Big-model-based electronic paper ghosting elimination method and apparatus

Abstract

A big-model-based electronic paper ghosting elimination method and apparatus. The method comprises: acquiring first image data of a first image and second image data of a second image (101), wherein the second image is an image frame subsequent to the first image; on the basis of the first image data, calculating first influence information between each first pixel point in the first image and each first influence pixel corresponding thereto (102); on the basis of all the first influence information corresponding to each first pixel point, generating a target compensation data table corresponding to the first pixel point (103); on the basis of the second image data, determining second influence information between each second pixel point in the second image and all second influence pixels corresponding thereto (104); and on the basis of all the second influence information and the target compensation data tables, executing a ghosting elimination operation on the second image data to obtain a ghosting-eliminated image of the second image (105).

Description

Electronic paper afterimage elimination method and device based on large model Technical Field

[0001] The present application relates to the field of image data processing technology, and in particular to a method and device for eliminating electronic paper afterimages based on a large model. Background Art

[0002] Due to the mutual influence between pixel particles in electronic paper ink screens, there will often be afterimages of the previous frame during the development process, which reduces the user's reading experience. The current industry solution to eliminate afterimages is to use a global clear refresh mode with three stages of "full black-full white-grayscale" to achieve afterimage elimination of electronic paper. However, the disadvantage of this method is that the refresh time is long and the screen will flicker during the refresh process. If this method is applied to devices that require frequent refreshes, the user's reading experience will be reduced. Another processing method is to divide the screen into multiple areas and only refresh the areas that have changed. The disadvantage of this processing method is that in most cases, the screen image changes across the entire screen, so the efficiency of this processing method is relatively low. It can be seen that it is particularly important to provide a method that can achieve afterimage elimination on electronic paper screens while improving the screen refresh speed and afterimage elimination efficiency.

[0003] Summary of the Invention

[0004] The technical problem to be solved by this application is to provide a method and device for eliminating electronic paper afterimages based on a large model, which can eliminate electronic paper screen afterimages while improving the screen refresh speed and afterimage elimination efficiency.

[0005] In order to solve the above technical problems, the first aspect of the present application discloses a method for eliminating electronic paper afterimages based on a large model, the method comprising:

[0006] Acquire first image data of a first image and second image data of a second image; the second image is an image frame subsequent to the first image; the first image and the second image are both subjected to a preset grayscale image processing;

[0007] calculating, based on the first image data, first influence information between each first pixel point in the first image and each corresponding first influencing pixel thereof; and generating, based on all the first influence information corresponding to each first pixel point, a target compensation data table corresponding to the first pixel point; the target compensation data table including compensation data corresponding to each first influence information;

[0008] determining, according to the second image data, second influence information between each second pixel point in the second image and all corresponding second influencing pixels thereof;

[0009] According to all the second impact information and the target compensation data table, an afterimage elimination operation is performed on the second image data to obtain an afterimage eliminated image of the second image.

[0010] As an optional implementation manner, in the first aspect of the present application, calculating, based on the first image data, first influence information between each first pixel point in the first image and each corresponding first influencing pixel thereof includes:

[0011] For each first pixel point in the first image, determining a plurality of influencing directions that have an afterimage influence on the first pixel point;

[0012] Taking the coordinates of the first pixel point as the center, determining the first influencing pixel corresponding to each of the influencing directions;

[0013] Perform a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point;

[0014] The impact calculation operation includes grayscale difference calculation, particle offset and impact factor conversion.

[0015] As an optional implementation manner, in the first aspect of the present application, performing a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain first influence information between the first pixel point and all the corresponding first influencing pixels includes:

[0016] Inputting the pixel coordinates of each of the first influencing pixels and the pixel coordinates of the first pixel point into a preset grayscale difference calculation formula to obtain a grayscale difference between each of the first influencing pixels and the first pixel point;

[0017] For each of the grayscale differences corresponding to the first influencing pixels, offset according to a predetermined offset bit amount to obtain an offset result corresponding to each of the first influencing pixels;

[0018] Calculating, based on a plurality of preset influencing factors, an influence conversion value corresponding to each of the first influencing pixels and the first pixel point;

[0019] The offset results and the impact conversion values ​​corresponding to all the first influencing pixels are determined as first impact information between the first pixel point and all the first influencing pixels corresponding to it.

[0020] As an optional implementation manner, in the first aspect of the present application, generating a target compensation data table corresponding to each first pixel point based on all the first impact information corresponding to the first pixel point includes:

[0021] For each of the first pixel points, performing a correction operation on each of the impact conversion values ​​of the first pixel point according to a preset correction parameter to obtain a correction result corresponding to each of the impact conversion values;

[0022] storing the correction result corresponding to each first-affected pixel at the storage address represented by the offset result corresponding to the first-affected pixel, to obtain a compensation table corresponding to all the first-affected pixels;

[0023] performing image sticking training correction on a training image set according to the compensation table to obtain a training correction result of the image set;

[0024] When the training correction result indicates that the preset afterimage correction standard is reached, the compensation table is determined to be a target compensation data table.

[0025] As an optional implementation manner, in the first aspect of the present application, the method further includes:

[0026] When the training correction result indicates that the preset ghosting correction standard is not met, repeatedly adjusting the correction parameters to update the correction result and the compensation table;

[0027] After each update of the compensation table, the operation corresponding to the image afterimage training correction is triggered to be performed on the training image set according to the preset AI simulation algorithm combined with the compensation table to obtain the training correction result of the image set, until it is determined that the training correction result after a certain update indicates that the afterimage correction standard is met, and the compensation table corresponding to the training correction result after a certain update is determined to be the target compensation data table.

[0028] As an optional implementation, in the first aspect of the present application, performing image sticking training correction on a training image set according to the compensation table to obtain a training correction result for the image set includes:

[0029] Selecting any frame of image from the training image set as a training image; performing an image afterimage correction operation on the training image according to the compensation table to obtain a corrected image of the training image;

[0030] Dividing the training image and the correction image into the same preset number of sub-images respectively;

[0031] Taking the sub-image as a unit, calculating the target difference value between the training image and the correction image under the sub-image; and calculating the square average of all the target difference values ​​to obtain the target comparison value between the training image and the correction image;

[0032] The target difference value includes a grayscale difference value and a similarity difference value, and the target comparison value includes a first square root corresponding to the grayscale difference value and a second square root corresponding to the similarity difference value;

[0033] The first square root and the second square root are multiplied by their corresponding weighted averages respectively and the averages are calculated to obtain a weighted average result of the training image and the correction image as a training correction result.

