Automated video creation-oriented script creation method based on large language model

Abstract

The invention relates to an automatic video creation-oriented script creation method based on a large language model. The method comprises the following steps of S1, inputting a text by a user; s2, the large language model analyzes texts and the like input by the user, and a complete script is automatically generated; s3, iteratively generating role description according to the complete script in combination with appearance requirements, and designing role cues meeting image large model requirements according to the role description; s4, generating a role three-view by using the role cue word and the image large model; s5, automatically creating a sub-lens script based on the complete script, the detailed figure appearance description and the sub-lens requirement; s6, converting the sub-shot script into a video prompt word meeting the requirement of a video generation large model; and S7, inputting the generated role three views and video cues into the video generation large model, and outputting a final script creation video. According to the invention, the video creation process is simplified, so that the user can quickly generate high-quality video content from simple text input.

Description

Technical Field

[0001] The present invention relates to the technical field of script creation, and in particular to a script creation method for automated video creation based on a large language model. Background Art

[0002] Traditional video production usually requires collaboration among professional screenwriters, directors and production teams. From script writing and character design to shooting and editing, the process is complicated and time-consuming.

[0003] In recent years, although some AI-based script creation tools have emerged, such as using large language models to generate scripts, these tools usually cannot automatically generate complete script structures, nor can they support complex multi-role and multi-scene interactions. At the same time, the scripts generated by these tools are not oriented to the field of video automation creation, and the generated scripts cannot be directly applied to video creation; further, the quality of the generated scripts and the quality of the generated videos cannot be guaranteed. In addition, it is difficult for existing video generation tools to directly obtain information from the script to generate visual storyboards and video content.

[0004] Therefore, it is necessary to provide a script creation method for automated video creation based on a large language model to simplify the video creation process and enable users to quickly generate high-quality video content from simple text input. Summary of the invention

[0005] The purpose of the present invention is to provide a script creation method for automated video creation based on a large language model to simplify the video creation process and enable users to quickly generate high-quality video content from simple text input.

[0006] In order to solve the problems existing in the prior art, the present invention provides a script creation method for automated video creation based on a large language model, which is characterized by comprising the following steps:

[0007] S1: The user inputs a text describing the basic idea of ​​video creation;

[0008] S2: The large language model parses the text input by the user and automatically generates a complete script based on the default parameter settings or the parameter settings modified by the user and referring to a large amount of materials;

[0009] S3: Based on the generated complete script and the default or user-set appearance requirements, iteratively generate character descriptions, and design character prompts that meet the requirements of the image model based on the character descriptions;

[0010] S4: Generate three views of the character using the character prompt words and the image model;

[0011] S5: Automatically create storyboards based on the complete script, detailed character descriptions, and storyboard requirements. The storyboards set the scene, character dynamics, and dialogue for each shot.

[0012] S6: Convert the storyboard script into video prompt words that meet the requirements of the video generation model;

[0013] S7: Input the generated three-view character image and video prompt words into the video generation model, and output the final script creation video.

[0014] Optionally, in the script creation method for automated video creation based on a large language model, the default parameter settings include: output structure settings, number of characters and settings, number of scenes settings, and communication channel settings.

[0015] Optionally, in the script creation method for automated video creation based on a large language model, the complete script includes: a title, theme, main characters, character relationships, scenes, and plot.

[0016] Optionally, in the script creation method for automated video creation based on a large language model, the large image model is Midjourney or Stable Diffusion.

[0017] Optionally, in the script creation method for automated video creation based on a large language model, the storyboard requirements include: output structure settings, storyboard lens requirements, scene requirements, camera movement requirements, storyboard style requirements and duration requirements.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] (1) It significantly simplifies the video creation process, allowing users to quickly generate high-quality video content from simple text input. Through the powerful capabilities of the big model, the automatically generated scripts and storyboard designs have a high level of creativity and coherence, which improves creation efficiency and lowers the technical threshold.

[0020] (2) The automatically generated three-view images and high-quality videos effectively reduce the manual involvement of character design and video editing, making video production more automated and intelligent. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flowchart of a script creation method provided by an embodiment of the present invention;

[0022] Figure 2 A flowchart of script generation provided by an embodiment of the present invention;

[0023] Figure 3A flow chart of generating role prompt words provided by an embodiment of the present invention;

[0024] Figure 4 A flowchart of generating video prompt words provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The specific implementation of the present invention will be described in more detail below in conjunction with the schematic diagram. The advantages and features of the present invention will become clearer based on the following description. It should be noted that the drawings are all in a very simplified form and are not in exact proportions, and are only used to facilitate and clearly assist in explaining the purpose of the embodiments of the present invention.

[0026] Hereinafter, if the method described herein includes a series of steps, the order in which the steps are presented herein is not necessarily the only order in which the steps may be performed, and some of the steps described may be omitted and / or some other steps not described herein may be added to the method.

[0027] Traditional video production usually requires the cooperation of professional screenwriters, directors and production teams. From script writing and character design to shooting and editing, the process is complicated and time-consuming. In recent years, although some AI-based script creation tools have emerged, such as using large language models to generate scripts, these tools usually cannot automatically generate complete script structures, nor can they support complex multi-role and multi-scene interactions. At the same time, the scripts generated by these tools are not oriented to the field of video automation creation, and the generated scripts cannot be directly applied to video creation; further, the quality of the generated scripts and the quality of the generated videos cannot be guaranteed. In addition, it is difficult for existing video generation tools to directly obtain information from the script to generate visual storyboards and video content.

[0028] In order to solve the problems existing in the prior art, the present invention provides a script creation method for automated video creation based on a large language model, such as Figure 1 As shown, the script creation method includes the following steps:

[0029] S1: The user inputs a text describing the basic idea of ​​video creation.

