Image dynamic enhancement processing method for flat panel display module

Abstract

The application discloses a flat panel display module image dynamic enhancement processing method, and belongs to the technical field of flat panel display modules. The flat panel display module original image to be subjected to image dynamic enhancement processing is acquired according to a camera module. An image dynamic enhancement model is constructed according to deep learning. The flat panel display module original image is subjected to image dynamic enhancement processing based on the image dynamic enhancement model, and the enhanced image of the flat panel display module is determined. The application solves the problem that the existing flat panel display module cannot perform dynamic enhancement processing on the flat panel display image, resulting in poor flat panel display effect and reduced user experience. The application constructs an image dynamic enhancement model according to deep learning and in combination with the flat panel display module historical image subjected to image dynamic enhancement processing. The flat panel display module original image is subjected to image dynamic enhancement processing based on the image dynamic enhancement model, the flat panel display image can be subjected to dynamic enhancement processing, and the flat panel display effect and user experience are improved.

Description

Technical Field

[0001] This invention relates to the field of flat panel display module technology, specifically to a method for dynamic image enhancement processing of flat panel display modules. Background Technology

[0002] Flat panel display modules are crucial components in modern electronic devices, widely used in smartphones, televisions, automotive displays, smart homes, and medical equipment. They are display components that integrate liquid crystal display devices, connectors, integrated circuits, and structural components, converting electronic signals into visual signals to present high-quality images to users.

[0003] Existing flat panel display modules cannot dynamically enhance the displayed images, resulting in poor display quality and a reduced user experience. Summary of the Invention

[0004] The purpose of this invention is to provide a method for dynamic image enhancement processing of flat panel display modules, which can dynamically enhance the images displayed on the flat panel, improve the display effect and user experience, and solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] Image dynamic enhancement processing methods for flat panel display modules include:

[0007] The original image of the tablet display module to be dynamically enhanced is obtained from the camera module.

[0008] An image dynamic enhancement model is constructed based on deep learning. The original image of the flat panel display module is then dynamically enhanced based on the image dynamic enhancement model to determine the enhanced image of the flat panel display module.

[0009] Preferably, an image dynamic enhancement model is constructed, and the following operations are performed:

[0010] Collect historical images of the flat panel display module and perform dynamic image enhancement processing on the historical images of the flat panel display module, including grayscale correction, contrast enhancement, and spatial and frequency domain enhancement;

[0011] The historical images of the flat panel display module after dynamic image enhancement are divided into training set and test set.

[0012] The deep learning model is trained based on the training set, enabling it to autonomously learn image dynamic enhancement behavior from the training set and perform dynamic enhancement processing on the original images of the flat panel display module, thus determining the image dynamic enhancement model.

[0013] The performance of the image dynamic enhancement model is tested and evaluated based on the test set, the model test evaluation results are determined, and the image dynamic enhancement model is adjusted and optimized based on the model test evaluation results to determine the optimal image dynamic enhancement model.

[0014] Preferably, the original image of the flat panel display module is subjected to dynamic image enhancement processing based on the image dynamic enhancement model, and the following operations are performed:

[0015] Deploy the image dynamic enhancement model by placing it in the image dynamic enhancement processing environment of the flat panel display module;

[0016] The original image of the flat panel display module is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement model analyzes the original image of the flat panel display module and automatically performs image dynamic enhancement processing on the original image of the flat panel display module, thereby determining the enhanced image of the flat panel display module.

[0017] Preferably, the historical images of the flat panel display module are dynamically enhanced, including:

[0018] The historical images of the flat panel display module are processed according to Gamma correction to adjust the grayscale distribution of the historical images, correct nonlinear brightness distortion, and improve the nonlinear response characteristics of the flat panel display module. In particular, by fitting the gamma curve in real time, the brightness level of the historical images of the flat panel display module is dynamically adjusted to ensure that the display effect of the historical images of the flat panel display module remains consistent under different lighting conditions.

[0019] Based on the Gamma correction in the anti-Gamma correction and the Gamma correction in the historical image of the flat panel display module, the true grayscale is restored, achieving linear brightness performance and a unified display effect. In the flat panel display module, linear grayscale brightness performance is achieved by adjusting the relationship between grayscale voltage and brightness. By applying a power function transformation to the historical image of the flat panel display module, grayscale distortion in the transmission of the historical image of the flat panel display module is compensated, and loss of low grayscale details is avoided.

[0020] Histogram equalization optimizes the grayscale distribution of historical images of the flat panel display module to a near-normal distribution, ensuring the overall brightness uniformity of the historical images of the flat panel display module. For dynamic scenes, local histogram enhancement is used to adjust the grayscale range separately for dark or bright areas, reducing local overexposure or underexposure caused by global processing.

[0021] Preferably, the dynamic enhancement processing of historical images of the flat panel display module further includes:

[0022] Adaptive contrast enhancement is used to perform regional contrast stretching on the historical image of the flat panel display module, which enhances the contrast of the historical image of the flat panel display module, expands the brightness range of the historical image of the flat panel display module, and improves the details in dark areas and the sense of layering in bright areas. Specifically, the historical image of the flat panel display module is divided into multiple sub-blocks, and the gray-level mean and variance are calculated for each sub-block. The gain coefficient is dynamically adjusted to enhance texture details and enrich the contrast of the region, while suppressing noise in flat areas.

[0023] Based on nonlinear grayscale transformation, an S-curve is used to transform historical images of the flat panel display module to expand the grayscale range of interest. Specifically, a steep slope transformation is used for low grayscale areas to enhance details in dark areas, while a gentle slope is used for high grayscale areas to avoid overexposure.

[0024] Preferably, the dynamic enhancement processing of historical images of the flat panel display module further includes:

[0025] The historical images of the flat panel display module are processed by median filtering to remove salt-and-pepper noise while retaining edge information, thus achieving smoothing of the historical images of the flat panel display module.

[0026] The Sobel operator is used to sharpen and enhance the historical images of the flat panel display module, and the edge gradient of the historical images of the flat panel display module is extracted to reduce noise interference while detecting edge features.

