Powerpoint generation method

Abstract

The invention relates to the technical field of automatic PowerPoint generation, in particular to a PowerPoint generation method. Obtaining a reference presentation and an input document, wherein the reference presentation comprises a structure slide and a content slide; performing functional clustering on the reference presentation file to obtain one or more structural slide groups and one or more content slide groups; analyzing the structural slide group and the content slide group to obtain a content mode frame, wherein the content mode frame comprises a content region, region description and filling requirements; extracting the input document to obtain an outline of a to-be-generated presentation file; and constructing a code instruction based on the outline and the content mode framework, and further generating a new presentation file. According to the method and the device, the defect of insufficient consistency of visual styles and contents in the prior art can be overcome, and the overall aesthetic property and professional sense of the presentation file are greatly improved, so that the generated presentation file is more coherent in logic level and visual communication.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic presentation generation, and in particular to a presentation generation method Background Art

[0002] Existing automatic presentation generation methods can be divided into two categories: text content-based generation methods and layout generation methods based on preset templates:

[0003] 1. Presentation generation based on text content

[0004] Text-based presentation generation methods usually use large language models (such as GPT-3 or GPT-4) to extract key information from input documents and generate logically coherent text content. This type of method ensures that the generated presentation has good content logic and information transmission effect by developing dedicated datasets and training models, while performing well in improving the diversity of text content. To further improve the layout of the content, some methods also introduce additional modules, such as automatically arranging the text and image elements of the presentation through object locators, or using similarity-based strategies to improve the accuracy of text and images, thereby improving the overall effect.

[0005] 2. Presentation generation based on preset templates

[0006] This technical solution uses preset layout templates to achieve automatic layout of content. Usually, preset templates contain commonly used layout elements such as titles, pictures, text boxes, etc., and their purpose is to quickly meet users' basic display needs. Because the preset templates have been considered for visual design, they can ensure aesthetics to a certain extent and are easy to use, greatly reducing the time for manual design. This method provides a fast and efficient solution for presentation generation by selecting appropriate templates for content filling.

[0007] In the field of automatic presentation generation, existing technologies mainly adopt text-based generation and template-based presentation generation methods. Although these technologies have certain advantages in content logic and generation efficiency, they have some inherent defects in technical principles and means, which limit their performance in practical applications.

[0008] 1. Lack of visual consistency and diversity

[0009] The presentation generation method based on text generation mainly relies on a large language model to extract key information from the input document and generate the corresponding text content. Although this method can ensure the logic of the content, it ignores the visual design requirements and lacks effective control over complex visual layouts. The inherent defect brought by this design mechanism is that it is impossible to achieve fine visual control when generating presentations, resulting in the lack of attractiveness and consistency of the generated presentations. For example, in complex display scenarios, a single text generation cannot support the visual coordination and aesthetics of a multi-page presentation, which affects the effect of information communication.

[0010] 2. Insufficient technical efficiency and adaptability

[0011] Existing technologies often have difficulty adapting to changes efficiently when dealing with complex content (such as those containing multimedia elements, charts, etc.). For example, template-based methods require multiple manual interventions and adjustments when adapting to different content and layouts, which increases operational complexity and costs and cannot meet users' needs for rapid generation. Summary of the invention

[0012] To solve the above problems, the present invention proposes a presentation manuscript generating method, the steps of which include:

[0013] Obtaining a reference presentation and an input document, wherein the reference presentation includes a structure slide and a content slide;

[0014] Functionally clustering the reference presentation to obtain one-to-many structure slide groups and one-to-many content slide groups;

[0015] Analyzing the structure slide group and the content slide group to obtain a content model framework, wherein the content model framework includes content areas, area descriptions, and filling requirements;

[0016] Extracting the input document to obtain an outline of a presentation to be generated;

[0017] Code instructions are constructed based on the outline and the content model framework to generate a new presentation.

[0018] Furthermore, functional clustering is performed based on the text features of the reference presentation to obtain a structural slide set.

[0019] Furthermore, functional clustering is performed on the reference presentations by image similarity to obtain a content slide group.

[0020] Further, the content area includes a title part and / or a subtitle part and / or a logo part and / or a footer link and / or license information; and the area description includes text and / or an image.

[0021] Furthermore, the outline includes one or more entries, and each entry corresponds to a slide to be generated.

[0022] Furthermore, one of the entries includes a reference presentation, an index of an input document chapter, and a title and description of the slide to be generated.

[0023] Furthermore, the new presentation is generated by a large language model.