[0030] S2: Figure 2 As shown, the large language model parses the text input by the user and automatically generates a complete script based on the default parameter settings or the parameter settings modified by the user and referring to the massive materials;

[0031] The default parameter settings include: output structure settings, character number and settings, scene number settings, and communication channel settings. A complete script includes: the title, theme, main characters, character relationships, scenes, and plots. Usually, the content of a complete script changes according to the text input by the user. Massive materials select some available materials (such as character references, plot references, etc.) from the material library and provide them to the large language model to ensure the quality and storytelling of the generated content.

[0032] For example, the user inputs: I want a script about love, the main characters are a man and a woman, they are in a relationship, and it includes their love story. According to the default parameter settings, appropriate reference materials are selected and input into the big language model, and the output format is specified, requiring the big language model to output the generated script according to the format.

[0033] S3: Figure 3 As shown in the figure, the generated complete script is combined with the default or user-set appearance requirements to iteratively generate character descriptions, which are input into the large language model. The large language model rewrites the character prompts, i.e., character prompts (multiple character prompts are created according to actual conditions), which meet the requirements of the image large model. The image large model can be Midjourney (MJ) or Stable Diffusion (SD). The character prompts enable the image large model to generate character settings and descriptions with high character differentiation and recognition, ensuring the character effects during three-view generation and video generation.

[0034] For example, character 1 has been generated, and then the appearance description of character 1, the appearance requirements of character 2, and the generated script are input into the large language model together to generate the appearance description of character 2.

[0035] S4: Generate three views of the character using the character prompt words and the image model.

[0036] S5: Figure 4 As shown, based on the complete script, detailed description of character appearance and storyboard requirements, a storyboard script is automatically created. The storyboard script sets the scene, character dynamics and dialogue of each shot, and at the same time makes each storyboard conform to the various settings in the script and meet the detailed description of the character's appearance.

[0037] Among them, the storyboard requirements include: output structure settings, storyboard lens requirements, scene requirements, camera movement requirements, storyboard style requirements and duration requirements. The storyboard requirements change according to the text input by the user and the changes in the script.

[0038] S6: Convert the storyboard script into video prompt words that meet the requirements of the video generation model, namely, video prompts. Multiple video prompt words are created according to actual conditions.

[0039] S7: Input the generated three-view character image and video prompt words into the video generation model, and output the final script creation video.

[0040] In one embodiment, a user inputs a sentence through a mobile phone application: "A little girl is looking for the legendary elves in the forest." The large language model first parses this sentence and generates a detailed plot, including the background story of the little girl, the forest environment setting, the relationship between the characters, the legend of the elves and the plot twist.

[0041] At the same time, according to the plot, a detailed description of the characters' appearance is generated to make each character distinctive and distinctive.

[0042] Then, use the generated character settings to create character prompts suitable for MJ or SD, and generate three views of the little girl and the elf. After the script is generated, the storyboard script is automatically designed according to the generated script, including the layout of the forest scene, character interaction, emotional expression, etc. According to the storyboard, generate high-quality video prompts that can be used by the video generation large model. Finally, based on the three views of the character and the high-quality video prompts, generate a script creation video. The video production process is greatly simplified, and a high-quality video can be quickly output to present a complete storyline and visual effects.

[0043] In summary, compared with the prior art, the present invention has the following advantages:

[0044] (1) It significantly simplifies the video creation process, allowing users to quickly generate high-quality video content from simple text input. Through the powerful capabilities of the big model, the automatically generated scripts and storyboard designs have a high level of creativity and coherence, which improves creation efficiency and lowers the technical threshold.

[0045] (2) The automatically generated three-view images and high-quality videos effectively reduce the manual involvement of character design and video editing, making video production more automated and intelligent.

[0046] The above is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any technician in the relevant technical field, without departing from the scope of the technical solution of the present invention, makes any form of equivalent replacement or modification to the technical solution and technical content disclosed in the present invention, which does not depart from the content of the technical solution of the present invention and still falls within the protection scope of the present invention.

Claims

1. A script creation method for automated video creation based on a large language model, characterized in that: The following steps are involved: S1: The user inputs a text describing the basic idea of ​​video creation; S2: The large language model parses the text input by the user and automatically generates a complete script based on the default parameter settings or the parameter settings modified by the user and referring to a large amount of materials; S3: Based on the generated complete script and the default or user-set appearance requirements, iteratively generate character descriptions, and design character prompts that meet the requirements of the image model based on the character descriptions; S4: Generate three views of the character using the character prompt words and the image model; S5: Automatically create storyboards based on the complete script, detailed character descriptions, and storyboard requirements. The storyboards set the scene, character dynamics, and dialogue for each shot. S6: Convert the storyboard script into video prompt words that meet the requirements of the video generation model; S7: Input the generated three-view character image and video prompt words into the video generation model, and output the final script creation video.

2. The script creation method for automated video creation based on a large language model as claimed in claim 1, characterized in that: The default parameter settings include: output structure settings, character quantity and settings, scene quantity settings, and communication channel settings.

3. The script creation method for automated video creation based on a large language model as claimed in claim 1, characterized in that: A complete script includes: title, theme, main characters, character relationships, scenes, and plot.

4. The script creation method for automated video creation based on a large language model as claimed in claim 1, characterized in that: The image model is Midjourney or Stable Diffusion.

5. The script creation method for automated video creation based on a large language model as claimed in claim 1, characterized in that: The storyboard requirements include: output structure setting, storyboard lens requirements, scene requirements, camera movement requirements, storyboard style requirements and duration requirements.

Images