A PDF image quick viewing system supporting cross-platform

Abstract

The present invention discloses a PDF image quick viewing system supporting cross-platform, which belongs to the field of image data processing, and includes: a PDF parsing module, which is used to parse and identify the image format of the PDF image and generate character data; a point processing module, which is used to capture and annotate key points and perform cross-platform processing to generate a second PDF image; a difference detection module, which is used to calculate the similarity coefficient and perform automatic comparison to generate a difference detection result; an image filling module, which is used to automatically fill the second PDF image according to the detection result to generate a third PDF image; and an intelligent search module, which is used to perform cross-platform intelligent search using character data and the third PDF image. The present application solves the technical problem that PDF images in the prior art cannot realize cross-platform quick viewing and intelligent search, and achieves the technical effect of realizing cross-platform quick viewing and intelligent search of PDF images, and improving the efficiency of PDF image processing and information retrieval.

Description

Technical Field

[0001] The invention relates to the field of image data processing, and in particular to a PDF image quick viewing system supporting cross-platforms. Background Art

[0002] With the rapid development of information technology, PDF has become an important format for document storage and exchange. However, there are still many deficiencies in PDF image processing and application in the prior art. First, the poor compatibility of PDF image formats on different platforms makes it difficult to view and apply them across platforms. When switching between different devices, problems such as PDF files being unable to open and display abnormalities may occur, affecting the user experience. Secondly, the massive amount of text, graphics and other information contained in PDF images is difficult to extract and retrieve quickly and accurately, resulting in low efficiency in intelligent search of PDF images, making it difficult to quickly locate the required information from PDF documents. Therefore, PDF images in the prior art cannot achieve cross-platform quick viewing and intelligent search. Summary of the invention

[0003] The present application provides a PDF image quick viewing system that supports cross-platform, aiming to solve the technical problem in the prior art that PDF images cannot achieve cross-platform quick viewing and intelligent search.

[0004] The present application discloses a system for quickly viewing a PDF image supporting cross-platform, comprising: a PDF parsing module, used to parse a first PDF image to determine a plurality of PDF image formats, recognize the first PDF image according to the plurality of PDF image formats, and generate a plurality of character data; a point processing module, used to traverse the first PDF image to capture in real time, mark points according to the capture results, determine a plurality of key points, perform cross-platform processing based on the plurality of key points, and generate a second PDF image; a difference detection module, used to perform scaling calculation based on the first PDF image and the second PDF image to obtain a structural similarity coefficient, automatically compare and detect the first PDF image and the second PDF image according to the structural similarity coefficient, and generate a drawing difference detection result; an image filling module, used to automatically fill the second PDF image according to the drawing difference detection result, and generate a third PDF image; and an intelligent search module, used to perform cross-platform intelligent search and viewing on the first PDF image based on the plurality of character data and the third PDF image.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] The first PDF image is parsed by a PDF parsing module to determine multiple PDF image formats, and the first PDF image is recognized according to these formats to generate multiple character data and extract text information from the PDF image, thereby laying a foundation for subsequent intelligent search. The first PDF image is traversed by a point processing module for real-time capture, and points are marked according to the capture results to determine multiple key points. Cross-platform processing is performed based on these key points to generate a second PDF image, thereby achieving compatibility processing of PDF images on different platforms and ensuring that they can be quickly viewed across platforms. The difference detection module is used to perform scaling calculations based on the first PDF image and the second PDF image to obtain a structural similarity coefficient, and the two PDF images are automatically compared and detected based on the coefficient. Generate drawing difference detection results, efficiently compare different versions of PDF images, and locate the differences therein; on this basis, automatically fill the second PDF image according to the drawing difference detection results through the image filling module to generate a third PDF image, automatically correct and synchronize the differences of the PDF image, and ensure the integrity and consistency of its content; use the intelligent search module to perform cross-platform intelligent search and viewing of the first PDF image based on multiple character data and the third PDF image; realize the technical solution of efficient retrieval and positioning of PDF image content, solve the technical problem that PDF images in the prior art cannot realize cross-platform fast viewing and intelligent search, and achieve the technical effect of realizing cross-platform fast viewing and intelligent search of PDF images and improving the efficiency of PDF image processing and information retrieval.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic diagram of a structure of a system supporting cross-platform PDF image quick viewing is provided for an embodiment of the present application;

[0009] Figure 2 A flowchart of generating image analysis results in a PDF image quick viewing system supporting cross-platform is provided for an embodiment of the present application.

[0010] Explanation of the reference numerals: PDF parsing module 11 , point processing module 12 , difference detection module 13 , image filling module 14 , intelligent search module 15 . DETAILED DESCRIPTION

[0011] The overall idea of ​​the technical solution provided by this application is as follows:

[0012] The embodiment of the present application provides a PDF image quick viewing system that supports cross-platform. First, the first PDF image is parsed by the PDF parsing module to determine multiple PDF image formats, and the first PDF image is identified according to these formats to generate multiple character data, laying the foundation for subsequent intelligent search. Secondly, the point processing module is used to traverse the first PDF image for real-time capture, and the point is marked according to the capture result, multiple key points are determined, and cross-platform processing is performed based on these key points to generate a second PDF image, so as to achieve compatibility processing of PDF images under different platforms. Thirdly, the difference detection module is used to perform scaling calculation based on the first PDF image and the second PDF image to obtain the structural similarity coefficient, and the two PDF images are automatically compared and detected according to the coefficient, and the drawing difference detection result is generated, and the PDF images of different versions are efficiently compared to locate the difference parts therein. Then, the second PDF image is automatically filled according to the drawing difference detection result by the image filling module to generate a third PDF image, and the difference parts of the PDF image are automatically corrected and synchronized to ensure the integrity and consistency of its content. Afterwards, the intelligent search module is used to perform cross-platform intelligent search and viewing of the first PDF image based on multiple character data and the third PDF image, so as to realize efficient retrieval and positioning of the PDF image content, achieve cross-platform quick viewing and intelligent search of PDF images, and enhance the technical effect of user experience.