[0024] The present invention also provides a presentation manuscript generating system, comprising:

[0025] A data acquisition module, used for acquiring a reference presentation and an input document, wherein the reference presentation includes a structure slide and a content slide;

[0026] A functional clustering module, used for functionally clustering the reference presentations to obtain one to many structure slide groups and one to many content slide groups;

[0027] A framework building module, used for analyzing the structure slide group and the content slide group to obtain a content model framework, wherein the content model framework includes content areas, area descriptions and filling requirements;

[0028] An outline generation module, used to extract the input document and obtain an outline of a presentation to be generated;

[0029] The presentation generation module is used to construct code instructions based on the outline and the content model framework to generate a new presentation.

[0030] The present invention further proposes an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program includes instructions for executing the above method.

[0031] The present invention also provides a storage medium storing a computer program, wherein the computer program implements the above method when executed by a computer.

[0032] The beneficial effects of the present invention are as follows:

[0033] The present invention can extract the structure and layout information of the reference presentation and organize it into a functional cluster accordingly, which ensures that a suitable existing presentation can be selected as the editing object when generating a new presentation. Finally, the content mode of each presentation is extracted to ensure that the generated content is consistent with the requirements. By introducing the content mode of the presentation, the present invention can overcome the defects of the prior art in that the visual style and content are not consistent enough, greatly improve the overall aesthetics and professionalism of the presentation, and make the generated presentation more coherent in terms of logical hierarchy and visual communication.

[0034] The present invention enables a large language model to generate code instructions to modify the content in a reference presentation, greatly reducing the need for manual intervention. This approach can automatically perform content editing operations, improve generation efficiency, and support automated processing of a variety of media content. Finally, external feedback is provided when the large language model generates erroneous codes to help the large language model achieve self-correction. Through this innovative technical means, the present invention solves the problems of low operating efficiency and high labor costs in the prior art when generating complex presentations, and is suitable for a variety of application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of an embodiment of the present invention.

[0036] Figure 2 This is an example diagram of a reference presentation for an embodiment of the present invention.

[0037] Figure 3 This is an example diagram of a new presentation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present invention are described clearly and completely below. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention. Figure 1 shown.

[0039] S1 Get reference presentation and input documents

[0040] In this step, a reference presentation is obtained, and the selection of the reference presentation is based on the needs of the user.

[0041] In this step, an input document is obtained and used as the input of the presentation to be generated.

[0042] S2 Reference Presentation Analysis

[0043] This step functionally clusters the reference presentations and divides them into two categories based on their functional attributes: one is to support the presentation structure, such as the opening page, transition page, and closing page of the presentation; the other is to convey specific content, such as the title page of the presentation containing bullet points. To achieve effective presentation clustering, this step uses different clustering algorithms to classify based on the text or visual features of the slides.

[0044] For the structural slides, the large language model (LLM) is used to infer the functional role of each slide, and they are grouped according to their unique text features to obtain the structural slide group.

[0045] For content slides, this step uses a hierarchical clustering method based on image similarity to process them, and infers the layout pattern of each cluster through a multimodal large language model (MLLM). The algorithm is based on the similarity matrix between slides, combined with the set similarity threshold, and effectively groups slides through a hierarchical clustering method. First, the algorithm receives the slide similarity matrix consisting of content slides and a 0.65 similarity threshold as input, and initializes an empty cluster set. In the iterative process, the algorithm continuously selects the two slides with the greatest similarity from the similarity matrix. If any of the two slides already belongs to an existing cluster, they are merged into the cluster; if they do not belong to any cluster yet, a new cluster is created and the two slides are added to it. This process continues until the maximum similarity in the similarity matrix is ​​less than the threshold, and finally returns the set of all clusters to obtain the content slide group.

[0046] Through this algorithm, the slides in the reference presentation can be effectively clustered to assist in the generation of subsequent presentations.

[0047] S3 content pattern analysis

[0048] This step is used to identify and summarize the common content and structure in the slides, and generate the corresponding content schema to provide detailed content layout guidance for the generation of new presentations. By analyzing the content type and structure of the reference presentation, the system can automatically generate a content schema framework that subdivides the various content areas of the slides into specific content types and their descriptions.

[0049] This content model structure contains multiple well-defined content areas, such as title, subtitle, logo, footer link, and license information; each area has a corresponding description, data type, such as text or image, and content filling requirements. The content model framework standardizes the function and presentation requirements of each content area so that the appropriate content type and data can be automatically matched when generating a new presentation.

[0050] This step ensures the consistency of the presentation content structure, enables the system to efficiently and accurately fill in the content, and improves the logical hierarchy of the generated presentation and the coherence of information transmission. Through this structured content model, the system can adapt to different content requirements and realize intelligent generation and layout optimization of presentation content.