[0034] As an optional implementation, in the first aspect of the present application, performing an afterimage removal operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage-removed image of the second image includes:

[0035] According to all the second impact information, indexing the target compensation data corresponding to each second pixel in the target compensation data table;

[0036] Performing a data merging operation on the target compensation data corresponding to each second pixel point and the second image data to obtain a data merging result of all the second pixel points as an afterimage-eliminating image of the second image;

[0037] The afterimage elimination image is used to be transmitted to a display module, and the display module performs a afterimage elimination refresh operation on the second image currently being refreshed using the afterimage elimination image.

[0038] The second aspect of the present application discloses a device for eliminating afterimages of electronic paper based on a large model, the device comprising:

[0039] an acquisition module, configured to acquire first image data of a first image and second image data of a second image; the second image being an image frame subsequent to the first image; and the first image and the second image both undergoing a preset grayscale image processing;

[0040] a calculation module, configured to calculate, based on the first image data, first influence information between each first pixel point in the first image and each corresponding first influencing pixel;

[0041] a generating module, configured to generate a target compensation data table corresponding to each first pixel point based on all the first impact information corresponding to the first pixel point; the target compensation data table includes compensation data corresponding to each first impact information;

[0042] a determining module, configured to determine, based on the second image data, second influence information between each second pixel point in the second image and all corresponding second influencing pixels;

[0043] An afterimage elimination module is used to perform an afterimage elimination operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image.

[0044] As an optional implementation, in the second aspect of the present application, the calculation module calculates, based on the first image data, the first influence information between each first pixel point in the first image and each corresponding first influencing pixel thereof, specifically including:

[0045] For each first pixel point in the first image, determining a plurality of influencing directions that have an afterimage influence on the first pixel point;

[0046] Taking the coordinates of the first pixel point as the center, determining the first influencing pixel corresponding to each of the influencing directions;

[0047] Perform a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point;

[0048] The impact calculation operation includes grayscale difference calculation, particle offset and impact factor conversion.

[0049] As an optional implementation manner, in the second aspect of the present application, the calculation module performs a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain the first influence information between the first pixel point and all the corresponding first influencing pixels, specifically including:

[0050] Inputting the pixel coordinates of each of the first influencing pixels and the pixel coordinates of the first pixel point into a preset grayscale difference calculation formula to obtain a grayscale difference between each of the first influencing pixels and the first pixel point;

[0051] For each of the grayscale differences corresponding to the first influencing pixels, offset according to a predetermined offset bit amount to obtain an offset result corresponding to each of the first influencing pixels;

[0052] Calculating, based on a plurality of preset influencing factors, an influence conversion value corresponding to each of the first influencing pixels and the first pixel point;

[0053] The offset results and the impact conversion values ​​corresponding to all the first influencing pixels are determined as first impact information between the first pixel point and all the first influencing pixels corresponding to it.

[0054] As an optional implementation, in the second aspect of the present application, the generation module generates a target compensation data table corresponding to each first pixel point based on all the first impact information corresponding to the first pixel point, specifically including:

[0055] For each of the first pixel points, performing a correction operation on each of the impact conversion values ​​of the first pixel point according to a preset correction parameter to obtain a correction result corresponding to each of the impact conversion values;

[0056] storing the correction result corresponding to each first-affected pixel at the storage address represented by the offset result corresponding to the first-affected pixel, to obtain a compensation table corresponding to all the first-affected pixels;

[0057] performing image sticking training correction on a training image set according to the compensation table to obtain a training correction result of the image set;

[0058] When the training correction result indicates that the preset afterimage correction standard is reached, the compensation table is determined to be a target compensation data table.

[0059] As an optional implementation manner, in the second aspect of the present application, the generation module generates a target compensation data table corresponding to each first pixel point based on all the first impact information corresponding to the first pixel point, further comprising:

[0060] When the training correction result indicates that the preset ghosting correction standard is not met, repeatedly adjusting the correction parameters to update the correction result and the compensation table;

[0061] After each update of the compensation table, the operation corresponding to the image afterimage training correction is triggered to be performed on the training image set according to the preset AI simulation algorithm combined with the compensation table to obtain the training correction result of the image set, until it is determined that the training correction result after a certain update indicates that the afterimage correction standard is met, and the compensation table corresponding to the training correction result after a certain update is determined to be the target compensation data table.

[0062] As an optional implementation, in the second aspect of the present application, the generation module performs image sticking training correction on the training image set according to the compensation table, and the method of obtaining the training correction result of the image set specifically includes:

[0063] Selecting any frame of image from the training image set as a training image; performing an image afterimage correction operation on the training image according to the compensation table to obtain a corrected image of the training image;

[0064] Dividing the training image and the correction image into the same preset number of sub-images respectively;

[0065] Taking the sub-image as a unit, calculating the target difference value between the training image and the correction image under the sub-image; and calculating the square average of all the target difference values ​​to obtain the target comparison value between the training image and the correction image;

[0066] The target difference value includes a grayscale difference value and a similarity difference value, and the target comparison value includes a first square root corresponding to the grayscale difference value and a second square root corresponding to the similarity difference value;

[0067] The first square root and the second square root are multiplied by their corresponding weighted averages respectively and the averages are calculated to obtain a weighted average result of the training image and the correction image as a training correction result.

[0068] As an optional implementation, in the second aspect of the present application, the afterimage removal module performs an afterimage removal operation on the second image data according to all the second impact information and the target compensation data table, and obtains the afterimage-removed image of the second image in a manner specifically including:

[0069] According to all the second impact information, indexing the target compensation data corresponding to each second pixel in the target compensation data table;

[0070] Performing a data merging operation on the target compensation data corresponding to each second pixel point and the second image data to obtain a data merging result of all the second pixel points as an afterimage-eliminating image of the second image;

[0071] The afterimage elimination image is used to be transmitted to a display module, and the display module performs a afterimage elimination refresh operation on the second image currently being refreshed using the afterimage elimination image.

