A page loading method, system, device, storage medium and electronic equipment

Abstract

The application provides a page loading method, a teaching system and a page loading device, and relates to the technical field of education. The method comprises the following steps: obtaining page layout information and question information of a to-be-loaded page, and dividing the to-be-loaded page into a public information area and a question area; performing layer layering on the to-be-loaded page to obtain a public text layer corresponding to the public information area, a question layer corresponding to the question area and a background texture layer corresponding to the to-be-loaded page; performing layer compression respectively to obtain public text loading content, question loading content and background texture loading content; encapsulating the question loading content according to a question reading sequence to obtain a question loading content unit, and generating a target loading instruction of the to-be-loaded page according to a target loading sequence. According to the application, different layers are compressed in a targeted manner, and a loading instruction is generated according to the target loading sequence, so that the picture quality of page loading is improved, the waiting time of the user is reduced, and the use experience of the user is improved.

Description

Technical Field

[0001] This application relates to the field of educational technology, and in particular to a page loading method, system, apparatus, storage medium, and electronic device. Background Technology

[0002] In scenarios with high-frequency interactive needs for test papers or assignments, such as intelligent grading, online education, and remote marking, high-speed loading of test papers or assignments can ensure the normal conduct of teaching.

[0003] Currently, existing technology involves globally compressing the page information of the entire paper roll or assignment, and then progressively rendering the paper roll or assignment as a unit through global scanning when loading the test paper.

[0004] However, when using this method, if the compression level of the global compression process is low, the paper roll or assignment will become clearer as a whole during the loading process, resulting in a slow overall loading speed and a long waiting time for the user. If the compression level of the global compression process is high, the loading speed of the paper roll or assignment will be faster, but the quality of the loaded image will decrease. This makes it impossible to balance the user's waiting time and image quality at the same time, thus affecting the user experience. Summary of the Invention

[0005] In view of this, this application provides a page loading method, a teaching system, and a page loading device. The main purpose is to improve the existing technology where, under low compression levels of global compression processing, the paper roll or assignment changes from blurry to clear during loading, resulting in slow overall loading speed and long reading wait time for users; while under high compression levels of global compression processing, the loading speed of the paper roll or assignment increases, but the quality of the loaded image decreases, making it impossible to simultaneously balance user wait time and image quality, thus affecting the user experience.

[0006] Firstly, this application provides a page loading method, including: Obtain the page layout information and question information of the page to be loaded, and divide the page to be loaded into a public information area and a question area according to the page layout information and the question information; Based on the pixel information corresponding to the public information area and the question area respectively, the page to be loaded is layered to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The common text layer, the title layer, and the background texture layer are compressed to obtain the common text loading content, title loading content, and background texture loading content of the page to be loaded. The loading content of the questions is encapsulated according to the reading order of the questions to obtain the loading content unit corresponding to the question area. Based on the target loading order among the common text loading content, the loading content unit of the questions and the background texture loading content, a target loading instruction for the page to be loaded is generated. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

[0007] Optionally, the step of layering the page to be loaded according to the pixel information corresponding to the public information area and the question area respectively, to obtain a public text layer corresponding to the public information area, a question layer corresponding to the question area, and a background texture layer corresponding to the page to be loaded, includes: Based on the layer contrast features between the page background and the page body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels of the public information area are layered, and the layer corresponding to the foreground pixels of the public information area is determined as the public text layer; Based on the layer contrast features of the page background and the main body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels of the question area are layered. The layer corresponding to the foreground pixels of the question stem text in the question area is determined as the question stem text layer, the layer corresponding to the foreground pixels of the question stem graphic in the question area is determined as the question stem graphic layer, and the layer corresponding to the foreground pixels of the handwriting area in the question area is determined as the handwriting layer. The public text layer of the public information area, the question stem text layer, the question stem graphic layer, and the handwritten layer in the question area are marked as areas to be filled in the corresponding positions on the page to be loaded. The areas to be filled are then filled by region interpolation to obtain the background texture layer corresponding to the background pixel information.

[0008] Optionally, the step of compressing the common text layer, the question layer, and the background texture layer respectively to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded includes: Based on the pixel density of the page to be loaded, the connected components of the question stem text layer in the common text layer and the question layer are extracted, and the characters in the common text layer and the question stem text layer are split based on the connected components to obtain at least one character pixel unit, as well as the coordinate information and size information of the at least one character pixel unit in the page to be loaded. The at least one character pixel unit is subjected to character similarity matching, and the character pixel units that meet the similarity matching conditions are clustered into multiple character classes based on the matching results, so that characters with similar character shapes in the at least one character pixel unit are compressed; Shape analysis is performed on the character pixel units in the multiple character classes to determine the character reference templates, and the character reference templates are used to form a global pixel dictionary; According to the reading order of the questions, the global pixel dictionary is segmented to obtain the basic character segment corresponding to the basic question reading unit and the new character segment corresponding to the new question reading unit. The characters in the new character segment are the characters added by the new question reading unit relative to the basic question reading unit. By comparing the pixel differences between the at least one character pixel unit and the character reference template, and analyzing the characters in the at least one character pixel unit that have missing or deformed strokes, the text residual data is obtained. The coordinate and size information of the at least one character pixel unit, the basic character segment, and the newly added character segment are encapsulated to obtain common text loading content and title text loading content. The text residual data is determined as delayed text loading content, which is loaded after the common text loading content and the title text loading content.

[0009] Optionally, the step of compressing the common text layer, the question layer, and the background texture layer respectively to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded includes: Edge detection is performed on the question stem image layer corresponding to the question layer to obtain the original image of the question stem layer; The original image is evaluated for vector features to determine whether there is a target image in the question image layer that can be compressed into vector data. The target image that can be compressed into vector data is then fitted with vector data to obtain vector data. The vector data is rasterized and rendered to obtain a vector rendered graphic. The pixel differences between the vector rendered graphic and the original graphic are compared, and the set of difference pixels corresponding to the original graphic is obtained based on the pixel differences between the vector rendered graphic and the original graphic. The pixels in the set of differing pixels are compressed to obtain the vector residual data between the original image and the vector data; Numerical analysis is performed on the vector residual data. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the graphics loading content. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the title graphics loading content. The vector residual data corresponding to the set of difference pixels that meet the residual value conditions is determined as the delayed graphics loading content. The handwriting layer is losslessly compressed to obtain handwriting loading content, and the background of the background texture layer is compressed at a target compression rate to obtain background texture loading content. The detailed texture in the background texture layer is determined as the delayed background texture loading content. The background texture loading content may include the compressed solid color background and the background size.

[0010] Optionally, the step of encapsulating the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generating a target loading instruction for the page to be loaded based on the target loading order among the public text loading content, the question loading content unit, and the background texture loading content, wherein the target loading instruction is used by the client to load the page to be loaded based on the target loading instruction, including: The text content, graphic content, and handwritten content of the question loading content are encapsulated according to the question reading order to obtain the question loading content unit corresponding to the question area; The target loading order is determined to be loading the background texture loading content, the public text loading content, the title loading content unit, and the delayed loading content in sequence. The delayed loading content includes the delayed text loading content, the delayed graphic loading content, and the delayed background texture loading content. Based on the background texture loading content and the common text loading content, an initial loading instruction for the page to be loaded is generated. The initial loading instruction is used to instruct the client to prioritize loading the solid color background, background size, and common text loading content of the page to be loaded. Based on the reading order of the questions and the target question loading content unit within the user's reading range, a main loading instruction for the question loading content of the target question loading content unit is generated. The main loading instruction is used to instruct the client to load the question loading content of the target question loading content unit after loading the initial loading instruction. Based on the delayed text loading content and delayed image loading content corresponding to the target question loading content unit, a question detail loading instruction corresponding to the target question loading content unit is generated. The question detail loading instruction is used to instruct the client to perform differential adjustment of residual pixels on the text and graphics in the target question loading content unit based on the character reference template and the vector data after loading the main body loading instruction, so as to perform detail residual repair on the text and graphics of the target question loading content unit. Based on the delayed background texture loading content corresponding to the background texture layer, a background detail loading instruction for the page to be loaded is generated. The background detail loading instruction is used to instruct the client to restore the details of the background texture of the page to be loaded after loading the question detail loading instruction.

