Intelligent matching conversion method and system from HTML to PPTX document

Abstract

The application discloses a kind of intelligent matching conversion methods and systems from HTML to PPTX document, method includes: obtaining HTML document and carrying out parsing, determine hierarchy, style and semantic structure, and extract element;Extract the visual features, structural features, semantic features and functional features of element;Classify elements according to categories using convolutional neural networks, and use a matching model based on the Transformer architecture to perform similarity matching to obtain the best matching result;Convert the element to a PPTX element, layout the PPTX element, and convert the PPTX element to an animation effect based on the functional features of the element;Render and standardize the converted and laid out PPTX elements to generate a PPTX document. Through the technical solution of the present application, the element matching accuracy is significantly improved, the layout conversion quality is greatly improved, the complete style information is preserved, and the conversion efficiency is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent matching and conversion method and system for converting HTML documents to PPTX documents. Background Technology

[0002] Currently, there are several main ways to convert HTML to PPTX:

[0003] 1. Manual conversion method: Users copy the HTML content, manually paste it into the PPTX document, and then adjust the format, layout, and element positions one by one. This method requires a lot of manual work, is inefficient, and makes it difficult to ensure format consistency.

[0004] 2. Semi-automatic conversion tools: such as Microsoft Word's "Save As" function or Adobe Acrobat's PDF conversion function, first convert HTML to an intermediate format (such as PDF or DOCX), and then to PPTX. This indirect conversion method can lead to information loss and formatting errors.

[0005] 3. Rule-based automatic conversion tools: such as some online conversion websites and desktop software, which use preset tag mapping rules (such as converting HTML tags) to automatically convert HTML tags to HTML tags. <h1>The tag corresponds to the title style of PPTX. The conversion is performed by mapping tags to the corresponding text styles. These tools typically use a simple one-to-one mapping relationship and lack intelligent judgment capabilities.

[0006] Among them, the most representative is the rule-based automatic conversion tool, whose basic process includes four steps: HTML parsing, element extraction, tag mapping, and PPTX generation. This technology achieves basic format conversion by establishing a fixed mapping table between HTML tags and PPTX elements.

[0007] The existing HTML to PPTX conversion technologies mentioned above have the following main drawbacks:

[0008] (1) Inaccurate element matching: Most existing technologies use simple tag mapping, which cannot recognize the complex nested structure and semantic relationship in HTML, resulting in the converted PPTX elements not matching the original HTML elements.

[0009] (2) Poor layout conversion effect: The flow layout of HTML and the slide layout of PPTX are quite different. Existing technology is difficult to achieve adaptive layout conversion, and problems such as content overflow and chaotic layout often occur.

[0010] (3) Severe loss of styles: CSS styles (especially complex selectors and responsive design) cannot be accurately converted to PPTX format, resulting in the loss of style information such as font, color, and spacing.

[0011] (4) Difficulty in handling interactive elements: Dynamic interactive elements in HTML (such as buttons, forms, and JavaScript effects) cannot be correctly identified and converted by existing technologies.

[0012] (5) Lack of intelligent decision-making ability: It is unable to automatically adjust the structure and layout of PPTX according to the semantics of the content, such as being unable to identify key content and set it as the slide title.

[0013] (6) Poor adaptability: For HTML documents and PPTX templates of different styles, the conversion rules need to be manually adjusted, lacking adaptability.

[0014] Statistics show that when using existing technologies to convert complex HTML documents, users need to spend an average of 3-5 times the original conversion time manually making adjustments, which seriously affects work efficiency. Summary of the Invention

[0015] To address the aforementioned issues, this invention provides an intelligent matching and conversion method and system for HTML to PPTX documents. It innovatively integrates visual, structural, semantic, and functional features through multimodal element feature fusion technology, achieving comprehensive representation of HTML elements. The Transformer-based cross-modal matching model utilizes an attention mechanism to capture the complex mapping relationship between HTML and PPTX elements, overcoming the limitations of traditional rule-based matching and significantly improving element matching accuracy. Adaptive layout conversion greatly enhances layout conversion quality, preserves complete style information, and significantly improves conversion efficiency.

