Adaptive microstructure optical film display screen and dynamic optimization method thereof

Abstract

The present invention discloses an adaptive microstructure optical film display screen and a dynamic optimization method thereof, which relates to the technical field of optical film display screens, and includes the following steps: obtaining data of a current frame image through a video stream, extracting features of the current frame image, and calculating the complexity of the current frame image based on the extracted image features; comparing the current frame with the previous frame image, obtaining dynamic change information between the two, extracting features of the obtained dynamic change information, and identifying the dynamic changes of two adjacent frames of images. The present invention obtains current frame image data through a video stream, extracts features such as detail density, calculates complexity and obtains dynamic change information, especially texture changes. A machine learning algorithm is used to intelligently evaluate complex image changes, predict the frequency of dynamic adjustment of color saturation, monitor and compare actual frequency and predicted frequency in real time, adjust system frequency, ensure color accuracy, reduce visual fatigue, and improve user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of optical film display screens, and in particular to an adaptive microstructure optical film display screen and a dynamic optimization method thereof. Background Art

[0002] Adaptive microstructure optical film display is a display technology that uses micron-level structures to optimize light propagation. It can effectively control and guide light by embedding tiny optical elements with specific structures and arrangements on the surface or inside the display, thereby improving the brightness, contrast and energy efficiency of the display. The core of this technology is to achieve precise control of the light propagation path through the design and adjustment of the microstructure, reduce light scattering and reflection losses, and enable the display to present clear and vivid images under different ambient light conditions. In addition, this microstructure optical film can also adapt to different display needs. By adjusting the microstructure parameters, different display effects can be achieved, such as anti-glare, brightening, and color enhancement, thereby enhancing the user's visual experience.

[0003] Dynamic optimization means that the adaptive microstructure optical film display can adjust and optimize itself according to the real-time external environment and usage conditions. For example, when the ambient light changes, the display can detect the intensity and direction of the ambient light through the built-in light sensor, and adjust the optical properties of the microstructure optical film in real time to maintain the best display effect. This dynamic optimization function not only improves the adaptability and intelligence of the display, but also effectively reduces energy consumption and extends the service life of the equipment. In the future, the dynamic optimization technology of the adaptive microstructure optical film display may be combined with artificial intelligence and Internet of Things technologies to further enhance its intelligence level and application scope, bringing a more outstanding visual experience and convenience of use.

[0004] The adaptive microstructure optical film display monitors the display content in real time through a built-in microprocessor and dynamically adjusts the color saturation of the microstructure optical film. The purpose is to optimize the display effect and enhance the user's visual experience. Specifically, different display content has different requirements for color saturation. For example, when watching videos or pictures, high saturation can enhance the vividness and visual impact of the color, while when reading text or processing documents, lower saturation is required to avoid visual fatigue. By dynamically adjusting the color saturation, the display can provide appropriate color performance in various usage scenarios, making the image more vivid and lifelike, and the text clearer and easier to read, thereby improving the overall viewing experience.

[0005] The prior art has the following deficiencies:

[0006] When watching videos, high saturation can enhance the vividness and visual impact of colors. However, when an abnormality occurs during the dynamic adjustment of color saturation, resulting in insufficient dynamic adjustment frequency of color saturation and failure to meet the high saturation requirements of the video, the display will be unable to adjust the color saturation in time, resulting in inaccurate presentation of colors in the video, color shift or distortion, and the picture losing its realism and appeal. Inaccurate color saturation will make the colors too dim or too glaring, causing visual fatigue when watching videos for a long time, reducing user comfort and experience satisfaction.

[0007] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0008] The purpose of the present invention is to provide an adaptive microstructure optical film display screen and a dynamic optimization method thereof, which obtains current frame image data through a video stream, extracts features such as detail density, calculates image complexity, obtains dynamic change information, especially texture changes, by comparing the current frame with the previous frame, integrates picture content and dynamic information, uses a machine learning algorithm to intelligently evaluate complex image changes, predicts the dynamic adjustment frequency of color saturation based on the evaluation results, and monitors the actual adjustment frequency in real time, compares and analyzes the predicted frequency with the actual frequency, calculates errors and adjusts the system frequency to ensure accurate color presentation, prevent colors from being too dark or glaring, reduce visual fatigue, and improve user experience, so as to solve the problems in the above-mentioned background technology.

[0009] In order to achieve the above object, the present invention provides the following technical solution: a method for dynamic optimization of an adaptive microstructure optical film display screen, comprising the following steps:

[0010] The data of the current frame image is obtained through the video stream, features of the current frame image are extracted, and the complexity of the current frame image is calculated based on the extracted image features;

[0011] Compare the current frame with the previous frame image, obtain the dynamic change information between the two, perform feature extraction on the obtained dynamic change information, and identify the dynamic change of two adjacent frames of images;

[0012] Comprehensive picture content information and dynamic change information, use machine learning algorithms to analyze the complexity and change trend of the current frame image, and perform intelligent evaluation of the complex changes of the current frame image;

[0013] According to the result of intelligent evaluation, the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained;

[0014] Real-time monitoring of the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment of color saturation;

[0015] The predicted dynamic adjustment frequency is compared and analyzed with the actual dynamic adjustment frequency to identify the difference between the two. The adjustment accuracy of the adjustment system is determined by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made; if the error exceeds the allowable range, the actual dynamic adjustment frequency is adjusted until the error is within the allowable range.

