The Beautiful Future
Facial Expression Retargeting from Human toAvatar Made Easy 본문
IEEE Computer Transactions Graphics on Visualization and 2020
DT: [7] Deformation transfer for triangle meshes, TOG 2004.
BS: [2] Facial retargeting with automatic range of motion alignment, TOG 2017.
Facial expression retargeting is a cross-domain problem.
blendshapes are not orthogonal: [6] Practice and theory of blendshape facial models, EuroG 2014.
two drawbacks of blendshape-based representation.
1. difficulty in representing expressions outside of the linear span.(exaggerated or unseen expressions)
2. creating the blendshapes for avatars still requires a tedious and time-consuming modeling process.
[17] Facewarehouse: A 3d facial expression database for visual computing, VCG 2013.
[32] Direct manipulation blendshapes, CGA 2010.
data-driven shape analysis methods
[43] Variational autoencoders for deforming 3d mesh models, CVPR 2018.
[44] Automatic unpaired shape deformation transfer, TOG 2018.
사용된 캐릭터
Mery, Chubby, Conan : Face rigs: meryprojet.com, Tri Nguyen, www.highend3d.com
Network 구조
'논문' 카테고리의 다른 글
| Face2Face (0) | 2023.03.03 |
|---|---|
| taobao Text/Speech-Driven Full-Body Animation (0) | 2023.02.20 |
| Attention Mesh: High-fidelity Face Mesh Prediction in Real-time (0) | 2022.12.28 |
| COMA (0) | 2022.09.23 |
| 3D Shape Regression for Real-time Facial Animation TOG2013 (0) | 2022.08.27 |
