Typesetting design method and system based on online content data dynamic style

Abstract

The invention discloses a typesetting design method and system based on online content data dynamic styles. The method comprises the following steps: S1, data acquisition; s2, performing feature analysis; s3, layout generation; s4, generating a style; s5, conflict optimization is carried out; s6, performing cross-platform output; s7, feedback optimization is carried out; the system comprises a data acquisition module (1), a feature analysis module (2), a layout generation module (3), a style configuration module (4), a conflict optimization module (5), an output module (6) and a feedback re-optimization module (7). The invention belongs to the technical field of intelligent typesetting.

Description

Technical Field

[0001] This invention relates to the field of intelligent typesetting technology, specifically to a typesetting design method and system based on dynamic styles of online content data. Background Technology

[0002] Online page layout is a crucial link in information delivery, widely used in areas such as corporate promotion, product display, and academic reports. Current layout methods primarily rely on manual operation or simple template application, which has significant drawbacks: First, design efficiency is low. A non-professional user needs an average of 8 hours to create a corporate product introduction page, while a professional designer needs 2-3 hours to complete a similar page, with over 60% of the time spent on style adjustments and text / image adaptation. Second, content and style adaptability is poor. Manual design often results in a disconnect between style and theme (e.g., using a cartoon style for technological products) and weakened core information (e.g., hiding key parameters in secondary areas), leading to an information delivery accuracy rate of only 65%. Third, cross-platform compatibility is insufficient. Traditional layout solutions have a style consistency rate of only 60% across mobile phones, tablets, and PCs, easily leading to text overlap and distorted charts. Fourth, there is a lack of dynamic optimization mechanisms. User feedback necessitates manual redesign, increasing repetitive workload by 40% and failing to adapt to personalized preferences.

[0003] Existing intelligent typesetting technologies mostly focus on simple typesetting of single-format content, failing to achieve deep linkage between content features and style rules, and lacking a closed-loop optimization mechanism, making it difficult to meet the modern typesetting needs of "efficiency, accuracy, and multi-terminal adaptation". Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a typesetting design method and system based on dynamic styles of online content data. Through intelligent analysis of content features, scenario-based style generation, automatic conflict optimization and feedback iteration, the typesetting design is automated, accurate and personalized, the design threshold is reduced and the efficiency of information transmission is improved.

[0005] The technical solution adopted in this invention is a layout design method based on dynamic styles of online content data, which includes the following steps: S1. Obtain online content data related to the page to be formatted, wherein the online content data includes text data, image data, and heterogeneous image data; S2. The online content data is analyzed by the content feature analysis module to identify the content theme, document sentiment, and information hierarchy; S3. Generate a content-aware layout framework based on the semantic analysis results of the content theme, document sentiment, information hierarchy, and user preference data. S4. Based on the preset semantic scene style model, combined with the content theme and user preference data, generate adaptive style rules that include style, pattern, color scheme and cross-terminal adaptive rules. S5. The conflict detection and re-optimization module performs conflict detection on the element relationship between the content-aware layout framework and the adaptive style rules, and adjusts the element layout or style parameters according to the detection results. S6. Outputs layout design schemes that are compatible with multiple terminals; S7. Receive user feedback data on the layout design scheme, adjust the adaptive style rules according to the feedback data, and optimize the layout effect.

[0006] Furthermore, in step S1, the text data includes topic description text, parameter description text, and user evaluation text; the image data includes product display images, data visualization charts, and scenario application images; and the heterogeneous data of the images includes the original chart dataset, image resolution parameters, image format information, and dynamic image frame rate parameters.

[0007] Furthermore, in step S2, the specific method for identifying information levels is as follows: the text keyword weights are calculated using the TF-IDF algorithm, and a content association graph is constructed by combining the semantic dependency analysis model. The online content data is divided into a core layer, a support layer, and an auxiliary layer. The core layer contains the core information conveyed by the page, the support layer provides evidence for the core information, and the auxiliary layer provides supplementary information.

[0008] Furthermore, in step S3, the user preference data includes historical layout style selection records, layout partition preference ratios, color scheme collection records, and terminal usage frequency statistics; the content-aware layout framework includes the area proportion of each information level, element arrangement order, and graphic combination logic.