[0013] After introducing the basic principles of the present application, the non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the drawings in the specification.

[0014] like Figure 1 As shown, the embodiment of the present application provides a system for quickly viewing PDF images across platforms, the system comprising:

[0015] The PDF parsing module 11 is used to parse the first PDF image to determine a plurality of PDF image formats, recognize the first PDF image according to the plurality of PDF image formats, and generate a plurality of character data.

[0016] Specifically, first, the PDF parsing module 11 analyzes the data structure and encoding method of the first PDF image to determine the different PDF image formats contained therein, and obtains multiple PDF image formats, such as text format, image format (such as JPEG, PNG, etc.), vector image format (such as SVG), etc. After determining multiple PDF image formats, the PDF parsing module 11 uses corresponding recognition algorithms to extract content in various formats. For example, for text format, optical character recognition technology is used to convert the text content in the image into editable text data; for image format, key graphic information is extracted through methods such as image segmentation and edge detection; for vector image format, vector image elements are parsed and related mathematical description information is obtained.

[0017] Through the above targeted recognition processing, the PDF parsing module 11 extracts corresponding character data from different format areas of the first PDF image to obtain multiple character data, such as text characters, image features or vector graphics parameters, which together constitute a digital representation of the first PDF image content.

[0018] The PDF parsing module 11 parses a variety of PDF image formats and performs accurate recognition, effectively obtaining key character information from the input first PDF image, and laying a foundation for realizing cross-platform quick viewing of PDF images.

[0019] The point processing module 12 is used to traverse the first PDF image to perform real-time capture, mark points according to the capture results, determine multiple key points, perform cross-platform processing based on the multiple key points, and generate a second PDF image.

[0020] Specifically, the point processing module 12 traverses the first PDF image and captures the key information in the image in real time. During the traversal process, the point processing module 12 adopts a specific capture algorithm, such as pixel scanning, region segmentation, etc., to accurately locate the important points in the first PDF image. The capture algorithm adaptively adjusts the capture strategy according to the content characteristics of the first PDF image to ensure the accuracy and comprehensiveness of the capture results. According to the capture results, the point processing module 12 marks the key points in the first PDF image. Specifically, the point processing module 12 analyzes the captured point information, such as coordinate position, color attributes, shape characteristics, etc., and combines the preset point marking rules to assign a unique identifier to each key point. The identifier can be a number, a letter or a combination thereof, which is used to uniquely identify and locate each key point.

[0021] Through the point marking process, the point processing module 12 determines multiple key points in the first PDF image. These key points include important content elements such as text blocks, image areas, table cells, etc., which have special significance in the structure and semantics of the first PDF image. Determining multiple key points provides a benchmark and basis for subsequent cross-platform processing. After obtaining multiple key points, the point processing module 12 uses these key point information for cross-platform processing to generate a second PDF image. Specifically, the point processing module 12 maps multiple key points to the coordinate system of the target platform, and converts and adjusts the key point coordinates according to the display characteristics of the target platform. Through the cross-platform mapping of multiple key points, the consistent display of the first PDF image on different platforms is ensured. During the cross-platform processing, the point processing module 12 simultaneously considers factors such as resolution and size of different platforms, and performs corresponding scaling, cropping or filling operations on the first PDF image to adapt to the display requirements of the target platform. After a series of cross-platform processing, the point processing module 12 generates a second PDF image that is consistent with the content of the first PDF image and meets the display specifications of the target platform.

[0022] The point processing module 12 traverses the first PDF image for real-time capture, performs point marking according to the capture results, determines multiple key points, and performs cross-platform processing based on the multiple key points to generate a second PDF image suitable for different platforms, thereby realizing a cross-platform quick viewing function for the first PDF image.

[0023] The difference detection module 13 is used to perform scaling calculation based on the first PDF image and the second PDF image to obtain a structural similarity coefficient, and automatically compare and detect the first PDF image and the second PDF image according to the structural similarity coefficient to generate a drawing difference detection result.

[0024] Specifically, first, the difference detection module 13 performs scaling calculation on the first PDF image and the second PDF image. The scaling calculation process has two aspects: one is resolution scaling, which adjusts the first PDF image and the second PDF image to the same resolution to eliminate the comparison error caused by the resolution difference; the other is size scaling, which scales the first PDF image and the second PDF image in proportion to their original sizes to make them visually comparable. The scaling calculation ensures the consistency and accuracy of the first PDF image and the second PDF image during comparison detection.

[0025] After completing the scaling calculation, the difference detection module 13 introduces a structural similarity coefficient as a quantitative indicator to measure the similarity between the first PDF image and the second PDF image. The structural similarity coefficient compares the pixels in the local window pixel by pixel based on the image's brightness, contrast, structure and other features to calculate the structural similarity between the two images. The structural similarity coefficient ranges from 0 to 1, where 0 indicates complete dissimilarity and 1 indicates complete similarity. The difference detection module 13 quantitatively evaluates the difference between the first PDF image and the second PDF image by calculating the structural similarity coefficient.

[0026] Based on the obtained structural similarity coefficient, the difference detection module 13 automatically compares and detects the first PDF image and the second PDF image. Specifically, when the structural similarity coefficient is lower than a preset threshold, the difference detection module 13 determines that there are significant differences between the two images, and triggers the difference recognition process, and accurately locates and identifies the difference areas between the first PDF image and the second PDF image through techniques such as pixel-by-pixel scanning and region segmentation. These difference areas may include changes in text content, addition or deletion of image elements, adjustment of layout structure, etc. The difference detection module 13 visualizes the identified difference areas and generates intuitive drawing difference detection results. For example, the difference areas are clearly identified on the first PDF image and the second PDF image by color marking, highlighting, etc. This visualized difference detection result facilitates rapid positioning and viewing of differences between drawings, and improves the efficiency and accuracy of difference detection.