[0016] Preferably, feature extraction is performed on the current frame image, and the extracted features include detail density in the image. The detail density of the image refers to the density of detail information contained in the image, which reflects the proportion of high-frequency components in the image. After acquisition, the detail density in the image is analyzed and processed to generate a detail density index. The steps for generating the detail density index are as follows:

[0017] Preprocess the acquired image, including grayscale and denoising, to reduce computational complexity and noise interference;

[0018] Use edge detection algorithms to extract edges in images;

[0019] The high-frequency components of the image are extracted using wavelet transform, including horizontal detail components, vertical detail components and diagonal detail components, and the horizontal detail components, vertical detail components and diagonal detail components are calibrated as cH, cV and cD respectively;

[0020] Calculate the energy of the high-frequency component as a quantitative indicator of detail. The calculation expression is:

[0021] E H =∑ i，j H i，j 2 , E V =∑ i，j cV i，j 2 , E D =∑ i，j cD i，j 2 , i and j represent the row index and column index in the image matrix respectively;

[0022] Calculate the edge density, the expression is: Where D E is the edge density, which indicates the proportion of edge pixels in the image, E i，j is the value of the i, jth pixel in the edge image, and N is the total number of pixels in the image;

[0023] Calculate the detail density index. The calculation expression is:

[0024] In the formula, α represents the weight parameter of high-frequency energy, β represents the weight parameter of edge density, and DDI Index Represents the detail density index.

[0025] Preferably, the specific steps of using the Canny edge detection algorithm to extract edges in an image are as follows:

[0026] Apply a Gaussian filter to smooth the grayscale image to reduce noise and small irrelevant details;

[0027] Calculate the gradient strength and direction of each pixel in the image;

[0028] Perform non-maximum suppression on the gradient image, retain the gradient of the local maximum value, and suppress other gradient values;

[0029] Two thresholds are applied to distinguish strong edges from weak edges, a high threshold is used to mark strong edges, and a low threshold is used to mark possible edges;

[0030] Through the edge connection process, all weak edges are connected to strong edges. If a weak edge pixel is connected to a strong edge pixel, it is retained as an edge, otherwise it is suppressed. By traversing the image, the weak edge pixels are associated with the strong edge pixels to ensure the continuity of the edge.

[0031] Output a binary image, where a pixel value of 1 represents an edge and a pixel value of 0 represents a non-edge.

[0032] Preferably, the steps of using Haar wavelet transform to extract high-frequency components of an image, including horizontal detail components, vertical detail components, and diagonal detail components, are as follows:

[0033] Select Haar wavelet as wavelet filter;

[0034] The Haar wavelet filter is decomposed into a low-pass filter and a high-pass filter. The low-pass filter is used to capture the low-frequency components of the image and reflect the overall structure and large-scale characteristics of the image; the high-pass filter is used to capture the high-frequency components of the image and reflect the details and edge information of the image.

[0035] Perform a low-pass and high-pass filtering operation on each line of the image, obtain the horizontal low-frequency component of the image after low-pass filtering, and obtain the horizontal high-frequency component of the image after high-pass filtering;

[0036] The image after row-direction filtering is subjected to low-pass and high-pass filtering in the column direction, the horizontal low-frequency component is subjected to high-pass filtering to obtain the vertical detail component of the image, the horizontal high-frequency component is subjected to low-pass filtering to obtain the horizontal detail component of the image; the horizontal high-frequency component is subjected to high-pass filtering to obtain the diagonal detail component of the image;

[0037] The low-frequency and high-frequency components obtained in the above steps are combined into sub-band images, namely, horizontal detail components, vertical detail components and diagonal detail components. The horizontal detail component reflects the horizontal detail changes of the image, the vertical detail component reflects the vertical detail changes of the image, and the diagonal detail component reflects the diagonal detail changes of the image.

[0038] Preferably, the current frame is compared with the previous frame image to obtain the dynamic change information between the two, and the specific steps are as follows:

[0039] Get the image data of the current frame and the previous frame through the video stream:

[0040] Before dynamic change analysis, the images of the current frame and the previous frame are preprocessed;

[0041] Use the optical flow algorithm or frame difference method to calculate the dynamic change information between the current frame and the previous frame.

[0042] Preferably, feature extraction is performed on the acquired dynamic change information, and the extracted dynamic change feature includes the texture change of two adjacent frames of images. After the texture change of two adjacent frames of images is analyzed and processed, a dynamic texture change index is generated, and the steps for generating the dynamic texture change index are as follows:

[0043] The texture feature extraction method is used to extract texture features from the preprocessed image, and the texture feature vectors of the current frame image and the previous frame image are respectively expressed as T t and T t-1 To express;

[0044] Calculate the difference in texture features between the current frame and the previous frame, and use the Euclidean distance to measure the difference between feature vectors. The calculation expression is: In the formula,

[0045] ΔT represents the difference in texture features between two adjacent frames of images, n represents the dimension of the feature vector, and k represents the kth feature component in the texture feature vector;

[0046] The local texture change is calculated at each pixel position (x, y) of the image to obtain a local texture change matrix. Each element in the local texture change matrix represents the degree of local texture change at that position. The calculation expression is:

[0047] In the formula,

[0048] ΔT local (x, y) represents the local texture change matrix, where x represents the horizontal coordinate of the pixel and y represents the vertical coordinate of the pixel;

[0049] Calculate the dynamic texture change index. The calculation expression is:

[0050] Where DT_Id represents the dynamic texture change index, M and N are the width and height of the image respectively.

[0051] Preferably, the specific steps of extracting texture features from the preprocessed image using the local binary pattern are as follows:

[0052] Step 1: For each pixel, take the pixel as the center, take the pixels around it as the neighborhood, and use the selected pixels to calculate the local binary pattern;

[0053] Step 2: For the central pixel, compare each pixel in its neighborhood with the central pixel. If the grayscale value of the neighborhood pixel is greater than or equal to the grayscale value of the central pixel, the value of the neighborhood pixel is recorded as 1; otherwise, it is recorded as 0, so that each neighborhood pixel is assigned a binary value.