[0011] Secondly, this application provides a page loading method applied to a client, including: Decode the target loading instructions generated by the server to obtain the target loading order and target loading content corresponding to the pages to be loaded; Based on the user's reading intention for the page to be loaded, the loading order of the question loading content is adjusted to obtain the question reading loading order, which is used to load the question loading content according to the reading order of the question reading units; The target content is loaded according to the target loading order and the title reading loading order.

[0012] Optionally, adjusting the loading order of the questions based on the user's reading intent of the page to be loaded, to obtain the question reading loading order, includes: Based on the user's touch interaction behavior on the page to be loaded, the user's reading intention is predicted to obtain the user's target reading intention. The touch interaction behavior includes screen swiping behavior, screen clicking behavior, screen long-term dwell behavior, and screen back swipe behavior. Based on the target reading intent, identify the question reading units within the current reading range, determine the question loading content corresponding to the question reading unit as the priority loading question, and determine the question loading content corresponding to the question reading unit outside the current reading range as the delayed loading question; The priority loading questions are sorted according to the reading order of the question reading units to determine the loading order of the priority loading questions; Based on the target reading intent and the loading order of the priority loading questions, the loading order of the loading content of the questions is adjusted to obtain the question reading loading order.

[0013] Optionally, the method further includes: When the touch interaction behavior is the screen swiping behavior, the user's target reading intention is determined to be the search intention, the loading of questions is paused and delayed, and only the common text loading content is loaded; When the touch interaction behavior is the screen click behavior, the user's target reading intention is determined as the click intention, and the priority loading questions corresponding to the question reading units within the current reading range are identified and loaded. When the touch interaction behavior is a long screen dwell behavior, the user's target reading intention is determined to be a dwell reading intention, and the priority loading question corresponding to the question reading unit within the current reading range, the text residual data and vector residual data corresponding to the priority loading question are identified and loaded. When the touch interaction behavior is the screen swipe behavior, the user's target reading intention is determined to be the swipe intention, and the historical priority loading questions, the text residual data and vector residual data corresponding to the question reading units within the historical reading range are identified and loaded.

[0014] Thirdly, this application provides a page loading device for use on a server, comprising: The acquisition module is configured to acquire the page layout information and question information of the page to be loaded, and divide the page to be loaded into a public information area and a question area according to the page layout information and the question information; The layering module is configured to perform layering on the page to be loaded according to the pixel information corresponding to the public information area and the question area respectively, to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The compression module is configured to compress the common text layer, the title layer and the background texture layer respectively to obtain the common text loading content, title loading content and background texture loading content of the page to be loaded; The generation module is configured to encapsulate the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generate the target loading instruction of the page to be loaded according to the target loading order among the public text loading content, the question loading content unit and the background texture loading content. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

[0015] Fourthly, this application provides a page loading device for use on a client side, comprising: The decoding module is configured to decode the target loading instructions generated by the server to obtain the target loading order and target loading content corresponding to the page to be loaded; The adjustment module is configured to adjust the loading order of the question loading content based on the user's reading intention of the page to be loaded, so as to obtain the question reading loading order, which is used to load the question loading content according to the reading order of the question reading units; The loading module is configured to load the target content according to the target loading order and the title reading loading order.

[0016] Using the above technical solution, this application provides a page loading method, teaching system, and page loading device, comprising: acquiring page layout information and question information of the page to be loaded, and dividing the page to be loaded into a common information area and a question area according to the page layout information and question information; performing layer layering on the page to be loaded according to the pixel information corresponding to the common information area and the question area respectively, to obtain a common text layer corresponding to the common information area, a question layer corresponding to the question area, and a background texture layer corresponding to the page to be loaded; performing layer compression on the common text layer, question layer, and background texture layer respectively, to obtain common text loading content, question loading content, and background texture loading content of the page to be loaded; encapsulating the question loading content according to the question reading order to obtain question loading content units corresponding to the question area, and generating a target loading instruction for the page to be loaded according to the target loading order among the common text loading content, question loading content units, and background texture loading content, wherein the target loading instruction is used by the client to load the page to be loaded based on the target loading instruction. Compared with existing technologies, this application achieves structured differentiation of page content by dividing the public information area into a question information area; it layers the page area according to the pixel information of the area and performs targeted compression on different layers to obtain the loading content corresponding to different layers, thereby achieving reasonable compression of page content and improving the image quality of page loading; by encapsulating question units and generating loading instructions according to the target loading order, it can improve loading speed and reduce user waiting time while ensuring image quality, thus improving the user experience. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a page loading method provided in an embodiment of this application is shown; Figure 2 A flowchart illustrating a page loading method provided in an embodiment of this application is shown; Figure 3 A flowchart illustrating a page loading method provided in an embodiment of this application is shown; Figure 4 A flowchart illustrating a page loading method provided in an embodiment of this application is shown; Figure 5 This illustration shows a schematic diagram of the structure of a page loading device provided in an embodiment of this application; Figure 6 This illustration shows a schematic diagram of the structure of a page loading device provided in an embodiment of this application; Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0020] The embodiments of this application will now be described in more detail with reference to the accompanying drawings. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0021] To address the technical problem of existing technologies where, under low global compression levels, the paper roll or assignment becomes increasingly clear during loading, resulting in slow overall loading speed and long reading wait times for users, and conversely, while high global compression levels increase loading speed, degrade image quality, this embodiment provides a page loading method, such as... Figure 1 As shown, applied to the server side, this method includes: Step 101: Obtain the page layout information and question information of the page to be loaded, and divide the page to be loaded into a common information area and a question area according to the page layout information and question information.

[0022] In this embodiment of the application, the page to be loaded can be a test paper or homework image page after image acquisition and preprocessing. For example, the page to be loaded in this embodiment of the application can specifically be a high-definition original image of a test paper or homework in JPEG, PNG, or HEIC format obtained by taking a photo with a mobile phone, scanning with a scanner, or cropping with a tablet.

[0023] In the embodiments of this application, the page to be loaded can first undergo preprocessing operations. Preprocessing may include image acquisition and format normalization, brightness equalization and global illumination correction, perspective correction and geometric deformation repair, and adaptive denoising preprocessing. Format normalization can uniformly decode the original image into a color image with H×W resolution and correct the orientation. Brightness equalization can use adaptive histogram equalization (CLAHE) or Gamma correction to eliminate shadows and light spots. Perspective correction can use edge detection to locate physical corner points and repair tilt deformation through perspective transformation. Adaptive denoising can use bilateral filtering or non-local mean denoising to preserve the sharpness of text edges.

[0024] In this embodiment, the page layout information may be the layout structure of the page to be loaded, the bounding box coordinates of the regions, and the hierarchy information. The page layout information can be used to determine the position and range of different functional areas within the page. For example, the page layout information in this embodiment may specifically include the JSON format bounding box coordinate information of the exam paper title area, question type description area, page number area, question stem area, graphic area, and handwritten answer area.

[0025] In this embodiment of the application, the question information may be the identifier, content type, semantic attributes, and reading order index information of each question on the page to be loaded. The question information can be used to distinguish questions from public information and determine the logical order of question loading and rendering. For example, the question information in this embodiment of the application may specifically include question number, question type, question stem text type, graphic type, handwritten mark identifier, and reading order index determined according to visual habits from top to bottom and left to right.

[0026] In this embodiment, the public information area can be a region on the page to be loaded that contains general explanatory content that does not change with the questions. The public information area can be used to display globally common content on the page. For example, in this embodiment, the public information area can specifically be the page area containing the exam title, question type descriptions, page number annotations, and exam instructions.

[0027] In this embodiment of the application, the question area can be an area containing the question content on the page to be loaded. For example, the question area in this embodiment of the application can specifically be a page area containing question numbers, printed question text, printed question graphics, and handwritten answers by students.

[0028] In this embodiment of the application, the preprocessed page to be loaded can be input into a multimodal large model or a lightweight layout analysis network. The model may not perform image matting operations, but only output semantic region annotation results based on reading order. The annotation may include region classification tags, question-level fine-grained decomposition, and reading order assembly. The region classification tags can identify and divide the public information area and the question area. The question-level fine-grained decomposition can identify the question number, printed question stem text area, printed question stem graphic area, and handwritten mark area for a single question. The reading order assembly can assign a global reading order index to all areas. Based on the annotation results, the page to be loaded is divided into public information area and question area.