[0016] To achieve the above objectives, this invention provides a method for intelligent matching and conversion of HTML to PPTX documents, comprising:

[0017] Obtain an HTML document, parse the HTML document, determine the hierarchical structure, style and semantic structure of the HTML document, and extract elements of a preset type;

[0018] Based on the analysis results, the visual features of the element are extracted; based on the hierarchical structure, the structural features of the element are extracted; and the semantic and functional features of the element are extracted.

[0019] The elements are classified according to category using a convolutional neural network, and the elements are matched with the preset PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result.

[0020] According to the best matching result, the elements in the HTML document are converted into PPTX elements. The PPTX elements are laid out according to the visual features, structural features and semantic structure of the HTML document. The PPTX elements are then animated according to their functional features.

[0021] The transformed and laid-out PPTX elements are rendered and standardized to generate a PPTX document.

[0022] In the above technical solution, preferably, the HTML document is parsed to determine its hierarchical structure, style, and semantic structure, and elements of a preset type are extracted. The specific process includes:

[0023] An improved DOM tree construction algorithm is used to parse the hierarchical structure of the HTML document and identify the nesting relationships of the HTML document;

[0024] The inline styles, internal style sheets, and external style sheets of the HTML document are parsed according to the style rule set.

[0025] The semantic structure of the HTML document is parsed and identified, including heading levels, paragraph relationships, and key content.

[0026] Elements are extracted from the HTML document, including text content, images, tables, and list types.

[0027] In the above technical solution, preferably, the visual features of the element are extracted, the structural features of the element are extracted according to the hierarchical structure, and the semantic and functional features of the element are extracted. The specific process includes:

[0028] The visual attribute features of the element are extracted based on the rendering results. The visual attribute features include size, color, font and spacing.

[0029] Extract the position, hierarchy, and sibling element distribution of elements in the DOM tree based on the hierarchical structure of the HTML document;

[0030] Extract semantic information features of the elements, including text content, keywords, and sentiment.

[0031] Extract the functional features of the element, including interactive attributes, event bindings, and dynamic behaviors.

[0032] In the above technical solution, preferably, a convolutional neural network is used to classify the elements according to categories, and a matching model based on the Transformer architecture is used to perform similarity matching of the elements in a preset PPTX element feature library to obtain the best matching result. The specific process includes:

[0033] The elements are classified using an element classifier based on a convolutional neural network into text, images, tables, lists, and charts.

[0034] A matching model based on the Transformer architecture is used to perform similarity matching on the elements in a preset PPTX element feature library, and the PPTX element feature with the highest matching degree with the element is obtained as the best matching result.

[0035] In the above technical solution, preferably, according to the best matching result, the elements in the HTML document are converted into PPTX elements, the PPTX elements are laid out according to the visual features, structural features, and semantic structure of the HTML document, and the PPTX elements are animated according to the functional features of the elements. The specific process includes:

[0036] A style conversion engine based on bidirectional mapping converts elements in the HTML document into PPTX element features corresponding to the best matching result.

[0037] The HTML document is divided into a preset number of slides based on its semantic structure.

[0038] A layout recommendation model is used to recommend the PPTX layout with the highest matching degree based on the corresponding category and quantity of the elements;

[0039] The PPTX element features are laid out according to the visual features, structural features and semantic structure of the elements in the HTML document;

[0040] The size and position of the elements are automatically adjusted based on the size of the slide and the number of elements therein;

[0041] Based on the functional characteristics of the element, the animation effects of the HTML document are converted into animation effects of the PPTX element characteristics.

[0042] In the above technical solution, preferably, the three-level layout conversion mechanism of page level, block level and element level is to convert the elements in the HTML document into the PPTX element features.