[0054] Step 3: Arrange the binary values ​​obtained in step 2 in a clockwise or counterclockwise order to form a binary number, and then convert the binary number into a decimal number, which is the LBP value of the central pixel;

[0055] The LBP feature map of the entire image is statistically analyzed, and the frequency of occurrence of each LBP value is calculated to form an LBP histogram. The histogram is used as the texture feature descriptor of the image to reflect the distribution of different patterns in the image.

[0056] Preferably, the detail density index DDI generated by analyzing and processing the current frame image is obtained. Index After analyzing and processing the texture changes of two adjacent frames of images to generate the dynamic texture change index DT_Id, the detail density index DDI Index The dynamic texture variation index DT_Id is input into the pre-trained machine learning model, and the complex variation coefficient Variation is generated based on the model. Coefficient , through the complex variation coefficient Variation Coefficient Intelligent evaluation of complex changes in image content.

[0057] Preferably, the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained, and the specific steps are as follows:

[0058] Get the detail density index DDI generated by analyzing the current frame image Index , the dynamic texture change index DT_Id generated by analyzing the texture changes of two adjacent frames of images, and the complex change coefficient Variation generated based on the machine learning model Coefficient, defines the basic adjustment frequency when the color saturation is dynamically adjusted. The expression defined is:

[0059] Base Frequency =f(DDI Index , DT_Id), where Base Frequency Indicates the basic adjustment frequency, which is determined by the detail density index DDI Index and the function f of the dynamic texture change index DT_Id;

[0060] Based on the basic frequency, the complex variation coefficient is Cogfficient To control the basic adjustment frequency, the expression for control is:

[0061] Adjusted Freqency =Base Frequency ·(1+u·Variation Coefficient ), where

[0062] Adjusted Frequency It represents the actual frequency after the basic adjustment frequency is regulated, and u represents the adjustment coefficient, which determines the amplitude of the complex change coefficient regulating the basic adjustment frequency;

[0063] Calculate the predicted dynamic adjustment frequency, the calculation expression is:

[0064] In the formula,

[0065] Predicted Frequency Indicates the predicted dynamic adjustment frequency, represents the smoothing factor, which is used to adjust the effect of color error on the adjustment frequency. ΔE represents the average color error between the current frame and the previous frame. E max Indicates the maximum permissible color error.

[0066] Adaptive microstructure optical film display screen, including image feature extraction module, dynamic change analysis module, intelligent evaluation module, predicted frequency calculation module, real-time monitoring module and frequency comparison and control module;

[0067] The image feature extraction module obtains the data of the current frame image through the video stream, extracts the features of the current frame image, and calculates the complexity of the current frame image based on the extracted image features;

[0068] The dynamic change analysis module compares the current frame with the previous frame image, obtains the dynamic change information between the two, extracts features from the obtained dynamic change information, and identifies the dynamic change conditions of two adjacent frames of images;

[0069] The intelligent evaluation module integrates the picture content information and dynamic change information, uses machine learning algorithms to analyze the complexity and change trend of the current frame image, and performs intelligent evaluation on the complex changes of the current frame image;

[0070] A predicted frequency calculation module, which obtains the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted according to the result of the intelligent evaluation;

[0071] A real-time monitoring module monitors in real time the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment process of color saturation;

[0072] The frequency comparison and control module compares and analyzes the predicted dynamic adjustment frequency with the actual dynamic adjustment frequency, identifies the difference between the two, and determines the adjustment accuracy of the adjustment system by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made; if the error exceeds the allowable range, the actual dynamic adjustment frequency is adjusted until the error is within the allowable range.

[0073] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0074] The present invention obtains data of a current frame image through a video stream, extracts features including detail density, and calculates the complexity of the current frame image based on these features. Subsequently, by comparing the current frame with the previous frame image, dynamic change information of two adjacent frames of images is obtained, and feature extraction is performed on these dynamic change information, especially texture changes of two adjacent frames of images, and comprehensive picture content information and dynamic change information are used to intelligently evaluate the complexity and change trend of the current frame image using a machine learning algorithm to further evaluate the complex changes of the image picture. Based on the intelligent evaluation result, the frequency of dynamic adjustment of color saturation of the current frame image is predicted, and the actual dynamic adjustment frequency is monitored in real time. By comparing and analyzing the predicted frequency with the actual frequency, the error between the two is calculated, and the adjustment frequency of the system is adjusted according to the error to ensure that the adjustment frequency of color saturation is within an allowable range, thereby maintaining accurate color presentation and realism, preventing colors from being too dim or glaring, reducing visual fatigue when watching videos for a long time, and improving user comfort and experience satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0076] Figure 1The present invention is a flow chart of the method for dynamically optimizing the adaptive microstructure optical film display screen.

[0077] Figure 2 It is a schematic diagram of a module of the adaptive microstructure optical film display screen of the present invention. DETAILED DESCRIPTION

[0078] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0079] The present invention provides Figure 1 The adaptive microstructure optical film display screen dynamic optimization method shown includes the following steps:

[0080] The data of the current frame image is obtained through the video stream, features of the current frame image are extracted, and the complexity of the current frame image is calculated based on the extracted image features;

[0081] Video streams are composed of continuous frames. To obtain these frames, a video decoder can be used to decompose video files or real-time video streams into individual image frames. This step ensures that the content information of the current frame image can be accessed and analyzed by subsequent processing steps; the purpose of feature extraction is to obtain descriptive information of the image, reflect its complexity and content characteristics, and provide basic data for subsequent complexity evaluation and dynamic adjustment; highly complex images usually contain more details and changes, which requires a higher dynamic adjustment frequency to maintain the accuracy of color saturation.

