BROWSE
Detail
Conference Paper
Automatic Generation of Animatable 3D Personalized Model based on Multi-view Images
Cited - time in 
Share 
Share
- Authors
- Seong-Jae Lim, Ho-Won Kim, Jin Sung Choi, Bon-Ki Koo
- Issue Date
- 2014-07
- Citation
- International Conference on Modeling, Simulation and Visualization Methods (MSV) 2014, pp.1-6
- Language
- English
- Type
- Conference Paper
- Abstract
- We propose a fully-automatic 3D model generating method of individual people from multi-view images. An automatically constructed 3D model can be deformed and animated by controlling the joint. An animatable 3D personalized model of individual people is generated by transferring the joint-skeleton structure and appearance of a generic 3D human model to an individual 3D volumetric model reconstructed from multi-view images. We automatically estimate the joint position of individual people by computing a weight function consisting of the kinematic joint-skeleton structure of the generic 3D human model and anthropometric information. Our generic 3D human model approximates the whole body using sweep surfaces. The vertices on the object boundary are bound to the sweep surfaces and follow their deformation. Thus, the animatable 3D individual model can be generated by transferring the sweep surface of the generic 3D model to the individual 3D volumetric model.
- KSP Keywords
- 3D Model, 3d Human Model, Automatic generation, Generating method, Joint position, Kinematic joint, Personalized model, Skeleton structure, Sweep surfaces, Whole-body, model-based