[0043] In the above technical solution, preferably, the converted and laid-out PPTX elements are rendered and standardized, and the specific process includes:

[0044] The PPTX element is rendered according to the style information of the HTML document;

[0045] The font, color, and spacing format of the PPTX elements are standardized and uniformized.

[0046] In the above technical solution, preferably, the intelligent matching and conversion method from HTML to PPTX documents further includes:

[0047] An optimization model based on reinforcement learning is adopted to dynamically adjust the similarity matching strategy of the matching model according to historical conversion results and user feedback.

[0048] In the above technical solution, preferably, the PPTX element feature library is used to store and update the matching rules and conversion parameters of the element and the PPTX element features, and supports user-defined extensions.

[0049] This invention also proposes an intelligent matching and conversion system for HTML to PPTX documents, applying the intelligent matching and conversion method for HTML to PPTX documents disclosed in any of the above technical solutions, including:

[0050] The HTML document parsing module is used to obtain an HTML document, parse the HTML document, determine the hierarchical structure, style and semantic structure of the HTML document, and extract elements of a preset type.

[0051] The element feature extraction module is used to extract the visual features of the element based on the parsing results, extract the structural features of the element based on the hierarchical structure, and extract the semantic and functional features of the element.

[0052] The element similarity matching module is used to classify the elements according to categories using a convolutional neural network, and to perform similarity matching of the elements in a preset PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result.

[0053] The element layout conversion module is used to convert the elements in the HTML document into PPTX elements according to the best matching result, to lay out the PPTX elements according to the visual features, structural features and semantic structure of the HTML document, and to convert the PPTX elements into animation effects according to the functional features of the elements.

[0054] The PPTX document generation module is used to render and standardize the converted and laid-out PPTX elements to generate PPTX documents.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0056] (1) Significantly improved element matching accuracy: semantic-level element matching is achieved through AI models, with an accuracy of over 92%, which is about 45% higher than traditional rule-based methods.

[0057] (2) The quality of layout conversion is greatly improved: the adaptive layout algorithm makes the qualified rate of the converted PPTX document reach more than 85%, reducing manual adjustment work by more than 80%.

[0058] (3) Complete retention of style information: The style conversion engine can retain more than 90% of the key style information, which solves the problem of serious style loss in the existing technology.

[0059] (4) Significantly improved conversion efficiency: The average time to process a 100-page complex HTML document is reduced from 2-3 hours under existing technology to 10-15 minutes, with an efficiency improvement of 8-12 times.

[0060] (5) User experience optimization: Provide an intuitive preview of conversion results and a simple adjustment interface to reduce the user threshold.

[0061] (6) Wide adaptability: Supports more than 95% of common HTML elements and CSS styles, is compatible with mainstream PPTX templates, and adapts to the conversion needs of different industries and scenarios. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating an intelligent matching and conversion method for HTML to PPTX documents according to an embodiment of the present invention.

[0063] Figure 2 This is a flowchart illustrating the AI ​​element matching process disclosed in one embodiment of the present invention;

[0064] Figure 3 This is a schematic diagram of the element layout transformation process disclosed in one embodiment of the present invention. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] The present invention will now be described in further detail with reference to the accompanying drawings:

[0067] like Figure 1 As shown, a smart matching conversion method for HTML to PPTX documents provided by the present invention includes:

[0068] Obtain the HTML document, parse the HTML document, determine the hierarchical structure, style and semantic structure of the HTML document, and extract elements of the preset type;

[0069] Based on the analysis results, extract the visual features of the elements, extract the structural features of the elements based on the hierarchical structure, and extract the semantic and functional features of the elements.

[0070] Convolutional neural networks are used to classify elements according to categories, and a matching model based on the Transformer architecture is used to perform similarity matching of elements in a pre-defined PPTX element feature library to obtain the best matching result.

[0071] Based on the best matching result, the elements in the HTML document are converted into PPTX elements. The PPTX elements are laid out according to the visual features, structural features and semantic structure of the HTML document. The PPTX elements are then animated according to their functional features.