[0082] The feature extraction of the current frame image includes the detail density in the image. The detail density of the image refers to the density of detail information contained in the image, which reflects the proportion of high-frequency components (subtle changes and textures) in the image. The higher the detail density, the more details, textures and edges there are in the image, and it looks more complex visually. After acquisition, the detail density in the image is analyzed and processed to generate a detail density index. The steps for generating the detail density index are as follows:

[0083] Preprocess the acquired image, including grayscale and denoising, to reduce computational complexity and noise interference;

[0084] Use edge detection algorithms (such as Canny) to extract edges in the image;

[0085] The specific steps for extracting edges in an image using the Canny edge detection algorithm are as follows:

[0086] Apply a Gaussian filter to the grayscale image for smoothing to reduce noise and small irrelevant details. The Gaussian filter smoothes the image through a convolution operation, making the edge detection process less sensitive to noise.

[0087] Calculate the gradient strength and direction of each pixel in the image. The Sobel operator is usually used to calculate the horizontal and vertical gradients. Through these two gradient values, the gradient magnitude and gradient direction of each pixel can be determined.

[0088] Perform non-maximum suppression on the gradient image, retain the gradient of the local maximum value, and suppress other gradient values. This step ensures that only true edge points are retained and other non-edge points are suppressed by checking the neighboring pixels in the gradient direction.

[0089] Two thresholds are applied to distinguish strong edges from weak edges. The high threshold is used to mark strong edges, and the low threshold is used to mark possible edges. In this process, all gradients above the high threshold are marked as strong edges, and gradients between the high and low thresholds are marked as weak edges.

[0090] Through the edge connection process, all weak edges are connected to strong edges. If a weak edge pixel is connected to a strong edge pixel, it is retained as an edge, otherwise it is suppressed. This process ensures the continuity of the edge by traversing the image and associating weak edge pixels with strong edge pixels.

[0091] The final output is a binary image, where a pixel value of 1 indicates an edge and a pixel value of 0 indicates a non-edge. This binary image clearly marks the edge locations in the image for use in subsequent image processing steps.

[0092] Use wavelet transform (such as Haar wavelet) to extract high-frequency components of the image, including horizontal detail components, vertical detail components and diagonal detail components, and calibrate the horizontal detail components, vertical detail components and diagonal detail components as cH, cV and cD respectively;

[0093] The steps of using Haar wavelet transform to extract the high-frequency components of an image, including horizontal detail components, vertical detail components, and diagonal detail components, are as follows:

[0094] Select Haar wavelet as the wavelet filter. Haar wavelet is the simplest wavelet filter with filter coefficients of 1 / 2, -1 / 21 / 2, -1 / 21 / 2, -1 / 2, which is suitable for fast calculation.

[0095] The Haar wavelet filter is decomposed into a low-pass filter and a high-pass filter. The low-pass filter is used to capture the low-frequency components of the image, reflecting the overall structure and large-scale characteristics of the image; the high-pass filter is used to capture the high-frequency components of the image, reflecting the details and edge information of the image.

[0096] Perform a low-pass and high-pass filtering operation on each line of the image. After low-pass filtering, the horizontal low-frequency component of the image is obtained, and after high-pass filtering, the horizontal high-frequency component of the image is obtained. The low-frequency component represents the smooth part of the image, and the high-frequency component represents the detail changes of the image in the horizontal direction.

[0097] The image after row filtering is subjected to low-pass and high-pass filtering in the column direction. The horizontal low-frequency component is subjected to high-pass filtering to obtain the vertical detail component of the image. The horizontal high-frequency component is subjected to low-pass filtering to obtain the horizontal detail component of the image; the horizontal high-frequency component is subjected to high-pass filtering to obtain the diagonal detail component of the image.

[0098] The low-frequency and high-frequency components obtained in the above steps are combined into sub-band images, namely, horizontal detail components, vertical detail components and diagonal detail components. The horizontal detail component reflects the horizontal detail changes of the image, the vertical detail component reflects the vertical detail changes of the image, and the diagonal detail component reflects the diagonal detail changes of the image.

[0099] Calculate the energy of the high-frequency component as a quantitative indicator of detail. The calculation expression is:

[0100] E H =∑ i，j H i，j 2 , E V =∑ i，j cV i，j 2 , E D =∑ i，j cD i，j 2 , i and j represent the row index and column index in the image matrix respectively;

[0101] Calculate the edge density, the expression is: Where D E is the edge density, which indicates the proportion of edge pixels in the image, E i，j is the value of the i, jth pixel in the edge image, and N is the total number of pixels in the image;

[0102] Calculate the detail density index. The calculation expression is:

[0103] In the formula, α represents the weight parameter of high-frequency energy, β represents the weight parameter of edge density, and DDI Index Represents the detail density index.

[0104] It can be seen from the detail density index that the larger the value of the detail density index generated after analyzing the detail density in the image, the more complex the image is, and vice versa. Specifically, the detail density index is a comprehensive indicator generated by analyzing the high-frequency components and edge density features of the image. The high-frequency component reflects the details and texture information in the image, while the edge density reflects the edge and contour information in the image. When the detail density index value is large, it means that the image contains more details, textures, and edges, which means that the image complexity is high; when the detail density index value is small, it means that the image contains fewer details and edges, and the image is simpler.

[0105] Compare the current frame with the previous frame image, obtain the dynamic change information between the two, perform feature extraction on the obtained dynamic change information, and identify the dynamic change of two adjacent frames of images;

[0106] Compare the current frame with the previous frame to obtain the dynamic change information between the two. The specific steps are as follows:

[0107] Get the image data of the current frame and the previous frame through the video stream;

[0108] Similarly, obtaining these frames can use a video decoder to decompose the video file or real-time video stream into individual image frames.

