Design categorization based on text input and design generation method with illustration

Abstract

To provide a text categorization and design generation system which forms a text into a format suitable for generating a slide and easily selecting a slide design template suitable for a content of the text, and a design generation method.SOLUTION: A design generation system 1 comprises a design selection part which stores design templates for multiple categorizations regulated in advance. The design generation system 1 comprises a prompt generation part which holds a prompt group by which categorization of a text and output of a structuralized text can be acquired from a language processing model on the basis of intermediate output from the language processing model and the input text. The design generation system 1 further comprises a slide generation part which generates a design on the basis of output acquired from the language processing model by the prompt generation part and the design template acquired from the design selection part.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a design generation system and a design generation method. [Background technology]

[0002] There are known techniques for easily creating presentation materials and techniques for automatically generating designs.

[0003] For example, Patent Document 1 discloses a system for automatically creating presentation materials by structuring input text into text information suited to the presentation materials.

[0004] Furthermore, Patent Document 2 places emphasis on the function of generating designs according to scenes and concepts. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2003-076678 A [Patent Document 2] Patent 7106085 Summary of the Invention [Problem to be solved by the invention]

[0006] Patent Document 1 outputs text information appropriate for presentation materials from input text, but the user must select the template. Patent Document 2 places emphasis on the function of generating designs appropriate for scenes and concepts, but the "illustration," which is one of the design elements of presentation materials, may not necessarily promote understanding of the contents of the text. In this specification, "illustration" refers to information on the combination and arrangement of the color, shape, and pattern of an object in a certain range with text, and in a certain format, has the effect of promoting understanding of the text. "Design" refers to information on an object that is visually identifiable by having color, shape, pattern, and information related thereto, as well as display information including the occupied area of ​​the information on the object and the non-occupied area of ​​the information on the object in a certain range, and includes information on the "illustration."

[0007] The present invention aims to take advantage of recent developments in language processing models in the field of artificial intelligence to automatically generate designs that promote understanding of the content of the text by classifying the input text and matching illustrations that correspond to the content of the text. [Means for solving the problem]

[0008] In order to achieve the above-mentioned object, the design generation system of the present invention comprises a text input unit having a function of inputting text, a prompt generation unit having a plurality of input questions to a language processing model for classifying and structuring text, a design selection unit having a plurality of templates selected based on the answers of the language processing model and image data to be inserted into slides, and a slide generation unit that creates slides based on the information output by the language processing model, the prompt generation unit, and the design selection unit. Effect of the Invention

[0009] According to the design generation system and design generation method of the present invention, it is possible to generate a design that promotes understanding of the text by including object shapes and arrangements that correspond to the content of the text. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a schematic diagram showing a design generation system according to an embodiment of the present invention; [Diagram 2] A diagram showing an example of a flow for identifying a category in the prompt generation unit. [Diagram 3] FIG. 1 is a diagram showing an example of a flow for shaping the structure of a text in a prompt generation unit. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the design generation system according to the present embodiment will be described with reference to the drawings.

[0012] As shown in FIG. 1, the design generation system of this embodiment includes a text input unit with a function for inputting text, a prompt generation unit with a plurality of input questions to a language processing model for classifying and structuring the text input from the text input unit, a design selection unit with a plurality of templates selected based on the answers of the language processing model and image data to be inserted into slides, and a slide generation unit that creates slides based on the information output by the language processing model, the prompt generation unit, and the design selection unit. For example, in the text input section, you can write "Our company's goals for the third term" in the title, "Timewitch's goals for the third term are to achieve annual sales of 300 million yen and to launch Alice" in the conclusion, and "1Q: Since many employees worked hard at the end of the second term, this will be a period to improve the atmosphere within the company by encouraging employees to take paid leave / 2Q : While resolving the negative aspects of the sales flow that were highlighted in the second phase, we will begin defining requirements for the slide creation AI. / 3Q: We will try out new sales policies and make minor adjustments to optimize them. The slide creation AI will complete the requirements definition and system integration diagram. / 4Q: Let's say you typed in, "We will immediately allocate sales costs to generate sales. We will also complete the alpha version of the slide creation AI."