[0027] The difference detection module 13 realizes fine difference recognition between the first PDF image and the second PDF image through scaling calculation, structural similarity coefficient calculation and automatic comparison detection, generates intuitive drawing difference detection results, and provides a basis for subsequent automatic filling of the second PDF image.

[0028] The image filling module 14 is used to automatically fill the second PDF image according to the drawing difference detection result to generate a third PDF image.

[0029] Specifically, the image filling module 14 automatically fills the second PDF image according to the drawing difference detection result generated by the difference detection module 13. The drawing difference detection result clearly identifies the difference area between the first PDF image and the second PDF image, providing the image filling module 14 with an accurate target area and filling basis.

[0030] First, the image filling module 14 analyzes the drawing difference detection results and extracts the position, size, shape and other information of the difference area to be filled. Then, the image filling module 14 adaptively selects the appropriate filling algorithm and strategy according to the characteristics of the difference area. For example, for the difference area of ​​text content, the image filling module 14 uses OCR technology to identify the text of the corresponding area in the first PDF image and fills it into the corresponding position of the second PDF image; for the difference area of ​​image elements, the image filling module 14 extracts the corresponding image fragments from the first PDF image and seamlessly merges them into the second PDF image; for the difference area of ​​layout structure, the image filling module 14 refers to the layout information of the first PDF image to rearrange and adjust the second PDF image to maintain consistency with the first PDF image. In the filling process, the image filling module 14 makes full use of the content information of the first PDF image to ensure that the filling result is highly consistent with the first PDF image. At the same time, the image filling module 14 considers the coordination between the filling area and the original content of the second PDF image, and through smooth transition, edge optimization and other technologies, the filled image has a natural and coherent visual effect.

[0031] After the automatic filling process of the image filling module 14, a filled third PDF image is generated. The third PDF image seamlessly fills the difference area between the first PDF image and the second PDF image on the basis of retaining the original content of the second PDF image, thereby achieving complete synchronization of the contents of the two images. The third PDF image is highly consistent with the first PDF image and is adapted to the display requirements of the target platform, providing convenience for fast viewing of PDF images across platforms.

[0032] The intelligent search module 15 is used to perform a cross-platform intelligent search and viewing on the first PDF image based on the multiple character data and the third PDF image.

[0033] Specifically, the intelligent search module 15 comprehensively utilizes the multiple character data extracted by the PDF parsing module 11 and the third PDF image generated by the image filling module 14 to implement a cross-platform intelligent search and viewing function for the first PDF image.

[0034] First, the intelligent search module 15 uses multiple character data as search indexes to establish an efficient search engine. Among them, the search engine uses text retrieval algorithms, such as full-text indexing, inverted indexing, etc., to optimize the organization and indexing of multiple character data to support fast and accurate keyword searches. Users can retrieve relevant PDF image content from multiple character data by entering search conditions such as keywords and phrases. When the user searches and views the first PDF image on the target platform, the intelligent search module 15 passes the user's search request to the search engine. The search engine matches and retrieves in the indexes of multiple character data according to the search conditions, and quickly locates the PDF image area containing the search keywords. The search results are presented to the user in the form of highlighting, link jumping, etc., which is convenient for the user to quickly browse and locate the content of interest.

[0035] In order to provide a consistent search experience across platforms, the intelligent search module 15 uses the third PDF image to visualize the search results. Since the third PDF image has been processed by the image filling module 14, it is highly consistent with the first PDF image in terms of content and vision. Therefore, when the user views the third PDF image on the target platform, he can obtain the same search result display effect as viewing the first PDF image on the original platform. In addition, the intelligent search module 15 can also provide other search functions, such as fuzzy matching, semantic understanding, relevance sorting, etc., to further improve the accuracy of the search and user experience. For example, through fuzzy matching, even if the search keywords entered by the user contain typos, synonyms, etc., the intelligent search module 15 can correctly identify the user's search intent and return relevant search results; through semantic understanding, the intelligent search module 15 analyzes the contextual semantics of the user's search conditions and provides more accurate and comprehensive search results; through relevance sorting, the intelligent search module 15 intelligently sorts the search results according to the degree of relevance between the search results and the user's search conditions, and presents the most relevant content to the user first.

[0036] Through the intelligent search module 15, a search engine based on multiple character data is constructed, and combined with the visual presentation of the third PDF image, a cross-platform intelligent search and viewing of the first PDF image is realized. The user can quickly retrieve and locate the key content in the first PDF image through the convenient search function, and obtain a viewing experience consistent with the original platform.

[0037] Furthermore, the embodiment of the present application also includes:

[0038] The first PDF image is traversed according to the multiple PDF image formats to perform edge detection and determine multiple geometric features; the first PDF image is point-captured based on the multiple geometric features to generate the capture result; a shape matching algorithm is used to perform point clustering analysis based on the capture result to generate multiple point categories, and multiple point labels are generated according to the multiple point categories; the capture result is point-labeled based on the multiple point labels to determine the multiple key points.

[0039] In a feasible implementation, first, the point processing module 12 performs edge detection on the first PDF image format by format according to the multiple PDF image formats determined by the PDF parsing module 11. Specifically, the point processing module 12 uses corresponding edge detection algorithms, such as the Canny algorithm, the Sobel algorithm, etc., according to different PDF image formats, to extract edges of different types of content such as text, graphics, and tables in the first PDF image. Through edge detection, the point processing module 12 determines multiple geometric features in the first PDF image, such as the rectangular border of the text block, the contour edge of the image, the cell boundary of the table, etc., to provide important reference information for subsequent point capture. Then, the point processing module 12 performs point capture on the first PDF image based on multiple geometric features. The point processing module 12 identifies key points in the first PDF image by analyzing the properties of the geometric features such as the position, size, and shape. For example, for the rectangular border of the text block, the point processing module 12 extracts its corner points as key points; for the contour edge of the image, the point processing module 12 selects its feature points (such as inflection points, extreme points, etc.) as key points; for the cell boundary of the table, the point processing module 12 uses the intersection as the key point. Through point capture, the point processing module 12 generates a set of capture results, which includes the key point information of the first PDF image.

