A method and system for predicting the brightness of a curved area of an OLED display screen

Abstract

The application relates to a curved surface area brightness prediction method and system of an OLED display screen, and the method comprises the following steps: acquiring an original brightness map of the OLED display screen, and extracting curved surface area brightness data from the original brightness map; pre-processing the curved surface area brightness data to obtain a plurality of groups of row vectors, and combining the plurality of groups of row vectors into a matrix; taking the matrix as training input data of a neural network, simultaneously converting the curved surface area brightness data through a preset method, and taking the converted curved surface area brightness data as training output data of the neural network; setting training parameters of the neural network, training the neural network with the set training parameters, and obtaining a trained neural network; and predicting the brightness of the curved surface area of the OLED display screen to be predicted through the trained neural network. The predicted curved surface area brightness is more balanced and delicate, and the brightness data distribution of the pixel points is also more gentle.

Description

Technical Field

[0001] This invention relates to the field of OLED screen display compensation technology, and in particular to a method and system for predicting the brightness of curved areas of an OLED display screen. Background Technology

[0002] OLED (Organic Light-Emitting Diode) screens manufactured in factories inevitably have some degree of display unevenness (mura), which affects their display quality. The flat panel display industry refers to the method of eliminating mura as Demura. Utilizing Demura systems can improve product display quality, thereby enhancing product competitiveness. The first step in Demura technology is to capture screen images and extract the brightness data corresponding to each LED to obtain a corresponding brightness map. Because the sides of a curved screen are not on the same horizontal line as the main screen, the brightness data of the curved area in the captured image is relatively lower than that of the main screen area, thus requiring secondary conversion and correction.

[0003] Existing technologies mainly use traditional image filtering methods for processing and calculation. First, the average brightness data of the central area of ​​the screen is calculated. Then, the ratio between the brightness data of each point in the original brightness map and the average brightness data is calculated. The original brightness map is then filtered by a preset filtering kernel. Finally, the filtered image is multiplied by the previously obtained ratio to obtain the final brightness map, thus completing the conversion of brightness data in the curved area.

[0004] While existing technologies can temporarily meet production requirements, they are not suitable for curved screens with large curvatures. Furthermore, the brightness data of the curved area calculated by existing technologies still differs from the brightness data of the main screen, which may affect the final compensation effect of the screen. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the problems in the prior art that the brightness conversion of the curved area of ​​the OLED display is not applicable to curved screens with large curvature and the brightness conversion effect is poor.

[0006] To address the aforementioned technical problems, this invention provides a method for predicting the brightness of curved areas in an OLED display screen, comprising:

[0007] Step S1: Obtain the original brightness map of the OLED display and extract the brightness data of the curved area from the original brightness map;

[0008] Step S2: Preprocess the brightness data of the curved area to obtain several sets of row vectors, and merge the several sets of row vectors into a matrix;

[0009] Step S3: Use the matrix as the training input data of the neural network, and at the same time, convert the surface region brightness data through a preset method, and use the converted surface region brightness data as the training output data of the neural network.

[0010] Step S4: Set the training parameters of the neural network and train the neural network with the set training parameters to obtain a trained neural network;

[0011] Step S5: Use the trained neural network to predict the brightness of the curved area of ​​the OLED display screen to be predicted.

[0012] In one embodiment of the present invention, step S2 involves preprocessing the brightness data of the curved area to obtain several sets of row vectors, and merging the several sets of row vectors into a matrix. The method includes:

[0013] The brightness data of the curved area is rearranged into a row vector of 1 row and m columns, which is used as the first row of training input data for the neural network.

[0014] Calculate the difference in grayscale value between each pixel and the pixel to its right in the brightness data of the curved area, and rearrange them into a row vector of 1 row and m columns, which will be used as the second row of training input data for the neural network.

[0015] Calculate the average gray value of each pixel in the brightness data of the curved area and the pixels in its preset neighborhood range, and rearrange them into a row vector of 1 row and m columns, which is used as the third row of the training input data of the neural network. The preset neighborhood range is an eight-neighborhood with a step size of one pixel.

[0016] The original brightness map is normalized, and the normalized surface region brightness data is rearranged into a row matrix of 1 row and m columns, which is used as the fourth row of input data for the neural network.

