Electronic paper display mapping method based on visual perception and dynamic clustering

Abstract

The invention relates to an electronic paper display mapping method based on visual perception and dynamic clustering, and belongs to the technical field of display. According to the method, firstly, learning factors and particle weights of a particle swarm optimization algorithm are dynamically adjusted, a gray segmentation threshold is optimized in combination with K-means clustering, and the accuracy and convergence efficiency of gray distribution are remarkably improved; secondly, establishing a dynamic error diffusion system based on visual perception, designing a frequency domain visual weighted filter and a spatial domain sensitivity function by combining a human visual perception theory, dynamically compensating quantization errors and enhancing edge details; and meanwhile, a Floyd-Steinberg error diffusion algorithm is improved, and a double image effect is inhibited in combination with a snakelike scanning path. According to the electronic paper display mapping method based on visual perception and dynamic clustering, the electronic paper image display quality is effectively improved, particularly when a high-gray-scale image is displayed, details can be reserved to the maximum extent, and the problems of gray-scale distortion and edge blurring in traditional electronic paper display are solved.

Description

Technical Field

[0001] The present invention belongs to the field of display technology, and particularly relates to an electronic paper display mapping method based on visual perception and dynamic clustering. Background Art

[0002] With the continuous development of electronic paper technology, electronic paper, as an innovative display technology with low power consumption and long-term display stability, has been widely used in many fields such as e-books, smart tags, and advertising boards. Although electronic paper has many advantages, when displaying grayscale images, the grayscale levels of electronic paper are limited, and problems such as image distortion and edge blurring are likely to occur, especially in the display of high-dynamic-range images. These problems seriously affect the application effect of electronic paper in some scenarios with high-precision display requirements. Currently, the display technology of electronic paper mainly adopts traditional error diffusion algorithms and threshold-based image binarization methods, but these methods cannot effectively handle complex images and image content with rich details. How to improve the grayscale accuracy and display quality of electronic paper images under complex lighting conditions and high-dynamic-range scenarios has become a difficult problem that needs to be solved urgently in the current technology. Summary of the Invention

[0003] The purpose of the present invention is to provide an electronic paper display mapping method based on visual perception and dynamic clustering. This method significantly improves the quality of electronic paper in the display of high-dynamic-range images by optimizing the grayscale distribution and introducing a visual optimization strategy based on the human visual system, and solves problems such as grayscale distortion and edge blurring.

[0004] To achieve the above purpose, the technical solution of the present invention is: an electronic paper display mapping method based on visual perception and dynamic clustering, including:

[0005] Step S1: Perform dynamic clustering analysis on the image to be displayed to determine the optimal grayscale segmentation threshold: According to the grayscale distribution characteristics of the pixel values of the image, build a dynamic particle swarm optimization (PSO), divide the pixel values into 16 categories as the initial centroids of K-Means clustering, and perform clustering analysis on the pixel values to generate 16 cluster centers;

[0006] Step S2: Based on the clustering result, combine the human visual perception theory, optimize the quantization error distribution, and construct a refined electronic paper display mapping model: Use the 16 cluster centers obtained in Step S1 as the quantization thresholds of the error diffusion optimization module, build a dynamic error diffusion system, and perform error diffusion processing pixel by pixel in combination with the human visual perception theory;

[0007] Step S3: Adapt the processed image to the hardware characteristics of the electronic paper and display it: Map the image data processed in Step S2 into an encoding format that can be displayed by the electronic paper, and transmit it to the electronic paper display screen for display.

[0008] In an embodiment of the present invention, step S1 includes:

[0009] Step S11: Extract the gray-scale distribution and edge information of the image, calculate the local gray-scale change amount ΔG through a sliding window edge , and calculate the edge amplitude G(x, y) using the Sobel operator; the local gray-scale change amount ΔG edge The calculation formula is as follows:

[0010]

[0011] where I represents the pixel value of the image at coordinates (x, y), i and j represent the offset indices in the local neighborhood window; N represents the total number of pixels in the window area (i.e., the neighborhood size); μ represents the mean value of all pixels in the window area.