[0040] Next, the point processing module 12 uses a shape matching algorithm to perform point clustering analysis on the capture results. Specifically, the point processing module 12 compares the geometric feature similarity between different key points, such as distance, angle, shape, etc., through shape matching algorithms such as template matching, feature point matching, and contour matching, and classifies key points with similar features into the same category. Through point clustering analysis, the point processing module 12 generates multiple point categories, each of which contains key points with similar geometric features. After that, the point processing module 12 generates corresponding point tags according to multiple point categories, and annotates the capture results based on these point tags. Among them, the point tag is a predefined symbol, icon or text mark used to represent key points of different categories. The point processing module 12 associates the point tag with the key points in the capture result to generate point data with annotation information. Through point annotation, the point processing module 12 finally determines multiple key points in the first PDF image and gives clear semantic attributes and category information to each key point.

[0041] Through the point processing module 12, accurate extraction and classification of key points of the first PDF image are achieved. This fine-grained point processing method improves the accuracy and reliability of cross-platform PDF image quick viewing.

[0042] Furthermore, the embodiment of the present application also includes:

[0043] The method comprises the steps of: performing a platform framework analysis on the target platform according to the multiple key points to determine multiple reference points; performing a cross-platform coordinate transformation on the first PDF image according to the multiple reference points to generate multiple transformation points; mapping the multiple transformation points to the target platform according to the multiple PDF image formats to determine a cross-platform initial PDF image; comparing the first PDF image with the initial PDF image, performing an image overlay analysis according to the comparison result to generate an image analysis result; and adjusting the initial PDF image based on the image analysis result to generate the second PDF image.

[0044] In a feasible implementation, in the process of performing cross-platform processing based on multiple key points to generate the second PDF image, the point processing module 12 first performs a platform framework analysis on the target platform according to the multiple key points. Specifically, the point processing module 12 considers factors such as the display device characteristics, operating system environment, and user interface style of the target platform, and determines multiple reference points adapted to the target platform in combination with the distribution of key points. These reference points have fixed positions and semantic attributes in the coordinate system of the target platform, providing a basis for subsequent cross-platform coordinate conversion. Then, the point processing module 12 performs cross-platform coordinate conversion on the first PDF image according to the multiple reference points. The point processing module 12 establishes a coordinate mapping relationship between the first PDF image and the target platform, and converts the coordinates of the key points in the first PDF image into corresponding coordinates in the coordinate system of the target platform. Through the coordinate conversion, the point processing module 12 generates a set of new point data, namely, multiple conversion points, which reflect the position and layout information of the first PDF image on the target platform.

[0045] Next, the point processing module 12 maps the multiple conversion points to the target platform according to the multiple PDF image formats determined by the PDF parsing module 11, and generates a cross-platform initial PDF image. Specifically, the point processing module 12 uses corresponding mapping algorithms and strategies according to different PDF image formats to map the conversion points to the image drawing interface of the target platform. For example, for text format, the point processing module 12 maps the conversion points to the text rendering position of the target platform; for graphic format, the point processing module 12 maps the conversion points to the graphic drawing coordinates of the target platform; for table format, the point processing module 12 maps the conversion points to the table cell position of the target platform. Through the formatted point mapping, the point processing module 12 generates an initial PDF image corresponding to the content of the first PDF image, which has been adapted to the display environment of the target platform. However, due to the differences between different platforms, there may be a certain visual deviation between the initial PDF image and the first PDF image. In order to further improve the consistency of the cross-platform PDF image, the point processing module 12 compares the first PDF image with the initial PDF image, and performs image overlay analysis based on the comparison results. Specifically, the point processing module 12 uses layer overlay technology to overlap the first PDF image and the initial PDF image, and determines the difference area between the two images by comparing the pixel differences between the two images. The point processing module 12 uses different colors, transparency and other visual effects to highlight the difference area and generate intuitive image analysis results. After that, the point processing module 12 fine-tunes and optimizes the initial PDF image based on the image analysis results to generate the final second PDF image. The point processing module 12 makes local adjustments to the corresponding area in the initial PDF image according to the position and size of the difference area, such as position translation, size scaling, content alignment, etc., so that it is as consistent as possible with the corresponding area of ​​the first PDF image. After a series of fine-tuning operations, the point processing module 12 obtains a second PDF image that is highly similar to the first PDF image, realizing cross-platform PDF image quick viewing.

[0046] Through the point processing module 12, a second PDF image with high cross-platform consistency and good visual quality is generated, which fully considers the characteristics and differences of different platforms and effectively solves the compatibility and consistency problems of cross-platform PDF image quick viewing.

[0047] Further, such as Figure 2 As shown, the embodiment of the present application also includes:

[0048] The first PDF image and the initial PDF image are subjected to image grayscale processing to generate a first PDF grayscale image and an initial PDF grayscale image; the first PDF image and the initial PDF image are subjected to image registration according to the first PDF grayscale image and the initial PDF grayscale image to generate an image alignment result; based on the image alignment result, an absolute difference between the first PDF grayscale image and the initial PDF grayscale image is calculated for comparison to generate the comparison result, wherein the comparison result includes an absolute difference; and according to the absolute difference, the first PDF image and the initial PDF image are subjected to transparency superposition to generate the image analysis result.

