multiplex circuit overvoltage fault identification system
CN120216947A
Sensor-based Multiplexing Circuit Overvoltage Fault Identification System
CN120216947B
Estimation method and system for correlation of pollutant data with skewed distribution and undetected values
CN120561444A