[0027] Image filtering processing is performed on the historical images of the flat panel display module based on Butterworth filtering to enhance the high-frequency components of the historical images of the flat panel display module and reduce the ringing effect;

[0028] Image filtering is performed on the historical images of the flat panel display module based on homomorphic filtering to separate the illumination component and reflection component of the historical images of the flat panel display module. The uneven illumination is improved by suppressing low frequencies and enhancing high frequencies.

[0029] Preferably, the performance of the image dynamic enhancement model is tested and evaluated based on the test set, and the following operations are performed:

[0030] The test set is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement performance of the model is tested and evaluated based on the test set. The model test evaluation index data is determined, and the model test evaluation result is determined based on the model test evaluation index data.

[0031] Among them, the model test evaluation index data is compared with the preset model test evaluation index thresholds, the matching degree between the model test evaluation index data and the model test evaluation index thresholds is analyzed, and the model test evaluation result is judged based on the matching degree between the model test evaluation index data and the model test evaluation index thresholds.

[0032] When the model test evaluation index data matches the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement model has good image dynamic enhancement performance, that is, it can perform dynamic enhancement processing on the original image of the flat panel display module.

[0033] When the model test evaluation index data does not match the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement performance of the image dynamic enhancement model is poor, that is, it cannot perform dynamic enhancement processing on the original image of the flat panel display module.

[0034] Preferably, the image dynamic enhancement model is adjusted and optimized based on the model test and evaluation results, and the following operations are performed:

[0035] When the image dynamic enhancement model has poor image dynamic enhancement performance and cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are continuously adjusted, and the image dynamic enhancement model after parameter adjustment is iteratively optimized.

[0036] The performance of the adjusted and optimized image dynamic enhancement model was tested and evaluated again to determine whether the adjusted and optimized image dynamic enhancement model could perform dynamic enhancement processing on the original image of the flat panel display module.

[0037] When the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, the optimal image dynamic enhancement model is directly determined.

[0038] If the adjusted and optimized image dynamic enhancement model cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are adjusted and optimized again until the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, thus determining the optimal image dynamic enhancement model.

[0039] Preferably, the brightness level of historical images of the flat panel display module is dynamically adjusted by fitting the gamma curve in real time, including:

[0040] Obtain any frame from the historical images of the flat panel display module as the image to be processed;

[0041] Based on the SLIC superpixel algorithm, the image to be processed is clustered into several superpixel regions, and each superpixel region is used as a basic processing unit.

[0042] Calculate the luminance variance of pixels within each superpixel region;

[0043] The target neighborhood range of each pixel within each superpixel region is determined based on the brightness variance; the local brightness mean of each pixel within the superpixel region is calculated based on the target neighborhood range.

[0044] Obtain the global brightness difference index between the image to be processed and the previous frame image;

[0045] The number of target consecutive frames of the image to be processed is determined based on the global brightness difference index.

[0046] Based on the number of target consecutive frames, the historical local brightness average value of each target consecutive frame corresponding to each superpixel region is obtained, and a historical local brightness average value sequence is constructed.

[0047] For each superpixel region in the image to be processed, the historical average brightness of all pixels within the superpixel region is aggregated to generate a region-level time-brightness sequence.

[0048] Construct historical sample groups and current frame sample groups for each superpixel region based on the historical local brightness mean sequence and the region-level time-brightness sequence of each superpixel region.

[0049] The historical sample group and the current frame sample group of each superpixel region are merged, and the gamma curve is fitted by logarithmic linearization and least squares method to generate the gamma value and brightness correction coefficient of each superpixel region.

[0050] The brightness correction algorithm is determined based on the gamma value and brightness correction coefficient of the superpixel region;

[0051] Superpixel regions are classified based on the Otsu algorithm;

[0052] The brightness value of the superpixel region is dynamically adjusted based on the brightness correction algorithm and the category of the superpixel region;

[0053] Iterate through all frames in the historical images of the flat panel display module to dynamically adjust the brightness of the historical images.

[0054] Preferably, the historical image of the flat panel display module is divided into multiple sub-blocks, the grayscale mean and variance are calculated for each sub-block, the gain coefficient is dynamically adjusted, texture details are enhanced and the contrast of the area is enriched, while noise in flat areas is suppressed, including:

[0055] Select any image from the historical images of the flat panel display module as the image to be adjusted;

[0056] Divide the image to be adjusted into several first sub-blocks on an equal basis;

[0057] Obtain the grayscale value of each pixel in each first sub-block; arbitrarily select one first sub-block as the target sub-block;

[0058] Calculate the grayscale difference between any two pixels in the target sub-block to obtain the first difference; use the square of the first difference as the grayscale fluctuation value to obtain several grayscale fluctuation values.

[0059] Based on the maximum and minimum grayscale fluctuation values ​​among several grayscale fluctuation values, the region scaling factor of the target sub-block is determined;

[0060] The side length of the target sub-block is adjusted based on the region scaling factor to obtain the second sub-block;

[0061] Obtain the average grayscale value and fluctuation coefficient of the pixels in the second sub-block;

[0062] The initial gain coefficient of the corrected second sub-block is calculated based on the gray-scale mean and fluctuation coefficient.

[0063] Calculate the gradient direction entropy and LBP mode ratio of pixels in the second sub-block, and construct the texture complexity index;

[0064] The initial gain coefficient of the second sub-block is corrected based on the texture complexity index, and the corrected gain coefficient of the second sub-block is determined.

[0065] The second sub-block is cropped diagonally to determine two sets of diagonal sub-regions, and the gray-scale distribution difference between the diagonal sub-regions is calculated.

[0066] The gain coefficient of the second sub-block is further modified based on the difference in grayscale distribution to determine the target gain coefficient of the second sub-block.

[0067] The second sub-block is enhanced based on the target gain coefficient;

[0068] Iterate through all the first sub-blocks to complete the enhancement of the image to be adjusted;

[0069] Traverse the historical images of the flat panel display module and complete the enhancement of the historical images of the flat panel display module.