[0049] In a preferred embodiment, when the point processing module 12 compares the first PDF image with the initial PDF image and performs image superposition analysis based on the comparison result to generate an image analysis result, first, the point processing module 12 performs image grayscale processing on the first PDF image and the initial PDF image. Specifically, the point processing module 12 converts the first PDF image into a first PDF grayscale image and converts the initial PDF image into an initial PDF grayscale image by removing the color information in the image. Grayscale processing can simplify the color complexity of the image, highlight the light and dark contrast and structural information of the image, and provide a more accurate and efficient data basis for subsequent image comparison and analysis. Then, the point processing module 12 performs image registration on the first PDF image and the initial PDF image according to the structural features of the first PDF grayscale image and the initial PDF grayscale image. Specifically, the point processing module 12 finds the corresponding relationship between the two images by analyzing the geometric information such as feature points, edges, and contours of the two grayscale images, and aligns the two images to the same coordinate space through transformation operations such as translation, rotation, and scaling. Image registration can eliminate the position deviation and geometric distortion caused by the cross-platform conversion process, and ensure the pixel-level consistency between the first PDF image and the initial PDF image during comparison and analysis.

[0050] After completing the image registration, the point processing module 12 calculates the absolute difference between the first PDF grayscale image and the initial PDF grayscale image based on the image alignment result, and performs pixel-by-pixel comparative analysis. Specifically, the point processing module 12 obtains a difference image by calculating the absolute value of the difference between the grayscale values ​​of corresponding pixels in the two grayscale images. Each pixel value in the difference image represents the grayscale difference degree between the first PDF image and the initial PDF image at that position. By analyzing the difference image, the point processing module 12 accurately locates and quantifies the content difference between the two PDF images, and generates a comparison result containing the absolute difference. Afterwards, the point processing module 12 superimposes the transparency of the first PDF image and the initial PDF image according to the absolute difference in the comparison result, and generates an intuitive image analysis result. Specifically, the point processing module 12 dynamically adjusts the transparency of the first PDF image and the initial PDF image according to the size of the absolute difference. The area with a large difference has a lower transparency (such as 30%), and the area with a small difference has a higher transparency (such as 70%). Then, the point processing module 12 superimposes and fuses the two PDF images with different transparency to generate a new image. In the superimposed image, the area with greater differences has darker colors, and the area with smaller differences has lighter colors. Through transparency superposition, the image analysis results intuitively show the distribution and degree of differences between the first PDF image and the initial PDF image in a visual way.

[0051] Through the point processing module 12, a refined comparative analysis of the first PDF image and the initial PDF image is achieved to generate a highly intuitive and quantitative image analysis result. This image overlay analysis method can accurately detect and locate the visual differences of PDF images during the cross-platform conversion process, providing a reliable basis for subsequent image optimization and adjustment, and ensuring the high consistency and reliability of the cross-platform PDF image quick viewing.

[0052] Furthermore, the embodiment of the present application also includes:

[0053] The first PDF image and the second PDF image are scaled in resolution to generate a first scaling result; the first PDF image and the second PDF image are scaled in size to generate a second scaling result; a local image window is set according to the multiple key points, and a pixel-by-pixel sliding comparison is performed on the first PDF image and the second PDF image through the local image window to generate a pixel comparison result; and a spatial structure calculation is performed on the first PDF image and the second PDF image according to the pixel comparison result to determine a structural similarity coefficient, where the structural similarity coefficient has a value range of 0-1.

[0054] In a feasible implementation manner, in the process of performing scaling calculation based on the first PDF image and the second PDF image to obtain the structural similarity coefficient, the difference detection module 13 firstly performs resolution scaling on the first PDF image and the second PDF image. Specifically, the difference detection module 13 performs proportional resampling processing on the first PDF image and the second PDF image according to the preset target resolution. Through resolution scaling, the pixel density of the two PDF images is unified, eliminating the inconsistency of image quality and detail performance caused by the resolution difference. After the resolution scaling, the difference detection module 13 generates the first PDF image and the second PDF image with consistent resolution, that is, the first scaling result. Then, the difference detection module 13 performs size scaling on the first PDF image and the second PDF image. Specifically, the difference detection module 13 performs proportional scaling processing on the first PDF image and the second PDF image according to the preset target size. Through size scaling, the physical sizes of the two PDF images are unified, eliminating the inconsistency of image layout and relative position caused by the size difference. After size scaling, the difference detection module 13 generates the first PDF image and the second PDF image with consistent size, that is, the second scaling result.

[0055] After completing the resolution scaling and the size scaling, the difference detection module 13 sets the corresponding image local windows in the first PDF image and the second PDF image according to the multiple key points generated by the point processing module 12. Specifically, the difference detection module 13 takes each key point as the center and determines the local area in the two PDF images according to the preset window size and shape. Then, the difference detection module 13 performs a pixel-by-pixel sliding comparison on the first PDF image and the second PDF image through the image local windows. Specifically, the difference detection module 13 slides the local window of one image on the other image at the pixel level, calculates the sum of the grayscale differences of all pixels in the window, and obtains the matching metric of the window position. By continuously sliding the local window, the difference detection module 13 generates a pixel comparison result containing the matching metrics of all window positions. Afterwards, the difference detection module 13 performs spatial structure calculation on the first PDF image and the second PDF image according to the pixel comparison result to obtain the structural similarity coefficient. Specifically, the difference detection module 13 evaluates the similarity of the two PDF images in local structure and spatial layout by analyzing the matching metric distribution of different window positions in the pixel comparison result. The difference detection module 13 comprehensively considers the statistical characteristics (such as mean, variance, histogram, etc.) and spatial distribution characteristics (such as connectivity, symmetry, regularity, etc.) of the window matching metric, and obtains a normalized structural similarity coefficient through weighted calculation. The structural similarity coefficient ranges from 0 to 1, where 0 indicates that the structures of the two PDF images are completely dissimilar, and 1 indicates that the structures of the two PDF images are exactly the same. The structural similarity coefficient quantifies the structural consistency and similarity between the first PDF image and the second PDF image after cross-platform conversion.

[0056] The structural similarity evaluation of the first PDF image and the second PDF image is realized by the difference detection module 13, and an objective and quantitative structural similarity coefficient is obtained. This scaling calculation and similarity evaluation method can fully consider factors such as resolution, size and local structure of PDF images during cross-platform conversion, provide a reliable metric for subsequent difference detection and optimization, and ensure a high degree of consistency and user experience for fast viewing of PDF images across platforms.