[0009] Furthermore, in step S4, the semantic scene style model is trained using a historical layout dataset labeled with content theme tags, style tags, style parameters, and effect scores; the styles include business-like, technologically vibrant, and minimalist styles; and the cross-terminal adaptive rules include element scaling ratios, spacing adjustment coefficients, and content folding logic under different screen sizes.

[0010] Furthermore, in step S5, the collision detection includes three types of detection content: Element position conflict detection is used to identify problems such as overlapping text and images and content overflowing the boundaries; Style consistency conflict detection is used to identify inconsistencies in font, color scheme, and spacing between elements at the same information level; Data correlation conflict detection is used to identify problems where chart data does not match text descriptions.

[0011] Furthermore, in step S7, the feedback data includes text modification instructions, style adjustment suggestions, layout optimization requirements, and effect scores; the specific process of adjusting the adaptive style rules is as follows: classify and label the feedback data, extract the style adjustment parameter thresholds, and update the preference weight matrix of the semantic scene style model.

[0012] A typesetting and design system based on dynamic styles of online content data, used to implement the above method, includes: The data acquisition module is used to acquire online content data related to the page to be formatted; The feature analysis module is used to parse the online content data and identify the content theme, document sentiment, and information hierarchy; The layout generation module is used to generate a content-aware layout framework based on semantic analysis results and user preference data. The style configuration module is used to generate adaptive style rules based on the semantic scenario style model; The conflict optimization module is used to detect and adjust conflicts in the relationship between elements in the layout and style. The output module is used to output layout design schemes that are compatible with multiple terminals; and The feedback and optimization module is used to receive user feedback data, adjust adaptive style rules, and optimize the layout.

[0013] Furthermore, the system also includes a model training module for iteratively training the semantic scene style model using a historical layout dataset; the historical layout dataset contains 1200 samples, each sample including original content data, manual layout scheme, style parameter annotations, and user effect evaluation.

[0014] Furthermore, the data acquisition module supports three data access methods: local file upload, cloud document library import, and third-party platform API interface integration. The API interface supports data interaction with enterprise content management systems and data analysis platforms, and automatically synchronizes and updates online content data. The feature analysis module uses a pre-trained BERT model for content topic recognition and calculates the sentiment tendency of documents by combining a sentiment dictionary with machine learning algorithms. The sentiment tendency is quantified as a score of 0-10, where 0 represents serious and professional, and 10 represents lively and vivid.

[0015] The beneficial effects of this invention are as follows: This invention acquires online content data of the page to be formatted through a data acquisition module, covering text (topic description, parameter description, etc.), images (product images, charts, etc.), and heterogeneous image data (raw chart data, resolution, etc.), supporting local upload, cloud import, and API integration. The feature analysis module uses the BERT model to identify content themes, calculates sentiment scores using a sentiment dictionary combined with machine learning, and uses TF-IDF and semantic dependency analysis to divide information into core, support, and auxiliary layers. The layout generation module combines semantic analysis results with user preferences (historical style, layout proportions, etc.) to generate the area proportion and arrangement logic of each information layer, with the core layer accounting for no less than 35% to highlight key information. The style configuration module calls the trained semantic scene style model, matches the theme and preferences to generate style, color scheme, font, and other rules, and simultaneously generates terminal adaptation logic (scaling and folding rules for different screens). The conflict optimization module detects three types of conflicts: position, style, and data, automatically adjusts layout spacing or corrects style parameters, achieving a conflict resolution rate of no less than 90%. The output module generates HTML5, PDF, and other format solutions, adapting to desktop, tablet, and mobile terminals. The feedback and optimization module receives user feedback, analyzes and adjusts parameters, and updates model weights to achieve dynamic upgrades of style rules. This significantly improves layout efficiency, with the automated process reducing the design cycle from 8 hours to less than 30 minutes and the manual operation rate from 100% to less than 8%. Non-professional users can directly generate professional-grade solutions. It achieves deep adaptation between content and style, and intelligent matching based on themes and levels improves the accuracy of information delivery by more than 90%. It has strong cross-platform adaptability, supporting 3 formats and 3 terminals, with a style consistency rate of more than 90%, solving the terminal adaptation chaos of traditional design. In addition, it builds a dynamic optimization closed loop, iterating the model through user feedback. After 3 rounds of optimization, the core information recognition efficiency has been improved by more than 50%, and the personalization adaptability has been significantly enhanced. Furthermore, it adapts to multiple scenario needs, covering various layout scenarios such as corporate publicity and product display, without the need to develop special tools for different scenarios, highlighting its practicality and scalability. Attached Figure Description

