Self-Supervised 3D Model Generation From 2D Image Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating high-quality 3D models from 2D images is challenging due to the lack of depth information, leading to high realization costs and inefficiencies in existing methods.
Innovation Solution
A method utilizing a symmetric conversion module to process 2D features from different viewing angles, enabling 3D reconstruction through self-supervision, which generates 3D models by converting 2D features into symmetrical viewing angles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional methods are used to generate 3D models from 2D images, then the process can be completed, but the accuracy is insufficient and the cost is high
Solution Approach 1:
The patent applies self-service by using the 2D image itself to provide supervisory signals for 3D model generation. The symmetric conversion module generates symmetric 2D features from the input 2D image, which then serve as supervision for the 3D reconstruction process, eliminating the need for external ground truth data or complex annotation systems
Solution Approach 2:
The patent implements feedback through the symmetric conversion module that generates supervisory signals from the input 2D image. These symmetric 2D features are fed back into the 3D model generation process to guide and correct the reconstruction, creating a self-correcting system that improves accuracy without requiring additional data
2Manufacturing precision
If extensive data learning is performed to improve 3D model generation, then the model accuracy improves, but the time required and computational cost increase
Solution Approach 1:
The system performs self-service by generating its own supervisory signals from the input 2D image through the symmetric conversion module. This eliminates the need for extensive pre-training on large datasets, as the model learns from the symmetric features generated during the reconstruction process itself
Solution Approach 2:
The symmetric conversion module performs preliminary action by pre-computing symmetric 2D features from the input image before the main 3D reconstruction process. This preliminary feature extraction and symmetric feature generation reduces the computational burden during the actual model generation, saving time and resources
Data Source
AI summary
Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for generating a three-dimensional (3D) model. The method includes generating two-dimensional (2D) features of a 2D image on the basis of performing feature extraction on the 2D image. The method further includes generating, on the basis of the 2D features by using a symmetric conversion module, 2D features subjected to symmetric conversion, wherein the symmetric conversion module is configured to generate 2D features of an object in the 2D image in different viewing angles. The method further includes generating a 3D model of the 2D image on the basis of the 2D features subjected to symmetric conversion. By means of embodiments of the present disclosure, the generation of the 3D model can be supervised by using the symmetric conversion module, thus achieving generation of a 3D model with a better effect.