We propose a novel 3D Dynamic Morphable Model which utilizes the Dynamic 3D FACS Dataset (D3DFACS) [1]. Compared to the technique proposed in [1], the model provides a more accurate global representation and improved methods for aligning independent sequences of facial expressions. Firstly, we find the rigid correspondence between raw meshes by using Coherent Point Drift (CPD) [2]. In order to obtain more precise correspondences on UV texture maps, we then use a novel hybrid registration approach combining Thin-plate Splines (TPS) and optical flow [3]. We compare this approach of alignment of independent dynamic sequences against standard optical flow methods proposed in [1]. After both rigid and nonrigid alignment, a 3D dynamic morphable model is constructed using Principle Component Analysis (PCA). We aim to use the model to assess the ability of 3D dynamic morphable models on speech replications for talking head and other animation applications.
[1]
Berthold K. P. Horn,et al.
Determining Optical Flow
,
1981,
Other Conferences.
[2]
Fred L. Bookstein,et al.
Thin-Plate Splines and the Atlas Problem for Biomedical Images
,
1991,
IPMI.
[3]
Thomas Vetter,et al.
A morphable model for the synthesis of 3D faces
,
1999,
SIGGRAPH.
[4]
William A. P. Smith,et al.
3D morphable face models revisited
,
2009,
CVPR.
[5]
Andriy Myronenko,et al.
Point Set Registration: Coherent Point Drift
,
2009,
IEEE Transactions on Pattern Analysis and Machine Intelligence.
[6]
William A. P. Smith,et al.
A Linear Approach to Face Shape and Texture Recovery using a 3D Morphable Model
,
2010,
BMVC.
[7]
Adrian Hilton,et al.
A FACS valid 3D dynamic action unit database with applications to 3D dynamic morphable facial modeling
,
2011,
2011 International Conference on Computer Vision.