[0051] S4 outline generation

[0052] In the outline generation step, we first construct the outline and main content of the presentation by extracting key information from the input document. This process plays a vital role in ensuring that the presentation is clearly structured and that information is delivered smoothly. The outline generation is divided into multiple entries, each of which corresponds to a slide. The entry content includes the reference presentation, the index of the input document chapter, and the title and description of the slide to be generated. Specifically, the system uses Qwen2 to streamline the input document, splitting the long content into multiple chapters, each of which contains a title and content paragraphs, and then generates entries. This process can help the large language model (LLM) process only the necessary information, thereby reducing context complexity and improving generation efficiency.

[0053] S5 Presentation Generation

[0054] Under the guidance of the outline, the slide generation process generates a new presentation by iteratively editing the reference presentation. To achieve efficient interaction between the large language model and the presentation, the present invention provides a set of Python APIs for interacting with specific elements in the presentation, including adding, deleting and editing presentation elements, as shown in Table 1. In addition, to facilitate model understanding, the present invention parses the presentation and represents it in HTML format, which facilitates LLM to parse the structure of the presentation more intuitively. For each slide, the content of the new slide is generated in JSON format under the guidance of the content model framework according to the content of the corresponding entry.

[0055] Next, LLM will receive the slide content to be generated, the HTML representation of the reference presentation, and detailed API documentation to generate executable editing operations, with each operation step represented by a line of executable code. These code instructions will then be executed in the interactive execution environment (REPL), and the system will detect errors and provide feedback in real time during the execution process, such as syntax errors or incorrect references to elements in the reference presentation, to help the model self-correct. Through this self-correction mechanism, the system can continuously optimize editing operations using intermediate results to ensure the robustness of the generation process and the high quality of the presentation content.

[0056] Table 1

[0057] Function Name describe del_span Delete specific passages of text. del_image Deletes an image element. clone_paragraph Create a copy of an existing paragraph. replace_span Replaces the contents of a specific paragraph of text. replace_image Replace the path of the image element

[0058] The reference presentation used in the embodiment of the present invention and the presentation finally generated are as follows: Figure 2 and Figure 3 shown.

[0059] Through the innovative design of the above steps, the present invention has demonstrated significant technical effects in automatic presentation generation. In order to further verify the effectiveness of each module and the causes of errors under different conditions, the present invention conducted an ablation experiment and compared the performance under different configurations. The specific experimental settings and technical effect analysis are as follows:

[0060] Ablation Experiment Setup

[0061] In order to further explore the independent contribution of each step in the present invention and its impact on the overall performance of the system, the present invention designs three ablation configurations, each removing one of the key steps, as follows:

[0062] 1. w / o code rendering

[0063] In this configuration, the system's HTML code rendering is replaced with PPTC-style formatting, the HTML code rendering mechanism is removed, and the simple PPTC-style format is used directly for layout display. This setting is used to verify the role of the code rendering mechanism in content generation.

[0064] 2. w / o error feedback

[0065] In this configuration, the feedback-driven self-correction mechanism is removed, making it impossible for the system to automatically adjust and optimize the generated presentation based on the feedback information from the API. This setting is used to evaluate the contribution of the error correction mechanism to the generation accuracy and content coherence.

[0066] 3.w / o layout summary

[0067] In this configuration, a slide is randomly selected as a blueprint, rather than selecting a suitable presentation template for layout inference based on the method of the present invention. This setting is used to verify the effectiveness of the layout induction module in ensuring visual consistency.

[0068] These ablation experiment configurations were tested on a combination of Qwen2 processing text input and GPT-4o processing visual input to compare the contribution of each component in the overall system performance. The method of the present invention is PPTAgent.

[0069] Qwen2 is an advanced large language model that focuses on natural language processing tasks and is capable of generating high-quality text content. Qwen2 is good at extracting key information from input documents, generating coherent text paragraphs, and performing well in multi-round dialogue and content generation tasks.

[0070] GPT-4o is an optimized version of GPT-4 that can accept and process visual information, allowing it to perform well in multimodal tasks. In addition to text generation and comprehension capabilities, GPT-4o can also combine visual information (such as images, charts, etc.) with text information to achieve cross-modal content generation and analysis. In the task of automated presentation generation, GPT-4o is responsible for generating text with clear logic and coherent content, and is able to explain or supplement the input visual content. This enables GPT-4o to have higher accuracy and adaptability in structured information extraction and content generation, ensuring that the generated presentation meets both logical requirements and has a style and information accuracy consistent with the visual content.

[0071] The ablation experiment results are shown in Table 2.

[0072] Table 2

[0073]

[0074] FID is a metric that measures the similarity between generated content and reference samples in feature space. It evaluates the visual quality of the generated effect by comparing the statistical features of the generated content and the reference content. The lower the FID value, the more similar the visual style of the generated content is to the reference content, thus meeting the requirements of visual consistency. FID is usually used for image generation tasks, but it is also suitable for evaluating visual consistency in presentation generation.