[0072] The third aspect of the present application discloses another device for eliminating residual image of electronic paper based on a large model, the device comprising:

[0073] a memory storing executable program code;

[0074] a processor coupled to the memory;

[0075] The processor calls the executable program code stored in the memory to execute the large model-based electronic paper afterimage elimination method disclosed in the first aspect of the present application.

[0076] The fourth aspect of the present application discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute the large model-based electronic paper afterimage elimination method disclosed in the first aspect of the present application.

[0077] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0078] In an embodiment of the present application, a method for eliminating electronic paper afterimages based on a large model is provided, the method including: obtaining first image data of a first image and second image data of a second image; the second image is a subsequent frame image of the first image; the first image and the second image are both subjected to preset grayscale image processing; according to the first image data, calculating the first impact information between each first pixel point in the first image and each first influencing pixel corresponding to it; and generating a target compensation data table corresponding to the first pixel point based on all the first impact information corresponding to each first pixel point; the target compensation data table includes compensation data corresponding to each first impact information; according to the second image data, determining the second impact information between each second pixel point in the second image and all the second influencing pixels corresponding to it; performing an afterimage elimination operation on the second image data based on all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image. It can be seen that the implementation of this application can automatically analyze and calculate all the first impact information corresponding to the previous frame image (first image) of the currently refreshed image (second image) according to the set image processing process, thereby generating a target compensation data table based on the first impact information. The target compensation data table can specifically be a mask compensation table; then, based on the mask compensation table, the currently refreshed image frame is subjected to afterimage elimination. In other words, this application predicts the distribution of afterimages of the refreshed frame image by analyzing the features of the previous and next frame images; then, the mask compensation table required for afterimage elimination is generated through pre-trained parameters, and afterimage elimination is performed using the mask compensation table, thereby improving the efficiency of afterimage elimination of the electronic paper screen and improving the afterimage elimination effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0080] FIG1 is a flow chart of a method for eliminating afterimages of an electronic paper based on a large model disclosed in an embodiment of the present application;

[0081] FIG2 is a flow chart of another method for eliminating electronic paper afterimages based on a large model disclosed in an embodiment of the present application;

[0082] FIG3 is a schematic structural diagram of an electronic paper afterimage elimination device based on a large model disclosed in an embodiment of the present application;

[0083] FIG4 is a schematic structural diagram of another electronic paper afterimage elimination device based on a large model disclosed in an embodiment of the present application;

[0084] FIG5 is a schematic structural diagram of a central particle and its particle cloud disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0085] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0086] The terms "first," "second," and so on in the specification and claims of this application and the accompanying drawings are used to distinguish between different objects, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or end comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or end.

[0087] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0088] The present application discloses a method and device for eliminating afterimages of an electronic paper based on a large model, which can automatically analyze and calculate all first impact information corresponding to the previous frame image (first image) of the currently refreshed image (second image) according to a set image processing process, thereby generating a target compensation data table based on the first impact information, and the target compensation data table can specifically be a mask compensation table; then, based on the mask compensation table, afterimage elimination is performed on the currently refreshed image frame. That is, the present application predicts the distribution of afterimages of the refreshed frame image by analyzing the features of the previous and next frame images; then, the mask compensation table required for afterimage elimination is generated through pre-trained parameters, and afterimage elimination is performed through the mask compensation table, thereby improving the efficiency of afterimage elimination of the electronic paper screen and improving the afterimage elimination effect. The following are detailed descriptions.

[0089] Example 1

[0090] Please refer to Figure 1, which is a flow chart of a method for eliminating electronic paper afterimages based on a large model disclosed in an embodiment of the present application, wherein the large model may be a MASK algorithm model, which is not limited in the embodiment of the present application. The method for eliminating electronic paper afterimages based on a large model described in Figure 1 can be applied to an electronic paper afterimage elimination device based on a large model. Specifically, the electronic paper afterimage elimination device may be an electrophoretic display device capable of implementing electrophoretic display technology, which is not limited in the embodiment of the present application. As shown in Figure 1, the method for eliminating electronic paper afterimages based on a large model may include the following operations:

[0091] 101. Acquire first image data of a first image and second image data of a second image; the second image is an image frame subsequent to the first image.

[0092] In the embodiment of the present application, both the first image and the second image undergo preset grayscale image processing.

[0093] In the embodiment of the present application, optionally, before executing step 101, the method further includes:

[0094] Acquire at least two frames of images that need to be displayed on a display device, where the two frames of images are adjacent in display order;

[0095] Performing grayscale image processing on the two frames of image to obtain grayscale image processing results of the two frames of image, wherein the preceding frame image is recorded as the first image and the succeeding frame image is recorded as the second image; then triggering execution of step 101;

[0096] The first image and the second image may be automatically generated as N*N 8-bit 256 grayscale images according to a set image format by a computing device, where N≥3.

[0097] 102. Calculate, based on the first image data, first influence information between each first pixel point in the first image and each corresponding first influencing pixel.

[0098] In the embodiment of the present application, the first pixel in the first image may be an electrophoretic particle encapsulated in an electronic paper capsule. The electronic paper capsule may be positively charged, negatively charged, or neutral, and is a display unit of a basic pixel.

[0099] In the embodiment of the present application, the electronic paper capsule may be specifically encapsulated with electrophoretic particles using microencapsulation technology. After encapsulation, the electronic paper capsule corresponding to the first pixel is fixed and immobile. Furthermore, the electronic paper capsule is filled with an electrophoretic fluid containing a large number of electrophoretic particles (i.e., the first pixel). By applying an electric field force to all of the electrophoretic particles, all of the electrophoretic particles are able to move within the electronic paper capsule.

[0100] It should be noted that, in addition to the microcapsule technology, the packaging technology for the above-mentioned electronic paper capsule can also use microcup technology to encapsulate and fill electrophoretic particles; or it can be a method of forming a cofferdam around a single pixel point to manufacture a patterned structure of the cofferdam, so that the electrode unit of each pixel point is surrounded by a plasma isolation dam, which is not limited in the embodiments of the present application.

[0101] In the embodiment of the present application, the method of calculating the first influence information between each first pixel point in the first image and each corresponding first influencing pixel thereof according to the first image data in step 102 specifically includes:

[0102] For each first pixel point in the first image, determining a plurality of influencing directions that have an afterimage influence on the first pixel point;

[0103] Taking the coordinates of the first pixel point as the center, determine the first influencing pixel corresponding to each influencing direction;

[0104] Perform a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point;

[0105] The impact calculation operation includes grayscale difference calculation, particle offset and impact factor conversion.

