LSCM Based Non-rigid Registration for Craniofacial Surfaces

The Least Squares Conformal Maps (LSCM) is an approximation of the conformal mapping in the least-squares sense, and it can map the corresponding feature points on two 3D surfaces into the same 2D location. This paper proposes a non-rigid registration method for craniofacial surfaces based on LSCM parameterization. Firstly, craniofacial surfaces are normalized in pose and scale by using a unified coordinate system. Secondly, by pinning six landmarks, which include the outer corners of the eyes, two corners of the mouth, two side points of the nose wing, each craniofacial surface is mapped into a nearly equal 2D domain by using LSCM. Finally, an iso-parameter mesh of each craniofacial surface can be obtained by 2D to 3D mapping, which establishes a unique correspondence among different craniofacial surfaces. To evaluate the proposed method, the target surface is deformed into the reference surface using TPS algorithm with dense correspondences being control points, and then the sum of the distance between two correspondence point sets are computed, and vice versa. According to the average distance, the proposed method is compared with ICP and a TPS based method. The comparison shows that the proposed approach is more accurate and effective.

[1]  Paul J. Besl,et al.  A Method for Registration of 3-D Shapes , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Peter Eisert,et al.  Algorithms For Automatic And Robust Registration Of 3D Head Scans , 2010, J. Virtual Real. Broadcast..

[3]  Sen Wang,et al.  Conformal Geometry and Its Applications on 3D Shape Matching, Recognition, and Stitching , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Yongli Hu,et al.  A Dense Point-to-Point Alignment Method for Realistic 3D Face Morphing and Animation , 2009, Int. J. Comput. Games Technol..

[5]  Fred L. Bookstein,et al.  Principal Warps: Thin-Plate Splines and the Decomposition of Deformations , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Marc Levoy,et al.  Efficient variants of the ICP algorithm , 2001, Proceedings Third International Conference on 3-D Digital Imaging and Modeling.

[7]  Bruno Lévy,et al.  Least squares conformal maps for automatic texture atlas generation , 2002, ACM Trans. Graph..

[8]  Szymon Rusinkiewicz,et al.  Non-rigid range-scan alignment using thin-plate splines , 2004 .

[9]  Max A. Viergever,et al.  A survey of medical image registration , 1998, Medical Image Anal..

[10]  Peter Hammond,et al.  Automated Registration of 3D Faces using Dense Surface Models , 2003, BMVC.

[11]  Takeo Kanade,et al.  An Iterative Image Registration Technique with an Application to Stereo Vision , 1981, IJCAI.

[12]  Guillermo Sapiro,et al.  Conformal Surface Parameterization for Texture Mapping , 1999 .