[0057] Furthermore, the embodiment of the present application also includes:

[0058] When the value of the structural similarity coefficient is greater than or equal to 0 and less than 1, an image detection instruction is generated; a difference recognition module is activated through the image detection instruction, and the first PDF image is traversed according to the structural similarity coefficient based on the difference recognition module, and the second PDF image is marked according to the traversal result to determine multiple difference areas; it is judged whether there are difference areas greater than or equal to a preset area area threshold among the multiple difference areas, and if there are difference areas greater than or equal to the preset area area threshold among the multiple difference areas, the difference areas greater than or equal to the preset area area threshold are color-marked to generate multiple difference marked areas; visualization processing is performed based on the multiple difference marked areas to generate a difference heat map; and the difference heat map is added to the drawing difference detection result.

[0059] In a feasible implementation, when the difference detection module 13 automatically compares and detects the first PDF image and the second PDF image according to the structural similarity coefficient to generate the drawing difference detection result, first, the difference detection module 13 determines whether the value of the structural similarity coefficient is greater than or equal to 0 and less than 1. When the structural similarity coefficient falls within this interval, it means that there are certain differences in structure between the first PDF image and the second PDF image, and further difference detection and positioning are required. At this time, the difference detection module 13 generates an image detection instruction for triggering subsequent difference recognition operations. Then, the difference detection module 13 activates the built-in difference recognition module through the image detection instruction. The difference recognition module adaptively adjusts the accuracy and threshold of difference recognition according to the size of the structural similarity coefficient. Specifically, the difference recognition module traverses the first PDF image region by region according to the structural similarity coefficient, and identifies the region with significant differences from the second PDF image through methods such as sliding windows and pixel comparison. The difference recognition module marks the corresponding position on the second PDF image according to the traversal result to generate multiple difference regions.

[0060] Next, the difference detection module 13 determines whether there is a difference area whose area is greater than or equal to the preset area area threshold among the identified multiple difference areas. The preset area area threshold is determined based on factors such as the size, resolution and user visual perception of the PDF image, and is used to filter out the difference areas with smaller areas and less impact. If there are difference areas exceeding the threshold, the difference detection module 13 performs color identification on these areas to generate eye-catching difference marked areas. After generating the difference marked areas, the difference detection module 13 performs visualization processing on the multiple difference marked areas to generate an intuitive difference heat map. The difference heat map uses a continuous color spectrum to represent the intensity and distribution of the difference. Specifically, the difference detection module 13 calculates the difference intensity of each difference marked area based on the area size and structural similarity coefficient, and maps it to a preset color spectrum. Among them, the color spectrum adopts a continuous gradient of red-yellow-blue. The difference detection module 13 combines the position, shape and difference intensity information of the difference marked area to generate a difference heat map covering the entire PDF image. Afterwards, the difference detection module 13 adds the generated difference heat map to the drawing difference detection result as an intuitive visual representation. Among them, the drawing difference detection results also include quantitative indicators such as the location coordinates, area size, structural similarity coefficient of the difference area, as well as auxiliary information such as thumbnails of the difference area and local enlarged images.

[0061] The difference detection module 13 realizes automatic comparison and detection of the first PDF image and the second PDF image, and generates information-rich, intuitive and readable drawing difference detection results. This automated and intelligent difference detection method can quickly and accurately discover and locate the key differences generated by PDF images during cross-platform conversion.

[0062] Furthermore, the embodiment of the present application also includes:

[0063] Based on the multiple character data, fuzzy matching is performed on the third PDF image to obtain a matching result, wherein the matching result includes a matching score; matching analysis is performed on the multiple character data through a context window according to the matching result, and an expected matching score is set according to the analysis result; it is determined whether the matching score is greater than or equal to the expected matching score, and if the matching score is greater than or equal to the expected matching score, an image query instruction is generated; and the image query instruction allows cross-platform intelligent search and viewing of the first PDF image.

[0064] In a preferred embodiment, in the process of performing cross-platform intelligent search and viewing of the first PDF image based on multiple character data and the third PDF image, the intelligent search module 15 first performs fuzzy matching on the third PDF image generated by the image filling module 14 based on the multiple character data extracted by the PDF parsing module 11. Specifically, the intelligent search module 15 uses the multiple character data as search keywords and performs flexible matching in the text content of the third PDF image. Fuzzy matching uses a variety of technologies, such as regular expressions, wildcards, stem extraction, etc., to adapt to different expressions and variations of keywords. The intelligent search module 15 finds all text fragments related to the keyword in the third PDF image through fuzzy matching, and generates matching results including matching positions, matching lengths and matching scores. Among them, the matching score measures the similarity between each matching fragment and the keyword. Then, the intelligent search module 15 performs matching analysis on the multiple character data through the context window according to the matching results. Specifically, the intelligent search module 15 takes each matching fragment as the center and extracts a certain length of text before and after it as the context window. The context window provides context information of the matching fragment for evaluating the accuracy and relevance of the match. The intelligent search module 15 performs semantic analysis, keyword extraction, topic modeling, etc. on the text in the context window to calculate its relevance to multiple character data. Based on the result of the matching analysis, the intelligent search module 15 sets an expected matching score as a threshold for judging the matching quality.

[0065] Next, the intelligent search module 15 determines whether the matching score of each matching segment is greater than or equal to the expected matching score. If the matching score is greater than or equal to the expected matching score, it means that the matching segment has a high correlation and matching degree with the keyword, and is content that the user may be interested in. At this time, the intelligent search module 15 generates an image query instruction to mark the position and range of the matching segment in the third PDF image. After that, the intelligent search module 15 presents the content related to the user query on the third PDF image through the image query instruction, and realizes the cross-platform intelligent search and viewing of the first PDF image. Specifically, the intelligent search module 15 performs visual processing such as highlighting, underlining, and pop-up box prompts at the corresponding position of the third PDF image according to the image query instruction. Since the third PDF image is generated on the basis of the first PDF image through cross-platform conversion and optimization, the search results seen by the user on the third PDF image correspond to the content in the first PDF image one by one. This cross-platform intelligent search and viewing method provides users with a search experience consistent with the original platform.