[0106] In the embodiments of the present application, specifically, please refer to FIG5 , which is a schematic diagram of the structure of a central particle and its particle cloud disclosed in the embodiments of the present application. The coordinates of each selected first pixel are labeled x1y1; the eight influence directions corresponding to the first pixel are upper left, upper, upper right, left, right, lower left, lower, and lower right. Furthermore, the pixel coordinates of the first influence pixels corresponding to these eight influence directions are recorded as x0y0, x1y0, x2y0, x0y1, x2y1, x2y2, x1y2, and x2y2, respectively.

[0107] After determining the first influencing pixels corresponding to each influencing orientation, first influence information between each first pixel point and all the first influencing pixels corresponding to it is calculated according to a preset influence calculation process.

[0108] 103. Generate a target compensation data table corresponding to each first pixel point according to all first impact information corresponding to the first pixel point.

[0109] In the embodiment of the present application, the target compensation data table includes compensation data corresponding to each first impact information.

[0110] 104. Determine, based on the second image data, second influence information between each second pixel point in the second image and all corresponding second influencing pixels.

[0111] In the embodiment of the present application, the structure of the second pixel point is basically the same as that of the first pixel point (both are electronic paper capsules), and the embodiment of the present application is not described here in detail.

[0112] In the embodiment of the present application, similarly, the method for determining all second impact information corresponding to the second pixel point in step 104 is similar to the method for determining all first impact information corresponding to each first pixel point in step 102 .

[0113] Specifically, eight pixel points in the influencing directions around the second pixel point are selected and recorded as second influencing pixels. By performing the influence calculation operation on the pixel coordinates of the second pixel point and the pixel coordinates of each second influencing pixel, the second influence information between the second pixel point and all its corresponding second influencing pixels is obtained.

[0114] 105. Perform an afterimage elimination operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image.

[0115] In the embodiment of the present application, in actual application, after executing step 103, the generated target compensation data table can be loaded into the display device (such as the electrophoretic display device mentioned above) as refresh data, and the display device loads the WFM waveform file at the same time; at the same time, the display device receives the image refresh data, and then triggers the execution of step 104. Compared with the traditional drive, the use of the WFM waveform file adds a time-attenuated compensation drive. In addition, by applying the large model-based electronic paper afterimage elimination method described in the embodiment of the present application to the electrophoretic display device, the afterimage on the electrophoretic display device can be significantly eliminated, and the user's reading experience of the electrophoretic display device is improved; in addition, the embodiment of the present application is to refresh the entire screen image as a whole, and use the pre-generated mask compensation matrix table (target compensation data table) to determine the pixels that need to be refreshed and the refreshed data. No GC global cleanup is used, and no partitioned local refresh method is used.

[0116] In an embodiment of the present application, the WFM waveform file may be an OED electronic paper driving waveform file; wherein, the target compensation data table may be stored in the disk and memory in the form of a waveform file, which is not limited in the embodiment of the present application.

[0117] It can be seen that the implementation of the large-scale model-based electronic paper afterimage elimination method described in FIG1 can automatically analyze and calculate all first impact information corresponding to the previous frame image (first image) of the currently refreshed image (second image) according to the set image processing process, thereby generating a target compensation data table based on the first impact information. The target compensation data table can specifically be a mask compensation table; then, the currently refreshed image frame is subjected to afterimage elimination based on the mask compensation table. That is, the present application predicts the afterimage distribution of the refreshed frame image by analyzing the features of the previous and next frame images; then, the mask compensation table required for afterimage elimination is generated through pre-trained parameters, and afterimage elimination is performed through the mask compensation table, thereby improving the efficiency of afterimage elimination of the electronic paper screen and improving the afterimage elimination effect.

[0118] In an optional embodiment, the method of performing a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all first influencing pixels to obtain the first influence information between the first pixel point and all its corresponding first influencing pixels specifically includes:

[0119] Inputting the pixel coordinates of each first-affected pixel and the pixel coordinates of the first pixel into a preset grayscale difference calculation formula to obtain the grayscale difference between each first-affected pixel and the first pixel;

[0120] For each grayscale difference corresponding to the first affected pixel, offset it according to its predetermined offset bit amount to obtain an offset result corresponding to each first affected pixel;

[0121] Calculating the impact conversion value corresponding to each first impact pixel and the first pixel point according to a plurality of preset impact factors;

[0122] The offset results and the impact conversion values ​​corresponding to all first-influenced pixels are determined as first impact information between the first pixel point and all the first-influenced pixels corresponding to it.

[0123] In this optional embodiment, specifically, the pixel coordinates of the first pixel point pc are denoted as x1y1; p1-p8 are used to represent all first-affecting pixels; the pixel coordinates of each first-affecting pixel and the pixel coordinates of the first pixel point are input into a preset grayscale difference calculation formula. The specific grayscale difference calculation formula is: Wpn=Fw(pn,pc), [0,255]

[0124] Among them, pn corresponds to the pixel coordinates of the first influencing pixel of p1-p8, and pc is the pixel coordinates of the first pixel point.

[0125] In this optional embodiment, after the grayscale difference between each first influencing pixel and the first pixel is obtained by the above calculation, the particle cloud corresponding to the first pixel in different states is represented and stored using a numerical value. That is, the grayscale difference corresponding to each first influencing pixel is offset according to its predetermined offset bit amount. The offset result corresponding to each first influencing pixel specifically includes:

[0126] Take the top 2 bits of W1 and shift them left by 14 bits; take the top 2 bits of W2 and shift them left by 12 bits; take the top 2 bits of W3 and shift them left by 10 bits; take the top 2 bits of W4 and shift them left by 8 bits; take the top 2 bits of W5 and shift them left by 6 bits; take the top 2 bits of W6 and shift them left by 4 bits; take the top 2 bits of W7 and shift them left by 2 bits; take the top 2 bits of W8 and do not shift them. Then, combine the above shifted results of W1-W8 and use an OR operation. The resulting union is the numerical description of the particle cloud at the first pixel in a certain state, denoted as V. Each first pixel corresponds to the state of a particle cloud cluster, and the number of states of all particle clouds ranges from 0 to 65535.

