Synthesis of realistic human faces from CT image and color photos.
- Author:
Jun CHEN
1
;
Jie YANG
Author Information
1. Institute of Image Processing & Pattern Recognition, Shanghai Jiaotong University, Shanghai 200240, China.
- Publication Type:Journal Article
- MeSH:
Face;
diagnostic imaging;
Humans;
Imaging, Three-Dimensional;
methods;
Models, Biological;
Photography;
methods;
Radiographic Image Interpretation, Computer-Assisted;
methods;
Tomography, X-Ray Computed;
methods
- From:
Journal of Biomedical Engineering
2008;25(3):515-519
- CountryChina
- Language:Chinese
-
Abstract:
In order to predict the post-plastic-surgical effect of patients, we propose a method for synthesizing a high-definition realistic human face from three 2D digital color photos and 3D face model constructed from CT data. The simple genetic algorithm is used to estimate the projective parameters from 3D face model to 2D photos via the correspondence between manually selected 2D and 3D points. Finally, a high-definition realistic human face with detailed texture and smooth transition from front to side is synthesized utilizing the multi-texture mapping mechanism.