[0016] Figure 1 This is a simplified flowchart of the method of the present invention; Figure 2 This is a simplified block diagram of the system of the present invention. Detailed Implementation

[0017] like Figure 1 and Figure 2 As shown, this invention provides a layout design method based on dynamic styles of online content data, which includes the following steps: S1. Obtain online content data related to the page to be formatted, wherein the online content data includes text data, image data, and heterogeneous image data; S2. The online content data is analyzed by the content feature analysis module to identify the content theme, document sentiment, and information hierarchy; S3. Generate a content-aware layout framework based on the semantic analysis results of the content theme, document sentiment, information hierarchy, and user preference data. S4. Based on the preset semantic scene style model, combined with the content theme and user preference data, generate adaptive style rules that include style, pattern, color scheme and cross-terminal adaptive rules. S5. The conflict detection and re-optimization module performs conflict detection on the element relationship between the content-aware layout framework and the adaptive style rules, and adjusts the element layout or style parameters according to the detection results. S6. Outputs layout design schemes that are compatible with multiple terminals; S7. Receive user feedback data on the layout design scheme, adjust the adaptive style rules according to the feedback data, and optimize the layout effect.

[0018] Further, in step S1, the text data includes topic description text, parameter description text, and user review text; the image data includes product display images, data visualization charts, and scenario application images; the heterogeneous image data includes the original chart dataset, image resolution parameters, image format information, and dynamic image frame rate parameters. In step S2, the specific method for identifying information levels is as follows: the text keyword weights are calculated using the TF-IDF algorithm, and a content association graph is constructed using a semantic dependency analysis model, dividing the online content data into a core layer, a support layer, and an auxiliary layer; the core layer conveys the core information of the page, the support layer provides corroboration for the core information, and the auxiliary layer provides supplementary information. In step S3, the user preference data includes historical layout style selection records, layout partition preference ratios, color scheme collection records, and terminal usage frequency statistics; the content-aware layout framework includes the area proportion of each information level, the element arrangement order, and the logic of text and image combination. In step S4, the semantic scene style model is trained using a historical layout dataset labeled with content theme tags, style tags, style parameters, and effect scores. The styles include business-like formality, technological vibrancy, and minimalist freshness. The cross-terminal adaptive rules include element scaling ratios, spacing adjustment coefficients, and content folding logic for different screen sizes. In step S5, the conflict detection includes three types of detection content: Element position conflict detection is used to identify problems such as overlapping text and images and content overflowing the boundaries; Style consistency conflict detection is used to identify inconsistencies in font, color scheme, and spacing between elements at the same information level; Data correlation conflict detection is used to identify problems where chart data does not match text descriptions.

[0019] In step S7, the feedback data includes text modification instructions, style adjustment suggestions, layout optimization requirements, and effect scores. The specific process of adjusting the adaptive style rules is as follows: classify and label the feedback data, extract the style adjustment parameter thresholds, and update the preference weight matrix of the semantic scene style model.

[0020] A typesetting and design system based on dynamic styles of online content data, used to implement the above method, the system includes: Data acquisition module 1 is used to acquire online content data related to the page to be formatted; Feature analysis module 2 is used to parse the online content data and identify the content theme, document sentiment, and information hierarchy; Layout generation module 3 is used to generate a content-aware layout framework based on semantic analysis results and user preference data. Style configuration module 4 is used to generate adaptive style rules based on the semantic scene style model; Conflict optimization module 5 is used to detect and adjust conflicts in the relationship between elements in the layout and style; Output module 6 is used to output layout design schemes compatible with multiple terminals; and The feedback and optimization module 7 is used to receive user feedback data, adjust adaptive style rules, and optimize the layout effect.