[0127] In this optional embodiment, after calculating the information of 0-65535 different cloud states, it is necessary to consider multiple influencing factors to calculate the influence value of each first influencing pixel and the first pixel point; wherein, the multiple influencing factors may include the distance, relative angle, and driving voltage between each first influencing pixel and the first pixel point; further, the distance between each first influencing pixel and the first pixel point is negatively correlated, that is, the farther the distance between the two pixel points is, the smaller the influence relationship is; the relative angle between the two pixel points is negatively correlated, that is, the larger the relative angle between the two pixel points is, the smaller the influence relationship is; the two pixel points are positively correlated with the driving voltage, that is, the larger the driving voltage of the two pixel points is, the greater the influence relationship is.

[0128] Specifically, the calculation formula for actually calculating the influence value of each first influencing pixel and the first pixel point can be: Gpn-pc=Fg (influence factor)

[0129] The influencing factor corresponds to the distance, relative angle, and driving voltage between each first influencing pixel and the first pixel point mentioned above.

[0130] It can be seen that in this optional embodiment, the afterimage displayed on the electronic paper display screen is mainly due to the mutual influence between pixel particles. In this optional embodiment, the influence of any pixel point on the surrounding pixel points is digitally calculated, and then described and stored in the form of particle cloud values, which is convenient for subsequent calls and calculations.

[0131] In this optional embodiment, further, generating a target compensation data table corresponding to each first pixel point based on all first impact information corresponding to the first pixel point includes:

[0132] For each first pixel point, performing a correction operation on each impact conversion value of the first pixel point according to a preset correction parameter to obtain a correction result corresponding to each impact conversion value;

[0133] storing the correction result corresponding to each first-affected pixel at the storage address represented by the offset result corresponding to the first-affected pixel, and obtaining a compensation table corresponding to all first-affected pixels;

[0134] Performing image ghosting training correction on the training image set according to the compensation table to obtain a training correction result of the image set;

[0135] When the training correction result indicates that the preset ghost correction standard is reached, the compensation table is determined to be a target compensation data table.

[0136] In an optional embodiment, after all first impact information corresponding to each pixel is calculated, it is necessary to adjust all first impact information using a correction parameter. The specific correction calculation formula may be: D = Fd (correction parameter, Gpn - pc)

[0137] Among them, Gpn-pc corresponds to the numerical values ​​of all the above-mentioned first influence information. Since the above-mentioned particle cloud value V has 0-65535 different cloud cluster states, the Gpn-pc corresponds to the influence numerical values ​​(that is, the first influence information) of 0-65535 different cloud cluster states; that is, all the first influence information under each cloud cluster state is used to refer to the influence value of the cloud cluster composed of all the first influencing pixels around the first pixel point on the central pixel (the first pixel point).

[0138] In an optional embodiment, the correction parameters are calculated by a designed AI model, or are manually adjusted by setting the parameters; D corresponds to the compensation data ultimately required to eliminate afterimages of the current frame image refreshed in real time (that is, the correction results of each of the above-mentioned factors affecting the conversion value).

[0139] In this optional embodiment, after calculating the correction result / compensation data D corresponding to each first influencing pixel, the compensation data D is used as the actual pixel information and stored at the storage address represented by the offset result V corresponding to the first influencing pixel, and a compensation table corresponding to all first influencing pixels is preliminarily obtained.

[0140] Furthermore, after generating the compensation table, the image afterimage elimination effect of the compensation table can be trained by simulating a big data image set through AI. If the training correction result meets the afterimage correction standard, the latest compensation table corresponding to the current correction parameters will be determined as the target compensation data table that can be put into practical application.

[0141] In this optional embodiment, the method further includes:

[0142] When the training correction result indicates that the preset ghosting correction standard is not met, the correction parameters are repeatedly adjusted to update the correction result and the compensation table;

[0143] After each update of the compensation table, the above-mentioned training correction of image afterimages is triggered to be performed on the training image set according to the preset AI simulation algorithm combined with the compensation table to obtain the operation corresponding to the training correction result of the image set, until it is determined that the training correction result after a certain update indicates that the afterimage correction standard is met, and the compensation table corresponding to the training correction result after a certain update is determined to be the target compensation data table.

[0144] In this optional embodiment, if the training correction result does not meet the afterimage correction standard, the correction parameters can be continuously adjusted through AI or manually. The corresponding compensation table will be updated accordingly with the adjustment of the correction parameters. The image afterimage elimination effect of the compensation table is trained using AI to simulate a big data image set, and the corresponding training correction results will also be updated accordingly.

[0145] It can be seen that in this optional embodiment, after the corresponding compensation table is preliminarily generated through the first impact information of each first pixel point, a corresponding correction training process is set up, which can repeatedly train and correct the compensation table, thereby improving the generation accuracy and practicality of the final target compensation data table.

[0146] In yet another optional embodiment, the method of performing image sticking training correction on the training image set according to the compensation table to obtain the training correction result of the image set specifically includes:

[0147] Select any frame of image from the training image set and record it as a training image; perform an image ghosting correction operation on the training image according to the compensation table to obtain a corrected image of the training image;

[0148] The training image and the calibration image are divided into the same preset number of sub-images respectively;

[0149] Taking the sub-image as the unit, calculate the target difference value between the training image and the correction image under the sub-image; and calculate the square average of all target difference values ​​to obtain the target comparison value between the training image and the correction image;

[0150] The target difference value includes a grayscale difference value and a similarity difference value, and the target comparison value includes a first square root corresponding to the grayscale difference value and a second square root corresponding to the similarity difference value;

[0151] The first square root and the second square root are multiplied by their respective corresponding weighted averages and the average is calculated to obtain the weighted average result of the training image and the correction image as the training correction result.

[0152] In this optional embodiment, any frame image is selected from the AI ​​simulated big data image set as a test image (training image) before image afterimage correction is performed through the compensation table, and the image after image afterimage correction is performed through the compensation table is recorded as a correction image. The grayscale values ​​of the training image and the correction image are measured by an optical instrument, and then the grayscale values ​​and similarities of the training image and the correction image are compared through AI, and finally the required training correction result is obtained by calculating the square root and the weighted average result.