[0072] The transformed and laid-out PPTX elements are rendered and standardized to generate a PPTX document.

[0073] In this implementation, a multimodal element feature fusion technology is used to innovatively integrate visual, structural, semantic and functional features to achieve a comprehensive representation of HTML elements. The Transformer-based cross-modal matching model uses an attention mechanism to capture the complex mapping relationship between HTML and PPTX elements, breaking through the limitations of traditional rule matching and significantly improving the element matching accuracy. The adaptive layout transformation greatly improves the layout transformation quality, retains complete style information, and significantly improves the transformation efficiency.

[0074] Specifically, it improves the matching accuracy of HTML elements to PPTX elements, achieves semantic-level intelligent matching, realizes adaptive conversion from HTML layout to PPTX slide layout, maintains the readability and aesthetics of the content, preserves the style information of HTML documents to the maximum extent, reduces style loss during the conversion process, enhances the algorithm's adaptability to different styles of HTML documents and PPTX templates, and reduces manual intervention.

[0075] In the above embodiments, preferably, the HTML document is parsed to determine its hierarchical structure, style, and semantic structure, and elements of a preset type are extracted. The specific process includes:

[0076] An improved DOM tree construction algorithm is used to parse the hierarchical structure of HTML documents and identify the nesting relationships within them.

[0077] Parse the inline styles, internal style sheets, and external style sheets of the HTML document based on the style rule set;

[0078] Parse and identify the semantic structure of HTML documents, including heading levels, paragraph relationships, and key content;

[0079] Extract elements from an HTML document. Elements include text content, images, tables, and list types.

[0080] In the above embodiments, preferably, the visual features of the elements are extracted, the structural features of the elements are extracted according to the hierarchical structure, and the semantic and functional features of the elements are extracted. The specific process includes:

[0081] Visual attribute features of elements are extracted based on the rendering results. These visual attribute features include visual attributes such as size, color, font, and spacing.

[0082] Extract the position, hierarchy, and sibling element distribution of elements in the DOM tree based on the hierarchical structure of the HTML document;

[0083] Extract semantic information features of elements, including text content, keywords, and sentiment.

[0084] Extract the functional characteristics of elements, including interactive attributes, event bindings, and dynamic behaviors.

[0085] During implementation, feature extraction employs a deep learning model to fuse various features into a unified high-dimensional feature vector.

[0086] like Figure 2 As shown, in the above embodiment, preferably, a convolutional neural network is used to classify elements according to categories, and a matching model based on the Transformer architecture is used to perform similarity matching of elements in a preset PPTX element feature library to obtain the best matching result. The specific process includes:

[0087] An element classifier based on a convolutional neural network (CNN) is used to classify elements into categories such as text, images, tables, lists, and charts.

[0088] A matching model based on the Transformer architecture is adopted to perform similarity matching calculations on elements in a pre-set PPTX element feature library, and the PPTX element feature with the highest matching degree with the element is obtained as the best matching result.

[0089] like Figure 3 As shown, in the above embodiment, preferably, elements in the HTML document are converted into PPTX elements according to the best matching result. The PPTX elements are then laid out according to the visual, structural, and semantic features of the HTML document, and animation effects are applied to the PPTX elements according to their functional features. The specific process includes:

[0090] A style conversion engine based on bidirectional mapping converts elements in an HTML document into PPTX element features in the best matching result.

[0091] Based on the semantic structure of HTML documents, including content semantics and length, the system automatically divides HTML content into a preset number of slides, achieving semantic-driven slide division and topic extraction.

[0092] A layout recommendation model is used to recommend the best PPTX layout with the highest matching degree based on the corresponding category and quantity of elements, including title page, content page and chart page, etc.

[0093] The PPTX element features are laid out according to the visual features, structural features, and semantic structure of the elements in the HTML document;

[0094] Automatically adjust the size and position of elements based on the slide's dimensions and the number of elements within it;

[0095] Generate PPTX document layout schemes that adapt to different screen sizes based on the responsive design features of HTML documents;

[0096] Based on the functional characteristics of the elements, the animation effects of the HTML document are converted into animation effects of the PPTX elements.