[0029] Step 102: Based on the pixel information corresponding to the public information area and the question area, the page to be loaded is layered to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded.

[0030] In this embodiment of the application, pixel information can be the color, brightness, grayscale value, and contrast value of each pixel in the page to be loaded. Pixel information can be used to distinguish between the foreground reading content and the background texture information of the page. For example, the pixel information in this embodiment of the application can specifically be high-contrast pixel data between the light background of the test paper image and the dark text, graphics, and handwriting.

[0031] In the embodiments of this application, layer layering can be used to split the page to be loaded into independent layers according to content semantics and functional type. Layer layering can be used to achieve independent compression, independent transmission, and independent loading of different content. For example, the layer layering in the embodiments of this application can specifically split the page into a skeleton layer, a printed public text layer, a printed title text layer, a printed title graphic layer, a handwritten mark layer, and a background texture layer.

[0032] In this embodiment, the public text layer can be a foreground layer carrying printed public text within a public information area. The public text layer can also include exam paper layout metadata information. The public text layer can be used to store pixel information of printed public text such as exam paper titles, question type descriptions, and page numbers, and can also prioritize loading exam paper layout information with low data volume. For example, the public text layer in this embodiment can specifically include basic information of each area of ​​the exam paper, boundary coordinates, question number positions, coordinates of the question stem area and handwritten area, and various types of printed public text pixel data.

[0033] In this embodiment of the application, the question layer can be a collection of foreground layers within the question area that carry the core reading content of the question. The question layer can be used to encapsulate question-related content such as question stem text, question stem graphics, handwriting marks, etc.

[0034] In the embodiments of this application, the background texture layer can be a pure background layer after removing all foreground reading content from the page to be loaded. The background texture layer can be used to store non-core reading information such as paper background color, paper fiber texture, light and shadow, back text, ink spots and dust spots.

[0035] In the embodiments of this application, the high pixel contrast feature of light background and dark subject in the test paper image can be used to extract local pixels of the marked area coordinates, thereby separating the main content layer from the background texture layer. The main content layer may include a printed question text layer, a printed question graphic layer, a printed public text layer, and a handwritten trace layer.

[0036] Step 103: Compress the common text layer, the title layer, and the background texture layer separately to obtain the common text loading content, title loading content, and background texture loading content of the page to be loaded.

[0037] In this embodiment, layer compression can be targeted at the content of different layers. Layer compression can improve data transmission efficiency and reduce loading latency in weak network environments. For example, layer compression in this embodiment may specifically include text pixel dictionary deduplication compression, graphic vectorization compression, lossless compression of handwritten features, and high-compression-rate background compression.

[0038] In the embodiments of this application, the public text loading content can be lightweight loading data obtained by compressing the public text layer with a pixel dictionary. The public text loading content can be used by the client to quickly render the public text information of the page.

[0039] In this embodiment of the application, the question loading content can be the core loading data obtained by compressing the question layer through text compression, image compression, and handwriting compression. The question loading content can be used by the client to render the complete semantic information of the question.

[0040] In the embodiments of this application, the background texture loading content can be the background data obtained after the background texture layer is compressed at an extremely high compression rate, and the background texture loading content can be used by the client to quickly render the page base background.

[0041] In this embodiment, pixel-level dictionary deduplication and local residual compensation compression can be used for the text layer, vectorization compression can be used for the graphics layer, lossless compression can be used for the handwriting layer, and extremely high compression ratio compression can be used for the background layer. It can achieve atomic progressive loading based on the user's cognitive order with almost no increase in file data volume, and solve the problems of unreasonable bitrate allocation and low compression efficiency of traditional unified compression algorithms.

[0042] Step 104: Encapsulate the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generate the target loading instruction of the page to be loaded based on the target loading order among the common text loading content, the question loading content unit and the background texture loading content.

[0043] The target loading directive is used by the client to load the page to be loaded based on the target loading directive.

[0044] In this embodiment of the application, the question loading content unit can be an independent code stream block after the question loading content is indexed and encapsulated according to the question reading order. The question loading content unit can be a logical unit (LU), and the logical unit can be a complete semantic code stream block that can be independently distributed and rendered by the client.

[0045] In this embodiment, the target loading order can be a progressive transmission order based on visual cognitive patterns and reading habits, prioritizing semantics and deferred to details. For example, the target loading order in this embodiment can specifically be the loading order of the initialization layer, the main reading information layer, the precision repair layer, and the visual enhancement layer. The initialization layer loads common text content such as page layout information and background content; the main reading information layer loads the title content; the precision repair layer loads text residual data and vector residual data; and the visual enhancement layer loads the deferred background content.

[0046] In the embodiments of this application, the target loading instruction may be a set of instructions including the target loading order, the loading content and the rendering rules. The target loading instruction can be used to instruct the client to complete the progressive loading of the page in sequence.

[0047] In the embodiments of this application, heterogeneous data blocks can be reassembled in sequence according to reading priority to construct a progressive transmission protocol that prioritizes semantics and delays details. The server can provide a lightweight status feedback interface to receive the current viewport target title number returned by the client, thereby realizing dynamic priority queue loading.

[0048] Compared with existing technologies, this embodiment achieves structured differentiation of page content by dividing the public information area into a question information area; it layers the page area according to the pixel information of the area and performs targeted compression on different layers to obtain the loading content corresponding to different layers, thereby achieving reasonable compression of page content and improving the image quality of page loading; by encapsulating question units and generating loading instructions according to the target loading order, it can improve loading speed and reduce user waiting time while ensuring image quality, thus improving the user experience.

[0049] As an optional approach, when performing the step of "layering the page to be loaded according to the pixel information corresponding to the public information area and the question area respectively, to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded", the following methods can be used, but are not limited to: Figure 2 As shown, the method includes: Step 201: Based on the layer contrast characteristics between the page background and the main body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels of the common information area are layered, and the layer corresponding to the foreground pixels of the common information area is determined as the common text layer.

[0050] In this embodiment of the application, a local adaptive binarization algorithm can be used to divide the coordinate frame of the public information area into layers corresponding to the foreground pixels and the background pixels, and the layer corresponding to the foreground pixels can be determined as the public text layer. The public text layer can carry public printed text information such as the exam paper title, question type description, and page number.

[0051] Step 202: Based on the layer contrast features of the page background and the main body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels in the question area are layered. The layer corresponding to the foreground pixels of the question stem text in the question area is determined as the question stem text layer, the layer corresponding to the foreground pixels of the question stem graphic in the question area is determined as the question stem graphic layer, and the layer corresponding to the foreground pixels of the handwriting area in the question area is determined as the handwriting layer.

[0052] In this embodiment, the question text layer can be a layer within the question area that includes printed question text, mathematical formulas, and printed symbols. The question text layer can be used for text pixel dictionary construction and differential compression.

[0053] In the embodiments of this application, the question stem graphic layer can be a layer including geometric figures, charts, and images within the question area, and the question stem graphic layer can be used for geometric vectorization compression.

[0054] In this embodiment of the application, the handwriting layer can be a layer in the question area that includes student answers and teacher correction marks, and the handwriting layer can be used for lossless compression of handwriting features.

[0055] For the embodiments of this application, a contrast separation method consistent with that of the public information area can be adopted to divide the question area into foreground pixels. The layer corresponding to the foreground pixels of the question stem text is determined as the question stem text layer, the layer corresponding to the foreground pixels of the question stem graphic is determined as the question stem graphic layer, and the layer corresponding to the foreground pixels of the handwriting area is determined as the handwriting layer. The question stem text layer, the question stem graphic layer, and the handwriting layer can constitute the question layer.

[0056] Step 203: Mark the corresponding positions of the public text layer in the public information area, the question stem text layer, the question stem graphic layer, and the handwritten layer in the page to be loaded as areas to be filled, and perform area interpolation filling on the areas to be filled to obtain the background texture layer corresponding to the background pixel information.