[0021] The system also includes a model training module 8, used for iterative training of the semantic scene style model using a historical layout dataset. This historical layout dataset contains 1200 samples, each including original content data, manual layout schemes, style parameter annotations, and user evaluations. The data acquisition module 1 supports three data access methods: local file upload, cloud document library import, and third-party platform API interface integration. The API interface supports data interaction with enterprise content management systems and data analysis platforms, automatically synchronizing and updating online content data. The feature analysis module 2 uses a pre-trained BERT model for content topic recognition and calculates document sentiment tendency using a combination of a sentiment dictionary and machine learning algorithms. Sentiment tendency is quantified as a score of 0-10, where 0 represents serious and professional, and 10 represents lively and engaging.

[0022] The present invention will now be described in more detail with reference to specific embodiments.

[0023] Taking the "layout of a product introduction page for smart home security cameras" as an example, the specific implementation process is as follows: Data Acquisition: Through API integration with the enterprise content management system, online content data is acquired, including text (800 words of product introduction, 15 core parameters such as 98% facial recognition accuracy, 0.3s remote response time, and 3 user reviews), images (4 product appearance images at 1920×1080, 2 function demonstration GIFs, and 3 revenue trend charts), and heterogeneous data (raw data from CSV charts, PNG images, and GIFs at 24fps).

[0024] Feature Analysis: The feature analysis module identifies the content theme as "Introduction to Smart Home Security Camera Products", with a sentiment score of 3 (professional and reliable); Information hierarchy: Core layer (98% face recognition accuracy, 0.3s remote response time, 15m night vision distance), Support layer (motion detection, voice intercom function module, installation process), Auxiliary layer (10 years of security experience of the brand, distribution of after-sales service outlets nationwide).

[0025] Layout generation: The user preference is "tech minimalist style" and "core parameters are displayed on the left and text on the right". The generated layout framework is as follows: the top 10% is the product title area, the left 40% is the core layer area (the upper half is the parameter chart and the lower half is the key indicator text), the right 35% is the support layer area (functional module card arrangement), and the bottom 15% is the auxiliary layer area (brand and after-sales information).

[0026] Style generation: Semantic scene style model matches theme and preferences, generation rules: color scheme: space gray #1A1A2E + tech blue #4CC9F0, font: Source Han Sans (core data 18px, body text 14px, auxiliary text 12px), charts use line charts with data labels; adaptive rules: mobile element scaling 0.7, spacing adjustment coefficient 0.8, tablet scaling 0.9, coefficient 0.95, PC 1.0.

[0027] Conflict optimization: The detection found that the right-side function module card and the bottom after-sales information overlapped vertically by 20px. The support layer area height was automatically compressed by 5%, and the auxiliary layer area was moved up to resolve the position conflict. At the same time, one issue of inconsistency between the "motion detection distance" text and chart data was corrected (the text was 10m, the chart was 12m, and the automatic synchronization was 12m).

[0028] Cross-platform output: Outputs HTML5, PDF, and Figma formats, with no text overlap or chart distortion issues when displayed on a 15.6-inch PC (1920×1080), a 10.9-inch tablet (2360×1640), or a 6.7-inch mobile phone (2400×1080).

[0029] Feedback optimization: After receiving user feedback that "the font size of the core parameters is too small and it is recommended to increase it to 20px", the feedback optimization module analyzed and adjusted the parameters, updating the font size of the core data from 18px to 20px, and simultaneously updating the model preference weights. After optimization, the test showed that the average time for users to identify the core parameters was shortened from 12s to 6s, and the satisfaction rate increased to 92%.

[0030] This invention shortens the typesetting design cycle from 8 hours to within 30 minutes through intelligent content feature analysis and scenario-based style adaptation, improves the accuracy of information transmission to 93%, and achieves a cross-platform style consistency rate of 99%. At the same time, it achieves dynamic optimization through feedback iteration, improves the core information recognition efficiency by 50%, lowers the design threshold, adapts to the needs of multiple scenarios, and effectively improves the efficiency of information transmission and user experience.