[0012] Step S12: Dynamically adjust the learning factor and particle weight of the PSO, generate the initial clustering center through the optimized dynamic particle swarm optimization PSO, and provide a high-precision initial value for K-means; the specific formula for the whole process is as follows:

[0013]

[0014] where, represents the current position of particle i. represents the velocity of particle i in the t-th generation, ω represents the inertia weight, which is used to control the movement inertia of the particle, affect the convergence speed and balance the global and local search capabilities, c 1 , c 2 represent the learning factors, which respectively control the influence of the particle on the individual best position and the global best position g t , r 1 , r 2 are random numbers, which are used to increase the randomness of particle search.

[0015] Set the number of particles n = 50, the search space is the gray value range [0, 255], and update c edge and G(x, y) in real time: 1 c 2 :

[0016] c 1 = 1.5 + 0.2·ΔG edge , c 2 = 1.8 + 0.15·G(x, y)

[0017] The particle weight distribution is:

[0018] ω = 0.3·ΔG edge + 0.7·G(x, y)

[0019] Step S13: Use the initial center optimized by dynamic particle swarm optimization (PSO) as the initial center of K-means {μ 1 , μ 2 …, μ 16 Use the initial center optimized by PSO to accelerate the convergence of K-means, divide the pixels of the high-gray-level image into 16 clustering clusters, calculate the Euclidean distance from each pixel to each center, assign it to the nearest cluster, update the cluster center, terminate after 50 iterations, and output the gray-level threshold of each cluster.

[0020] In an embodiment of the present invention, step S2 includes:

[0021] Step S21: Construct a visual weighting filter H(u, v) to quantify the visual error in the HVS sensitive region; where W EPD (x, y) is the weight of the e-paper, set to 1.2 in the gray-scale transition region and 0.8 in the flat region.

[0022]

[0023] Where (u, v) represents the spatial frequency coordinates of the image in the frequency domain, and (x, y) represents the pixel coordinates in the e-paper display screen.

[0024] Step S22: Improve the Floyd-Steinberg error diffusion algorithm and combine it with the Sobel operator to enhance edge details; the specific formula is as follows:

[0025]

[0026] f'(x, y) = f * (x, y) + β · G * (x, y)

[0027] Where f(x, y) represents the original input image, f * (x, y) is the image after error diffusion, k and l are the coordinate offsets of adjacent pixels, used to traverse and calculate the influence in the error diffusion process, is the error diffusion weight matrix, e(x - i, y - j) represents the error value diffused to adjacent pixels, G * (x, y) represents the gradient information after edge enhancement processing, β represents the edge enhancement weight, used to control the influence of edge enhancement on the final image, and f′(x, y) represents the image after edge enhancement.

[0028] Step S23: Dynamically adjust the error diffusion weight W(i, j) according to the hardware characteristics of the e-paper to suppress the ghosting effect;

[0029] f final(x, y) = [f'(x, y) + γ·V(x, y)]·F EPD (x, y)

[0030] Where V(x, y) represents the visual perception error compensation term, which is adjusted in combination with the human visual model (HVS), γ represents the visual compensation coefficient, controlling the intensity of visual error correction, and F EPD (x, y) represents the optimization weight of the electronic paper, making the error diffusion result more in line with the display characteristics of the electronic paper, and f final (x, y) represents the finally output optimized image.

[0031] Step S24: Optimize the pixel processing order using a "snake-like" scanning path. For odd rows, scan from left to right, and for even rows, scan from right to left, reducing the refresh delay of adjacent pixels and avoiding the generation of afterimages.

[0032] In an embodiment of the present invention, step S3 includes:

[0033] Step S31: Encode the result obtained in step S3 according to the mapping relationship from a 16-bit image to a display code, and convey the encoded result to the image buffer module of the system;

[0034] Step S32: Initialize the main control chip, the driving chip, and the electronic paper, and store the display code in the main control chip;

[0035] Step S33: Retrieve the image data from the image buffer module, convey it to the driving control module of the system, and drive the electronic paper to complete image display.

[0036] In an embodiment of the present invention, the parameters of the dynamic particle swarm optimization PSO are set as follows: the number of particles N = 50, the search space is [0, 255], the number of iterations is 20 times, and the termination condition is that the change amount of the cluster center < 1.

[0037] In an embodiment of the present invention, the method is applicable to electronic paper devices using electrophoretic display technology.

[0038] The present invention also provides an electronic paper display mapping system based on visual perception and dynamic clustering, including a memory, a processor, and computer program instructions stored on the memory and capable of being run by the processor. When the processor runs the computer program instructions, it can implement the method steps as described in any of the above.

