Electronic Device Neural Network Image Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image editing tools require numerous steps and tools to combine two different images, making the process time-consuming and difficult, especially when the boundary between the background and object regions is unclear, and users may need to start over if they are not satisfied with the composite image.
Innovation Solution
An electronic device equipped with neural networks that automatically detects objects in an input image, generates a composite image by combining the input image with a target object image, and allows users to update the target object image based on user input, using features of the synthesis region to ensure a natural and satisfactory composite.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual image editing tools are used to combine images, then users can achieve precise control over the composite image, but the process requires many steps and tools making it time-consuming
Solution Approach 1:
The system performs preliminary actions by automatically detecting objects and generating candidate composite images before the user makes final selections. The neural network pre-processes the images by identifying objects, removing backgrounds, and creating multiple candidate composites, so the user only needs to review and select rather than manually execute each editing step.
Solution Approach 2:
The system enables self-service by allowing users to simply select a target image and synthesis location, then the neural network automatically completes the complex tasks of object detection, background removal, and composite image generation. The system serves itself by using AI to perform the tedious manual operations that would otherwise require expert knowledge and time.
2Manufacturing precision
If layer masks and brush tools are used to remove backgrounds, then users can achieve precise background removal, but the operation becomes difficult especially when boundaries are unclear
Solution Approach 1:
The patent replaces the mechanical manual operation of brush tools and layer masks with an automated neural network system. The AI model automatically detects object boundaries and removes backgrounds without requiring users to manually paint or mask, substituting complex manual mechanical operations with intelligent automated processing that handles unclear boundaries more effectively.
3Ease of operation
If users manually select target images and perform all operations, then they can control the composite image creation process, but they need to start over if not satisfied with the result
Solution Approach 1:
The system performs preliminary actions by generating multiple candidate composite images automatically before the user makes final selections. By pre-generating various options with different objects and compositions, the system eliminates the need to start over - users can simply review the pre-generated candidates and select their preferred result.
4Productivity
If automated neural networks are used to generate composite images, then the process time is significantly reduced, but the device complexity increases
Solution Approach 1:
The patent introduces an intermediary neural network system that mediates between the simple user input (selecting target image and location) and the complex task of composite image generation. The neural network acts as an intelligent intermediary that handles all the complex processing internally, allowing the user interface to remain simple while the backend performs sophisticated automated image manipulation.
Data Source
AI summary
The electronic device includes a memory storing one or more instructions, and a processor configured to execute the one or more instructions to: receive a selection of a location of a synthesis region in an input image, obtain a target object image to be located in the synthesis region of the input image, by using at least one object detected in the input image, generate a composite image by combining the input image with the target object image, and control a display of the electronic device to display the composite image.