[0031] Finally, it should be emphasized that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A typesetting design method based on dynamic styles of online content data, characterized in that, The method includes the following steps: S1. Obtain online content data related to the page to be formatted, wherein the online content data includes text data, image data, and heterogeneous image data; S2. The online content data is analyzed by the content feature analysis module to identify the content theme, document sentiment, and information hierarchy; S3. Generate a content-aware layout framework based on the semantic analysis results of the content theme, document sentiment, information hierarchy, and user preference data. S4. Based on the preset semantic scene style model, combined with the content theme and user preference data, generate adaptive style rules that include style, pattern, color scheme and cross-terminal adaptive rules. S5. The conflict detection and re-optimization module performs conflict detection on the element relationship between the content-aware layout framework and the adaptive style rules, and adjusts the element layout or style parameters according to the detection results. S6. Outputs layout design schemes that are compatible with multiple terminals; S7. Receive user feedback data on the layout design scheme, adjust the adaptive style rules according to the feedback data, and optimize the layout effect.

2. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S1, the text data includes topic description text, parameter description text, and user evaluation text; the image data includes product display images, data visualization charts, and scenario application images; and the heterogeneous image data includes the original chart dataset, image resolution parameters, image format information, and dynamic image frame rate parameters.

3. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S2, the specific method for identifying information levels is as follows: the text keyword weights are calculated using the TF-IDF algorithm, and a content association graph is constructed by combining the semantic dependency analysis model. The online content data is divided into a core layer, a support layer, and an auxiliary layer. The core layer contains the core information conveyed by the page, the support layer provides evidence for the core information, and the auxiliary layer provides supplementary information.

4. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S3, the user preference data includes historical layout style selection records, layout partition preference ratios, color scheme collection records, and terminal usage frequency statistics; the content-aware layout framework includes the area proportion of each information level, element arrangement order, and graphic combination logic.

5. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S4, the semantic scene style model is trained using a historical layout dataset labeled with content theme tags, style tags, style parameters, and effect scores. The styles include business-like, technologically vibrant, and minimalist styles. The cross-terminal adaptive rules include element scaling ratios, spacing adjustment coefficients, and content folding logic for different screen sizes.

6. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S5, the collision detection includes three types of detection content: Element position conflict detection is used to identify problems such as overlapping text and images and content overflowing the boundaries; Style consistency conflict detection is used to identify inconsistencies in font, color scheme, and spacing between elements at the same information level; Data correlation conflict detection is used to identify problems where chart data does not match text descriptions.

7. The typesetting design method based on dynamic styles of online content data according to claim 1, characterized in that, In step S7, the feedback data includes text modification instructions, style adjustment suggestions, layout optimization requirements, and effect scores. The specific process of adjusting the adaptive style rules is as follows: classify and label the feedback data, extract the style adjustment parameter thresholds, and update the preference weight matrix of the semantic scene style model.

8. A typesetting and design system based on dynamic styles of online content data, used to implement the method described in any one of claims 1 to 7, characterized in that, The system includes: The data acquisition module (1) is used to acquire online content data related to the page to be formatted; The feature analysis module (2) is used to parse the online content data and identify the content theme, document sentiment and information hierarchy; The layout generation module (3) is used to generate a content-aware layout framework based on semantic analysis results and user preference data; The style configuration module (4) is used to generate adaptive style rules based on the semantic scene style model; The conflict optimization module (5) is used to detect and adjust the conflict relationships between elements in the layout and style. Output module (6) is used to output a layout design scheme compatible with multiple terminals; and The feedback and optimization module (7) is used to receive user feedback data, adjust adaptive style rules, and optimize the layout effect.

9. The system according to claim 8, characterized in that, The system also includes a model training module (8), which is used to iteratively train the semantic scene style model using a historical layout dataset; the historical layout dataset contains 1200 sets of samples, each set of samples includes original content data, manual layout scheme, style parameter annotation and user effect evaluation.

10. The system according to claim 8, characterized in that, The data acquisition module (1) supports three data access methods: local file upload, cloud document library import, and third-party platform API interface docking. The API interface supports data interaction with enterprise content management system and data analysis platform, and automatically synchronizes and updates online content data. The feature analysis module (2) uses a pre-trained BERT model to identify content topics and calculates the sentiment tendency of documents by combining sentiment dictionary and machine learning algorithm. The sentiment tendency is quantified as a score of 0-10, where 0 represents serious and professional and 10 represents lively and vivid.

Images