[0097] In the above embodiments, preferably, the three-level layout conversion mechanism of page level, block level and element level converts elements in HTML document into PPTX element features to ensure the rationality and aesthetics of the layout.

[0098] In the above embodiments, preferably, the converted and laid-out PPTX elements are rendered and normalized, and the specific process includes:

[0099] Render the PPTX element according to the style information of the HTML document;

[0100] Standardize the font, color, and spacing format of PPTX elements to ensure consistency in document style.

[0101] During implementation, images are compressed and the structure is optimized to reduce the file size of PPTX documents, thus achieving document optimization processing.

[0102] In the above embodiments, preferably, the intelligent matching conversion method from HTML to PPTX documents further includes:

[0103] An optimization model based on reinforcement learning is adopted to dynamically adjust the similarity matching strategy of the matching model according to historical conversion results and user feedback, thereby improving the algorithm's adaptability.

[0104] In the above embodiments, preferably, the PPTX element feature library is used to store and update the matching rules and conversion parameters of elements and PPTX element features, and supports user-defined extensions.

[0105] This invention also proposes an intelligent matching and conversion system for HTML to PPTX documents, applying the intelligent matching and conversion method for HTML to PPTX documents disclosed in any of the above embodiments, including:

[0106] The HTML document parsing module is used to obtain HTML documents, parse them, determine their hierarchical structure, style and semantic structure, and extract elements of preset types.

[0107] The element feature extraction module is used to extract the visual features of elements based on the parsing results, extract the structural features of elements based on the hierarchical structure, and extract the semantic and functional features of elements.

[0108] The element similarity matching module is used to classify elements according to categories using a convolutional neural network and to perform similarity matching of elements in a preset PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result.

[0109] The element layout conversion module is used to convert elements in an HTML document into PPTX elements according to the best matching results. It lays out the PPTX elements according to the visual, structural and semantic features of the HTML document, and performs animation effects on the PPTX elements according to their functional features.

[0110] The PPTX document generation module is used to render and standardize the converted and laid-out PPTX elements to generate PPTX documents.

[0111] The intelligent matching and conversion system for HTML to PPTX documents disclosed in the above embodiments has modules whose functions correspond to the steps of the intelligent matching and conversion method for HTML to PPTX documents disclosed in the above embodiments. In the implementation process, the above embodiments are referred to for operation, and will not be repeated here.

[0112] The intelligent matching and conversion method and system for HTML to PPTX documents disclosed in the above embodiments will be specifically described through the following examples.

[0113] Example 1: Simple HTML Page Conversion

[0114] Input: A simple HTML page containing headings, paragraphs, lists, and images.

[0115] html

[0116] <!DOCTYPE html> <title> Example Documentation< / title> <style>h1 { color:#333; font-size: 24px;}p { font-size: 14px; line-height: 1.5;}.important {color: red;}< / style> < / h1> <h1> Overview of Artificial Intelligence Development< / h1> Artificial intelligence is a branch of computer science that studies how to enable computers to simulate human intelligent behavior. <p class=""important”"> Artificial intelligence can be mainly divided into the following research directions: Machine Learning Natural Language Processing Computer Vision Expert System <img src=""ai.jpg”" alt=""人工智能概念图”"> .

[0117] Conversion process:

[0118] 1. The HTML parsing module parses the document structure and identifies headings, paragraphs, lists, and image elements;

[0119] 2. The element feature extraction module extracts the visual and semantic features of each element;

[0120] 3. The AI ​​element matching module will <h1>Matches as PPTX title, Matched as the main text, Matches a list of bulleted characters. Match as an image;

[0121] 4. The layout conversion module divides the content into one slide, selecting the "Title + Content" layout;

[0122] 5. The PPTX generation module renders elements according to style information and generates a PPTX document.