[0013] The classification question storage unit of the prompt generation unit has a plurality of questions for the language processing model to identify which category the text input from the text input unit belongs to. The answer results and classification numbers are tagged so that the answers to the plurality of questions converge to one classification.

[0014] The classification question storage unit in the prompt generation unit can divide the question flow into patterns by obtaining the result of the first question in the text classification identification flow and changing the next question depending on the result. For example, in the case of the text entered as an example, it is desirable to generate a design that arranges items from 1Q to 4Q in chronological order, and this is classified as a "timeline" type.

[0015] The prompt generator includes a text structure question repository that contains a text structure format for each text classification, which is tagged according to the classification number. For example, if the text entered in the example is classified as a timeline, a format that can organize and express the headings and contents at each point in time is prepared in advance, and the entered text is edited to fit the text format of the format. Specifically, it can be structured as follows: "1Q: Reorganize the working environment / · Many employees worked hard at the end of the second period· This period will be used to improve the atmosphere in the company by encouraging employees to take paid leave, 2Q: Improve the sales flow / · Eliminate the negative aspects of the sales flow that were highlighted in the second period· Start defining requirements for the slide creation AI, 3Q: Optimize the sales policy and complete the slide creation AI / · Try out new sales policies and make minor adjustments to optimize· Complete the requirements definition and system integration diagram for the slide creation AI, 4Q: Recover sales and launch the slide creation AI / · Allocate sales costs all at once to generate sales· Complete the alpha version of the slide creation AI."

[0016] In the image question storage section of the prompt generation section, a portion of the structured text information required for image selection is extracted and stored in the image question storage section. For example, in the case of the text entered as an example, it is desirable to select images that match the implementation details at each point in time from 1Q to 4Q, so extract the text necessary for image selection from the text edited to match the format in the previous section. Specifically, the following can be extracted: "Rebuilding the working environment, improving sales flow, optimizing sales policy, recovering sales, and launching slide creation AI."

[0017] In the prompt generation unit, the image question storage unit has a plurality of questions for determining an image to be inserted into a slide. An image is selected according to the answer to the question. For example, in the case of the text entered as an example, it is considered that an image expressing the working environment would be selected for "reorganizing the working environment," an image expressing the flow would be selected for "improving the sales flow," an image expressing sales would be selected for "optimizing sales policy," and an image expressing money or AI would be selected for "recovering sales and launching slide creation AI."

[0018] The design selection section has multiple illustrated templates for each category, tagged with information such as the number of items derived from the text classification and text structure. For example, for the text entered as an example, a template with four items from 1Q to 4Q in a timeline format is selected.

[0019] The design selection unit has a plurality of images to be inserted into slides. For example, in the case of text entered as an example, an image expressing a working environment, an image expressing a flow, an image expressing sales, an image expressing money or AI, among the images previously stored in the design selection unit, corresponding ones are inserted into the slide.

[0020] The slide generator generates slides having designs with illustrations based on the template data and image data determined by the design selector and the structured text data acquired by the prompt generator.

Claims

1. A design generation system comprising: a text input unit with a function for inputting text; a prompt generation unit having a plurality of input questions to a language processing model for classifying and structuring text; a design selection unit having a plurality of templates selected based on the answers of the language processing model and image data to be inserted into slides; and a slide generation unit that creates slides based on the information output by the language processing model, the prompt generation unit, and the design selection unit.

2. 2. The design generation system according to claim 1, wherein the text input unit is capable of accepting text input by a user that is to be used as a basis for generating a design.

3. A design generation system that generates slides by simply entering the text you want to put on a slide into the text input section, dividing it into title, conclusion, and body, and selecting the required file format.

4. The text as described in claim 3 can be input in any format, such as bullet points or prose including paragraphs, and a design generation system is provided which selects a design according to the format and content of the input text and generates slides.

5. A design generation system that counts the number of characters in the title, conclusion, and body of each slide, and guides you to the visually optimal number of characters when generating slides.

6. A design generation system that takes text input and summarises or supplements the text according to the user's choices, guiding the user to easily understandable content when generating slides.

7. A design generation system that, when text is input, selects an icon or image based on the content of the input text and generates a slide.

Images