Simultaneous Determination of Registration and Deformation Parameters among 3D Range Images
The conventional registration algorithms are mostly concerned with the rigidbody transformation parameters between a pair of 3D range images. Our proposed framework aims to determine, in a unified manner, not only such rigid transformation parameters but also various deformation parameters, assuming that the deformation we handle here is strictly defined by some parameterized formulation derived from the deformation mechanism. In this point, our proposed framework is different from the deformation researched in such field as the medical imaging.
Similar to other conventional registration algorithms, our algorithm is formulated as a minimization problem of the squared distance sum between the corresponding points among a pair of range images. While the conventional registration algorithms mainly minimize this sum concerned about 6 parameters (3 translation and 3 rotation parameters), the evaluation function in our proposed algorithm includes those deformation parameters as well. Our proposed algorithm can be applied to a wide range of application areas of computer vision, in particular, shape modelling and shape analysis. In this chapter, we describe how we formulated such an algorithm, implemented it, and evaluated its performance.
KeywordsGround Truth Deformation Parameter Range Image Rotation Parameter Registration Algorithm
Unable to display preview. Download preview PDF.
- C. V. Stewart, C. L. Tsai, and A. Perera. A view-based approach to registration: Theory and application to vascular image regisrtaion. In Proceedings of International Conference on Information Processing in Medical Imaging (IPMI), pages 475-486, 2003.Google Scholar
- J. and N. Ayache. Rigid and affine registration of smooth surfaces using differential properties. In Proceedings of Third European Conference on Computer Vision (ECCV’94), pages 397-406, 1994.Google Scholar
- A. Guéziec, X. Pennec, and N. Ayache. Medical image registration using geometric hashing. 4(4):29-41, 1997.Google Scholar
- Eric Bardinet, Laurent D. Cohen, and Nicholas Ayache. A parametric deformable model to fit unstructured 3d data. 71(1):39-54, 1998.Google Scholar
- P. R. Andresen and M. Nielsen. Non-rigid registration by geometryconstrained diffusion. In Proceedings of Medical Image Computing and Computer-Assisted Intervention (MICCAI’99), pages 533-543, 1999.Google Scholar
- G. Turk and M. Levoy. Zipped polygon meshes from range images. In ACM SIGGRAPH Proceedings, pages 311-318, July 1994.Google Scholar
- David Simon. Fast and Accurate Shape-Based Registration. PhD thesis, School of Computer Science, Carnegie Mellon University, 1996.Google Scholar
- Szymon Rusinkiewicz and Marc Levoy. Efficient varinats of the icp algorithm. In Proceedings of the 3rd International Conference on 3D Digital Imaging and Modeling, pages 145-152, May 2001.Google Scholar
- P. Neugebauer. Geometrical cloning of 3d objects via simultaneous registration of multiple range images. In Proceedings of International Conference on Shape Modeling and Application, pages 130-139, March 1997.Google Scholar
- A.E. Johnson and S. Kang. Registration and integration of textured 3-d data. In Proceedings of International Conference on 3D Digital Imaging and Modeling, pages 234-241, May 1997.Google Scholar
- Mark D. Wheeler. Automatic Modeling and Localization for Object Recognition. PhD thesis, School of Computer Science, Carnegie Mellon University, 1996.Google Scholar
- Mark Daniel Sebastian Thrun and Dirk H ahnel. Scan alignment and 3-d surface modelling with a helicopter platform. In The 4th Int. Conf. on Field and Service Robotics, July 14-16, 2003.Google Scholar
- Ryan Miller and Omead Amidi. 3-d site mapping with the cmu autonomous helicopter. June 1998.Google Scholar
- Jana Visnovcova, Li Zhang, and Armin Gruen. Generating a 3d model of a bayon tower using non-metric imagery. In Proc. of Int. Workshop Recreating the Past -Visualization and Animation of Cultural Heritage, 2001.Google Scholar