[0070] Compared with the prior art, the beneficial effects of the present invention are:

[0071] This invention acquires the original image of a tablet display module to be dynamically enhanced using a camera module. It collects historical images of the tablet display module and performs dynamic image enhancement processing on these historical images. Based on deep learning and combined with the historical images of the tablet display module after dynamic image enhancement, it constructs an image dynamic enhancement model. Based on the image dynamic enhancement model, it performs dynamic image enhancement processing on the original image of the tablet display module to determine the enhanced image of the tablet display module. This invention can dynamically enhance the tablet display image, improving the tablet display effect and user experience. Attached Figure Description

[0072] Figure 1 This is a flowchart of the image dynamic enhancement processing method for the flat panel display module of the present invention. Detailed Implementation

[0073] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] To address the issue that existing flat panel display modules cannot dynamically enhance the displayed image, resulting in poor display quality and a reduced user experience, please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution:

[0075] Image dynamic enhancement processing methods for flat panel display modules include:

[0076] The original image of the tablet display module to be dynamically enhanced is obtained from the camera module.

[0077] An image dynamic enhancement model is constructed based on deep learning. The original image of the flat panel display module is then dynamically enhanced based on the image dynamic enhancement model to determine the enhanced image of the flat panel display module.

[0078] In this embodiment, an image dynamic enhancement model is constructed, and the following operations are performed:

[0079] Collect historical images of the flat panel display module and perform dynamic image enhancement processing on the historical images of the flat panel display module, including grayscale correction, contrast enhancement, and spatial and frequency domain enhancement;

[0080] Specifically, the historical images of the flat panel display module are processed according to Gamma correction to adjust the grayscale distribution of the historical images, correct nonlinear brightness distortion, and improve the nonlinear response characteristics of the flat panel display module. In particular, by fitting the gamma curve in real time, the brightness level of the historical images of the flat panel display module is dynamically adjusted to ensure that the display effect of the historical images of the flat panel display module remains consistent under different lighting conditions.

[0081] Based on the Gamma correction in the anti-Gamma correction and the Gamma correction in the historical image of the flat panel display module, the true grayscale is restored, achieving linear brightness performance and a unified display effect. In the flat panel display module, linear grayscale brightness performance is achieved by adjusting the relationship between grayscale voltage and brightness. By applying a power function transformation to the historical image of the flat panel display module, grayscale distortion in the transmission of the historical image of the flat panel display module is compensated, and loss of low grayscale details is avoided.

[0082] Histogram equalization optimizes the grayscale distribution of historical images of the flat panel display module to a near-normal distribution, ensuring the overall brightness uniformity of historical images of the flat panel display module. For dynamic scenes, local histogram enhancement is used to adjust the grayscale range separately for dark or bright areas, reducing local overexposure or underexposure caused by global processing.

[0083] Among them, the historical image of the flat panel display module is subjected to regional contrast stretching according to adaptive contrast enhancement, which enhances the contrast of the historical image of the flat panel display module, expands the brightness range of the historical image of the flat panel display module, and improves the details in dark areas and the sense of layering in bright areas. Specifically, the historical image of the flat panel display module is divided into multiple sub-blocks, and the gray-level mean and variance are calculated for each sub-block. The gain coefficient is dynamically adjusted to enhance texture details and enrich the contrast of the region, while suppressing noise in flat areas.

[0084] Based on nonlinear grayscale transformation, an S-curve is used to transform historical images of the flat panel display module to expand the grayscale range of interest. Specifically, a steep slope transformation is used for low grayscale areas to enhance details in dark areas, while a gentle slope is used for high grayscale areas to avoid overexposure.

[0085] Among them, the historical image of the flat panel display module is processed by median filtering to remove salt and pepper noise in the historical image of the flat panel display module while retaining edge information, thereby achieving smoothing of the historical image of the flat panel display module.

[0086] The Sobel operator is used to sharpen and enhance the historical images of the flat panel display module, and the edge gradient of the historical images of the flat panel display module is extracted to reduce noise interference while detecting edge features.

[0087] Image filtering processing is performed on the historical images of the flat panel display module based on Butterworth filtering to enhance the high-frequency components of the historical images of the flat panel display module and reduce the ringing effect;

[0088] Image filtering is performed on the historical images of the flat panel display module based on homomorphic filtering to separate the illumination component and reflection component of the historical images of the flat panel display module. The uneven illumination is improved by suppressing low frequencies and enhancing high frequencies.

[0089] The historical images of the flat panel display module after dynamic image enhancement are divided into training set and test set.

[0090] The deep learning model is trained based on the training set, enabling it to autonomously learn image dynamic enhancement behavior from the training set and perform dynamic enhancement processing on the original images of the flat panel display module, thus determining the image dynamic enhancement model.

[0091] The performance of the image dynamic enhancement model is tested and evaluated based on the test set, the model test evaluation results are determined, and the image dynamic enhancement model is adjusted and optimized based on the model test evaluation results to determine the optimal image dynamic enhancement model.

[0092] In this embodiment, the performance of the image dynamic enhancement model is tested and evaluated based on the test set, and the following operations are performed:

[0093] The test set is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement performance of the model is tested and evaluated based on the test set. The model test evaluation index data is determined, and the model test evaluation result is determined based on the model test evaluation index data.

[0094] Among them, the model test evaluation index data is compared with the preset model test evaluation index thresholds, the matching degree between the model test evaluation index data and the model test evaluation index thresholds is analyzed, and the model test evaluation result is judged based on the matching degree between the model test evaluation index data and the model test evaluation index thresholds.

[0095] When the model test evaluation index data matches the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement model has good image dynamic enhancement performance, that is, it can perform dynamic enhancement processing on the original image of the flat panel display module.

[0096] When the model test evaluation index data does not match the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement performance of the image dynamic enhancement model is poor, that is, it cannot perform dynamic enhancement processing on the original image of the flat panel display module.

[0097] Specifically, the performance of the image dynamic enhancement model was tested and evaluated based on the test set. The model test evaluation results are shown in Table 1.