[0066] Through the intelligent search module 15, a cross-platform intelligent search and viewing function based on multiple character data and the third PDF image is realized. This intelligent search method makes full use of the results of PDF parsing and image filling, can quickly and accurately find the content that the user is interested in, and provides a consistent search experience on different platforms, thereby improving the efficiency and convenience of cross-platform PDF image quick viewing.

[0067] Furthermore, the embodiment of the present application also includes:

[0068] If the matching score is less than the expected matching score, a multi-directional tracing instruction is generated, and the multi-directional tracing instruction includes multiple adjustment conditions; a multi-dimensional tracing analysis is performed on the third PDF image through the multi-directional tracing instruction to generate a multi-dimensional abnormal data source; the third PDF image is analyzed according to the multi-dimensional abnormal data source in combination with the first PDF image to generate an image tracing matrix; the multiple adjustment conditions are combined based on the image tracing matrix to generate a multi-condition combination result, and a cross-platform intelligent search and viewing is performed on the first PDF image according to the multi-condition combination result.

[0069] In a preferred embodiment, when the intelligent search module 15 determines whether the matching score is greater than or equal to the expected matching score, if the intelligent search module 15 finds that the matching score of a certain matching segment is less than the expected matching score, it indicates that the segment has a low correlation and matching degree with the keyword, and there may be differences or distortions introduced during the cross-platform conversion process. At this time, the intelligent search module 15 generates a multi-directional traceability instruction to start an in-depth traceability analysis process. The multi-directional traceability instruction contains a series of adjustment conditions for guiding the traceability analysis, such as text encoding format, image compression quality, font style, typesetting layout, etc. According to the multi-directional traceability instruction, the intelligent search module 15 performs a multi-dimensional traceability analysis on the third PDF image. Specifically, the intelligent search module 15 starts from multiple dimensions such as text content, image quality, and format compatibility to trace the potential causes of the low matching score. For example, the intelligent search module 15 checks whether the text in the third PDF image has garbled characters, encoding errors, etc.; analyzes whether the image's resolution, compression ratio, color mode, and other attributes affect the recognition of keywords; and evaluates whether the format conversion, font mapping, layout adjustment, and other operations used in the cross-platform conversion process introduce differences. Through multi-dimensional traceability analysis, the intelligent search module 15 generates a set of multi-dimensional abnormal data sources, recording various factors and clues that lead to matching failures.

[0070] Then, the intelligent search module 15 further analyzes and processes the third PDF image in combination with the first PDF image according to the multidimensional abnormal data source. Specifically, the intelligent search module 15 uses the first PDF image as a reference, compares and analyzes the abnormal conditions of the third PDF image in different dimensions, and generates an image traceability matrix. The image traceability matrix is ​​a multidimensional data structure, in which each element represents the degree of difference between the third PDF image and the first PDF image in a certain position and a certain attribute. Through the image traceability matrix, the intelligent search module 15 comprehensively and finely evaluates the conversion quality and restoration degree of the third PDF image. Afterwards, the intelligent search module 15 combines multiple adjustment conditions in the multi-directional traceability instruction based on the image traceability matrix to generate a multi-condition combination result. The multi-condition combination result includes adjustment strategies of different dimensions and granularities, such as improving image resolution, changing text encoding format, adjusting font mapping relationship, optimizing layout typesetting, etc. The intelligent search module 15 dynamically adjusts the conversion parameters and processing flow of the third PDF image according to the multi-condition combination result, and re-executes the cross-platform intelligent search viewing operation. By optimizing the combination of multiple conditions, the intelligent search module 15 can effectively improve the matching success rate and ensure that users obtain a consistent, high-quality search experience on different platforms.

[0071] The intelligent search module 15 realizes adaptive correction of differences and distortions introduced in the cross-platform conversion process. This intelligent correction mechanism can dynamically optimize the conversion quality of the third PDF image and improve the accuracy of search matching. Even when the initial matching score is low, it can be optimized through a combination of multiple conditions, and finally realize efficient and accurate cross-platform PDF image search and viewing, which enhances the adaptability and robustness in dealing with complex and changeable cross-platform environments, and provides users with a more reliable and intelligent service experience.

[0072] In summary, the cross-platform PDF image quick viewing system provided by the embodiment of the present application has the following technical effects:

[0073] The PDF parsing module is used to parse the first PDF image to determine multiple PDF image formats, identify the first PDF image according to multiple PDF image formats, generate multiple character data, extract text information from the PDF image, and lay the foundation for subsequent intelligent search. The point processing module is used to traverse the first PDF image for real-time capture, mark the points according to the capture results, determine multiple key points, perform cross-platform processing based on multiple key points, generate a second PDF image, and achieve compatibility processing of PDF images on different platforms to ensure that they can be quickly viewed across platforms. The difference detection module is used to perform scaling calculations based on the first PDF image and the second PDF image to obtain a structural similarity coefficient, automatically compare and detect the first PDF image and the second PDF image according to the structural similarity coefficient, generate a drawing difference detection result, efficiently compare different versions of PDF images, and locate the difference parts therein. The image filling module is used to automatically fill the second PDF image according to the drawing difference detection results, generate a third PDF image, automatically correct and synchronize the difference parts of the PDF image, and ensure the integrity and consistency of its content. The intelligent search module is used to perform cross-platform intelligent search and viewing of the first PDF image based on multiple character data and the third PDF image, so as to achieve efficient retrieval and positioning of PDF image content and improve user work efficiency.

[0074] Any steps of the system described above may be stored as computer instructions or programs in an unlimited computer memory, and may be called and recognized by an unlimited computer processor to implement any system in the embodiments of the present application, without any unnecessary restrictions.

[0075] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to the selection of multiple elements individually or in whole. Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these changes and variations.