[0075] Fidelity refers to whether the generated presentation is faithful to the input content or template in terms of content and structure. High fidelity means that the generated presentation is highly consistent with the input content in terms of information delivery and visual layout, avoiding content deviation or layout errors. This metric is used to evaluate the accuracy and consistency of the generated results.

[0076] G-Eval is a comprehensive evaluation metric that scores the organization, clarity, and coherence of generated content. The score range is usually 1 to 5, and a higher score indicates that the generated presentation content is superior in organization and logical coherence. G-Eval is suitable for evaluating the quality of text generation, visual layout, etc.

[0077] Experimental results show that the proposed method has better performance in visual consistency index FID (Fréchet Inception Distance) than the method without content mode framework. Specifically, the addition of content mode framework reduces the FID of the generated presentation from 3.48 to 4.32, indicating that the visual style of the generated presentation is closer to the reference presentation.

[0078] In addition, the present invention also performs well in the success rate (SR) of the presentation generation task. For example, the success rate of the method without using the content model framework drops to 73.0%, and there is also a decline in indicators such as fidelity and G-Eval.

[0079] SR (Success Rate) refers to the proportion of presentations that the system successfully completes in the presentation generation task that meet the expected content and structure. The higher the success rate, the fewer errors or failures the system encounters during the generation process, and the greater the stability and generation capability of the system. This indicator is expressed as a percentage and is used to measure the reliability of automated generation.

[0080] The method of the present invention converts the generated presentation code into HTML format, so that LLM can accurately understand and adjust the element layout in the presentation; at the same time, the error feedback mechanism allows the method to capture and correct generation errors during the generation process, thereby reducing manual intervention and significantly improving generation efficiency and accuracy.

[0081] Experimental data show that both HTML code rendering and error feedback mechanism have a significant impact on the success rate and generation quality of the method. The success rate of the method without code rendering dropped from 84.8% to 44.4%, and the G-Eval score dropped from 3.72 to 2.64, indicating that the lack of code rendering will significantly affect the organization and readability of the content. The success rate of the method without the error feedback mechanism dropped from 84.8% to 56.4%, and the G-Eval score dropped from 3.72 to 2.97, indicating that the error feedback mechanism also plays an important role in improving generation accuracy and coherence. These two features together ensure that the final generated presentation meets high quality standards in terms of content organization, visual effects, and coherence.

[0082] The above implementation modes are only used to illustrate the technical solutions of the present invention rather than to limit the same. A person skilled in the art may modify or make equivalent substitutions for the technical solutions of the present invention without departing from the spirit and scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for generating a presentation, comprising the steps of: Obtaining a reference presentation and an input document, wherein the reference presentation includes a structure slide and a content slide; Functionally clustering the reference presentation to obtain one to multiple structure slide groups and one to multiple content slide groups; analyzing the structure slide group and the content slide group to obtain a content model framework, wherein the content model framework includes content areas, area descriptions, and filling requirements; Extracting the input document to obtain an outline of a presentation to be generated; Code instructions are constructed based on the outline and the content model framework to generate a new presentation.

2. The method according to claim 1, characterized in that: Functional clustering is performed based on the text features of the reference presentation to obtain a structural slide set.

3. The method according to claim 1, characterized in that The reference presentations are functionally clustered by image similarity to obtain content slide groups.

4. The method according to claim 1, characterized in that: The content area includes a title part and / or a subtitle part and / or a logo part and / or a footer link and / or license information; the area description includes text and / or an image.

5. The method according to claim 1, characterized in that The outline includes one or more entries, and each entry corresponds to a slide to be generated.

6. The method according to claim 5, characterized in that The entry includes a reference presentation, an index of an input document section, and a title and description of the slide to be generated.

7. The method according to claim 1, characterized in that The new presentation is generated by a large language model.

8. A presentation generation system, comprising: A data acquisition module, used for acquiring a reference presentation and an input document, wherein the reference presentation includes a structure slide and a content slide; A functional clustering module, used for functionally clustering the reference presentations to obtain one to many structure slide groups and one to many content slide groups; A framework building module, used for analyzing the structure slide group and the content slide group to obtain a content model framework, wherein the content model framework includes content areas, area descriptions and filling requirements; An outline generation module, used to extract the input document and obtain an outline of a presentation to be generated; The presentation generation module is used to construct code instructions based on the outline and the content model framework to generate a new presentation.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, the computer program is configured to be executed by the processor, and the computer program comprises instructions for executing the method according to any one of claims 1 to 7.

10. A storage medium storing a computer program, wherein when the computer program is executed by a computer, the method according to any one of claims 1 to 7 is implemented.

Images