[0057] In this embodiment, the corresponding position of the extracted foreground pixels in the original image can be marked as the pixel to be filled. An image inpainting technique based on the Fast Marching Method or the Navier-Stokes equation is used to smoothly interpolate and transition inward using the paper texture and color information of the edge of the pixel to be filled, so as to ensure that the filled area blends naturally with the surrounding paper. The inpainted image can be a background texture layer, which can remove high-frequency details and support extremely high compression ratio encoding.

[0058] Optionally, when performing the step of "compressing the common text layer, question layer, and background texture layer separately to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded," the following method can be used, but is not limited to: extracting the connected components of the question stem text layer in the common text layer and question layer based on the pixel density of the page to be loaded, and splitting the characters in the common text layer and question stem text layer based on the connected components to obtain at least one character pixel unit, as well as the coordinate information and size information of at least one character pixel unit in the page to be loaded; performing character similarity matching on at least one character pixel unit, and clustering the character pixel units that meet the similarity matching conditions into multiple character classes according to the matching results, so that characters with similar shapes in at least one character pixel unit are compressed; and further processing the character pixel units in the multiple character classes... The system performs shape analysis to determine character baseline templates and assembles these templates into a global pixel dictionary. Following the question reading order, the global pixel dictionary is segmented to obtain basic character segments corresponding to basic question reading units and new character segments corresponding to newly added question reading units. The characters in the new character segments are those added by the new question reading units relative to the basic question reading units. Pixel differences are compared between at least one character pixel unit and the character baseline template to identify characters with missing or deformed strokes, resulting in text residual data. The coordinate and size information of at least one character pixel unit, along with the basic and new character segments, are encapsulated to obtain common text loading content and question text loading content. The text residual data is designated as delayed text loading content, which is loaded after the common and question text loading content.

[0059] In the embodiments of this application, a character pixel unit can be an independent pixel block after being split by connected component analysis; for example, a character pixel unit can be a single Chinese character, English letter, punctuation mark, or mathematical formula component.

[0060] In this embodiment, the global pixel dictionary can be a pixel template library composed of character baseline templates and global index IDs. The global pixel dictionary can be used to implement global deduplication and compression of text pixels.

[0061] In the embodiments of this application, the text residual data can be the set of pixels that differ from the original character pixel unit by XOR or pixel subtraction between the original character pixel unit and the reference template. The text residual data can be used to non-destructively restore details such as printing defects, broken strokes, and burrs in text.

[0062] For the embodiments of this application, character similarity matching may employ, but is not limited to, cross-entropy, Hamming distance, and Hausdorff distance algorithms; In this embodiment of the application, the global pixel dictionary can be incrementally segmented according to logical units. Let D be the global pixel dictionary. This is the kth logical unit arranged in reading order (i.e., the title loading content unit in the embodiments of this application). For rendering Required character baseline template ID set For incremental character fragmentation, the first logical unit The incremental character fragmentation of subsequent logic units can be as shown in Formula 1, where, It can represent rendering The required set of character baseline template IDs, where i can represent the current logical unit ID and k can represent the total number of logical units in the page to be loaded.

[0063] (Formula 1) For example, if the page to be loaded is a student's exam paper, and the stem of the first question includes the characters A, B, C, and D, then the incremental fragment of the first question contains the characters A, B, C, and D. If the stem of the second question includes the characters A, B, C, D, E, F, and G, then the incremental fragment of the second question contains the newly added characters E, F, and G.

[0064] As an optional approach, when performing the step of "compressing the common text layer, question layer, and background texture layer separately to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded," the following methods can be used, but are not limited to: performing edge detection on the question stem image layer corresponding to the question layer to obtain the original image of the question stem layer; performing vector feature evaluation on the original image to evaluate whether there is a target image in the question stem image layer that can be compressed into vector data, and performing image vector fitting on the target image that can be compressed into vector data to obtain vector data; performing rasterization rendering on the vector data to obtain a vector rendered image; comparing the pixel differences between the vector rendered image and the original image, and obtaining the difference corresponding to the original image based on the pixel difference between the vector rendered image and the original image. The process involves several steps: First, a set of differing pixels is created. Then, pixels in this set are compressed to obtain the vector residual data between the original graphic and the vector data. Numerical analysis is performed on the vector residual data. Vector data corresponding to sets of differing pixels that meet the residual value conditions are identified as the graphic loading content. Vector data corresponding to sets of differing pixels that meet the residual value conditions are identified as the title graphic loading content. Vector residual data corresponding to sets of differing pixels that meet the residual value conditions are identified as the delayed graphic loading content. Finally, lossless compression of handwriting features is performed on the handwriting layer to obtain the handwriting loading content. The background texture layer is compressed at the target compression rate to obtain the background texture loading content. Detailed textures in the background texture layer are identified as the delayed background texture loading content. The background texture loading content can include the compressed solid color background and the background size.

[0065] In the embodiments of this application, vector data can be geometric description information of a graphic after mathematical fitting, and vector data can be the coordinates and attribute information of a straight line, arc, polygon, or Bézier curve.

[0066] In this embodiment, the vector residual data can be the set of pixels that differ between the original graphic and the vector-rendered graphic. The vector residual data can be used to restore the printing bleed and line burr details of the graphic.

[0067] In the embodiments of this application, edge detection may employ, but is not limited to, the Canny operator, and vector fitting may employ, but is not limited to, the Douglas-Peucker algorithm, Hough transform, and Bezier curve fitting.

[0068] For the embodiments of this application, numerical analysis of vector residual data can be performed by numerical analysis of the proportion of vector residual data in the sum of vector data and vector residual data; the residual numerical condition can be met if the proportion of vector residual data in the sum of vector data and vector residual data does not exceed the residual proportion threshold.

[0069] Optionally, an adaptive compression ratio evaluation and fallback mechanism can be used to determine whether a graphic can undergo volumetric vector compression. Specifically, this can include: the volume of the vectorization scheme equals the sum of the volume of the vector data and the volume of the vector residual data; the volume of the traditional bitmap scheme equals the volume of the original graphic under standard compression. If the volume of the vectorization scheme is less than the product of the volume of the traditional bitmap scheme and the compression benefit threshold, the vectorization compression scheme is applied to the graphic, retaining the vector data and vector residual data; otherwise, it falls back to traditional bitmap compression. The compression benefit threshold is typically set to 0.8 to 0.9, and the total volume of the vectorization scheme is at least 10% to 20% smaller than that of the traditional bitmap scheme, meaning the vector residual data is 10% of the sum of the vector data and the vector residual data.

[0070] In this embodiment, the background texture layer can be compressed with a large downsampling or an extremely high quantization step size. When the network environment is extremely poor, only the pure white background color value and the canvas size can be sent, and the detailed texture can be used as the content to be loaded after the background texture.

[0071] As an optional approach, when executing the process of "encapsulating the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generating a target loading instruction for the page to be loaded based on the target loading order among the common text loading content, the question loading content unit, and the background texture loading content, the target loading instruction is used by the client to load the page to be loaded based on the target loading instruction," the following method can be used, but is not limited to: encapsulating the question text loading content, graphic loading content, and handwritten loading content in the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area; determining the target loading order as loading background texture loading content, common text loading content, question loading content unit, and delayed loading content in sequence, wherein the delayed loading content includes delayed text loading content, delayed graphic loading content, and delayed background texture loading content; and generating an initial loading instruction for the page to be loaded based on the background texture loading content and the common text loading content, wherein the initial loading instruction is used to instruct the client to load the page to be loaded first. The system loads solid color backgrounds, background sizes, and common text content. Based on the reading order of the questions and the user's reading range, it generates a main loading instruction for the target question content unit, instructing the client to load the question content after loading the initial loading instruction. Based on the delayed text and image loading content corresponding to the target question content unit, it generates a question detail loading instruction, instructing the client to perform differential adjustment of residual pixels on the text and graphics in the target question content unit based on the character base template and vector data after loading the main loading instruction, thus performing detail residual repair on the text and graphics. Finally, based on the delayed background texture loading content corresponding to the background texture layer, it generates a background detail loading instruction for the page to be loaded, instructing the client to restore the background texture details of the page to be loaded after loading the question detail loading instruction.