[0017] Create a new empty image matrix of the same size as the surface region. Set the value of each point in the first empty image matrix to the mean gray value of the non-zero region of the normalized brightness map. Then rearrange the matrix into a row vector of 1 row and m columns, which will be used as the fifth row of input data for the neural network.

[0018] Create a new empty image matrix of the same size as the surface region. Set the value of each point in the second empty image matrix to the column number of that point. Then rearrange the matrix into a row vector of 1 row and m columns, which will be used as the sixth row of training input data for the neural network.

[0019] The six sets of row vectors are merged into a 6-row, m-column matrix.

[0020] In one embodiment of the present invention, in step S3, the brightness data of the curved area is calculated using a preset method, and the calculated brightness data of the curved area is used as the training output data of the neural network. The method includes:

[0021] Calculate the average brightness data of the central region of the d*d region in the original brightness map;

[0022] Calculate the proportional relationship between the brightness data of each point in the original brightness map and the average brightness data;

[0023] The original brightness map is filtered by a preset filter kernel, and the filtered image is multiplied by the obtained proportional relationship to obtain the final brightness map, thus completing the conversion of the brightness data of the curved area. Finally, the converted brightness data of the curved area is rearranged into a row vector of 1 row and m columns and used as the training output data of the neural network.

[0024] In one embodiment of the present invention, setting the training parameters of the neural network in step S4 includes:

[0025] The neural network is a backpropagation (BP) neural network, which includes an input layer, hidden layers, and an output layer.

[0026] The number of neurons in the hidden layer of the neural network is set according to the training input data and training output data.

[0027] Set the training function, number of iterations, maximum error value, and learning rate of the neural network.

[0028] In one embodiment of the present invention, the number of neurons in the hidden layer of the neural network is set according to the training input data and the training output data, using the following formula:

[0029]

[0030] Where h represents the number of hidden layer neurons, n represents the number of input layer neurons, m represents the number of output layer neurons, and a represents a constant and a∈[1,10].

[0031] In one embodiment of the present invention, step S1, which involves extracting surface region brightness data from the original brightness map, includes: extracting surface region brightness data corresponding to the three RGB color channels from the original brightness map.

[0032] To address the aforementioned technical problems, this invention provides a brightness prediction system for curved areas of an OLED display screen, comprising:

[0033] Extraction module: used to obtain the original brightness map of the OLED display and extract the brightness data of the curved area from the original brightness map;

[0034] Preprocessing module: used to preprocess the brightness data of the curved area to obtain several sets of row vectors, and merge the several sets of row vectors into a matrix;

[0035] Training data construction module: used to use the matrix as training input data for the neural network, and to convert the surface region brightness data through a preset method, and use the converted surface region brightness data as training output data for the neural network;

[0036] Parameter setting module: used to set the training parameters of the neural network, and to train the neural network with the set training parameters to obtain a trained neural network;

[0037] Prediction module: Used to predict the brightness of the curved area of ​​the OLED display screen to be predicted using a trained neural network.

[0038] In one embodiment of the present invention, the preprocessing module preprocesses the brightness data of the curved area to obtain several sets of row vectors, and merges the several sets of row vectors into a matrix, including:

[0039] The first data construction unit is used to rearrange the brightness data of the curved area into a row vector of 1 row and m columns, which serves as the first row of training input data for the neural network.

[0040] The second data construction unit is used to calculate the difference in gray value between each pixel and the pixel to its right in the brightness data of the curved area, and rearrange them into a row vector of 1 row and m columns, which serves as the second row of training input data for the neural network.

[0041] The third data construction unit is used to calculate the average gray value of each pixel in the brightness data of the curved area and the pixels in its preset neighborhood range, and rearrange them into a row vector of 1 row and m columns, which is used as the third row of training input data for the neural network. The preset neighborhood range is an eight-neighborhood with a step size of one pixel.

[0042] The fourth data construction unit is used to normalize the original brightness map and rearrange the normalized surface region brightness data into a row matrix of 1 row and m columns, which serves as the fourth row of input data for the neural network.

[0043] The fifth data construction unit is used to create a first empty image matrix with the same size as the curved area, set the value of each point in the first empty image matrix to the mean gray value of the non-zero region of the normalized brightness map, and then rearrange the matrix into a row vector of 1 row and m columns, which is used as the fifth row of input data for the neural network.

