Virtual Scene Generation Using Object Shape and Texture Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Artificial intelligence in current virtual scene generation schemes fails to learn comprehensive information of real objects, leading to insufficient perception and inaccurate virtual scene generation.
Innovation Solution
A method involving an electronic device that utilizes a virtual scene generation network to identify and generate virtual scenes by classifying pixel points, determining object categories and positions, and simulating virtual objects using a virtual scene simulator, combined with a virtual image generation network to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If artificial intelligence is used for virtual scene generation, then automation is improved, but perception accuracy of real objects deteriorates
Solution Approach 1:
The patent introduces an intermediary component (virtual scene generation network) that bridges the gap between automated AI processing and accurate object perception. This network acts as a mediator that processes real scene images to extract comprehensive object information, which then guides the virtual scene generation process, thereby maintaining both automation and perception accuracy.
Solution Approach 2:
The patent segments the virtual scene generation process into distinct functional modules: real scene image acquisition, object information extraction, virtual scene synthesis, and rendering. This segmentation allows each module to specialize in specific tasks, improving overall system accuracy while maintaining automation through coordinated operation of these specialized components.
2Manufacturing precision
If comprehensive object information is extracted, then virtual scene accuracy is improved, but processing complexity increases
Solution Approach 1:
The virtual scene generation network performs multiple functions simultaneously: it identifies object categories, determines spatial positions, extracts shape characteristics, and captures texture information from real scene images. This multi-functionality reduces the need for separate processing systems for each task, thereby improving virtual scene accuracy without proportionally increasing system complexity.
Solution Approach 2:
The patent combines multiple object recognition and scene understanding functions into an integrated virtual scene generation network. By merging category classification, position detection, shape extraction, and texture analysis into a unified system, the patent achieves comprehensive object information extraction while managing processing complexity through consolidated architecture.
Data Source
AI summary
A virtual scene generation method applied to an electronic device is provided. The electronic device identifies object prediction information of each real object in each of real scene images. A first virtual image corresponding to each real scene image is obtained according to the object prediction information of each real object. A second virtual image and a texture difference image corresponding to each real scene image are generated. A target image corresponding to each real scene image is generated according to the second virtual image and the texture difference image corresponding to each real scene image. Once a virtual scene generation model is generated based on the real scene images, the first virtual image, the second virtual image, the target image corresponding to each real scene image, a virtual scene corresponding to an image is obtained using the virtual scene generation model and the virtual scene simulator.