“Fast-automatic 3D face generation using a single video camera” by Kubo, Maejima and Morishima

Reconstructing a 3D face model only with a single camera without attaching landmarks and pattern projecting on a face is still a challenging task in computer vision and computer graphics. FaceGen [2011] is able to reconstruct 3D shape and texture from 2D single image based on 3D morphable model. Since the users have to determine of the facial feature points manually, the method does not work automatically. Moreover, it requires a couple of minutes to generate a 3D face model. Thus, the computational cost is not enough cheap for practical usage.

1. Choi, j., et.al.. 2010. 3D Face Reconstruction Using A Single or Multiple Views, IEEE International Conference on Pattern Recognition 2010, pp.3959–3962
2. Facegen. 2011, http://www.facegen.com
3. Hara, t., et.al.. 2011. Automatic 3D Face Generation from Video with Sparse Point Constraint and Dense Deformable Model, ACM SIGGRAPH ASIA 2011, Poster, no.18
4. Irie, a., et.al., 2011. Improvements to Facial Contour Detection by Hierarchical Fitting and Regression, The First Asian Conference on Pattern Recognition, pp.273–277
5. Maejima, a., and Morishima, s. 2008. Fast Plausible 3D Face Generation from a Single Photograph, ACM SIGGRAPH ASIA 2008, Poster, maejima.pdf