[0153] It can be seen that in this optional embodiment, the training correction result is comprehensively calculated by calculating the grayscale difference value and the similarity difference value, combined with the acquisition of weighted average data, which is beneficial to improving the accuracy of the determination of the training correction result and the practicality and accuracy of the target compensation data table finally obtained; in addition, in this optional embodiment, the overall calculation process is different from the fixed compensation mode of the existing algorithm. The calculation process uses a dynamic matrix weighting method to consider different grayscale effects (0-65535) on the affected pixels, and uses different compensation matrices to perform white value or black value compensation, which improves the applicability of the overall calculation process and the accuracy of the calculation results. The dynamic matrix is ​​an N*M matrix, N and M can be the same, N and M are both greater than 2, and is obtained by training the computer debugging model at different grayscales.

[0154] Example 2

[0155] Please refer to Figure 2, which is a flow chart of another method for eliminating electronic paper afterimages based on a large model disclosed in an embodiment of the present application, wherein the large model may be a MASK algorithm model, which is not limited in the embodiment of the present application. The method for eliminating electronic paper afterimages based on a large model described in Figure 2 can be applied to an electronic paper afterimage elimination device based on a large model. Specifically, the electronic paper afterimage elimination device may be an electrophoretic display device capable of implementing electrophoretic display technology, which is not limited in the embodiment of the present application. As shown in Figure 2, the method for eliminating electronic paper afterimages based on a large model may include the following operations:

[0156] 201. Acquire first image data of a first image and second image data of a second image; the second image is an image frame subsequent to the first image.

[0157] 202. Calculate, based on the first image data, first influence information between each first pixel point in the first image and each corresponding first influencing pixel.

[0158] 203 . Generate a target compensation data table corresponding to each first pixel point according to all first impact information corresponding to the first pixel point.

[0159] 204. Determine, based on the second image data, second influence information between each second pixel point in the second image and all of its corresponding second influencing pixels.

[0160] In the embodiment of the present application, for other descriptions of steps 201 to 204, please refer to the other specific descriptions of steps 101 to 104 in embodiment 1, which will not be repeated in this embodiment of the present application.

[0161] 205. According to all the second impact information, index the target compensation data corresponding to each second pixel point in the target compensation data table.

[0162] 206 . Perform a data merging operation on the target compensation data corresponding to each second pixel point and the second image data to obtain a data merging result of all second pixel points as an afterimage-eliminated image of the second image.

[0163] In an embodiment of the present application, the ghosting elimination image is used to be transmitted to the display module, which performs a ghosting elimination refresh operation on the currently refreshed second image using the ghosting elimination image. A cross-rendering technique may be used during display driving to further eliminate the influence between pixels.

[0164] It can be seen that the electronic paper afterimage elimination method based on the large model described in Figure 2 is implemented. After the first impact information of the first image is calculated and the target compensation data table is generated, when there is an actual refreshed second image, the pre-generated target compensation data table is used as an index to directly index the calculated second impact information of the second image in the target compensation data table, and the indexed target compensation data is merged with the second image data to achieve the afterimage elimination of the refreshed image. The setting of the pre-generated target compensation data table and the index call of the data improve the processing efficiency and accuracy of the afterimage elimination.

[0165] Example 3

[0166] Please refer to Figure 3, which is a structural diagram of an electronic paper afterimage elimination device based on a large model disclosed in an embodiment of the present application. Among them, the electronic paper afterimage elimination device based on a large model can be an electronic paper afterimage elimination terminal, device, system or server based on a large model. The server can be a local server, a remote server, or a cloud server (also known as a cloud server). When the server is a non-cloud server, the non-cloud server can communicate with the cloud server, which is not limited in the embodiment of the present application. As shown in Figure 3, the electronic paper afterimage elimination device based on a large model can include an acquisition module 301, a calculation module 302, a generation module 303, a determination module 304 and an afterimage elimination module 305, wherein:

[0167] The acquisition module 301 is used to acquire first image data of a first image and second image data of a second image; the second image is an image frame subsequent to the first image; and both the first image and the second image are subjected to a preset grayscale image processing.

[0168] The calculation module 302 is configured to calculate first influence information between each first pixel point in the first image and each first influence pixel corresponding thereto according to the first image data.

[0169] In the embodiment of the present application, the calculation module 302 calculates the first influence information between each first pixel in the first image and each corresponding first influencing pixel based on the first image data, specifically including:

[0170] For each first pixel point in the first image, determining a plurality of influencing directions that have an afterimage influence on the first pixel point;

[0171] Taking the coordinates of the first pixel point as the center, determine the first influencing pixel corresponding to each influencing direction;

[0172] Perform a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point;

[0173] The impact calculation operation includes grayscale difference calculation, particle offset and impact factor conversion.

[0174] The generating module 303 is configured to generate a target compensation data table corresponding to each first pixel point according to all first impact information corresponding to the first pixel point; the target compensation data table includes compensation data corresponding to each first impact information.

[0175] The determination module 304 is configured to determine, based on the second image data, second influence information between each second pixel point in the second image and all corresponding second influence pixels.

[0176] The afterimage elimination module 305 is configured to perform an afterimage elimination operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image.

[0177] It can be seen that the electronic paper afterimage elimination device based on the large model described in FIG3 can automatically analyze and calculate all the first influence information corresponding to the first image for the previous frame image (first image) of the currently refreshed image (second image) according to the set image processing process, thereby generating a target compensation data table based on the first influence information, and the target compensation data table can specifically be a mask compensation table; then, the currently refreshed image frame is subjected to afterimage elimination based on the mask compensation table. That is, the present application realizes the prediction of the afterimage distribution of the refreshed frame image by analyzing the features of the previous and next frame images; then, the mask compensation table required for afterimage elimination is generated through pre-trained parameters, and afterimage elimination is performed through the mask compensation table, thereby improving the efficiency of afterimage elimination of the electronic paper screen and improving the effect of afterimage elimination; in addition, the embodiment of the present application performs an overall refresh on the entire screen image, and uses a pre-generated mask compensation matrix table (target compensation data table) to determine the pixels that need to be refreshed and the refreshed data. No GC global cleanup is used, nor is a partitioned local refresh method.