[0123] Output: A PPTX document containing one slide, titled "Overview of Artificial Intelligence Development", with two paragraphs of text and a bulleted list in the body, and the image located in the lower right corner of the slide.

[0124] Example 2: HTML Transformation Including Tables and Complex Styles

[0125] Input: An HTML page containing complex tables and CSS styles

[0126] Conversion process:

[0127] (1) The HTML parsing module parses complex table structures and CSS styles;

[0128] (2) The element feature extraction module identifies the table header, data cells, and merged cells;

[0129] (3) The AI ​​element matching module matches HTML tables as PPTX tables while maintaining cell merging relationships;

[0130] (4) The layout conversion module recommends the "content only" layout based on the table size and adjusts the table size to fit the slides;

[0131] (5) The PPTX generation module retains the table borders, background colors and other styles, and generates PPTX documents.

[0132] Output: A PPTX slideshow containing tables. The table structure and style are basically the same as the original table in HTML, requiring no manual adjustments.

[0133] Example 3: Responsive HTML Page Transformation

[0134] Input: A responsive HTML page that displays different layouts on different devices.

[0135] Conversion process:

[0136] (1) The HTML parsing module identifies media queries and responsive rules;

[0137] (2) The element feature extraction module analyzes element features under different screen sizes;

[0138] (3) The AI ​​element matching module generates multiple matching schemes for different layouts;

[0139] (4) The layout conversion module generates two PPTX layout schemes adapted to standard screens and widescreens;

[0140] (5) The PPTX generation module generates the final PPTX document according to the user's selection.

[0141] Output: Users can choose between standard or widescreen PPTX documents, both of which maintain the best readability and aesthetics of the original HTML.

[0142] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention. < / h1>

Claims

1. A method for intelligent matching and conversion of HTML to PPTX documents, characterized in that, include: Obtain an HTML document, parse the HTML document, determine the hierarchical structure, style and semantic structure of the HTML document, and extract elements of a preset type; Based on the analysis results, the visual features of the element are extracted; based on the hierarchical structure, the structural features of the element are extracted; and the semantic and functional features of the element are extracted. The elements are classified according to category using a convolutional neural network, and the elements are matched with the preset PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result. According to the best matching result, the elements in the HTML document are converted into PPTX elements. The PPTX elements are laid out according to the visual features, structural features and semantic structure of the HTML document. The PPTX elements are then animated according to their functional features. The transformed and laid-out PPTX elements are rendered and standardized to generate a PPTX document. The specific process of laying out the PPTX elements based on the visual features, structural features, and semantic structure of the HTML document includes: Based on the semantic structure of the HTML document, the HTML content is divided into a preset number of slides; A layout recommendation model is used to recommend the PPTX layout with the highest matching degree based on the corresponding category and quantity of the elements; The PPTX element features are laid out according to the visual features, structural features, and semantic structure of the element in the HTML document.

2. The intelligent matching and conversion method for HTML to PPTX documents according to claim 1, characterized in that, The HTML document is parsed to determine its hierarchical structure, style, and semantic structure, and elements of a preset type are extracted. The specific process includes: An improved DOM tree construction algorithm is used to parse the hierarchical structure of the HTML document and identify the nesting relationships of the HTML document; The inline styles, internal style sheets, and external style sheets of the HTML document are parsed according to the style rule set. The semantic structure of the HTML document is parsed and identified, including heading levels, paragraph relationships, and key content. Elements are extracted from the HTML document, including text content, images, tables, and list types.

3. The intelligent matching and conversion method for HTML to PPTX documents according to claim 2, characterized in that, The visual features of the element are extracted, the structural features of the element are extracted based on the hierarchical structure, and the semantic and functional features of the element are extracted. The specific process includes: The visual features of the element are extracted based on the rendering results. The visual features include size, color, font, and spacing. Extract the position, hierarchy, and sibling element distribution of elements in the DOM tree based on the hierarchical structure of the HTML document; Extract semantic information features of the elements, including text content, keywords, and sentiment. Extract the functional features of the element, including interactive attributes, event bindings, and dynamic behaviors.

