Target Element Generation for Stylized Images Through Local Re-Drawing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality (AR) technologies result in uniform content generation for different users due to the use of consistent AR props, leading to a lack of richness in the generated content.
Innovation Solution
A method and apparatus that involves displaying a stylized image generated through target style conversion on an initial image, and allowing local re-drawing of a target region within the image using processing operations to generate a target element based on an initial element, enhancing the content's richness and conformity to a specific style.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If the same AR prop is used for shooting by different users, then the AR prop consistency is maintained, but the content richness is reduced
Solution Approach 1:
The patent segments the AR prop into a base prop and multiple replaceable target elements. The base prop maintains consistency across users, while the target elements can be varied to create different content. This is achieved by allowing users to select different target elements (e.g., different characters, objects, or symbols) to be superimposed on the same AR prop, thus maintaining prop consistency while enriching content diversity.
Solution Approach 2:
The patent introduces dynamic replaceability to the AR prop system. Instead of a fixed prop, the system allows dynamic substitution of target elements based on user selection. The prop structure remains stable, but the visual content can be dynamically changed by replacing target elements, enabling both consistency and versatility.
2Manufacturing precision
If target style conversion is performed on initial images, then the style conformity is improved, but the content diversity is reduced
Solution Approach 1:
The patent segments the image processing into style conversion and content generation parts. The style conversion applies a consistent target style to the background and overall composition, while the target elements can be varied independently. This allows the style to remain conformant while the content elements maintain diversity.
Solution Approach 2:
The patent applies local quality by maintaining different characteristics in different parts of the generated image. The background and overall style conform to the target style, while the target elements can be locally varied based on user preferences. This creates a unified style with diverse local content.
Data Source
AI summary
A target element generation method and apparatus, an electronic device, and a storage medium are provided. The method includes: displaying an editing page on which a stylized image is displayed, where the stylized image is an image generated by performing target style conversion on an initial image; and displaying, in a target region of the stylized image, a target element corresponding to a processing operation for the target region in response to the processing operation for the target region, where the target element is an element generated by re-drawing the target region based on an initial element corresponding to the processing operation.