[0109] Before dynamic change analysis, the images of the current frame and the previous frame are preprocessed;

[0110] The preprocessing steps may include operations such as image scaling, grayscale, and noise removal. Image scaling can unify the size of the frame, and grayscale converts the color image to a grayscale image to simplify the subsequent processing process. Noise removal is to reduce the interference information in the image and improve the accuracy of change detection.

[0111] Use optical flow algorithm or frame difference method to calculate the dynamic change information between the current frame and the previous frame;

[0112] The optical flow algorithm captures dynamic changes by calculating the motion vector of pixels in the image, while the frame difference method detects moving areas by calculating the pixel differences between two frames. Optical flow algorithms such as Lucas-Kanade or Farneback methods can provide more detailed motion information, while the frame difference method is simple and has low computational complexity, making it suitable for fast change detection.

[0113] The dynamic change information obtained is subjected to feature extraction. The extracted dynamic change features include the texture changes between two adjacent frames of images. The texture changes between two adjacent frames of images refer to the changes in surface details (such as texture patterns, roughness, fine structures, etc.) in the image due to object movement, lighting changes or other dynamic factors between the two frames of images. This change can be manifested as a change in the shape, direction, contrast or distribution of the texture, which is an important indicator for measuring the complexity and dynamics of the image. By detecting and analyzing texture changes, the detailed motion and dynamic features in the scene can be captured, so as to better understand and describe the changes in the image content. After analyzing and processing the texture changes between two adjacent frames of images, a dynamic texture change index is generated. The steps for generating the dynamic texture change index are as follows:

[0114] Use texture feature extraction methods (such as gray level co-occurrence matrix (GLCM), local binary pattern (LBP), etc.) to extract texture features from the preprocessed image, and use T t and T t-1 To express;

[0115] The specific steps of extracting texture features from the preprocessed image using local binary patterns (LBP) are as follows:

[0116] Step 1: For each pixel, take the pixel as the center, take the pixels around it as the neighborhood, and use the selected pixels to calculate the local binary pattern;

[0117] Select the neighborhood size for each pixel. Usually a 3x3 window is used, but a larger window can be selected as needed.

[0118] Step 2: For the central pixel, compare each pixel in its neighborhood with the central pixel. If the grayscale value of the neighborhood pixel is greater than or equal to the grayscale value of the central pixel, the value of the neighborhood pixel is recorded as 1; otherwise, it is recorded as 0, so that each neighborhood pixel is assigned a binary value.

[0119] Step 3: Arrange the binary values ​​obtained in step 2 in a clockwise or counterclockwise order to form a binary number, and then convert the binary number into a decimal number, which is the LBP value of the central pixel;

[0120] In this way, each pixel will get an LBP value to form an LBP feature map.

[0121] The LBP feature map of the entire image is counted, the frequency of occurrence of each LBP value is calculated, and the LBP histogram is formed. The histogram is used as the texture feature descriptor of the image to reflect the distribution of different modes in the image;

[0122] Each bucket of the histogram represents a local pattern, and the bucket value indicates the frequency of the pattern in the image.

[0123] Calculate the difference in texture features between the current frame and the previous frame, and use the Euclidean distance to measure the difference between feature vectors. The calculation expression is: In the formula,

[0124] ΔT represents the difference in texture features between two adjacent frames of images, n represents the dimension of the feature vector, and k represents the kth feature component in the texture feature vector;

[0125] The local texture change is calculated at each pixel position (x, y) of the image to obtain a local texture change matrix. Each element in the local texture change matrix represents the degree of local texture change at that position. The calculation expression is:

[0126] In the formula,

[0127] ΔT local (x, y) represents the local texture change matrix, where x represents the horizontal coordinate of the pixel and y represents the vertical coordinate of the pixel;

[0128] Calculate the dynamic texture change index. The calculation expression is:

[0129] Where DT_Id represents the dynamic texture change index, M and N are the width and height of the image respectively.

[0130] It can be seen from the dynamic texture change index that the larger the performance value of the dynamic texture change index generated after analyzing the texture changes of two adjacent frames of images, the more complex the texture changes of the two adjacent frames of images are, and vice versa, it means that the changes in the images are less complex. Specifically, the dynamic texture change index evaluates the degree of change of the image by measuring the difference in texture features between adjacent frames. When the dynamic texture change index is high, it means that there are a large number of detail changes, motion, and dynamic texture features in the image, reflecting the high complexity of the scene; and when the dynamic texture change index is low, it means that there are fewer changes in the image, the dynamic features are not obvious, and the complexity of the scene is low. This index provides a quantitative indicator to help analyze and understand the dynamic changes in image sequences.

[0131] Comprehensive picture content information and dynamic change information, use machine learning algorithms to analyze the complexity and change trend of the current frame image, and perform intelligent evaluation of the complex changes of the current frame image;

[0132] Get the detail density index DDI generated by analyzing the current frame image IndexAfter analyzing and processing the texture changes of two adjacent frames of images to generate the dynamic texture change index DT_Id, the detail density index DDI Index The dynamic texture variation index DT_Id is input into the pre-trained machine learning model, and the complex variation coefficient Variation is generated based on the model. Coefficient , through the complex variation coefficient Variation Coefficient Intelligent evaluation of complex changes in image content.

[0133] The machine learning model is not specifically limited here, and can achieve the detail density index DDI Index And dynamic texture change index DT_Id, and then conduct comprehensive analysis to generate complex change coefficient Variation Coefficient In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0134] Complex variation coefficient Variation Coefficient The resulting calculation formula is:

[0135] Where k1 and k2 are detail density index DDI respectively. Index And the preset proportional coefficient of the dynamic texture change index DT_Id, and k1 and k2 are both greater than 0.