[0178] In an optional embodiment, the calculation module 302 performs a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all first influencing pixels to obtain the first influence information between the first pixel point and all the first influencing pixels corresponding thereto, specifically including:

[0179] Inputting the pixel coordinates of each first-affected pixel and the pixel coordinates of the first pixel into a preset grayscale difference calculation formula to obtain the grayscale difference between each first-affected pixel and the first pixel;

[0180] For each grayscale difference corresponding to the first affected pixel, offset it according to its predetermined offset bit amount to obtain an offset result corresponding to each first affected pixel;

[0181] Calculating the impact conversion value corresponding to each first impact pixel and the first pixel point according to a plurality of preset impact factors;

[0182] The offset results and the impact conversion values ​​corresponding to all first-influenced pixels are determined as first impact information between the first pixel point and all the first-influenced pixels corresponding to it.

[0183] It can be seen that in an optional embodiment, the afterimage displayed on the electronic paper display screen is mainly due to the mutual influence between pixel particles. In this optional embodiment, the influence of any pixel point on the surrounding pixels is digitally calculated, and then described and stored in the form of particle cloud values, which is convenient for subsequent calls and calculations.

[0184] In another optional embodiment, the generation module 303 generates a target compensation data table corresponding to each first pixel point according to all first impact information corresponding to the first pixel point, specifically including:

[0185] For each first pixel point, performing a correction operation on each impact conversion value of the first pixel point according to a preset correction parameter to obtain a correction result corresponding to each impact conversion value;

[0186] storing the correction result corresponding to each first-affected pixel at the storage address represented by the offset result corresponding to the first-affected pixel, and obtaining a compensation table corresponding to all first-affected pixels;

[0187] Performing image ghosting training correction on the training image set according to the compensation table to obtain a training correction result of the image set;

[0188] When the training correction result indicates that the preset ghosting correction standard is reached, the compensation table is determined to be a target compensation data table.

[0189] In this optional embodiment, the method in which the generating module 303 generates the target compensation data table corresponding to each first pixel point according to all first impact information corresponding to the first pixel point specifically further includes:

[0190] When the training correction result indicates that the preset ghosting correction standard is not met, the correction parameters are repeatedly adjusted to update the correction result and the compensation table;

[0191] After each update of the compensation table, the above-mentioned training correction of image afterimages is triggered to be performed on the training image set according to the preset AI simulation algorithm combined with the compensation table to obtain the operation corresponding to the training correction result of the image set, until it is determined that the training correction result after a certain update indicates that the afterimage correction standard is met, and the compensation table corresponding to the training correction result after a certain update is determined to be the target compensation data table.

[0192] It can be seen that in an optional embodiment, after the corresponding compensation table is preliminarily generated through the first impact information of each first pixel point, a corresponding correction training process is set up, which can repeatedly train and correct the compensation table, thereby improving the generation accuracy and practicality of the final target compensation data table.

[0193] In another optional embodiment, the generation module 303 performs image ghosting training correction on the training image set according to the compensation table, and obtains the training correction result of the image set in a manner specifically including:

[0194] Select any frame of image from the training image set and record it as a training image; perform an image ghosting correction operation on the training image according to the compensation table to obtain a corrected image of the training image;

[0195] The training image and the calibration image are divided into the same preset number of sub-images respectively;

[0196] Taking the sub-image as the unit, calculate the target difference value between the training image and the correction image under the sub-image; and calculate the square average of all target difference values ​​to obtain the target comparison value between the training image and the correction image;

[0197] The target difference value includes a grayscale difference value and a similarity difference value, and the target comparison value includes a first square root corresponding to the grayscale difference value and a second square root corresponding to the similarity difference value;

[0198] The first square root and the second square root are multiplied by their respective corresponding weighted averages and the average is calculated to obtain the weighted average result of the training image and the correction image as the training correction result.

[0199] It can be seen that in this optional embodiment, the training correction result is comprehensively calculated by calculating the grayscale difference value and the similarity difference value, combined with the acquisition of weighted average data, which is beneficial to improving the accuracy of the determination of the training correction result and the practicality and accuracy of the target compensation data table finally obtained; in addition, in this optional embodiment, the overall calculation process is different from the fixed compensation mode of the existing algorithm. The calculation process uses a dynamic matrix weighting method to consider the different grayscale effects (0-65535) on the affected pixels, and uses different compensation matrices to perform white value or black value compensation, thereby improving the applicability of the overall calculation process and the accuracy of the calculation results.

[0200] In another optional embodiment, the afterimage removal module 305 performs an afterimage removal operation on the second image data according to all the second impact information and the target compensation data table to obtain the afterimage-removed image of the second image, specifically including:

[0201] According to all the second impact information, indexing the target compensation data corresponding to each second pixel point in the target compensation data table;

[0202] Performing a data merging operation on the target compensation data corresponding to each second pixel point and the second image data to obtain a data merging result of all second pixel points as a ghost-eliminating image of the second image;

[0203] The afterimage elimination image is used to be transmitted to the display module, and the display module performs a afterimage elimination refresh operation on the currently refreshed second image using the afterimage elimination image.

[0204] It can be seen that in this optional embodiment, after the first impact information of the first image is calculated and the target compensation data table is generated, when there is a second image that is actually refreshed, the pre-generated target compensation data table is used as an index to directly index the calculated second impact information of the second image in the target compensation data table, and the indexed target compensation data is merged with the second image data to achieve the elimination of the afterimage of the refreshed image. The setting of the pre-generated target compensation data table and the index call of the data improve the processing efficiency and accuracy of the afterimage elimination.

[0205] Example 4

[0206] Please refer to Figure 4, which is a schematic diagram of the structure of another large-scale model-based electronic paper afterimage elimination device disclosed in an embodiment of the present application. As shown in Figure 4, the large-scale model-based electronic paper afterimage elimination device may include:

[0207] A memory 401 storing executable program code;

[0208] a processor 402 coupled to the memory 401;

[0209] The processor 402 calls the executable program code stored in the memory 401 to execute the steps of the large model-based electronic paper afterimage elimination method described in the first embodiment of the present application or the second embodiment of the present application.

[0210] Example 5

[0211] An embodiment of the present application discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute the steps of the large model-based electronic paper afterimage elimination method described in Example 1 or Example 2 of the present application.

[0212] Example 6

[0213] An embodiment of the present application discloses a computer program product, which includes a non-transitory computer storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps of the large model-based electronic paper afterimage elimination method described in Example 1 or Example 2.

