Geometric Properties from Incomplete Data Dagstuhl Seminar

From 21.03.04 to 26.03.04, the Dagstuhl Seminar 04131 Geometric Properties from Incomplete Data was held in the International Conference and Research Center (IBFI), Schloss Dagstuhl. During the seminar, several participants presented their current research, and ongoing work and open problems were discussed. Abstracts of the presentations given during the seminar as well as abstracts of seminar results and ideas are put together in this paper. The rst section describes the seminar topics and goals in general. Links to extended abstracts or full papers are provided, if available. Con dence Measures for Variational Optic Flow Methods Andrés Bruhn (Universität Saarbrücken, D) It has often been criticised that there is no reliable con dence measure that allows to sparsify the result of a dense ow eld such that the remaining ow is more reliable. In this way it would be possible to compare the real quality of dense methods with the characteristics of local, nondense approaches. In our paper we investigate the usefulness of local image based con dence measures for global variational methods and present a ow based alternative that is simple and applicable to the entire class of energy minimising global optic ow techniques. Experimental evaluation on various images sequences will show that this con dence measure can give excellent results. Joint work of: Bruhn, Andrés; Weickert, Joachim Dagstuhl Seminar Proceedings 04131 Geometric Properties from Incomplete Data http://drops.dagstuhl.de/opus/volltexte/2006/465 2 R. Klette, R. Kozera, L. Noakes and J. Weickert Shape Recovery of Strictly Convex Solid from N-Views Simon Collings (Univ. of Western Australia Nedlands, AU) State-of-the-art methods for N-view shape recovery generally involve formulating it as constraint satisfaction problem in the discrete domain. In this talk it is shown that in the case where the scene is a smooth, strictly convex solid, this approach can be avoided. As an introduction to the ideas in this talk, some neat results are presented in the two dimensional case. In the three dimensional case it is shown that two photographs will yield two calculable points on the surface of the solid. A set of cameras will then produce a point cloud that can be interpolated to give an estimate of the original shape.