[0136] It can be seen from the complex variation coefficient that the greater the expression value of the detail density index generated after analyzing and processing the detail density in the current frame image, the greater the expression value of the dynamic texture variation index generated after analyzing and processing the texture changes of two adjacent frame images, that is, the greater the expression value of the generated complex variation coefficient, the greater the complex change of the current frame image, and vice versa.

[0137] According to the result of intelligent evaluation, the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained;

[0138] The predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained. The specific steps are as follows:

[0139] Get the detail density index DDI generated by analyzing the current frame image Index , the dynamic texture change index DT_Id generated by analyzing the texture changes of two adjacent frames of images, and the complex change coefficient Variation generated based on the machine learning model Coefficient , defines the basic adjustment frequency when the color saturation is dynamically adjusted. The expression defined is:

[0140] BaseFrequecy =f(DDI Index , Df_Id), where Base Frequency Indicates the basic adjustment frequency, which is determined by the detail density index DDI Index and the function f of the dynamic texture change index DT_Id;

[0141] Function f can be a linear combination, weighted summation or other complex functions. It is not specifically limited here and can be set according to actual needs to reflect the detail density index DDI Index and the impact of the dynamic texture change index DT_Id on the basic adjustment frequency;

[0142] Based on the basic frequency, the complex variation coefficient is Coefficient To control the basic adjustment frequency, the expression for control is:

[0143] Adjusted Frequency =Base Frequency ·(1+u·Variation Coefficient ), where

[0144] Adjusted Frequency It represents the actual frequency after the basic adjustment frequency is regulated, and u represents the adjustment coefficient, which determines the amplitude of the complex change coefficient regulating the basic adjustment frequency;

[0145] Calculate the predicted dynamic adjustment frequency, the calculation expression is:

[0146] In the formula,

[0147] Predicted Frequency Indicates the predicted dynamic adjustment frequency, represents the smoothing factor, which is used to adjust the effect of color error on the adjustment frequency. ΔE represents the average color error between the current frame and the previous frame. E max Indicates the maximum permissible color error;

[0148] This part ensures that the frequency of color saturation adjustment can respond to specific color change requirements in real time;

[0149] The average color error refers to the average value of the color difference between the current frame image and the previous frame image. Specifically, this value is calculated by comparing the color difference of each pixel in two adjacent frames, and then averaging the difference values ​​of all pixels. This error reflects the degree of color change of the image in the time dimension. A higher average color error indicates that the color of the image has changed significantly, and more frequent color adjustments may be required to maintain image quality.

[0150] The maximum allowed color error is a predefined threshold used to determine the tolerance range of image color changes. This value represents the visually acceptable color error range, and color changes beyond this range may affect the user's viewing experience. Setting the maximum allowed color error can help balance performance and quality during color adjustment, ensuring that processing speed and resource usage are optimized without significantly affecting image quality.

[0151] Real-time monitoring of the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment of color saturation;

[0152] Real-time monitoring of the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment of color saturation can be achieved through sensors and software algorithms embedded in the display control system. These sensors and algorithms can continuously collect data of the current frame image, analyze the number and frequency of color saturation adjustments, and feedback the actual dynamic adjustment frequency.

[0153] Compare and analyze the predicted dynamic adjustment frequency with the actual dynamic adjustment frequency to identify the difference between the two. Determine the adjustment accuracy of the adjustment system by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made. If the error exceeds the allowable range, adjust the actual dynamic adjustment frequency until the error is within the allowable range.

[0154] The prediction model calculates the color saturation dynamic adjustment frequency of the current frame image, which is an ideal frequency based on the complexity and change trend of the image content. Next, the system monitors the dynamic adjustment frequency actually applied to the current frame in real time, and compares and analyzes the actual dynamic adjustment frequency with the predicted frequency to identify the difference between the two. By calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency, the system can evaluate the accuracy of the adjustment process. If the error is within the pre-defined allowable range, it means that the adjustment system is working properly and no further adjustment is required; however, if the error exceeds the allowable range, it means that the actual dynamic adjustment frequency deviates from the predicted value, which may cause insufficient or excessive color saturation adjustment, thereby affecting the image quality. At this time, the system will regulate and correct the actual dynamic adjustment frequency according to the size and direction of the error until the error returns to the allowable range. This feedback control mechanism ensures that the frequency of color saturation adjustment can accurately respond to changes in image content, provide the best visual effect, and avoid image distortion or visual fatigue caused by improper adjustment frequency.

[0155] The present invention obtains data of the current frame image through a video stream, extracts features including detail density, and calculates the complexity of the current frame image based on these features. Subsequently, by comparing the current frame with the previous frame image, the dynamic change information of the two adjacent frames of images is obtained, and the features of these dynamic change information are extracted, especially the texture change of the two adjacent frames of images, and the picture content information and the dynamic change information are integrated. The complexity and change trend of the current frame image are intelligently evaluated using a machine learning algorithm to further evaluate the complex changes of the image screen. Based on the intelligent evaluation results, the frequency of dynamic adjustment of the color saturation of the current frame image is predicted, and the actual dynamic adjustment frequency is monitored in real time. By comparing and analyzing the predicted frequency with the actual frequency, the error between the two is calculated, and the adjustment frequency of the system is adjusted according to the error, ensuring that the adjustment frequency of the color saturation is within the allowable range, thereby maintaining the accurate presentation and realism of the color, preventing the color from being too dim or dazzling, reducing visual fatigue when watching videos for a long time, and improving the user's comfort and experience satisfaction.