[0039] The present invention also provides a computer-readable storage medium, on which computer program instructions capable of being run by the processor are stored. When the processor runs the computer program instructions, it can implement the method steps as described in any of the above.

[0040] Compared with the prior art, the present invention has the following beneficial effects: By designing an electronic paper display mapping method based on visual perception and dynamic clustering, the present invention effectively improves the display quality of electronic paper images. Especially when displaying high-gray-scale images, it can maximize the retention of details and solve the problems of gray-scale distortion and edge blurring in traditional electronic paper displays. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is the overall flowchart of an electronic paper display mapping method based on visual perception and dynamic clustering of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0042] The technical solution of the present invention will be specifically described below with reference to the drawings.

[0043] The present invention provides an electronic paper display mapping method based on visual perception and dynamic clustering, including:

[0044] Step S1: Perform dynamic clustering analysis on the image to be displayed to determine the optimal gray-scale segmentation threshold: According to the gray-scale distribution characteristics of the pixel values of the image, build a dynamic particle swarm optimization (PSO), divide the pixel values into 16 categories as the initial centroids of K-Means clustering, and perform clustering analysis on the pixel values to generate 16 cluster centers;

[0045] Step S2: Based on the clustering result, combine the human visual perception theory to optimize the quantization error distribution and construct a refined electronic paper display mapping model: Use the 16 cluster centers obtained in Step S1 as the quantization thresholds of the error diffusion optimization module, build a dynamic error diffusion system, and perform error diffusion processing pixel by pixel in combination with the human visual perception theory;

[0046] Step S3: Adapt the processed image to the hardware characteristics of the electronic paper and display it: Map the image data processed in Step S2 into an encoding format that can be displayed by the electronic paper and transmit it to the electronic paper display for display.

[0047] The following is the specific implementation process of the present invention.

[0048] An electronic paper display mapping method based on visual perception and dynamic clustering, the overall working process is as Figure 1 shown, including the following steps:

[0049] Step S1: Perform dynamic clustering analysis on the image to be displayed to determine the optimal gray-scale segmentation threshold: According to the gray-scale distribution characteristics of the pixel values of the image, build a dynamic particle swarm optimization (PSO), and divide the pixel values into 16 categories through an optimization algorithm as the initial centroids of K-Means clustering, and perform clustering analysis on the pixel values to generate 16 cluster centers;

[0050] Step S2: Based on the clustering results and combined with the human visual perception theory, optimize the quantization error distribution and construct a refined e-paper display mapping model. Use the 16 cluster centers obtained in Step S1 as the quantization thresholds of the error diffusion optimization module, build a dynamic error diffusion system, and perform error diffusion processing pixel by pixel in combination with the human visual perception theory.

[0051] Step S3: Adapt the processed image to the hardware characteristics of the e-paper and display it. Map the image data processed in Step S2 into an encoding format that can be displayed by the e-paper, and transmit it to the e-paper display screen for display.

[0052] In this embodiment, in Step S1, perform dynamic clustering analysis on the image to be displayed to determine the optimal gray-scale segmentation threshold. According to the gray-scale distribution characteristics of the pixel values of the image, build a dynamic particle swarm optimization (PSO) module, divide the pixel values into 16 categories through an optimization algorithm as the initial centroids of K-Means clustering, and perform clustering analysis on the pixel values to generate 16 cluster centers. The specific steps are as follows:

[0053] Step S11: Extract the gray-scale distribution and edge information of the image to be displayed, and calculate the local gray-scale change amount ΔG edge , and use the Sobel operator to calculate the edge amplitude G(x, y). The local gray-scale change amount ΔG edge The calculation formula is as follows:

[0054]

[0055] where I represents the pixel value of the image at the coordinate (x, y), i and j represent the offset indices in the local neighborhood window; N represents the total number of pixels in the window area (i.e., the neighborhood size); μ represents the mean value of all pixels in the window area.

[0056] Step S12: Dynamically adjust the learning factor and particle weight of the PSO algorithm, and generate the initial clustering centers through the optimized PSO to provide high-precision initial values for K-means. The specific formulas are as follows:

[0057]

[0058] where, represents the current position of particle i. represents the velocity of particle i in the t-th generation, ω represents the inertia weight, which is used to control the movement inertia of the particle, affect the convergence speed and balance the global and local search capabilities, c 1 , c 2 represent the learning factors, which respectively control the influence of the particle on the individual best position and the global best position g t , r 1 , r 2is a random number used to increase the randomness of particle search.