[0098] Table 1: Model Test Evaluation Results

[0099]

[0100] Therefore, by testing and evaluating the performance of the image dynamic enhancement model using the test set, the model test evaluation results can be determined, which facilitates the subsequent determination of the optimal image dynamic enhancement model.

[0101] In this embodiment, the image dynamic enhancement model is adjusted and optimized based on the model test and evaluation results, and the following operations are performed:

[0102] When the image dynamic enhancement model has poor image dynamic enhancement performance and cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are continuously adjusted, and the image dynamic enhancement model after parameter adjustment is iteratively optimized.

[0103] The performance of the adjusted and optimized image dynamic enhancement model was tested and evaluated again to determine whether the adjusted and optimized image dynamic enhancement model could perform dynamic enhancement processing on the original image of the flat panel display module.

[0104] When the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, the optimal image dynamic enhancement model is directly determined.

[0105] If the adjusted and optimized image dynamic enhancement model cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are adjusted and optimized again until the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, thus determining the optimal image dynamic enhancement model.

[0106] Specifically, the image dynamic enhancement model was adjusted and optimized based on the model testing and evaluation results. The adjustment and optimization details of the image dynamic enhancement model are shown in Table 2.

[0107] Table 2: Adjustment and optimization of the image dynamic enhancement model

[0108]

[0109]

[0110] Therefore, by adjusting and optimizing the image dynamic enhancement model based on the model test and evaluation results, the optimal image dynamic enhancement model can be determined. This allows for dynamic image enhancement processing of the original image of the flat panel display module based on the optimal model, thereby improving the display effect and user experience of the flat panel display.

[0111] In this embodiment, the original image of the flat panel display module is subjected to dynamic image enhancement processing based on the image dynamic enhancement model, and the following operations are performed:

[0112] Deploy the image dynamic enhancement model by placing it in the image dynamic enhancement processing environment of the flat panel display module;

[0113] The original image of the flat panel display module is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement model analyzes the original image of the flat panel display module and automatically performs image dynamic enhancement processing on the original image of the flat panel display module, thereby determining the enhanced image of the flat panel display module.

[0114] In summary, an image dynamic enhancement model is constructed based on deep learning and combined with historical images of the flat panel display module after dynamic image enhancement processing. Based on the image dynamic enhancement model, the original images of the flat panel display module are dynamically enhanced, thereby improving the display effect and user experience of the flat panel.

[0115] In this embodiment, the brightness level of historical images of the flat panel display module is dynamically adjusted by fitting the gamma curve in real time, including:

[0116] Obtain any frame from the historical images of the flat panel display module as the image to be processed;

[0117] Based on the SLIC superpixel algorithm, the image to be processed is clustered into several superpixel regions, and each superpixel region is used as a basic processing unit.

[0118] Calculate the luminance variance of pixels within each superpixel region;

[0119] The target neighborhood range of each pixel within each superpixel region is determined based on the brightness variance; the local brightness mean of each pixel within the superpixel region is calculated based on the target neighborhood range.

[0120] Obtain the global brightness difference index between the image to be processed and the previous frame image;

[0121] The number of target consecutive frames of the image to be processed is determined based on the global brightness difference index.

[0122] Based on the number of target consecutive frames, the historical local brightness average value of each target consecutive frame corresponding to each superpixel region is obtained, and a historical local brightness average value sequence is constructed.

[0123] For each superpixel region in the image to be processed, the historical average brightness of all pixels within the superpixel region is aggregated to generate a region-level time-brightness sequence.

[0124] Construct historical sample groups and current frame sample groups for each superpixel region based on the historical local brightness mean sequence and the region-level time-brightness sequence of each superpixel region.

[0125] The historical sample group and the current frame sample group of each superpixel region are merged, and the gamma curve is fitted by logarithmic linearization and least squares method to generate the gamma value and brightness correction coefficient of each superpixel region.

[0126] The brightness correction algorithm is determined based on the gamma value and brightness correction coefficient of the superpixel region;

[0127] Superpixel regions are classified based on the Otsu algorithm;

[0128] The brightness value of the superpixel region is dynamically adjusted based on the brightness correction algorithm and the category of the superpixel region;

[0129] Iterate through all frames in the historical images of the flat panel display module to dynamically adjust the brightness of the historical images.

[0130] In this embodiment, the SLIC superpixel algorithm is a commonly used image segmentation algorithm used to generate superpixels. A superpixel is a group of pixels with similar features, including color, texture, or brightness.

[0131] In this embodiment, determining the target neighborhood range of a pixel within each superpixel region based on the brightness variance includes: comparing the brightness variance with a preset brightness variance threshold; if the brightness variance is greater than or equal to the preset brightness variance threshold, it indicates that the superpixel region is rich in detail, and a neighborhood range of 3×3 can be taken; if the brightness variance is less than the preset brightness variance threshold, it indicates that the superpixel region is a smooth region, and a neighborhood range of 7×7 can be taken.

[0132] In this embodiment, obtaining the global brightness difference index between the image to be processed and the previous frame image includes: calculating the difference between the average brightness of all pixels in the image to be processed and the average brightness of all pixels in the previous frame image to obtain a first difference; taking the absolute value of the ratio of the first difference to the average brightness of all pixels in the previous frame image as the global brightness difference index of the image to be processed; comparing the global brightness difference index with a preset difference index threshold, where the threshold can be 10%. That is, if the global brightness difference index is less than 10%, it indicates that the image to be processed and the previous frame image may be a static scene, and the target consecutive frame number is 10 frames; if the global brightness difference index is greater than or equal to 10%, it indicates that the image to be processed and the previous frame image may be a dynamic scene, and the target consecutive frame number is 3 frames.

[0133] In this embodiment, the historical local brightness mean of each target continuous frame corresponding to each superpixel region is obtained based on the number of target continuous frames, and a historical local brightness mean sequence is constructed. For example: Superpixel region: contains 3 pixels (P1, P2, P3); Number of historical frames: (n = 3)(t0, t1, t2); Pixel-level sequence: P1: [80, 82, 85]; P2: [79, 81, 84]; P3: [81, 83, 86]; Region-level sequence calculation: t0: (80 + 79 + 81) / 3 = 80; t1: (82 + 81 + 83) / 3 = 82; t2: (85 + 84 + 86) / 3 = 85; Historical local brightness mean sequence: [80, 82, 85].