[0072] In this embodiment of the application, the question loading content unit can be a logic unit (LU), which can encapsulate the question text loading content, graphic loading content and handwritten loading content in the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area.

[0073] In the embodiments of this application, the target question loading content unit can be the question unit within the reading range when the user is reading the page to be loaded; for example, in the case of an exam paper on the page to be loaded, the user can jump to the tenth question after reading the first question, then the target question loading content unit can change from the first question to the tenth question.

[0074] For the embodiments of this application, the target loading order can be to load background texture loading content, public text loading content, title loading content unit, and delayed loading content in sequence. The delayed loading content includes delayed text loading content, delayed graphics loading content, and delayed background texture loading content.

[0075] In this embodiment, the client's loading of target loading instructions based on the target loading order can be divided into four stages. Stage 1 is loading the initial loading instruction, which loads the background texture loading content first, including a solid color background and canvas size, and then loads the common text loading content. Stage 2 is loading the main loading instruction, which loads the core content of the target question loading content unit after loading the initial loading instruction, including the question stem text loading content, question stem graphic loading content, and handwritten loading content. Stage 3 is loading the question detail loading instruction, which loads the delayed text loading content and delayed image loading content of the target question loading content unit based on the question detail loading instruction, thereby restoring the details of the text and graphics. Stage 4 is loading the delayed background texture loading content, which loads the delayed background texture loading content after loading the delayed text loading content and delayed image loading content, thereby restoring and filling the solid color background with detailed textures, thereby achieving visual perfection.

[0076] Optionally, the server can provide a lightweight status feedback interface that can receive the target question loading content units within the user's current reading range from the client. Based on the target question loading content units, the server can adjust the current loading order, prioritizing the loading of question units within the user's reading range. After loading all the question units within the user's reading range, the server can restore the initial reading order of the question loading content units. For example, restoring the initial reading order could be reading the first question from left to right, the second question from top to bottom, the third question, and so on.

[0077] Compared with existing technologies, this embodiment achieves separation of the main content of the page from the background by generating a background texture layer through foreground and background layering and interpolation filling; improves the efficiency of text data compression by processing characters in the text layer and constructing a dictionary; optimizes the compression effect of multiple types of layers by vectorizing graphics, lossless compression of handwriting, and high compression of the background, thereby improving the quality of the loaded screen; and achieves progressive loading of the page by generating loading instructions in stages.

[0078] This embodiment provides a page loading method, such as... Figure 3 As shown, applied to the client, the method includes: Step 301: Decode the target loading instruction generated by the server to obtain the target loading order and target loading content corresponding to the page to be loaded.

[0079] In this embodiment, decoding can be an operation of parsing and splitting the code stream and instructions transmitted by the server. Decoding can be used to extract the loading order, loaded content, and rendering rules. For example, the decoding operation in this embodiment can specifically parse out the target loading order and target loaded content such as background data, text data, logical unit data, and residual data.

[0080] In this embodiment of the application, the client can receive the target loading instruction and corresponding loading content sent by the server, prioritize parsing the background texture loading data and public text loading content, reserve a question placeholder area on the page canvas using JSON / SVG coordinate information, display basic information such as question number, and complete the rapid presentation of the test paper layout, providing a global coordinate system for the accurate placement of subsequent logical units.

[0081] Step 302: Based on the user's reading intention of the page to be loaded, adjust the loading order of the questions to obtain the question reading loading order. The question reading loading order is used to load the question content according to the reading order of the question reading units.

[0082] In this embodiment, the reading intent can be the user's operational intent based on touch interaction behavior, such as quick search, target locking, detailed reading, and backtracking correction. The reading intent can be used to dynamically adjust the loading priority of questions.

[0083] For the embodiments of this application, the question reading loading order can be a question loading order adjusted by combining viewport perception and reading intent. The question reading loading order can be used to achieve viewport priority and on-demand loading.

[0084] In this embodiment of the application, the client can monitor changes in the screen viewport, predict reading intent based on screen swiping dynamics and dwell state, and dynamically adjust the loading order of the question content.

[0085] Step 303: Load the target content according to the target loading order and the question reading loading order.

[0086] In this embodiment, the client can first load the background texture and common text of the page according to the initial loading instruction, and then load the main loading instruction according to the loading order of the question reading. It maintains a global pixel dictionary table, and continuously updates the dictionary as incremental dictionary fragments arrive. It renders text, graphics, and handwritten content by looking up the dictionary through the index. After the main question is rendered, it loads residual data to complete the detail repair, and loads the original background texture as needed to replace the initial background, thus realizing the transition from structured reading to realistic restoration.

[0087] Compared with existing technologies, this embodiment decodes the target loading instructions to obtain the target loading order and content, thereby parsing the loaded data; it adjusts the loading order of questions based on the user's reading intent to adapt page loading to the reading intent, thus improving the personalization and targeting of page loading; and it improves page loading efficiency by loading content in the adjusted order.

[0088] As an optional approach, when performing the task of "adjusting the loading order of questions based on the user's reading intent of the page to be loaded, thus obtaining the question reading loading order," the following methods can be used, but are not limited to: Figure 4 As shown, the method includes: Step 401: Based on the user's touch interaction behavior on the page to be loaded, predict the user's reading intention to obtain the user's target reading intention. The touch interaction behavior includes screen swiping behavior, screen clicking behavior, screen long-term dwell behavior, and screen back swipe behavior.

[0089] In the embodiments of this application, the client can listen to the user's touch interaction events such as touch down (TOUCH_DOWN), touch move (TOUCH_MOVE), and touch up (TOUCH_UP), and combine instantaneous speed, displacement vector, dwell time, and swipe direction to form a dynamic event stream to predict the user's target reading intention.

[0090] Step 402: Based on the target reading intention, identify the question reading units within the current reading range, determine the question loading content corresponding to the question reading unit as the priority loading question, and determine the question loading content corresponding to the question reading unit outside the current reading range as the deferred loading question.

[0091] In this embodiment of the application, the question reading unit may be a page reading area unit corresponding to the logic unit (LU), and the question reading unit may be used to divide the priority loading range within the viewport.

[0092] In this embodiment of the application, the client can identify the current viewport range of the screen based on the target reading intention, mark the question reading units within the viewport as priority loading questions, and mark those outside the viewport as delayed loading questions, thereby reducing invalid data transmission and rendering.

[0093] Step 403: Sort the priority loading questions according to the reading order of the question reading units to determine the loading order of priority loading questions.

[0094] In this embodiment, the priority loading questions in the viewport can be sorted from top to bottom and from left to right according to the question reading order index to determine the loading order of priority loading questions and ensure that the loading order conforms to the user's reading habits.

[0095] Step 404: Based on the target reading intention and the priority loading order of questions, adjust the loading order of the question loading content to obtain the question reading loading order.

[0096] In this embodiment of the application, the loading order of the overall loading content of the questions can be dynamically adjusted by combining the user's target reading intention and the sorting result of the priority loading of questions, so as to obtain the loading order of the questions that adapts to the user's reading behavior; the client can send the target question number of the current viewport to the lightweight status feedback interface of the server to trigger the server's dynamic queue loading logic.

[0097] As an optional approach, other methods may also be employed, but not limited to these, including: when the touch interaction behavior is a screen swipe, determining the user's target reading intent as a search intent, pausing and delaying the loading of questions, and loading only common text content; when the touch interaction behavior is a screen click, determining the user's target reading intent as a click intent, identifying and loading the priority loading questions corresponding to the question reading units within the current reading range; when the touch interaction behavior is a long screen dwell, determining the user's target reading intent as a dwell reading intent, identifying and loading the priority loading questions, text residual data, and vector residual data corresponding to the question reading units within the current reading range; when the touch interaction behavior is a screen swipe back, determining the user's target reading intent as a swipe back intent, identifying and loading the historical priority loading questions, text residual data, and vector residual data corresponding to the question reading units within the historical reading range.

[0098] In the embodiments of this application, the screen swiping behavior can be such that the instantaneous speed of multiple consecutive TOUCH_UP moments remains high. When a new TOUCH_DOWN is immediately followed by a large TOUCH_MOVE in the same direction before the view inertial swiping has stopped, the user's target reading intention can be determined. If the user's target reading intention is to search for the intended content, it can be determined that the current view is a low-confidence view, and the loading of the title can be paused and delayed, with only the common text loading content being loaded.