[0044] The sixth data construction unit is used to create a second empty image matrix with the same size as the surface region, set the value of each point in the second empty image matrix to the column number of the point, and then rearrange the matrix into a row vector of 1 row and m columns, which serves as the sixth row of training input data for the neural network.

[0045] The merging unit is used to merge the six groups of row vectors into a 6-row, m-column matrix.

[0046] To address the aforementioned technical problems, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the curved area brightness prediction method for the OLED display described above.

[0047] To address the aforementioned technical problems, the present invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for predicting the brightness of curved areas of an OLED display screen.

[0048] The technical solution of the present invention has the following advantages compared with the prior art:

[0049] This invention is applicable to curved screens with large curvature. The brightness data of the curved area predicted by this invention can be kept as close as possible to the brightness data of the main screen.

[0050] This invention creatively constructs 6 row vectors to train a BP neural network, which enables the BP neural network to reach the optimal prediction state.

[0051] This invention predicts that the brightness of the curved areas of OLED displays will be more balanced and delicate, and the brightness data distribution of pixels will be more gradual.

[0052] This invention is highly practical and suitable for large-scale promotion. Attached Figure Description

[0053] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0054] Figure 1 This is a flowchart of the method of the present invention;

[0055] Figure 2 This is a schematic diagram of the BP neural network structure in an embodiment of the present invention;

[0056] Figure 3 This is a flowchart illustrating the method for constructing training input data and training output data during the BP neural network training process in this embodiment of the invention.

[0057] Figure 4 This is a diagram showing the experimental comparison results in an embodiment of the present invention. Detailed Implementation

[0058] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0059] Example 1

[0060] Because the curved areas on both sides of the OLED display are not at the same horizontal level as the main screen, the collected brightness data is lower, affecting subsequent brightness compensation. To solve this problem, this invention proposes a method for predicting the brightness of the curved areas of an OLED display. Figure 1 As shown, it includes:

[0061] Step S1: Obtain the original brightness map of the OLED display and extract the brightness data of the curved area from the original brightness map;

[0062] Step S2: Preprocess the brightness data of the curved area to obtain several sets of row vectors, and merge the several sets of row vectors into a matrix;

[0063] Step S3: Use the matrix as the training input data of the neural network, and at the same time, convert the surface region brightness data through a preset method, and use the converted surface region brightness data as the training output data of the neural network.

[0064] Step S4: Set the training parameters of the neural network and train the neural network with the set training parameters to obtain a trained neural network;

[0065] Step S5: Use the trained neural network to predict the brightness of the curved area of ​​the OLED display screen to be predicted.

[0066] Further, in step S2, the brightness data of the curved area is preprocessed to obtain several sets of row vectors, and the several sets of row vectors are merged into a matrix. The method includes:

[0067] 1. Rearrange the brightness data of the curved area in the original brightness map of the screen into a row vector of 1 row and m columns, and use it as the first row of training input data for the BP neural network;

[0068] 2. Calculate the difference in grayscale value between each pixel in the curved region of the original screen brightness image and the pixel to its right, and rearrange them into a row vector of 1 row and m columns, which will be used as the second row of training input data for the BP neural network.

[0069] 3. Calculate the average gray value of each pixel in the curved area of ​​the original screen brightness map and the average gray value of the pixels within one pixel range (if the pixel is an edge pixel, expand one pixel at the edge and copy the gray value of the edge pixel), and rearrange them into a row vector of 1 row and m columns, which will be used as the third row of training input data for the BP neural network.

[0070] 4. Normalize the original brightness map, and rearrange the normalized surface region brightness data into a 1xm matrix, which will be used as the fourth row of input data for the BP neural network.

[0071] 5. Create a new empty image matrix with the same size as the surface region. Set the value of each point in the first empty image matrix to the mean gray value of the non-zero region of the normalized brightness map. Then rearrange the matrix into a row vector of 1 row and m columns and use it as the fifth row of input data for the BP neural network.

[0072] 6. Create a new second empty image matrix with the same size as the surface region. Set the value of each point in the second empty image matrix to the column number of that point. Then rearrange the matrix into a row vector of 1 row and m columns, and use it as the sixth row of training input data for the BP neural network.