[0059] Set the number of particles n = 50, and the search space is the grayscale value range [0, 255]. According to ΔG edge and G(x, y), update c 1 , c 2 :

[0060] c 1 = 1.5 + 0.2·ΔG edge c 2 = 1.8 + 0.15·G(x, y)

[0061] The particle weight distribution is as follows:

[0062] ω = 0.3·ΔG edge + 0.7·G(x, y)

[0063] Step S13: Take the 16 particle positions output by PSO as the initial centers {μ 1 , μ 2 …, μ 16} of K-means. Use the initial centers optimized by PSO to accelerate the convergence of K-means. Divide the pixels of the high grayscale image into 16 clustering clusters, calculate the Euclidean distance from each pixel to each center, assign it to the nearest cluster, update the cluster center, terminate after 50 iterations, and output the grayscale thresholds of each cluster.

[0064] Next, in step S2: Based on the clustering results, combined with the human visual perception theory, optimize the quantization error distribution and construct a refined electronic paper display mapping model: Take the 16 cluster centers obtained in step S1 as the quantization thresholds of the error diffusion optimization module, build a dynamic error diffusion system, and perform error diffusion processing pixel by pixel in combination with the human visual perception theory, which specifically includes the following steps:

[0065] Step S21: Construct a visual weighting filter H(u, v) to enhance the HVS sensitive area, where W EPD (x, y) is the weight of the electronic paper, set the gray transition area to 1.2, and the flat area to 0.8.

[0066]

[0067] where (u, v) represents the spatial frequency coordinates of the image in the frequency domain, and (x, y) represents the pixel coordinates in the electronic paper display screen.

[0068] Step S22: Improve the Floyd-Steinberg error diffusion algorithm and enhance the edge details in combination with the Sobel operator. The specific formula is as follows:

[0069]

[0070] f'(x, y) = f * (x, y) + β · G * (x, y)

[0071] Where f(x, y) represents the original input image, and f * (x, y) is the image after error diffusion. k and l are the coordinate offsets of adjacent pixels, used to traverse and calculate the influence during the error diffusion process. is the error diffusion weight matrix. e(x - i, y - j) represents the error value diffused to adjacent pixels. G * (x, y) represents the gradient information after edge enhancement processing. β represents the edge enhancement weight, used to control the influence degree of edge enhancement on the final image. f′(x, y) represents the image after edge enhancement.

[0072] Step S23: Dynamically adjust the error diffusion weight according to the e - paper hardware characteristics to suppress the ghosting effect;

[0073] f final (x, y) = [f'(x, y) + γ · V(x, y)] · F EPD (x, y)

[0074] Where V(x, y) represents the visual perception error compensation term, which is adjusted in combination with the human visual model (HVS). γ represents the visual compensation coefficient, controlling the intensity of visual error correction. F EPD (x, y) represents the e - paper optimization weight, making the error diffusion result more in line with the display characteristics of the e - paper. f final (x, y) represents the finally output optimized image.

[0075] Step S24: Optimize the pixel processing order using a "snake - like" scanning path. For odd - numbered rows, scan from left to right, and for even - numbered rows, scan from right to left, reducing the refresh delay of adjacent pixels and avoiding the generation of afterimages.

[0076] Furthermore, in step S3: Adapt the processed image to the e - paper hardware characteristics and display it: Map the image data processed in step S2 into an encoding format that can be displayed by the e - paper, and transmit it to the e - paper display screen for display, which specifically includes the following steps:

[0077] Step S31: According to the mapping relationship from 16 - level images to display encodings, encode the result obtained in step S3, and send the encoded result to the image cache module of the system;

[0078] Step S32: Initialize the main control chip, the driving chip, and the e - paper. Store the display encoding in the main control chip, and call the initialization waveform to refresh the display screen to the all - white state;

[0079] Step S33: Retrieve the image data from the cache module, compare the image data in the buffer with all the images to obtain the address of the drive waveform lookup table, complete the display of the first image, compare the image to be updated with the original image to obtain the address of the new waveform, and drive to complete the display of the new image.

[0080] The present invention also provides an e-paper display mapping system based on visual perception and dynamic clustering, including a memory, a processor, and computer program instructions stored on the memory and executable by the processor. When the processor runs the computer program instructions, the method steps described in any of the above can be implemented.