4. The intelligent matching and conversion method for HTML to PPTX documents according to claim 3, characterized in that, The elements are classified according to category using a convolutional neural network, and the elements are matched against a pre-defined PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result. The specific process includes: The elements are classified using an element classifier based on a convolutional neural network into text, images, tables, lists, and charts. A matching model based on the Transformer architecture is used to perform similarity matching on the elements in a preset PPTX element feature library, and the PPTX element feature with the highest matching degree with the element is obtained as the best matching result.

5. The intelligent matching and conversion method for HTML to PPTX documents according to claim 4, characterized in that, According to the best matching result, the elements in the HTML document are converted into PPTX elements. The PPTX elements are then laid out according to the visual, structural, and semantic features of the HTML document. Finally, the PPTX elements are animated according to their functional characteristics. The specific process includes: A style conversion engine based on bidirectional mapping converts elements in the HTML document into PPTX element features corresponding to the best matching result. The HTML document is divided into a preset number of slides based on its semantic structure. A layout recommendation model is used to recommend the PPTX layout with the highest matching degree based on the corresponding category and quantity of the elements; The PPTX element features are laid out according to the visual features, structural features and semantic structure of the elements in the HTML document; The size and position of the elements are automatically adjusted based on the size of the slide and the number of elements therein; Based on the functional characteristics of the element, the animation effects of the HTML document are converted into animation effects of the PPTX element characteristics.

6. The intelligent matching and conversion method for HTML to PPTX documents according to claim 5, characterized in that, A three-level layout transformation mechanism of page level, block level and element level is adopted to transform the elements in the HTML document into the PPTX element features.

7. The intelligent matching and conversion method for HTML to PPTX documents according to claim 5, characterized in that, The transformed and laid-out PPTX elements are then rendered and normalized. The specific process includes: The PPTX element is rendered according to the style information of the HTML document; The font, color, and spacing format of the PPTX elements are standardized and uniformized.

8. The intelligent matching and conversion method for HTML to PPTX documents according to claim 7, characterized in that, Also includes: An optimization model based on reinforcement learning is adopted to dynamically adjust the similarity matching strategy of the matching model according to historical conversion results and user feedback.

9. The intelligent matching and conversion method for HTML to PPTX documents according to claim 8, characterized in that, The PPTX element feature library is used to store and update the matching rules and conversion parameters of the elements and the PPTX element features, and supports user-defined extensions.

10. A smart matching and conversion system for HTML to PPTX documents, characterized in that, The method for intelligent matching and conversion of HTML to PPTX documents as described in any one of claims 1 to 9 includes: The HTML document parsing module is used to obtain an HTML document, parse the HTML document, determine the hierarchical structure, style and semantic structure of the HTML document, and extract elements of a preset type. The element feature extraction module is used to extract the visual features of the element based on the parsing results, extract the structural features of the element based on the hierarchical structure, and extract the semantic and functional features of the element. The element similarity matching module is used to classify the elements according to categories using a convolutional neural network, and to perform similarity matching of the elements in a preset PPTX element feature library using a matching model based on the Transformer architecture to obtain the best matching result. The element layout conversion module is used to convert the elements in the HTML document into PPTX elements according to the best matching result, to lay out the PPTX elements according to the visual features, structural features and semantic structure of the HTML document, and to convert the PPTX elements into animation effects according to the functional features of the elements. The PPTX document generation module is used to render and standardize the converted and laid-out PPTX elements to generate PPTX documents. The specific process by which the element layout conversion module lays out the PPTX elements based on the visual features, structural features, and semantic structure of the HTML document includes: Based on the semantic structure of the HTML document, the HTML content is divided into a preset number of slides; A layout recommendation model is used to recommend the PPTX layout with the highest matching degree based on the corresponding category and quantity of the elements; The PPTX element features are laid out according to the visual features, structural features, and semantic structure of the element in the HTML document.

Images