[0073] 7. Combine the six row vectors obtained above into a 6xm matrix, which will be used as the training input data for the BP neural network.

[0074] It is worth mentioning that the reason for constructing 6 row vectors in this embodiment is as follows:

[0075] The first row is the raw brightness, which is the most direct training feature;

[0076] The second row shows the grayscale difference on the right, which is used to establish the relationship between adjacent columns of data.

[0077] The third row is the mean value across eight neighborhoods, used to establish the relationship between each pixel and its surrounding data;

[0078] The fourth row contains normalized brightness data, which is used to apply brightness data to different gray levels;

[0079] The fifth row is the normalized mean, which is used to establish the relationship between local pixels and the overall screen data;

[0080] The sixth row represents the number of pixel columns. Since the brightness deviation in curved areas is mainly due to the different screen heights corresponding to each column of pixels, including the column number in the training allows for targeted training of pixels with different column numbers.

[0081] It is easy to see that the six row vectors constructed in this embodiment can effectively train the BP neural network, and the trained BP neural network can achieve the best prediction effect on the brightness of the curved area.

[0082] Further, in step S3, the brightness data of the curved area is calculated using a preset method, and the calculated brightness data of the curved area is used as the training output data of the neural network. The method includes:

[0083] Calculate the average brightness data of the central region of the d*d region in the original brightness map;

[0084] Calculate the proportional relationship between the brightness data of each point in the original brightness map and the average brightness data;

[0085] The original brightness map is filtered by a preset filter kernel, and the filtered image is multiplied by the obtained proportional relationship to obtain the final brightness map, thus completing the conversion of the brightness data of the curved area. Finally, the converted brightness data of the curved area is rearranged into a row vector of 1 row and m columns, and used as the training output data of the BP neural network.

[0086] For details on the specific process of constructing training input and output data during BP neural network training, please refer to [link to relevant documentation]. Figure 3 .

[0087] The neural network used in this embodiment is a backpropagation (BP) neural network, which includes an input layer, hidden layers, and an output layer; the formula for the number of neurons in the hidden layer is:

[0088]

[0089] Where h represents the number of hidden layer neurons, n represents the number of input layer neurons, m represents the number of output layer neurons, and a represents a constant and a∈[1,10]

[0090] In this embodiment, the number of input layer neurons is 6, and the number of output layer neurons is 1. Therefore, the number of hidden layer neurons is between 4 and 12. The optimal number of hidden layer neurons can be obtained by setting the number of neurons for each hidden layer and then verifying it experimentally. See details. Figure 2 .

[0091] The number of neurons in the hidden layer of the neural network is set according to the training input data and training output data.

[0092] Furthermore, in step S4, the training parameters of the neural network are set, including setting the training function, number of iterations, maximum error value, and learning rate of the neural network.

[0093] Further, in step S1, extracting surface region brightness data from the original brightness map includes: extracting surface region brightness data corresponding to the three RGB color channels from the original brightness map respectively.

[0094] The experimental results are analyzed as follows:

[0095] like Figure 4 As shown, the left image is the brightness prediction result of the prior art, and the right image is the brightness prediction result of the method of the present invention. To the naked eye, the two images are almost indistinguishable, but the brightness data transition in the right image appears smoother. Table 1 shows the variance comparison results of the curved area after processing six different gray levels of the test screen using the two methods. The smaller the variance value, the smoother the brightness data in that area. It is clear from the data that regardless of the gray level, the brightness prediction result of the method of the present invention is superior to that of the prior art in the curved area.

[0096] Table 1. Variance of surface regions at different gray levels converted using two methods.

[0097]

[0098] Example 2

[0099] This embodiment provides a brightness prediction system for curved areas of an OLED display, including:

[0100] Extraction module: used to obtain the original brightness map of the OLED display and extract the brightness data of the curved area from the original brightness map;

[0101] Preprocessing module: used to preprocess the brightness data of the curved area to obtain several sets of row vectors, and merge the several sets of row vectors into a matrix;

[0102] Training data construction module: used to use the matrix as training input data for the neural network, and to convert the surface region brightness data through a preset method, and use the converted surface region brightness data as training output data for the neural network;

[0103] Parameter setting module: used to set the training parameters of the neural network, and to train the neural network with the set training parameters to obtain a trained neural network;

[0104] Prediction module: Used to predict the brightness of the curved area of ​​the OLED display screen to be predicted using a trained neural network.