Claims

1. A PDF image quick viewing system supporting cross-platform, characterized in that: The system comprises: A PDF parsing module, the PDF parsing module is used to parse the first PDF image to determine multiple PDF image formats, recognize the first PDF image according to the multiple PDF image formats, and generate multiple character data; A point processing module, the point processing module is used to traverse the first PDF image to capture in real time, mark points according to the capture results, determine multiple key points, perform cross-platform processing based on the multiple key points, and generate a second PDF image; a difference detection module, the difference detection module being used to perform scaling calculation based on the first PDF image and the second PDF image to obtain a structural similarity coefficient, and automatically compare and detect the first PDF image and the second PDF image according to the structural similarity coefficient to generate a drawing difference detection result; An image filling module, the image filling module is used to automatically fill the second PDF image according to the drawing difference detection result to generate a third PDF image, including: for the difference area of ​​text content, using OCR technology to identify the text of the corresponding area in the first PDF image, and filling it into the corresponding position of the second PDF image; for the difference area of ​​image elements, extracting the corresponding image fragments from the first PDF image, and seamlessly integrating them into the second PDF image; for the difference area of ​​layout structure, referring to the layout information of the first PDF image, rearranging and adjusting the second PDF image to maintain consistency with the first PDF image; An intelligent search module is used to perform a cross-platform intelligent search and view on the first PDF image based on the multiple character data and the third PDF image.

2. A PDF image quick viewing system supporting cross-platform as claimed in claim 1, characterized in that: The point processing module includes the following execution steps: traversing the first PDF image according to the multiple PDF image formats to perform edge detection and determine multiple geometric features; Capturing points of the first PDF image based on the multiple geometric features to generate the capture result; Performing point cluster analysis according to the capture results using a shape matching algorithm to generate multiple point categories, and generating multiple point labels according to the multiple point categories; The capture results are point-labeled based on the multiple point tags to determine the multiple key points.

3. A PDF image quick viewing system supporting cross-platform as claimed in claim 1, characterized in that: The point processing module also includes the following execution steps: Performing a platform framework analysis on the target platform according to the multiple key points to determine multiple reference points; Performing cross-platform coordinate transformation on the first PDF image according to the multiple reference points to generate multiple transformation points; According to the multiple PDF image formats, mapping the multiple conversion points to the target platform to determine a cross-platform initial PDF image; Comparing the first PDF image with the initial PDF image, performing image superposition analysis according to the comparison result, and generating an image analysis result; The initial PDF image is adjusted based on the image analysis result to generate the second PDF image.

4. A PDF image quick viewing system supporting cross-platform as claimed in claim 3, characterized in that: The point processing module also includes the following execution steps: Performing image grayscale processing on the first PDF image and the initial PDF image to generate a first PDF grayscale image and an initial PDF grayscale image; Performing image registration on the first PDF image and the initial PDF image according to the first PDF grayscale image and the initial PDF grayscale image to generate an image alignment result; Based on the image alignment result, an absolute difference between the first PDF grayscale image and the initial PDF grayscale image is calculated for comparison, and the comparison result is generated, where the comparison result includes the absolute difference; According to the absolute difference, the first PDF image and the initial PDF image are subjected to transparency superposition to generate the image analysis result.

5. A PDF image quick viewing system supporting cross-platform as claimed in claim 1, characterized in that: The difference detection module includes the following execution steps: Scaling the resolution of the first PDF image and the second PDF image to generate a first scaling result; Scaling the first PDF image and the second PDF image to generate a second scaling result; Setting a local image window according to the multiple key points, and performing a pixel-by-pixel sliding comparison on the first PDF image and the second PDF image through the local image window to generate a pixel comparison result; A spatial structure calculation is performed on the first PDF image and the second PDF image according to the pixel comparison result to determine a structural similarity coefficient, where a value range of the structural similarity coefficient is 0-1.

6. A PDF image quick viewing system supporting cross-platform as claimed in claim 5, characterized in that: The difference detection module also includes the following execution steps: When the value of the structural similarity coefficient is greater than or equal to 0 and less than 1, an image detection instruction is generated; activating a difference recognition module through the image detection instruction, traversing the first PDF image according to the structural similarity coefficient based on the difference recognition module, marking the second PDF image according to the traversal result, and determining a plurality of difference regions; Determine whether there is a difference area greater than or equal to a preset area threshold value among the multiple difference areas, and if there is a difference area greater than or equal to the preset area threshold value among the multiple difference areas, color mark the difference area greater than or equal to the preset area threshold value to generate multiple difference marked areas; Performing visualization processing based on the multiple difference-annotated regions to generate a difference heat map; The difference heat map is added to the drawing difference detection result.

7. A PDF image quick viewing system supporting cross-platform as claimed in claim 1, characterized in that: The intelligent search module includes the following execution steps: Performing fuzzy matching on the third PDF image based on the plurality of character data to obtain a matching result, wherein the matching result includes a matching score; Performing matching analysis on the plurality of character data through a context window according to the matching result, and setting an expected matching score according to the analysis result; Determining whether the matching score is greater than or equal to the expected matching score, and if the matching score is greater than or equal to the expected matching score, generating an image query instruction; The image query instruction allows cross-platform intelligent search and viewing of the first PDF image.

8. A PDF image quick viewing system supporting cross-platform as claimed in claim 7, characterized in that: The intelligent search module also includes the following execution steps: If the matching score is less than the expected matching score, a multi-directional tracing instruction is generated, wherein the multi-directional tracing instruction includes a plurality of adjustment conditions; Performing multi-dimensional tracing analysis on the third PDF image through the multi-directional tracing instruction to generate a multi-dimensional abnormal data source; The third PDF image is processed according to the multi-dimensional abnormal data source in combination with the first PDF image to generate an image traceability matrix; The multiple adjustment conditions are combined based on the image source tracing matrix to generate a multi-condition combination result, and a cross-platform intelligent search and viewing is performed on the first PDF image according to the multi-condition combination result.

Images