Image generation method, electronic device, and storage medium
By generating and matching text descriptions for images and automatically updating the prompt text until the conditions are met, the problem of low image generation efficiency in existing technologies is solved, and a highly efficient image generation process is achieved.
CN116485943BActive Publication Date: 2026-06-16YUANLI JINZHI (CHONGQING) TECHNOLOGY CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YUANLI JINZHI (CHONGQING) TECHNOLOGY CO LTD
- Filing Date
- 2023-03-22
- Publication Date
- 2026-06-16
AI Technical Summary
⚠Technical Problem
Existing image generation models, such as Stable Diffusion, often produce images that do not match the original text input, requiring users to repeatedly adjust prompts, resulting in cumbersome and inefficient operations.
⚗Method used
By generating initial text descriptions for images, extracting target detection results of entity words, generating target text descriptions, matching them with prompt text, updating the prompt text until the conditions are met, and iteratively generating images.
🎯Benefits of technology
It improves the efficiency of image generation, avoids the tedious operation of manually adjusting prompt text for users, and ensures that the generated images meet expectations.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116485943B_ABST
Abstract
Embodiments of the present application provide an image generation method, an electronic device and a storage medium. The method comprises: generating at least one image corresponding to a prompt text; generating an initial textual description of each image; extracting entity words in the prompt text, and determining a target detection result of an object represented by each entity word in each image; generating a target textual description of each image according to the target detection result of each image and the initial textual description; matching the prompt text and the target textual description of each image to obtain a matching result; when there is no matching result satisfying a target condition, updating the prompt text according to the target textual description, and iteratively performing the above steps based on the updated prompt text until there is a matching result satisfying the target condition; and determining a target image from at least one image whose matching result satisfies the target condition. The embodiments of the present application can improve the generation efficiency of images.
Need to check novelty before this filing date? Find Prior Art