[0214] The device embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0215] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0216] Finally, it should be noted that the method and device for eliminating electronic paper afterimages based on a large model disclosed in the embodiments of the present application are only preferred embodiments of the present application, and are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A method for eliminating afterimage of electronic paper based on a large model, characterized in that: The method comprises: Acquire first image data of a first image and second image data of a second image; the second image is an image frame subsequent to the first image; the first image and the second image are both subjected to a preset grayscale image processing; According to the first image data, first influence information between each first pixel point in the first image and each first influence pixel corresponding thereto is calculated; and a target compensation data table corresponding to the first pixel point is generated according to all the first influence information corresponding to each first pixel point; the target compensation data table includes compensation data corresponding to each first influence information; Determine, according to the second image data, second influence information between each second pixel point in the second image and all second influence pixels corresponding to it; According to all the second impact information and the target compensation data table, an afterimage elimination operation is performed on the second image data to obtain an afterimage eliminated image of the second image.

2. The method for eliminating residual image of electronic paper based on a large model according to claim 1, characterized in that: The step of calculating, according to the first image data, first influence information between each first pixel point in the first image and each first influence pixel corresponding to the first pixel point, comprises: For each first pixel point in the first image, determining a plurality of influencing directions that have an afterimage influence on the first pixel point; Taking the coordinates of the first pixel point as the center, determining the first influencing pixel corresponding to each influencing orientation; Perform a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point; The impact calculation operation includes grayscale difference calculation, particle offset and impact factor conversion.

3. The method for eliminating afterimage of electronic paper based on a large model according to claim 1 or 2, characterized in that: The performing a preset influence calculation operation on the pixel coordinates of the first pixel point and the pixel coordinates of all the first influencing pixels to obtain first influence information between the first pixel point and all the first influencing pixels corresponding to the first pixel point includes: Inputting the pixel coordinates of each of the first influencing pixels and the pixel coordinates of the first pixel point into a preset grayscale difference calculation formula to obtain a grayscale difference between each of the first influencing pixels and the first pixel point; For each of the grayscale differences corresponding to the first influencing pixels, shifting is performed according to a predetermined shift bit amount to obtain a shift result corresponding to each of the first influencing pixels; Calculating, according to a plurality of preset influencing factors, an influencing conversion value corresponding to each of the first influencing pixels and the first pixel point; The offset results and the impact conversion values ​​corresponding to all the first impact pixels are determined as first impact information between the first pixel point and all the first impact pixels corresponding to it.

4. The method for eliminating residual image of electronic paper based on a large model according to claim 3, characterized in that: The step of generating a target compensation data table corresponding to each first pixel point according to all the first impact information corresponding to the first pixel point includes: For each of the first pixel points, a correction operation is performed on each of the impact conversion values ​​of the first pixel point according to a preset correction parameter to obtain a correction result corresponding to each of the impact conversion values; storing the correction result corresponding to each first-affected pixel at the storage address represented by the offset result corresponding to the first-affected pixel, to obtain a compensation table corresponding to all the first-affected pixels; According to the compensation table, performing training correction of image afterimage on the training image set to obtain a training correction result of the image set; When the training correction result indicates that the preset afterimage correction standard is reached, the compensation table is determined to be a target compensation data table.

5. The method for eliminating residual image of electronic paper based on a large model according to claim 4, characterized in that: The method further comprises: When the training correction result indicates that the preset afterimage correction standard is not reached, repeatedly adjusting the correction parameters to update the correction result and the compensation table; After each update of the compensation table, the operation corresponding to the training correction of image afterimage is performed on the training image set according to the preset AI simulation algorithm combined with the compensation table, and the training correction result of the image set is obtained, until it is determined that the training correction result after a certain update indicates that the afterimage correction standard is met, and the compensation table corresponding to the training correction result after a certain update is determined to be the target compensation data table.

6. The method for eliminating residual image of electronic paper based on a large model according to claim 4, characterized in that: The step of performing training correction of image afterimage on a training image set according to the compensation table to obtain a training correction result of the image set includes: Select any frame of image from the training image set and record it as a training image; perform an image afterimage correction operation on the training image according to the compensation table to obtain a corrected image of the training image; Dividing the training image and the correction image into the same preset number of sub-images respectively; Taking the sub-image as a unit, calculating the target difference value between the training image and the correction image under the sub-image; and calculating the square average of all the target difference values ​​to obtain the target comparison value between the training image and the correction image; The target difference value includes a grayscale difference value and a similarity difference value, and the target comparison value includes a first square root corresponding to the grayscale difference value and a second square root corresponding to the similarity difference value; The first square root and the second square root are multiplied by their corresponding weighted averages respectively and the averages are calculated to obtain a weighted average result of the training image and the correction image as a training correction result.

7. The method for eliminating residual image of electronic paper based on a large model according to claim 1, 2, 4, 5 or 6, characterized in that: The performing an afterimage elimination operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image includes: According to all the second impact information, indexing the target compensation data corresponding to each of the second pixels in the target compensation data table; Performing a data merging operation on the target compensation data corresponding to each second pixel point and the second image data to obtain a data merging result of all the second pixel points as an afterimage elimination image of the second image; The afterimage elimination image is used to be transmitted to a display module, and the display module performs a afterimage elimination refresh operation on the second image currently being refreshed by using the afterimage elimination image.

8. An electronic paper afterimage elimination device based on a large model, characterized in that: The device comprises: An acquisition module, used for acquiring first image data of a first image and second image data of a second image; the second image is an image of a subsequent frame of the first image; the first image and the second image are both subjected to a preset grayscale image processing; a calculation module, configured to calculate first influence information between each first pixel point in the first image and each first influence pixel corresponding to the first pixel point according to the first image data; A generating module, configured to generate a target compensation data table corresponding to each first pixel point according to all the first impact information corresponding to each first pixel point; the target compensation data table includes compensation data corresponding to each first impact information; a determination module, configured to determine, according to the second image data, second influence information between each second pixel point in the second image and all second influence pixels corresponding to it; An afterimage elimination module is used to perform an afterimage elimination operation on the second image data according to all the second impact information and the target compensation data table to obtain an afterimage eliminated image of the second image.

9. An electronic paper afterimage elimination device based on a large model, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the electronic paper afterimage elimination method based on a large model as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: The computer storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the electronic paper afterimage elimination method based on a large model as described in any one of claims 1-7.

Images