Face alignment via joint-AAM

In this paper, a joint active appearance model (joint-AAM) framework is proposed for face alignment. The object function consists of more than one active appearance model and some constraint items. It can be optimized through the efficient project-out inverse compositional (POIC) fitting algorithm. By transferring the low dimensional parameter space to the high one, the facial shape can be converged to the acceptable solution easier by joint-AAM comparing to single AAM, especially if the initial solutions locate on each side of the optimal solution. In multi-view case, different AAMs are jointed if the true view is far from the initial views. In single view case, different initial solutions of one AAM can be jointed to handle poor initialization or exaggerative expressions. Alternatively, 3D shape model is employed to impose stronger shape constraints on joint-AAM. A geometrical explanation is given to describe the reason of the robustness of the joint-AAM. The experiments demonstrate its accuracy, robustness and efficiency. The acronyms in this paper are listed in Tab. 1.

[1]  K. Walker,et al.  View-based active appearance models , 2000, Proceedings Fourth IEEE International Conference on Automatic Face and Gesture Recognition (Cat. No. PR00580).

[2]  Simon Baker,et al.  Active Appearance Models Revisited , 2004, International Journal of Computer Vision.

[3]  Simon Baker,et al.  2D vs. 3D Deformable Face Models: Representational Power, Construction, and Real-Time Fitting , 2007, International Journal of Computer Vision.

[4]  Lin Liang,et al.  AAM based face tracking with temporal matching and face segmentation , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[5]  Simon Baker,et al.  Lucas-Kanade 20 Years On: A Unifying Framework , 2004, International Journal of Computer Vision.

[6]  Yangsheng Wang,et al.  Real-Time 3D Face and Facial Action Tracking Using Extended 2D+3D AAMs , 2010, 2010 20th International Conference on Pattern Recognition.

[7]  Stan Sclaroff,et al.  Active blobs: region-based, deformable appearance models , 2003, Computer Vision and Image Understanding.

[8]  Stan Z. Li,et al.  Direct appearance models , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[9]  Jörgen Ahlberg,et al.  CANDIDE-3 - An Updated Parameterised Face , 2001 .

[10]  Takeo Kanade,et al.  A Generative Shape Regularization Model for Robust Face Alignment , 2008, ECCV.

[11]  Simon Prince,et al.  Face Pose Estimation in Uncontrolled Environments , 2009, BMVC.

[12]  Minh Hoai.  Local Minima Free Parameterized Appearance Models , 2008 .

[13]  Thomas Vetter,et al.  On compositional Image Alignment, with an application to Active Appearance Models , 2009, CVPR.

[14]  Demetri Terzopoulos,et al.  Snakes: Active contour models , 2004, International Journal of Computer Vision.

[15]  马勇,et al.  Real-Time Multi-View Face Detection and Pose Estimation Based on Cost-Sensitive AdaBoost , 2005 .

[16]  Timothy F. Cootes,et al.  On representing edge structure for model matching , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.

[17]  Timothy F. Cootes,et al.  Active Appearance Models , 1998, ECCV.

[18]  Timothy F. Cootes,et al.  Active Shape Models-Their Training and Application , 1995, Comput. Vis. Image Underst..

[19]  Xiaoming Liu,et al.  Generic Face Alignment using Boosted Appearance Model , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.