System for creating image generation prompts and method for creating image generation prompts
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- SUBARU CORP
- Filing Date
- 2024-12-12
- Publication Date
- 2026-06-24
Smart Images

Figure 2026103334000001_ABST
Abstract
Claims
1. A system for creating image generation prompts to be applied to a machine learning model trained to produce an output image, An input section configured to receive, respectively, first text data as a product and second text data as elements to be combined. A determination unit configured to determine the design similarity of content based on the first text data and the second text data, A prompt creation unit is configured to create the image generation prompt according to the similarity determined by the determination unit. Equipped with, The machine learning model is configured to output an output image obtained by multiplying the content of the first text data and the content of the second text data based on the image generation prompt created by the prompt creation unit. A system for creating image generation prompts.
2. The prompt generation unit changes the content of the image generation prompt according to the evaluation value indicating the degree of similarity. A system for creating an image generation prompt according to claim 1.
3. The prompt generation unit modifies the content of the text included in the image generation prompt according to the evaluation value. A system for creating an image generation prompt according to claim 2.
4. The prompt generation unit swaps the positions of the first text data and the second text data included in the image generation prompt according to the evaluation value. A system for creating an image generation prompt according to claim 2 or claim 3.
5. A method for creating an image generation prompt to be applied to a machine learning model that has been trained to produce an output image, When a first text data representing a product and a second text data representing elements to be combined are input, the design similarity regarding the content is determined based on the first text data and the second text data. Based on the determined similarity, the image generation prompt is created. Includes, The machine learning model is configured to output an output image in which the content of the first text data and the content of the second text data are multiplied together based on the created image generation prompt. How to create an image generation prompt.