[0134] In this embodiment, a historical sample group and a current frame sample group for each superpixel region are constructed based on the historical local brightness mean sequence and the regional time-brightness sequence of each superpixel region. This includes: extracting the local brightness mean of all pixels in the target continuous frame in the historical local brightness mean sequence of each superpixel region; extracting the regional brightness aggregation value of the region in the corresponding frame from the regional time-brightness sequence; and, using pixels as units, forming "historical feature-historical target" sample pairs with the above-mentioned pixel historical brightness mean, regional historical aggregation value and the corrected brightness value of the corresponding historical frame, and aggregating them to form a historical sample group.

[0135] Extract the mean local brightness of all pixels in each superpixel region in the current frame and the aggregated regional brightness value of each superpixel region in the current frame. Using pixels as units, combine the mean current brightness of pixels in each superpixel region, the current aggregated regional brightness value, and the original brightness value of the current frame to form a current feature-corrected target sample pair, and aggregate them to form a sample group for the current frame.

[0136] The historical sample groups and current frame sample groups of each superpixel region are aligned and merged according to the same pixel space dimension and feature dimension.

[0137] In this embodiment, for each superpixel region in the image to be processed, the historical average brightness of all pixels within the superpixel region is aggregated to generate a region-level time-brightness sequence, including: assuming the superpixel region R contains M pixels {P1, P2, ..., P...} M Extract pixel-level data from frame t (t∈[0,n-1]) and obtain the pixel P for each pixel. i The mean local luminance L(P) in frame t i ,t); Calculate the average brightness of the region: Generate the sequence: Iterate through all t to obtain the regional time-luminance sequence [L(R,t0),L(R,t1),...,L(R,t2)]. n-1 )).

[0138] In this embodiment, the historical sample group and the current frame sample group of each superpixel region are merged, and the gamma curve is fitted by logarithmic linearization and least squares method to generate the gamma value and brightness correction coefficient of each superpixel region; for each superpixel region, the gamma curve is fitted by logarithmic linearization and least squares method; 1. Model conversion: Take the natural logarithm of the gamma formula to convert it into a linear model:

[0139] ln(O)=ln(k)+γ·ln(I)

[0140] Let y = ln(0), x = ln(1), a = ln(k), b = γ, then the model simplifies to:

[0141] y = a + b·x

[0142] Sample fusion and parameter solving: merging historical sample pairs and current sample pair (Total n+M samples), solve for a and b using the least squares method:

[0143]

[0144] ( (Sample mean)

[0145] Parameter Restoration and Constraints: Gamma value: γ = b; Gain coefficient: k = e a Constraints: If

[0146] In this embodiment, the brightness correction algorithm is O = k·I γ O represents the corrected brightness value; k represents the gain coefficient; γ represents the gamma value; I represents the input brightness.

[0147] In this embodiment, the Otsu algorithm is an adaptive image thresholding method based on the maximum inter-class variance.

[0148] In this embodiment, the superpixel region is classified based on the Otsu algorithm, specifically into dark regions, bright regions, and intermediate regions; wherein, the dark region: O = min(k·I γ ×1.1, 0.95L max Bright area: O = min(k·I) γ 0.9L max ); Intermediate region: O = k·I γ O represents the corrected brightness value; k represents the gain coefficient; γ represents the gamma value; I represents the input brightness; L max This indicates the maximum safe brightness level.

[0149] The working principle and beneficial effects of the above technical solution are as follows: By performing spatiotemporal dual-dimensional adaptive processing, data-driven fitting, and regional fine-grained correction on historical images of the flat panel display module, a triangular balance of real-time performance, accuracy, and robustness is achieved; the number of superpixels and dynamic frames reduces the computational load, adapting to the real-time processing scenarios of flat panels / monitors; data-driven gamma fitting and regional perceptual correction significantly improve dark details, suppress overexposure in bright areas, and optimize image quality; adaptive neighborhood, number of dynamic frames, and Otsu classification cover diverse scenarios such as static images and dynamic videos, and have wide applicability.

[0150] In this embodiment, the historical image of the flat panel display module is divided into multiple sub-blocks. The grayscale mean and variance are calculated for each sub-block, and the gain coefficient is dynamically adjusted to enhance texture details and enrich regional contrast, while simultaneously suppressing noise in flat areas. This includes:

[0151] Select any image from the historical images of the flat panel display module as the image to be adjusted;

[0152] Divide the image to be adjusted into several first sub-blocks on an equal basis;

[0153] Obtain the grayscale value of each pixel in each first sub-block; arbitrarily select one first sub-block as the target sub-block;

[0154] Calculate the grayscale difference between any two pixels in the target sub-block to obtain the first difference; use the square of the first difference as the grayscale fluctuation value to obtain several grayscale fluctuation values.

[0155] Based on the maximum and minimum grayscale fluctuation values ​​among several grayscale fluctuation values, the region scaling factor of the target sub-block is determined;

[0156] The side length of the target sub-block is adjusted based on the region scaling factor to obtain the second sub-block;

[0157] Obtain the average grayscale value and fluctuation coefficient of the pixels in the second sub-block;

[0158] The initial gain coefficient of the corrected second sub-block is calculated based on the gray-scale mean and fluctuation coefficient.

[0159] Calculate the gradient direction entropy and LBP mode ratio of pixels in the second sub-block, and construct the texture complexity index;

[0160] The initial gain coefficient of the second sub-block is corrected based on the texture complexity index, and the corrected gain coefficient of the second sub-block is determined.

[0161] The second sub-block is cropped diagonally to determine two sets of diagonal sub-regions, and the gray-scale distribution difference between the diagonal sub-regions is calculated.

[0162] The gain coefficient of the second sub-block is further modified based on the difference in grayscale distribution to determine the target gain coefficient of the second sub-block.