[0099] In this embodiment, the screen click behavior can be detected during the screen swiping process. If the TOUCH_DOWN signal is captured and the subsequent TOUCH_MOVE displacement vector is close to 0, that is, the user stops the screen swiping by pressing the screen, then the target reading intention can be determined to be the click intention. The current reading range corresponding to the user's click action can be determined, and the priority loading content of the question reading unit corresponding to the question within the current reading range can be loaded.

[0100] In this embodiment of the application, if a continuous TOUCH_MOVE signal from the user is detected, and the calculated movement speed remains in the low-speed range (synchronized with hand speed, without inertial displacement), and the view displacement stops immediately after the TOUCH_UP signal is triggered, and there are no continuous sliding events in a short period of time, then it can be determined that the user's target reading intention is to stay and read, and the priority loading question, the text residual data and vector residual data corresponding to the question reading unit within the current reading range are identified and loaded.

[0101] In this embodiment of the application, a sliding displacement vector opposite to the previous historical sliding direction is detected. The ratio of the sliding displacement to the previous total displacement is calculated. If the ratio is less than 1, it is determined that the user is stationary for a short time after sliding back. It can be determined that the user's target reading intention is the intention to slide back. The historical priority loading questions, the text residual data and vector residual data corresponding to the question reading units within the historical reading range are identified and loaded.

[0102] Optionally, the client can also employ a transmission channel management strategy to handle situations where the user's reading range jumps. Specifically, the client can maintain a data stream monitor to calculate in real time the download progress ratio of the currently loading logical unit (the ratio of received bytes to the expected total bytes). When the user's reading focus jumps (e.g., swiping to jump to the tenth question while loading the second question), if the download progress ratio of the old task (the second question) is less than the set sunk cost threshold (e.g., 80%), the client can actively interrupt the data reception of the old unit. The released downlink bandwidth resources can be allocated to the new viewport target (the tenth question), and the interrupted logical unit can be moved to the suspend queue. If the download progress ratio of the old task is greater than or equal to the threshold, the old unit can continue to be downloaded in the background using the remaining bandwidth. When the user swipes back to the question later, they can read it directly, thus avoiding wasting the loading time cost already incurred and optimizing loading efficiency and user experience in weak network environments.

[0103] Compared with existing technologies, this embodiment achieves dynamic adjustment of the loading order of questions by predicting intent based on touch behavior and classifying loading priorities; and improves the adaptability between page loading and user reading intent by performing differentiated loading based on different touch behaviors.

[0104] Furthermore, as Figure 1 and Figure 2 To provide a specific implementation of the method shown, this embodiment offers a page loading device, such as... Figure 5 As shown, the device includes: an acquisition module 51, a layering module 52, a compression module 53, and a generation module 54.

[0105] The acquisition module 51 is configured to acquire the page layout information and question information of the page to be loaded, and divide the page to be loaded into a public information area and a question area according to the page layout information and question information; The layering module 52 is configured to perform layering on the page to be loaded according to the pixel information corresponding to the public information area and the question area, respectively, to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. Compression module 53 is configured to compress the common text layer, the title layer and the background texture layer respectively to obtain the common text loading content, title loading content and background texture loading content of the page to be loaded; The generation module 54 is configured to encapsulate the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generate the target loading instruction of the page to be loaded based on the target loading order among the common text loading content, the question loading content unit and the background texture loading content. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

[0106] In some examples of this embodiment, the layering module 32 is specifically configured to, based on the layer contrast features between the page background and the page body of the page to be loaded, layer the layers corresponding to the foreground pixels of the public information area and the layers corresponding to the background pixels, and determine the layers corresponding to the foreground pixels of the public information area as the public text layer; based on the layer contrast features between the page background and the page body of the page to be loaded, layer the layers corresponding to the foreground pixels of the question area and the layers corresponding to the background pixels, and determine the layers corresponding to the foreground pixels of the question stem text in the question area as the question stem text layer, the layers corresponding to the foreground pixels of the question stem graphics in the question area as the question stem graphics layer, and the layers corresponding to the foreground pixels of the handwriting area in the question area as the handwriting layer; mark the corresponding positions of the public text layer of the public information area, the question stem text layer, the question stem graphics layer, and the handwriting layer in the page to be loaded as areas to be filled, and perform area interpolation filling on the areas to be filled to obtain the background texture layer corresponding to the background pixel information.

[0107] In some examples of this embodiment, the compression module 33 is specifically configured to extract the connected components of the question stem text layer in the common text layer and the question stem text layer based on the pixel density of the page to be loaded, and to split the characters in the common text layer and the question stem text layer based on the connected components to obtain at least one character pixel unit, as well as the coordinate information and size information of at least one character pixel unit in the page to be loaded; to perform character similarity matching on at least one character pixel unit, and to cluster the character pixel units that meet the similarity matching conditions into multiple character classes according to the matching results, so that characters with similar shapes in at least one character pixel unit are compressed; to perform shape analysis on the character pixel units in the multiple character classes, to determine the character reference template, and to form a global pixel dictionary from the character reference template. According to the reading order of the questions, the global pixel dictionary is segmented to obtain the basic character segment corresponding to the basic question reading unit and the new character segment corresponding to the new question reading unit. The characters in the new character segment are the characters added by the new question reading unit relative to the basic question reading unit. The pixel difference of at least one character pixel unit and the character reference template is compared to analyze the characters with missing or deformed strokes in at least one character pixel unit to obtain the text residual data. The coordinate information and size information of at least one character pixel unit, the basic character segment and the new character segment are encapsulated to obtain the common text loading content and the question text loading content. The text residual data is determined as the delayed text loading content, which is loaded after the common text loading content and the question text loading content.

[0108] In some examples of this embodiment, the compression module 33 is further configured to perform edge detection on the question stem image layer corresponding to the question layer to obtain the original image of the question stem layer; perform vector feature evaluation on the original image to evaluate whether there is a target image in the question stem image layer that can be compressed into vector data, and perform image vector fitting on the target image that can be compressed into vector data to obtain vector data; perform raster rendering on the vector data to obtain a vector rendered image; compare the pixel differences between the vector rendered image and the original image, and obtain the set of difference pixels corresponding to the original image based on the pixel difference between the vector rendered image and the original image; compress the pixels in the set of difference pixels to obtain the original image and vector data. The vector residual data between the two layers is analyzed numerically. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the graphics loading content. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the graphics loading content for the question. The vector residual data corresponding to the set of difference pixels that meet the residual value conditions is determined as the delayed graphics loading content. The handwriting layer is losslessly compressed to obtain the handwriting loading content. The background of the background texture layer is compressed at the target compression rate to obtain the background texture loading content. The detailed texture in the background texture layer is determined as the delayed background texture loading content. The background texture loading content may include the compressed solid color background and the background size.

[0109] In some examples of this embodiment, the generation module 34 is specifically configured to encapsulate the question text loading content, graphic loading content, and handwritten loading content in the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area; determine the target loading order as loading background texture loading content, public text loading content, question loading content unit, and delayed loading content in sequence, where delayed loading content includes delayed text loading content, delayed graphic loading content, and delayed background texture loading content; generate an initial loading instruction for the page to be loaded based on the background texture loading content and public text loading content, the initial loading instruction being used to instruct the client to prioritize loading the solid color background, background size, and public text loading content of the page to be loaded; and generate the question loading content unit of the target question loading content unit based on the question reading order and the target question loading content unit within the user's reading range. The main loading instruction instructs the client to load the title content of the target title content unit after loading the initial loading instruction. Based on the delayed text and image loading content corresponding to the target title content unit, a title detail loading instruction is generated. This instruction instructs the client to perform differential adjustment of residual pixels on the text and graphics in the target title content unit based on the character base template and vector data after loading the main loading instruction, in order to perform detail residual repair on the text and graphics of the target title content unit. Based on the delayed background texture loading content corresponding to the background texture layer, a background detail loading instruction is generated for the page to be loaded. This instruction instructs the client to restore the background texture details of the page to be loaded after loading the title detail loading instruction.

