3D Face Model Reconstruction via Feature Point Superposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The construction of three-dimensional face models is hindered by incomplete images due to the obstruction and deformation caused by positioning brackets during scanning, which affects the quality and application of these models in fields like orthodontics and beauty treatments.
Innovation Solution
A method and equipment that involve obtaining front and full face information, generating corresponding planar images with feature points, and superimposing blocks to create a complete three-dimensional face model by aligning feature points, thereby addressing the image defects caused by positioning brackets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a positioning bracket is used to fix the user's head during scanning, then the accuracy of the captured image is improved, but the positioning bracket covers a part of the user's face, making the three-dimensional face model incomplete
Solution Approach 1:
The scanning process is divided into multiple time intervals: a first time interval for scanning the front face without the positioning bracket, and a second time interval for scanning the full face with the positioning bracket. This segmentation allows capturing different facial regions at different times, avoiding the conflict between positioning accuracy and image completeness.
Solution Approach 2:
The front face information is captured in advance during the first time interval before the positioning bracket is applied for the full face scan. This preliminary action ensures that the area that will be covered by the positioning bracket is already recorded, and can be used to supplement the final complete face model.
2Stability of the object's composition
If the positioning bracket covers part of the user's face during scanning, then the user's head is fixed for accurate scanning, but the scanned face image has a covered area that reduces image quality
Solution Approach 1:
The scanning process is divided into multiple time intervals: a first time interval for scanning the front face without the positioning bracket, and a second time interval for scanning the full face with the positioning bracket. This segmentation allows capturing different facial regions at different times, avoiding the conflict between positioning accuracy and image completeness.
Solution Approach 2:
The front face information captured in the first time interval serves as an intermediary to supplement the covered areas in the full face scan. By superimposing and fusing the front face information with the full face information, the missing or covered regions are reconstructed, improving the overall image quality.
Data Source
AI summary
A three-dimensional face model constructing method includes: obtaining front face information to establish a three-dimensional front face model and generate a front planar image corresponding to the three-dimensional front face model, wherein the front planar image includes first feature points; obtaining full face information to establish a three-dimensional full face model and generate a full face planar image corresponding to the three-dimensional full face model, wherein the three-dimensional full face model includes a specific portion and the full face planar image includes second feature points; superimposing a first block of the front plan image to a second block of the full face planar image according to the correspondence between the first feature points and the second feature points, to obtain a complete full face planar image; and generating a three-dimensional face model according to the complete full face planar image.