[0163] The second sub-block is enhanced based on the target gain coefficient;

[0164] Iterate through all the first sub-blocks to complete the enhancement of the image to be adjusted;

[0165] Traverse the historical images of the flat panel display module and complete the enhancement of the historical images of the flat panel display module.

[0166] In this embodiment, Where P represents the region scaling factor of the target sub-block; δ represents the grayscale fluctuation value in the target sub-block; max(δ) represents the maximum grayscale fluctuation value in the target sub-block; and min(δ) represents the minimum grayscale fluctuation value in the target sub-block.

[0167] In this embodiment, the fluctuation coefficient is the ratio of the standard deviation of the grayscale values ​​of the pixels in the second sub-block to the mean grayscale value.

[0168] In this embodiment, the initial gain coefficient includes:

[0169]

[0170] Where G1 represents the initial gain coefficient of the second sub-block; λ represents the reference coefficient; C v represents the fluctuation coefficient of the second sub-block; μ represents the average gray value of the second sub-block.

[0171] In this embodiment, the gradient direction entropy and LBP mode ratio of pixels in the second sub-block are calculated to construct a texture complexity index, including:

[0172] Gradient direction entropy: For each pixel (x,y) in the second sub-block, calculate the horizontal gradient G using the Sobel operator. x and vertical gradient G y ,formula:

[0173]

[0174] (* indicates convolution operation, I(x,y) is the pixel gray value within the region.) Calculate the gradient direction θ(x,y) = arctan2(G y G x The gradient direction is divided into 8 intervals (0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°), and the number of pixels n1, n2, ..., n8 in each interval is counted. The directional entropy GDE is calculated using the entropy formula of the probability distribution. Where, p iLet be the percentage of pixels in the i-th direction, and let entropy reflect the disorder of the directional distribution.

[0175] LBP consistent mode proportion: For each pixel in the second sub-block, using a 3×3 neighborhood as a window, calculate the grayscale difference between the center pixel and the neighboring pixels to generate a local binary mode: Among them, g c For the grayscale of the center pixel, g k Let be the gray level of the k-th neighboring pixel, and the result is an integer from 0 to 255, corresponding to 256 LBP modes; z represents the difference between the gray level of the neighboring pixel and the gray level of the center pixel, i.e., g k -g c ; s(z) represents the binarization mapping of grayscale differences; k represents the neighboring pixel number; define a consistent pattern: the number of transitions from "0→1" or "1→0" in the LBP value is ≤2 (e.g., 000000000, 00000001, 1111111110, etc., a total of 58); count the number of pixels with a consistent pattern within the region, N. umi Total number of pixels N total Calculate the percentage:

[0176] GDE (structural dimension) and LBP R The primitive dimension is mapped to a uniform scale (0-1) to construct the texture complexity index.

[0177] In this embodiment, the initial gain coefficient of the second sub-block is corrected based on the texture complexity index to determine the corrected gain coefficient of the second sub-block, including:

[0178] G2 = G1 × (1 + tanh(T))

[0179] Where G2 represents the correction gain coefficient of the second sub-block; tanh(T) represents mapping the texture complexity T (0 to +∞) to the interval (-1, +1);

[0180] In this embodiment, the second sub-block is cropped diagonally to determine two sets of diagonal sub-regions, and the grayscale distribution difference between the diagonal sub-regions is calculated, including:

[0181] For the region R (W×H), perform two sets of diagonal divisions; dividing line: the main diagonal (from the upper left corner (0, 0) to the lower right corner (W - 1, H - 1)). Sub - regions: R1 (upper triangle): satisfying the coordinate relationship y≥x; R2 (lower triangle): satisfying the coordinate relationship y < x; dividing line: the secondary diagonal (from the upper right corner (W - 1, 0) to the lower left corner (0, H - 1)). Sub - regions: R3 (right triangle): satisfying the coordinate relationship y + x≤W + H - 2; R4 (left triangle): satisfying the coordinate relationship y + x>W + H - 2; For each set of diagonal sub - regions (such as R1&R2, R3&R4), divide the gray - scale range (0~255) into K intervals (by default K = 32), count the number of pixels in each interval within the sub - region, and obtain histograms H1, H2 (corresponding to R, R2). Convert the histograms into probability distributions: (∑H1 is the total number of pixels in R1, ensuring ∑P = ∑Q = 1.) The KL divergence measures the "degree of difference" between two distributions, but it is asymmetric (D KL (P||Q)≠D KL (Q||P)); Therefore, use the symmetric KL divergence to eliminate the direction bias: Among them, the formula for the one - way KL divergence is: (∈ - 10 -8 , to avoid the denominator being 0; the logarithm base is 2, and the unit is "bit";) H1(i) represents the number of pixels in the i - th gray - scale interval of the diagonal sub - region R1; H2(i) represents the number of pixels in the i - th gray - scale interval of the diagonal sub - region R2; (∑H2 is the total number of pixels in R2; represents the gray - scale probability distribution of R1; represents the gray - scale probability distribution of R2;

[0182] Use the KL divergence D as the evaluation index for the gray - scale distribution difference degree.

[0183] In this embodiment, further correct the correction gain coefficient of the second sub - block based on the gray - scale distribution difference degree, and determine the target gain coefficient of the second sub - block;

[0184] G = G2×(0.5 + D / 2)

[0185] Among them, G represents the target gain coefficient of the second sub - block; D represents the gray - scale distribution difference degree corresponding to the second sub - block.

[0186] The working principle and beneficial effects of the above technical solution are as follows: By dynamically adjusting the gain coefficient by calculating the gray-level mean, variance, and other parameters of each sub-block, the texture details in the image can be enhanced in a targeted manner; for different sub-blocks, the gain coefficient is determined according to their own gray-level characteristics, so that the contrast of different regions in the image is reasonably adjusted; by cropping the second sub-block diagonally and calculating the gray-level distribution difference between the diagonal sub-regions to further correct the gain coefficient, it helps to better adjust the local contrast and make the overall image more layered; in flat areas, since their gray-level fluctuation values ​​are small, noise in the area can be suppressed by calculating the region scaling factor and other operations; because the gray-level values ​​of pixels in flat areas are relatively close, this targeted processing can avoid amplifying noise when enhancing the image, thereby improving the purity of the image.