[0105] Furthermore, the preprocessing module preprocesses the brightness data of the curved area to obtain several sets of row vectors, and merges these sets of row vectors into a matrix, including:

[0106] The first data construction unit is used to rearrange the brightness data of the curved area into a row vector of 1 row and m columns, which serves as the first row of training input data for the neural network.

[0107] The second data construction unit is used to calculate the difference in gray value between each pixel and the pixel to its right in the brightness data of the curved area, and rearrange them into a row vector of 1 row and m columns, which serves as the second row of training input data for the neural network.

[0108] The third data construction unit is used to calculate the average gray value of each pixel in the brightness data of the curved area and the pixels in its preset neighborhood range, and rearrange them into a row vector of 1 row and m columns, which is used as the third row of training input data for the neural network. The preset neighborhood range is an eight-neighborhood with a step size of one pixel.

[0109] The fourth data construction unit is used to normalize the original brightness map and rearrange the normalized surface region brightness data into a row matrix of 1 row and m columns, which serves as the fourth row of input data for the neural network.

[0110] The fifth data construction unit is used to create a first empty image matrix with the same size as the curved area, set the value of each point in the first empty image matrix to the mean gray value of the non-zero region of the normalized brightness map, and then rearrange the matrix into a row vector of 1 row and m columns, which is used as the fifth row of input data for the neural network.

[0111] The sixth data construction unit is used to create a second empty image matrix with the same size as the surface region, set the value of each point in the second empty image matrix to the column number of the point, and then rearrange the matrix into a row vector of 1 row and m columns, which serves as the sixth row of training input data for the neural network.

[0112] The merging unit is used to merge the six groups of row vectors into a 6-row, m-column matrix.

[0113] Example 3

[0114] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the curved area brightness prediction method for the OLED display screen described in Embodiment 1.

[0115] Example 4

[0116] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the steps of the curved area brightness prediction method for the OLED display screen described in Embodiment 1.

[0117] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0118] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0119] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0120] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0121] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0122] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for predicting the brightness of a curved area of ​​an OLED display screen, characterized in that: The application relates to a method for predicting the brightness of a curved area of an OLED display screen. The method comprises the following steps: S1: obtaining an original brightness map of the OLED display screen, and extracting curved area brightness data from the original brightness map; S2: preprocessing the curved area brightness data to obtain a plurality of row vectors, and combining the plurality of row vectors into a matrix; S3: taking the matrix as training input data of a neural network, and simultaneously calculating the curved area brightness data through a preset method to obtain the curved area brightness data as training output data of the neural network; S4: setting training parameters of the neural network, and training the neural network with the set training parameters to obtain a trained neural network; 2. The method of claim 1, wherein the method further comprises: S5: predicting the brightness of the curved area of the OLED display screen to be predicted through the trained neural network. In the step S2, the curved area brightness data is preprocessed to obtain a plurality of row vectors, and the plurality of row vectors are combined into a matrix, and the method comprises the following steps: The curved area brightness data is rearranged into a row vector with m columns as the first row of the training input data of the neural network; The difference between the gray value of each pixel point in the curved area brightness data and the gray value of the right side of the pixel point is calculated, and the difference is rearranged into a row vector with m columns as the second row of the training input data of the neural network; The average value of the gray values of the pixel points in the preset neighborhood range of each pixel point in the curved area brightness data is calculated, and the average value is rearranged into a row vector with m columns as the third row of the training input data of the neural network, wherein the preset neighborhood range is an eight-neighborhood with a step of one pixel; The original brightness map is normalized, and the normalized curved area brightness data is rearranged into a row matrix with m columns as the fourth row of the input data of the neural network; A first empty image matrix with the same size as the curved area is newly created, the value of each point in the first empty image matrix is set as the average value of the gray values of the non-zero area of the normalized brightness map, and the matrix is rearranged into a row vector with m columns as the fifth row of the input data of the neural network; A second empty image matrix with the same size as the curved area is newly created, the value of each point in the second empty image matrix is set as the column number of the point, and the matrix is rearranged into a row vector with m columns as the sixth row of the training input data of the neural network; 3. The method of claim 1, wherein the method further comprises: The six groups of row vectors are combined into a matrix with six rows and m columns. In the step S3, the curved area brightness data is calculated through a preset method, and the calculated curved area brightness data is taken as the training output data of the neural network, and the method comprises the following steps: The average value of the brightness data of a d*d center area of the original brightness map is calculated; The proportional relationship between the brightness data of each point in the original brightness map and the average value of the brightness data is calculated; The original brightness map is filtered through a preset filter kernel, and the filtered image is multiplied by the obtained proportional relationship to obtain a final brightness map, the calculation of the curved area brightness data is completed, and finally the calculated curved area brightness data is rearranged into a row vector with m columns and taken as the training output data of the neural network.