[0156] The present invention provides Figure 2 The adaptive microstructure optical film display screen shown includes an image feature extraction module, a dynamic change analysis module, an intelligent evaluation module, a predicted frequency calculation module, a real-time monitoring module, and a frequency comparison and control module;

[0157] The image feature extraction module obtains the data of the current frame image through the video stream, extracts the features of the current frame image, and calculates the complexity of the current frame image based on the extracted image features;

[0158] The dynamic change analysis module compares the current frame with the previous frame image, obtains the dynamic change information between the two, extracts features from the obtained dynamic change information, and identifies the dynamic change conditions of two adjacent frames of images;

[0159] The intelligent evaluation module integrates the picture content information and dynamic change information, uses machine learning algorithms to analyze the complexity and change trend of the current frame image, and performs intelligent evaluation on the complex changes of the current frame image;

[0160] A predicted frequency calculation module, which obtains the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted according to the result of the intelligent evaluation;

[0161] A real-time monitoring module monitors in real time the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment process of color saturation;

[0162] The frequency comparison and control module compares and analyzes the predicted dynamic adjustment frequency with the actual dynamic adjustment frequency, identifies the difference between the two, and determines the adjustment accuracy of the adjustment system by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made; if the error exceeds the allowable range, the actual dynamic adjustment frequency is adjusted until the error is within the allowable range.

[0163] The dynamic optimization method of the adaptive microstructure optical film display provided by the embodiment of the present invention is realized by the above-mentioned adaptive microstructure optical film display. The specific method and process of the adaptive microstructure optical film display are detailed in the embodiment of the dynamic optimization method of the adaptive microstructure optical film display, which will not be repeated here.

[0164] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0165] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0166] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for dynamic optimization of an adaptive microstructure optical film display screen, characterized in that: The following steps are involved: The data of the current frame image is obtained through the video stream, features of the current frame image are extracted, and the complexity of the current frame image is calculated based on the extracted image features; Compare the current frame with the previous frame image, obtain the dynamic change information between the two, perform feature extraction on the obtained dynamic change information, and identify the dynamic change of two adjacent frames of images; Comprehensive picture content information and dynamic change information, use machine learning algorithms to analyze the complexity and change trend of the current frame image, and perform intelligent evaluation of the complex changes of the current frame image; According to the result of intelligent evaluation, the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained; Real-time monitoring of the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment of color saturation; The predicted dynamic adjustment frequency is compared and analyzed with the actual dynamic adjustment frequency to identify the difference between the two. The adjustment accuracy of the adjustment system is determined by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made; if the error exceeds the allowable range, the actual dynamic adjustment frequency is adjusted until the error is within the allowable range.

2. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 1, characterized in that: The feature extraction of the current frame image includes the detail density in the image. The detail density of the image refers to the density of detail information contained in the image, which reflects the proportion of high-frequency components in the image. After acquisition, the detail density in the image is analyzed and processed to generate a detail density index. The steps for generating the detail density index are as follows: Preprocess the acquired image, including grayscale and denoising, to reduce computational complexity and noise interference; Use edge detection algorithms to extract edges in images; Wavelet transform is used to extract the high-frequency components of the image, including horizontal detail components, vertical detail components and diagonal detail components. The horizontal detail components, vertical detail components and diagonal detail components are calibrated as , , ; Calculate the energy of the high-frequency component as a quantitative indicator of detail. The calculation expression is: , , , i and j represent the row index and column index in the image matrix respectively; Calculate the edge density, the expression is: , where is the edge density, which indicates the proportion of edge pixels in the image. is the value of the i, jth pixel in the edge image, and N is the total number of pixels in the image; Calculate the detail density index. The calculation expression is: , where represents the weight parameter of high frequency energy, The weight parameter representing the edge density, Represents the detail density index.

3. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 2, characterized in that: The specific steps for extracting edges in an image using the Canny edge detection algorithm are as follows: Apply a Gaussian filter to smooth the grayscale image to reduce noise and irrelevant details; Calculate the gradient strength and direction of each pixel in the image; Perform non-maximum suppression on the gradient image, retain the gradient of the local maximum value, and suppress other gradient values; Two thresholds are applied to distinguish strong edges from weak edges, a high threshold is used to mark strong edges, and a low threshold is used to mark possible edges; Through the edge connection process, all weak edges are connected to strong edges. If a weak edge pixel is connected to a strong edge pixel, it is retained as an edge, otherwise it is suppressed. By traversing the image, the weak edge pixels are associated with the strong edge pixels to ensure the continuity of the edge. Output a binary image, where a pixel value of 1 represents an edge and a pixel value of 0 represents a non-edge.

4. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 2, characterized in that: The steps of using Haar wavelet transform to extract the high-frequency components of an image, including horizontal detail components, vertical detail components, and diagonal detail components, are as follows: Select Haar wavelet as wavelet filter; The Haar wavelet filter is decomposed into a low-pass filter and a high-pass filter. The low-pass filter is used to capture the low-frequency components of the image and reflect the overall structure and large-scale characteristics of the image; the high-pass filter is used to capture the high-frequency components of the image and reflect the details and edge information of the image. Perform a low-pass and high-pass filtering operation on each line of the image, obtain the horizontal low-frequency component of the image after low-pass filtering, and obtain the horizontal high-frequency component of the image after high-pass filtering; The image after row-direction filtering is subjected to low-pass and high-pass filtering in the column direction, the horizontal low-frequency component is subjected to high-pass filtering to obtain the vertical detail component of the image, the horizontal high-frequency component is subjected to low-pass filtering to obtain the horizontal detail component of the image; the horizontal high-frequency component is subjected to high-pass filtering to obtain the diagonal detail component of the image; The horizontal detail component, vertical detail component and diagonal detail component obtained in the above steps are combined into a sub-band image, the horizontal detail component reflects the horizontal detail change of the image, the vertical detail component reflects the vertical detail change of the image, and the diagonal detail component reflects the diagonal detail change of the image.

5. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 2, characterized in that: Compare the current frame with the previous frame to obtain the dynamic change information between the two. The specific steps are as follows: Get the image data of the current frame and the previous frame through the video stream; Before dynamic change analysis, the images of the current frame and the previous frame are preprocessed; Use the optical flow algorithm or frame difference method to calculate the dynamic change information between the current frame and the previous frame.

6. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 5, characterized in that: The obtained dynamic change information is subjected to feature extraction. The extracted dynamic change features include texture changes between two adjacent frames of images. After analyzing and processing the texture changes between two adjacent frames of images, a dynamic texture change index is generated. The steps for generating the dynamic texture change index are as follows: The texture feature extraction method is used to extract texture features from the preprocessed image, and the texture feature vectors of the current frame image and the previous frame image are respectively used and To express; Calculate the difference in texture features between the current frame and the previous frame, and use the Euclidean distance to measure the difference between feature vectors. The calculation expression is: , where Represents the difference in texture features between two adjacent frames of images, n represents the dimension of the feature vector, and k represents the kth feature component in the texture feature vector; At each pixel position in the image The local texture change is calculated at , and the local texture change matrix is ​​obtained. Each element in the local texture change matrix represents the degree of local texture change at that position. The calculation expression is: , where represents the local texture change matrix, where x represents the horizontal coordinate of the pixel and y represents the vertical coordinate of the pixel; Calculate the dynamic texture change index. The calculation expression is: , where Represents the dynamic texture change index, M and N are the width and height of the image respectively.

7. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 6, characterized in that: The specific steps of using local binary patterns to extract texture features from preprocessed images are as follows: Step 1: For each pixel, take the pixel as the center, take the pixels around it as the neighborhood, and use the selected pixels to calculate the local binary pattern; Step 2: For the central pixel, compare each pixel in its neighborhood with the central pixel. If the grayscale value of the neighborhood pixel is greater than or equal to the grayscale value of the central pixel, the value of the neighborhood pixel is recorded as 1; otherwise, it is recorded as 0, so that each neighborhood pixel is assigned a binary value. Step 3: Arrange the binary values ​​obtained in step 2 in a clockwise or counterclockwise order to form a binary number, and then convert the binary number into a decimal number, which is the LBP value of the central pixel; The LBP feature map of the entire image is statistically analyzed, and the frequency of occurrence of each LBP value is calculated to form an LBP histogram. The histogram is used as the texture feature descriptor of the image to reflect the distribution of different patterns in the image.

8. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 6, characterized in that: Get the detail density index generated by analyzing the current frame image The dynamic texture change index generated by analyzing the texture changes of two adjacent frames of images Then, the detail density index and dynamic texture change index Input into a pre-trained machine learning model to generate complex coefficients of variation based on the model , through the complex variation coefficient Intelligent evaluation of complex changes in image content.

9. The method for dynamic optimization of an adaptive microstructure optical film display screen according to claim 8, characterized in that: The predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted is obtained. The specific steps are as follows: Get the detail density index generated by analyzing the current frame image , the dynamic texture change index generated by analyzing the texture changes of two adjacent frames of images And complex coefficients of variation generated based on machine learning models , defines the basic adjustment frequency when the color saturation is dynamically adjusted. The expression defined is: , where Indicates the basic adjustment frequency, which is determined by the detail density index and dynamic texture change index The function f determines; Based on the basic frequency, by integrating the complex change coefficient To control the basic adjustment frequency, the expression for control is: , where It represents the actual frequency after the basic adjustment frequency is regulated, and u represents the adjustment coefficient, which determines the amplitude of the complex change coefficient regulating the basic adjustment frequency; Calculate the predicted dynamic adjustment frequency, the calculation expression is: , where Indicates the predicted dynamic adjustment frequency, Represents the smoothing factor, which is used to adjust the effect of color error on the adjustment frequency. Represents the average color error between the current frame and the previous frame, Indicates the maximum permissible color error.

10. An adaptive microstructure optical film display screen, characterized in that: It includes image feature extraction module, dynamic change analysis module, intelligent evaluation module, prediction frequency calculation module, real-time monitoring module and frequency comparison and control module; The image feature extraction module obtains the data of the current frame image through the video stream, extracts the features of the current frame image, and calculates the complexity of the current frame image based on the extracted image features; The dynamic change analysis module compares the current frame with the previous frame image, obtains the dynamic change information between the two, extracts features from the obtained dynamic change information, and identifies the dynamic change conditions of two adjacent frames of images; The intelligent evaluation module integrates the picture content information and dynamic change information, uses machine learning algorithms to analyze the complexity and change trend of the current frame image, and performs intelligent evaluation on the complex changes of the current frame image; A predicted frequency calculation module, which obtains the predicted dynamic adjustment frequency when the color saturation of the current frame image is dynamically adjusted according to the result of the intelligent evaluation; A real-time monitoring module monitors in real time the actual dynamic adjustment frequency of the current frame image in the video screen during the dynamic adjustment of color saturation; The frequency comparison and control module compares and analyzes the predicted dynamic adjustment frequency with the actual dynamic adjustment frequency, identifies the difference between the two, and determines the adjustment accuracy of the adjustment system by calculating the error between the predicted dynamic adjustment frequency and the actual dynamic adjustment frequency. If the error is within the allowable range, no adjustment is made; if the error exceeds the allowable range, the actual dynamic adjustment frequency is adjusted until the error is within the allowable range.

Images