[0187] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0188] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic image enhancement processing of a flat panel display module, characterized in that, include: The original image of the tablet display module to be dynamically enhanced is obtained from the camera module. An image dynamic enhancement model is constructed based on deep learning. Based on the image dynamic enhancement model, the original image of the flat panel display module is subjected to image dynamic enhancement processing to determine the enhanced image of the flat panel display module. This includes collecting historical images of flat panel display modules and performing dynamic image enhancement processing on these historical images; The historical image of the flat panel display module is divided into multiple sub-blocks. The grayscale mean and variance of each sub-block are calculated, and the gain coefficient is dynamically adjusted to enhance texture details and enrich the contrast of the area, while suppressing noise in flat areas. The initial gain coefficient is calculated based on the gray-scale mean and fluctuation coefficient, and the initial gain coefficient is corrected based on the texture complexity index to determine the corrected gain coefficient. The modified gain coefficient is further adjusted based on the difference in gray-level distribution to determine the target gain coefficient, and then enhancement is performed based on the target gain coefficient. The historical images of the flat panel display module are divided into multiple sub-blocks. The grayscale mean and variance are calculated for each sub-block, and the gain coefficient is dynamically adjusted to enhance texture details and enrich regional contrast, while suppressing noise in flat areas. This includes: Select any image from the historical images of the flat panel display module as the image to be adjusted; Divide the image to be adjusted into several first sub-blocks on an equal basis; Obtain the grayscale value of each pixel in each first sub-block; arbitrarily select one first sub-block as the target sub-block; Calculate the grayscale difference between any two pixels in the target sub-block to obtain the first difference; use the square of the first difference as the grayscale fluctuation value to obtain several grayscale fluctuation values. Based on the maximum and minimum grayscale fluctuation values ​​among several grayscale fluctuation values, the region scaling factor of the target sub-block is determined; The side length of the target sub-block is adjusted based on the region scaling factor to obtain the second sub-block; Obtain the average grayscale value and fluctuation coefficient of the pixels in the second sub-block; The fluctuation coefficient is the ratio of the standard deviation of the gray values ​​of pixels in the second sub-block to the mean gray value; The initial gain coefficient of the corrected second sub-block is calculated based on the gray-scale mean and fluctuation coefficient. Calculate the gradient direction entropy and the proportion of LBP consistent modes of pixels in the second sub-block, and construct the texture complexity index; Map the gradient direction entropy and the proportion of consistent LBP modes to a uniform scale (0~1) to construct a texture complexity index; The initial gain coefficient of the second sub-block is corrected based on the texture complexity index to determine the corrected gain coefficient of the second sub-block, including: in, This represents the initial gain coefficient of the second sub-block. This represents the correction gain coefficient for the second sub-block; express The second sub-block is cropped diagonally to determine two sets of diagonal sub-regions. The gray-scale distribution difference between the diagonal sub-regions is calculated, and the KL divergence D is used as the evaluation index of the gray-scale distribution difference. The gain coefficient of the second sub-block is further modified based on the difference in grayscale distribution to determine the target gain coefficient of the second sub-block. The second sub-block is enhanced based on the target gain coefficient; Iterate through all the first sub-blocks to complete the enhancement of the image to be adjusted; Traverse the historical images of the flat panel display module and complete the enhancement of the historical images of the flat panel display module; To build a dynamic image enhancement model, perform the following operations: Collect historical images of the flat panel display module and perform dynamic image enhancement processing on the historical images of the flat panel display module, including grayscale correction, contrast enhancement, and spatial and frequency domain enhancement; The historical images of the flat panel display module after dynamic image enhancement are divided into training set and test set. The deep learning model is trained based on the training set, enabling it to autonomously learn image dynamic enhancement behavior from the training set and perform dynamic enhancement processing on the original images of the flat panel display module, thus determining the image dynamic enhancement model. The performance of the image dynamic enhancement model is tested and evaluated based on the test set, the model test evaluation results are determined, and the image dynamic enhancement model is adjusted and optimized based on the model test evaluation results to determine the optimal image dynamic enhancement model.

2. The image dynamic enhancement processing method for a flat panel display module according to claim 1, characterized in that, Dynamic enhancement processing of historical images from flat panel display modules, including: The historical images of the flat panel display module are processed according to Gamma correction to adjust the grayscale distribution of the historical images, correct nonlinear brightness distortion, and improve the nonlinear response characteristics of the flat panel display module. In particular, by fitting the gamma curve in real time, the brightness level of the historical images of the flat panel display module is dynamically adjusted to ensure that the display effect of the historical images of the flat panel display module remains consistent under different lighting conditions. Based on the Gamma correction in the anti-Gamma correction and the Gamma correction in the historical image of the flat panel display module, the true grayscale is restored, achieving linear brightness performance and a unified display effect. In the flat panel display module, linear grayscale brightness performance is achieved by adjusting the relationship between grayscale voltage and brightness. By applying a power function transformation to the historical image of the flat panel display module, grayscale distortion in the transmission of the historical image of the flat panel display module is compensated, and loss of low grayscale details is avoided. Histogram equalization optimizes the grayscale distribution of historical images of the flat panel display module to a near-normal distribution, ensuring the overall brightness uniformity of the historical images of the flat panel display module. For dynamic scenes, local histogram enhancement is used to adjust the grayscale range separately for dark or bright areas, reducing local overexposure or underexposure caused by global processing.

