Nonlinear Shape Statistics in Mumford—Shah Based Segmentation
We present a variational integration of nonlinear shape statistics into a Mumford—Shah based segmentation process. The nonlinear statistics are derived from a set of training silhouettes by a novel method of density estimation which can be considered as an extension of kernel PCA to a stochastic framework.
The idea is to assume that the training data forms a Gaussian distribution after a nonlinear mapping to a potentially higher-dimensional feature space. Due to the strong nonlinearity, the corresponding density estimate in the original space is highly non–Gaussian. It can capture essentially arbitrary data distributions (e.g. multiple clusters, ring- or banana–shaped manifolds).
Applications of the nonlinear shape statistics in segmentation and tracking of 2D and 3D objects demonstrate that the segmentation process can incorporate knowledge on a large variety of complex real—world shapes. It makes the segmentation process robust against misleading information due to noise, clutter and occlusion.
KeywordsSegmentation shape learning nonlinear statistics density estimation Mercer kernels variational methods probabilistic kernel PCA
- 2.B. E. Boser, I. M. Guyon, and V. N. Vapnik. A training algorithm for optimal margin classifiers. In D. Haussler, editor, Proc. of the 5th Annual ACM Workshop on Comput. Learning Theory, pages 144–152, Pittsburgh, PA, 1992. ACM Press.Google Scholar
- 3.B. Chalmond and S. C. Girard. Nonlinear modeling of scattered multivariate data and its application to shape change. IEEE PAMI, 21(5):422–432, 1999.Google Scholar
- 4.T. Cootes and C. Taylor. Active shape model search using local grey-level models: A quantitative evaluation. In J. Illingworth, editor, BMVC, pages 639–648, 1993.Google Scholar
- 6.R. Courant and D. Hilbert. Methods of Mathematical Physics, volume 1. Inter-science Publishers, Inc., New York, 1953.Google Scholar
- 8.D. Cremers, C. Schnörr, and J. Weickert. Diffusion-snakes: Combining statistical shape knowledge and image information in a variational framework. In IEEE First Workshop on Variational and Level Set Methods, pages 137–144, Vancouver, 2001.Google Scholar
- 12.T. Heap and D. Hogg. Automated pivot location for the cartesian-polar hybrid point distribution model. In BMVC, pages 97–106, Edinburgh, UK, Sept. 1996.Google Scholar
- 13.T. Heap and D. Hogg. Improving specificity in pdms using a hierarchical approach. In BMVC, Colchester, UK, 1997.Google Scholar
- 15.M.E. Leventon, W.E.L. Grimson, and O. Faugeras. Statistical shape influence in geodesic active contours. In Proc. Conf. Computer Vis. and Pattern Recog., volume 1, pages 316–323, Hilton Head Island, SC, June 13–15, 2000.Google Scholar
- 16.B. Moghaddam and A. Pentland. Probabilistic visual learning for object detection. In Proc. IEEE Internat. Conf. on Comp. Vis., pages 786–793, 1995.Google Scholar
- 18.N. Paragios and R. Deriche. Coupled geodesic active regions for image segmentation: a level set approach. In D. Vernon, editor, ECCV, volume 1843 of LNCS, pages 224–240. Springer, 2000.Google Scholar
- 21.S. Roweis. Em algorithms for PCA and SPCA. In M. Jordan, M. Kearns, and S. Solla, editors, Advances in Neural Information Processing Systems 10, pages 626–632, Cambridge, MA, 1998. MIT Press.Google Scholar
- 22.B. Schölkopf, S. Mika, Smola A., G. Rätsch, and Müller K.-R. Kernel PCA pattern reconstruction via approximate pre-images. In L. Niklasson, M. Boden, and T. Ziemke, editors, ICANN, pages 147–152, Berlin, Germany, 1998. Springer.Google Scholar
- 24.M. Tipping. Sparse kernel principal component analysis. In Advances in Neural Information Processing Systems 13, Vancouver, Dec. 2001.Google Scholar
- 25.M.E. Tipping and C.M. Bishop. Probabilistic principal component analysis. Technical Report Woe-19, Neural Computing Research Group, Aston University, 1997.Google Scholar
- 26.M. Werman and D. Weinshall. Similarity and affine invariant distances between 2d point sets. IEEE PAMI, 17(8):810–814, 1995.Google Scholar
- 27.S.C. Zhu and A. Yuille. Region competition: Unifying snakes, region growing, and Bayes/MDL for multiband image segmentation. IEEE PAMI, 18(9):884–900, 1996.Google Scholar