[0081] The present invention also provides a computer-readable storage medium, on which computer program instructions executable by the processor are stored. When the processor runs the computer program instructions, the method steps described in any of the above can be implemented.

[0082] The above are the preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention in terms of the functions and effects produced belong to the protection scope of the present invention.

Claims

1. An electronic paper display mapping method based on visual perception and dynamic clustering, characterized in that: include: Step S1, dynamic clustering analysis is performed on the image to be displayed to determine the optimal grayscale segmentation threshold: the pixel values ​​of the image are divided into 16 categories according to the grayscale distribution characteristics, and the pixel values ​​are used as the initial centroid of K-Means clustering. Cluster analysis is performed on the pixel values ​​to generate 16 cluster centers; Step S2: Based on the clustering results and in combination with the theory of human visual perception, the quantization error distribution is optimized to build a refined electronic paper display mapping model: the 16 cluster centers obtained in step S1 are used as the quantization thresholds of the error diffusion optimization module, a dynamic error diffusion system is built, and error diffusion processing is performed pixel by pixel in combination with the theory of human visual perception; Step S3, adapting the processed image to the electronic paper hardware characteristics and displaying it: mapping the image data processed in step S2 into a coding format displayable by the electronic paper, and transmitting it to the electronic paper display screen for display.

2. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 1, characterized in that: Step S1 includes: Step S11: extract the grayscale distribution and edge information of the image, and calculate the local grayscale change ΔG through a sliding window edge , and use the Sobel operator to calculate the edge amplitude G(x,y); Step S12, dynamically adjust the learning factors c1, c2 and particle weight ω of PSO, generate initial cluster centers through the optimized dynamic particle swarm optimization PSO, and provide high-precision initial values ​​for K-means; Step S13: Divide the initial center optimized by dynamic particle swarm optimization (PSO) into 16 clusters, iteratively update the cluster center until convergence, and output the grayscale threshold.

3. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 2, characterized in that: In step S11, the local grayscale change ΔG edge The calculation formula is as follows: Where I represents the pixel value of the image at the coordinate (x, y), i and j represent the offset index in the local neighborhood window; N represents the total number of pixels in the window area, that is, the neighborhood size; μ represents the mean value of all pixels in the window area.

4. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 1, characterized in that: Step S2 includes: Step S21, constructing a visual weighted filter H(u,v) to quantify the visual error of the HVS sensitive area, where (u,v) represents the spatial frequency coordinates of the image in the frequency domain; Step S22, improving the Floyd-Steinberg error diffusion algorithm and combining it with the Sobel operator to enhance edge details; Step S23, dynamically adjusting the error diffusion weight according to the hardware characteristics of the electronic paper to suppress the ghosting effect; Step S24: Use a "snake" scanning path to optimize the pixel processing order, with odd rows from left to right and even rows from right to left, to reduce the refresh delay of adjacent pixels.

5. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 4, characterized in that: In step S21, the visual weighted filter H(u,v) is calculated as follows: Where (x, y) represents the pixel coordinates in the electronic paper display, W EPD (x,y) is the electronic paper weight.

6. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 1, characterized in that: Step S3 includes: Step S31, encoding the result obtained in step S3 according to the mapping relationship from the 16-order image to the display code, and transmitting the encoding result to the image cache module of the system; Step S32, initializing the main control chip, the driver chip and the electronic paper, and storing the display code in the main control chip; Step S33: fetch the image data from the image buffer module and transmit it to the drive control module of the system to drive the electronic paper to complete the image display.

7. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 1, characterized in that: The dynamic particle swarm optimization (PSO) parameters are set as follows: number of particles n = 50, search space [0, 255], number of iterations 20 times, and termination condition when the change in cluster center is <1.

8. The electronic paper display mapping method based on visual perception and dynamic clustering according to claim 1, characterized in that: The method is applicable to electronic paper devices using electrophoretic display technology.

9. An electronic paper display mapping system based on visual perception and dynamic clustering, characterized in that: The method comprises a memory, a processor and computer program instructions stored in the memory and capable of being executed by the processor. When the processor executes the computer program instructions, the method steps as claimed in any one of claims 1 to 8 can be implemented.

10. A computer-readable storage medium having stored thereon computer program instructions that can be executed by a processor, and when the processor executes the computer program instructions, the method steps according to any one of claims 1 to 8 can be implemented.

Images