[0110] Furthermore, as Figure 3 and Figure 4 To provide a specific implementation of the method shown, this embodiment offers a page loading device, such as... Figure 6 As shown, the device includes: a decoding module 61, an adjustment module 62, and a loading module 63.

[0111] Decoding module 61 is configured to decode the target loading instructions generated by the server to obtain the target loading order and target loading content corresponding to the page to be loaded; The adjustment module 62 is configured to adjust the loading order of the questions based on the user's reading intention of the page to be loaded, so as to obtain the question reading loading order. The question reading loading order is used to load the question content according to the reading order of the question reading units. Loading module 63 is configured to load the target content according to the target loading order and the question reading loading order.

[0112] In some examples of this embodiment, the adjustment module 62 is specifically configured to predict the user's reading intention based on the user's touch interaction behavior on the page to be loaded, thereby obtaining the user's target reading intention. The touch interaction behavior includes screen swiping, screen clicking, screen long-term dwell, and screen back swipe. Based on the target reading intention, the module identifies the question reading units within the current reading range, determines the question loading content corresponding to the question reading unit as the priority loading question, and determines the question loading content corresponding to the question reading units outside the current reading range as the delayed loading question. The priority loading questions are sorted according to the reading order of the question reading units to determine the loading order of the priority loading questions. Based on the target reading intention and the loading order of the priority loading questions, the loading order of the question loading content is adjusted to obtain the question reading loading order.

[0113] In some examples of this embodiment, the adjustment module 63 is further configured to: when the touch interaction behavior is a screen swipe, determine the user's target reading intent as a search intent, pause loading and delay loading of questions, and only load common text loading content; when the touch interaction behavior is a screen click, determine the user's target reading intent as a click intent, identify and load the priority loading questions corresponding to the question reading units within the current reading range; when the touch interaction behavior is a long screen stay, determine the user's target reading intent as a stay reading intent, identify and load the priority loading questions, text residual data, and vector residual data corresponding to the question reading units within the current reading range; when the touch interaction behavior is a screen swipe back, determine the user's target reading intent as a swipe back intent, identify and load the historical priority loading questions, text residual data, and vector residual data corresponding to the question reading units within the historical reading range.

[0114] Based on the above, Figure 1 , Figure 2 , Figure 3 and Figure 4 Accordingly, this embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figure 1 , Figure 2 , Figure 3 and Figure 4 The method shown.

[0115] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0116] like Figure 7 The diagram shown is a hardware structure schematic of an electronic device according to the present invention, comprising: At least one processor 701; and, A memory 702 is communicatively connected to at least one processor 701; wherein, The memory 702 stores instructions that can be executed by at least one processor, such that the instructions are executed by at least one processor to enable the at least one processor to perform the page loading method as described above.

[0117] Figure 7 Take the 701 processor as an example.

[0118] The electronic device may also include an input device 703 and a display device 704.

[0119] The processor 701, memory 702, input device 703, and display device 704 can be connected via a bus or other means. Figure 7 Taking the example of a connection between China and Israel via a bus.

[0120] The memory 702, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the page loading method in the embodiments of this application, for example, Figure 1 , Figure 2 , Figure 3 and Figure 4 The method flow is shown. The processor 701 executes various functional applications and data processing by running non-volatile software programs, instructions, and modules stored in the memory 702, thereby implementing the page loading method in the above embodiments.

[0121] Memory 702 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created according to the use of the page loading method, etc. Furthermore, memory 702 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 702 may optionally include memory remotely located relative to processor 701, and these remote memories may be connected to the means of performing the page loading method via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0122] The input device 703 can receive user clicks and generate signal inputs related to user settings and function control for page loading methods. The display device 704 may include a display screen or other display device.

[0123] One or more modules are stored in memory 702, and when run by one or more processors 701, the page loading method in any of the above method embodiments is executed.

[0124] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0125] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0126] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms, or it can be implemented by hardware. Compared with the prior art, by applying the solution of this embodiment, this embodiment achieves structured differentiation of page content by dividing the public information area and the question information area; by layering the page area according to the pixel information of the area, and performing targeted compression on different layers to obtain the loading content corresponding to different layers, it achieves reasonable compression of page content and improves the image quality of page loading; by encapsulating question units and generating loading instructions according to the target loading order, it can improve loading speed and reduce user waiting time while ensuring image quality, thus improving the user experience; by generating a background texture layer through foreground and background layering and interpolation filling, it achieves separation of the main content of the page from the page background; and by character processing and dictionary construction of the text layer, it improves the quality of text data. Compression efficiency; Optimizes compression effects for multiple layer types through graphic vectorization, lossless compression of handwriting, and high background compression, improving loading image quality; Achieves progressive page loading by generating loading instructions in stages; Parses loading data by decoding target loading instructions to obtain target loading order and content; Adapts page loading to reading intent by adjusting the loading order of questions based on user reading intent, enhancing the personalization and targeting of page loading; Improves page loading efficiency by loading content in the adjusted order; Dynamically adjusts the loading order of questions by predicting intent based on touch behavior and classifying loading priorities; Enhances the adaptability between page loading and user reading intent by performing differentiated loading based on different touch behaviors.

[0128] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0129] The above are merely specific embodiments of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to these embodiments, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A page loading method, characterized in that, Applied to the server side, including: Obtain the page layout information and question information of the page to be loaded, and divide the page to be loaded into a public information area and a question area according to the page layout information and the question information; Based on the pixel information corresponding to the public information area and the question area respectively, the page to be loaded is layered to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The common text layer, the title layer, and the background texture layer are compressed to obtain the common text loading content, title loading content, and background texture loading content of the page to be loaded. The loading content of the questions is encapsulated according to the reading order of the questions to obtain the loading content unit corresponding to the question area. Based on the target loading order among the common text loading content, the loading content unit of the questions and the background texture loading content, a target loading instruction for the page to be loaded is generated. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

2. The method according to claim 1, characterized in that, The step of layering the page to be loaded according to the pixel information corresponding to the public information area and the question area respectively, to obtain a public text layer corresponding to the public information area, a question layer corresponding to the question area, and a background texture layer corresponding to the page to be loaded, includes: Based on the layer contrast features between the page background and the page body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels of the public information area are layered, and the layer corresponding to the foreground pixels of the public information area is determined as the public text layer; Based on the layer contrast features of the page background and the main body of the page to be loaded, the layers corresponding to the foreground pixels and the layers corresponding to the background pixels of the question area are layered. The layer corresponding to the foreground pixels of the question stem text in the question area is determined as the question stem text layer, the layer corresponding to the foreground pixels of the question stem graphic in the question area is determined as the question stem graphic layer, and the layer corresponding to the foreground pixels of the handwriting area in the question area is determined as the handwriting layer. The public text layer of the public information area, the question stem text layer, the question stem graphic layer, and the handwritten layer in the question area are marked as areas to be filled in the corresponding positions on the page to be loaded. The areas to be filled are then filled by region interpolation to obtain the background texture layer corresponding to the background pixel information.

3. The method according to claim 2, characterized in that, The step of compressing the common text layer, the question layer, and the background texture layer respectively to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded includes: Based on the pixel density of the page to be loaded, the connected components of the question stem text layer in the common text layer and the question layer are extracted, and the characters in the common text layer and the question stem text layer are split based on the connected components to obtain at least one character pixel unit, as well as the coordinate information and size information of the at least one character pixel unit in the page to be loaded. The at least one character pixel unit is subjected to character similarity matching, and the character pixel units that meet the similarity matching conditions are clustered into multiple character classes based on the matching results, so that characters with similar character shapes in the at least one character pixel unit are compressed; Shape analysis is performed on the character pixel units in the multiple character classes to determine the character reference templates, and the character reference templates are used to form a global pixel dictionary; According to the reading order of the questions, the global pixel dictionary is segmented to obtain the basic character segment corresponding to the basic question reading unit and the new character segment corresponding to the new question reading unit. The characters in the new character segment are the characters added by the new question reading unit relative to the basic question reading unit. By comparing the pixel differences between the at least one character pixel unit and the character reference template, and analyzing the characters in the at least one character pixel unit that have missing or deformed strokes, the text residual data is obtained. The coordinate and size information of the at least one character pixel unit, the basic character segment, and the newly added character segment are encapsulated to obtain common text loading content and title text loading content. The text residual data is determined as delayed text loading content, which is loaded after the common text loading content and the title text loading content.

