Global alignment of MR images using a scale based hierarchical model
This paper proposes a novel automated method for global alignment of three dimensional MR images. The matching algorithm employed is closely related to a common constraint based tree searching algorithm , but uses a novel multi-resolution encoding of the search space to improve the search time and permit searching of curved surfaces. The algorithm uses the shape index defined by Koenderink  which provides the very useful property of invariance to uniform scale. The surfaces of the objects are extracted from the MR images automatically using a 3D deformable model . An intelligent mechanism is used for selecting unusual surface features that are common to both objects.
Unable to display preview. Download preview PDF.
- 1.W. E. L. Grimson and T. Lozano-Perez. Localizing overlapping parts by searching the interpretation tree. IEEE Transactions on Pattern Analysis and Machine Intelligence, 9(4):469–482, 1987.Google Scholar
- 2.J. Koenderink. Solid Shape. The MIT Press, 1991.Google Scholar
- 3.A. J. Bulpitt and N. D. Efford. An efficient 3D deformable model with a self-optimising topology. In Proceedings of the British Machine Vision Conference, volume 1, pages 37–46, 1995.Google Scholar
- 4.P. H. Mowforth and J. Zhengping. Model based tissue differentiation in MR brain images. In Proceedings of the Fifth Alvey Vision Conference, pages 67–73, September 1989.Google Scholar
- 5.W. R. Fright and A. D. Linney. Registration of 3-D head surfaces using multiple landmarks. IEEE Transactions on Medical Imaging, 12(3):515–520, September 1993.Google Scholar
- 6.S. Fletcher. Global Alignment of Medical Images. PhD thesis, Unversity of Leeds, School of Computer Studies, 1995.Google Scholar
- 7.R. B. Fisher. Performance comparison of ten variations on the interpretation-tree matching algorithm. In J-O. Eklundh, editor, Computer Vision — ECCV94, volume 1, pages 507–512, Stockholm, May 1994. Springer Verlag.Google Scholar
- 8.P. Fua and P. Sander. Reconstructing surfaces from unstructured 3D points. In Image Understanding Workshop, San Diego, California, January 1992.Google Scholar