4. The method of claim 1, wherein the method further comprises: The step S4 sets the training parameters of the neural network, including: The neural network adopts a BP neural network, and the BP neural network includes an input layer, a hidden layer and an output layer; According to the training input data and the training output data, the number of neurons in the hidden layer of the neural network is set; The training function, the number of iterations, the maximum error value and the learning rate of the neural network are set.

5. The method of claim 4, wherein the method further comprises: According to the training input data and the training output data, the number of neurons in the hidden layer of the neural network is set, and the formula is: Wherein, h represents the number of hidden layer neurons, n represents the number of input layer neurons, m represents the number of output layer neurons, and a represents a constant and a∈[1,10].

6. The method of claim 1, wherein the method further comprises: In the step S1, the curved area brightness data is extracted from the original brightness map, including: extracting the curved area brightness data corresponding to the RGB three color channels from the original brightness map.

7. A brightness prediction system for curved areas of an OLED display screen, characterized in that: Including: The extraction module is used to obtain the original brightness map of the OLED display screen, and extract the curved area brightness data from the original brightness map; The preprocessing module is used to preprocess the curved area brightness data to obtain a plurality of row vectors, and combine the plurality of row vectors into a matrix; The training data construction module is used to take the matrix as the training input data of the neural network, and simultaneously calculate the curved area brightness data by a preset method, and take the calculated curved area brightness data as the training output data of the neural network; The parameter setting module is used to set the training parameters of the neural network, and train the neural network with the set training parameters to obtain a trained neural network; The prediction module is used to predict the brightness of the curved area of the OLED display screen to be predicted by the trained neural network.

8. The curved area brightness prediction system for OLED display screens of claim 7, wherein: In the preprocessing module, the curved area brightness data is preprocessed to obtain a plurality of row vectors, and the plurality of row vectors are combined into a matrix, including: The first data construction unit is used to rearrange the curved area brightness data into a row vector of 1 row and m columns as the first row of the training input data of the neural network; The second data construction unit is used to calculate the gray value difference between each pixel point and the pixel point on the right side of the curved area brightness data, and rearrange it into a row vector of 1 row and m columns as the second row of the training input data of the neural network; The third data construction unit is used to calculate the average value of the gray values of the pixel points in the preset neighborhood range of each pixel point in the curved area brightness data, and rearrange it into a row vector of 1 row and m columns as the third row of the training input data of the neural network, wherein the preset neighborhood range is an eight-neighborhood with a step of one pixel; The fourth data construction unit is used to perform normalization operation on the original brightness map, and rearrange the normalized curved area brightness data into a row matrix of 1 row and m columns as the fourth row of the input data of the neural network; a fifth data construction unit, configured to newly create a first empty image matrix with the same size as the curved surface region, set the value of each point in the first empty image matrix as the average of the gray scale values of the non-zero region of the normalized luminance map, and then rearrange the matrix into a row vector with one row and m columns as the fifth row of input data of the neural network; a sixth data construction unit, configured to newly create a second empty image matrix with the same size as the curved surface region, set the value of each point in the second empty image matrix as the column number where the point is located, and then rearrange the matrix into a row vector with one row and m columns as the sixth row of training input data of the neural network; a merging unit, configured to merge the six groups of row vectors into a matrix with six rows and m columns.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the curved surface region luminance prediction method of the OLED display screen according to any one of claims 1 to 6.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to implement the steps of the curved surface region luminance prediction method of the OLED display screen according to any one of claims 1 to 6.

Images