4. The method according to claim 2, characterized in that, The step of compressing the common text layer, the question layer, and the background texture layer respectively to obtain the common text loading content, question loading content, and background texture loading content of the page to be loaded includes: Edge detection is performed on the question stem image layer corresponding to the question layer to obtain the original image of the question stem layer; The original image is evaluated for vector features to determine whether there is a target image in the question image layer that can be compressed into vector data. The target image that can be compressed into vector data is then fitted with vector data to obtain vector data. The vector data is rasterized and rendered to obtain a vector rendered graphic. The pixel differences between the vector rendered graphic and the original graphic are compared, and the set of difference pixels corresponding to the original graphic is obtained based on the pixel differences between the vector rendered graphic and the original graphic. The pixels in the set of differing pixels are compressed to obtain the vector residual data between the original image and the vector data; Numerical analysis is performed on the vector residual data. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the graphics loading content. The vector data corresponding to the set of difference pixels that meet the residual value conditions is determined as the title graphics loading content. The vector residual data corresponding to the set of difference pixels that meet the residual value conditions is determined as the delayed graphics loading content. The handwriting layer is subjected to lossless compression of handwriting features to obtain handwriting loading content, and the background of the background texture layer is compressed at a target compression rate to obtain background texture loading content. The detailed texture in the background texture layer is determined as delayed background texture loading content. The background texture loading content includes compressed solid color background and background size.

5. The method according to claim 1, characterized in that, The process involves encapsulating the loaded question content according to the question reading order to obtain a question loading content unit corresponding to the question area. Based on the target loading order among the common text loading content, the question loading content unit, and the background texture loading content, a target loading instruction for the page to be loaded is generated. This target loading instruction is used by the client to load the page to be loaded, including: The text content, graphic content, and handwritten content of the question loading content are encapsulated according to the question reading order to obtain the question loading content unit corresponding to the question area; The target loading order is determined to be loading the background texture loading content, the public text loading content, the title loading content unit, and the delayed loading content in sequence. The delayed loading content includes delayed text loading content, delayed graphic loading content, and delayed background texture loading content. Based on the background texture loading content and the common text loading content, an initial loading instruction for the page to be loaded is generated. The initial loading instruction is used to instruct the client to prioritize loading the solid color background, background size, and common text loading content of the page to be loaded. Based on the reading order of the questions and the target question loading content unit within the user's reading range, a main loading instruction for the question loading content of the target question loading content unit is generated. The main loading instruction is used to instruct the client to load the question loading content of the target question loading content unit after loading the initial loading instruction. Based on the delayed text loading content and delayed image loading content corresponding to the target question loading content unit, a question detail loading instruction corresponding to the target question loading content unit is generated. The question detail loading instruction is used to instruct the client to perform differential adjustment of residual pixels on the text and graphics in the target question loading content unit based on the character reference template and vector data respectively after loading the main loading instruction, so as to perform detail residual repair on the text and graphics of the target question loading content unit. Based on the delayed background texture loading content corresponding to the background texture layer, a background detail loading instruction for the page to be loaded is generated. The background detail loading instruction is used to instruct the client to restore the details of the background texture of the page to be loaded after loading the question detail loading instruction.

6. A page loading method, characterized in that, Applied to the client side, including: Decode the target loading instructions generated by the server to obtain the target loading order and target loading content corresponding to the pages to be loaded; Based on the user's reading intention for the page to be loaded, the loading order of the questions is adjusted to obtain the question reading loading order, which is used to load the question content according to the reading order of the question reading units; The target content is loaded according to the target loading order and the question reading loading order; The target loading instruction is that the server obtains the page layout information and question information of the page to be loaded, and divides the page to be loaded into a public information area and a question area according to the page layout information and the question information. Based on the pixel information corresponding to the public information area and the question area respectively, the page to be loaded is layered to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The common text layer, the title layer, and the background texture layer are compressed to obtain the common text loading content, title loading content, and background texture loading content of the page to be loaded. The loading content of the questions is encapsulated according to the reading order of the questions to obtain the loading content unit corresponding to the question area. Based on the target loading order among the common text loading content, the loading content unit of the questions and the background texture loading content, a target loading instruction for the page to be loaded is generated. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

7. The method according to claim 6, characterized in that, The process of adjusting the loading order of the questions based on the user's reading intent of the page to be loaded, to obtain the question reading loading order, includes: Based on the user's touch interaction behavior on the page to be loaded, the user's reading intention is predicted to obtain the user's target reading intention. The touch interaction behavior includes screen swiping behavior, screen clicking behavior, screen long-term dwell behavior, and screen back swipe behavior. Based on the target reading intent, identify the question reading units within the current reading range, determine the question loading content corresponding to the question reading unit as the priority loading question, and determine the question loading content corresponding to the question reading unit outside the current reading range as the delayed loading question; The priority loading questions are sorted according to the reading order of the question reading units to determine the loading order of the priority loading questions; Based on the target reading intent and the loading order of the priority loading questions, the loading order of the loading content of the questions is adjusted to obtain the question reading loading order.

8. The method according to claim 7, characterized in that, The method further includes: When the touch interaction behavior is the screen swiping behavior, the user's target reading intention is determined to be the search intention, the loading of questions is paused and delayed, and only the common text loading content is loaded; When the touch interaction behavior is the screen click behavior, the user's target reading intention is determined as the click intention, and the priority loading questions corresponding to the question reading units within the current reading range are identified and loaded. When the touch interaction behavior is a long screen dwell behavior, the user's target reading intention is determined to be a dwell reading intention, and the priority loading question corresponding to the question reading unit within the current reading range, the text residual data and vector residual data corresponding to the priority loading question are identified and loaded. When the touch interaction behavior is the screen swipe behavior, the user's target reading intention is determined to be the swipe intention, and the historical priority loading questions, the text residual data and vector residual data corresponding to the question reading units within the historical reading range are identified and loaded.

9. A page loading device, characterized in that, Applied to the server side, including: The acquisition module is configured to acquire the page layout information and question information of the page to be loaded, and divide the page to be loaded into a public information area and a question area according to the page layout information and the question information; The layering module is configured to perform layering on the page to be loaded according to the pixel information corresponding to the public information area and the question area respectively, to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The compression module is configured to compress the common text layer, the title layer and the background texture layer respectively to obtain the common text loading content, title loading content and background texture loading content of the page to be loaded; The generation module is configured to encapsulate the question loading content according to the question reading order to obtain the question loading content unit corresponding to the question area, and generate the target loading instruction of the page to be loaded according to the target loading order among the public text loading content, the question loading content unit and the background texture loading content. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

10. A page loading device, characterized in that, Applied to the client side, including: The decoding module is configured to decode the target loading instructions generated by the server to obtain the target loading order and target loading content corresponding to the page to be loaded; The adjustment module is configured to adjust the loading order of the questions based on the user's reading intention on the page to be loaded, thereby obtaining the question reading loading order, which is used to load the question content according to the reading order of the question reading units; The loading module is configured to load the target content according to the target loading order and the question reading loading order; The target loading instruction is that the server obtains the page layout information and question information of the page to be loaded, and divides the page to be loaded into a public information area and a question area according to the page layout information and the question information. Based on the pixel information corresponding to the public information area and the question area respectively, the page to be loaded is layered to obtain the public text layer corresponding to the public information area, the question layer corresponding to the question area, and the background texture layer corresponding to the page to be loaded. The common text layer, the title layer, and the background texture layer are compressed to obtain the common text loading content, title loading content, and background texture loading content of the page to be loaded. The loading content of the questions is encapsulated according to the reading order of the questions to obtain the loading content unit corresponding to the question area. Based on the target loading order among the common text loading content, the loading content unit of the questions and the background texture loading content, a target loading instruction for the page to be loaded is generated. The target loading instruction is used by the client to load the page to be loaded based on the target loading instruction.

Images