3. The image dynamic enhancement processing method for a flat panel display module according to claim 2, characterized in that, By dynamically adjusting the brightness level of historical images from the flat panel display module in real time through gamma curve fitting, including: Obtain any frame from the historical images of the flat panel display module as the image to be processed; Based on the SLIC superpixel algorithm, the image to be processed is clustered into several superpixel regions, and each superpixel region is used as a basic processing unit. Calculate the luminance variance of pixels within each superpixel region; The target neighborhood range of each pixel within each superpixel region is determined by comparing the brightness variance with a preset brightness variance threshold; the local average brightness of each pixel within the superpixel region is calculated based on the target neighborhood range. Obtaining the global brightness difference index between the image to be processed and the previous frame image includes: calculating the difference between the average brightness of all pixels in the image to be processed and the average brightness of all pixels in the previous frame image to obtain a first difference; and taking the absolute value of the ratio of the first difference to the average brightness of all pixels in the previous frame image as the global brightness difference index of the image to be processed. The number of target consecutive frames of the image to be processed is determined based on the global brightness difference index. Based on the number of target consecutive frames, the historical local brightness average value of each target consecutive frame corresponding to each superpixel region is obtained, and a historical local brightness average value sequence is constructed. For each superpixel region in the image to be processed, the historical local brightness average of all pixels within the superpixel region is aggregated to generate a region-level time-brightness sequence. Construct historical sample groups and current frame sample groups for each superpixel region based on the historical local brightness mean sequence and the region-level time-brightness sequence of each superpixel region. The historical sample group and the current frame sample group of each superpixel region are merged, and the gamma curve is fitted by logarithmic linearization and least squares method to generate the gamma value and brightness correction coefficient of each superpixel region. The brightness correction algorithm is determined based on the gamma value and brightness correction coefficient of the superpixel region; Superpixel regions are classified based on the Otsu algorithm; The brightness value of the superpixel region is dynamically adjusted based on the brightness correction algorithm and the category of the superpixel region; Iterate through all frames in the historical images of the flat panel display module to dynamically adjust the brightness of the historical images.

4. The image dynamic enhancement processing method for a flat panel display module according to claim 3, characterized in that, Based on the image dynamic enhancement model, the original image of the flat panel display module is dynamically enhanced by performing the following operations: Deploy the image dynamic enhancement model by placing it in the image dynamic enhancement processing environment of the flat panel display module; The original image of the flat panel display module is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement model analyzes the original image of the flat panel display module and automatically performs image dynamic enhancement processing on the original image of the flat panel display module, thereby determining the enhanced image of the flat panel display module.

5. The image dynamic enhancement processing method for a flat panel display module according to claim 4, characterized in that, Dynamic enhancement processing of historical images from flat panel display modules also includes: Adaptive contrast enhancement is used to perform regional contrast stretching on the historical image of the flat panel display module, which enhances the contrast of the historical image of the flat panel display module, expands the brightness range of the historical image of the flat panel display module, and improves the details in dark areas and the sense of layering in bright areas. Specifically, the historical image of the flat panel display module is divided into multiple sub-blocks, and the gray-level mean and variance are calculated for each sub-block. The gain coefficient is dynamically adjusted to enhance texture details and enrich the contrast of the region, while suppressing noise in flat areas. Based on nonlinear grayscale transformation, an S-curve is used to transform historical images of the flat panel display module to expand the grayscale range of interest. Specifically, a steep slope transformation is used for low grayscale areas to enhance details in dark areas, while a gentle slope is used for high grayscale areas to avoid overexposure.

6. The image dynamic enhancement processing method for a flat panel display module according to claim 5, characterized in that, Dynamic enhancement processing of historical images from flat panel display modules also includes: The historical images of the flat panel display module are processed by median filtering to remove salt-and-pepper noise while retaining edge information, thus achieving smoothing of the historical images of the flat panel display module. The Sobel operator is used to sharpen and enhance the historical images of the flat panel display module, and the edge gradient of the historical images of the flat panel display module is extracted to reduce noise interference while detecting edge features. Image filtering is performed on the historical images of the flat panel display module based on Butterworth filtering to enhance the high-frequency components of the historical images and reduce the ringing effect. Image filtering is performed on the historical images of the flat panel display module based on homomorphic filtering to separate the illumination component and reflection component of the historical images of the flat panel display module. The uneven illumination is improved by suppressing low frequencies and enhancing high frequencies.

7. The image dynamic enhancement processing method for a flat panel display module according to claim 6, characterized in that, The performance of the image dynamic enhancement model is evaluated based on the test set, and the following operations are performed: The test set is used as input data and fed into the image dynamic enhancement model. The image dynamic enhancement performance of the model is tested and evaluated based on the test set. The model test evaluation index data is determined, and the model test evaluation result is determined based on the model test evaluation index data. Among them, the model test evaluation index data is compared with the preset model test evaluation index thresholds, the matching degree between the model test evaluation index data and the model test evaluation index thresholds is analyzed, and the model test evaluation result is judged based on the matching degree between the model test evaluation index data and the model test evaluation index thresholds. When the model test evaluation index data matches the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement model has good image dynamic enhancement performance, that is, it can perform dynamic enhancement processing on the original image of the flat panel display module. When the model test evaluation index data does not match the model test evaluation index threshold, the model test evaluation result is that the image dynamic enhancement performance of the image dynamic enhancement model is poor, that is, it cannot perform dynamic enhancement processing on the original image of the flat panel display module.

8. The image dynamic enhancement processing method for a flat panel display module according to claim 7, characterized in that, Based on the model testing and evaluation results, the image dynamic enhancement model is adjusted and optimized, and the following operations are performed: When the image dynamic enhancement model has poor image dynamic enhancement performance and cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are continuously adjusted, and the image dynamic enhancement model after parameter adjustment is iteratively optimized. The performance of the adjusted and optimized image dynamic enhancement model was tested and evaluated again to determine whether the adjusted and optimized image dynamic enhancement model could perform dynamic enhancement processing on the original image of the flat panel display module. When the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, the optimal image dynamic enhancement model is directly determined. If the adjusted and optimized image dynamic enhancement model cannot perform dynamic enhancement processing on the original image of the flat panel display module, the parameters of the image dynamic enhancement model are adjusted and optimized again until the adjusted and optimized image dynamic enhancement model can perform dynamic enhancement processing on the original image of the flat panel display module, thus determining the optimal image dynamic enhancement model.

Images