%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi@80/2006/08.16.16.56
%2 sid.inpe.br/sibgrapi@80/2006/08.16.16.56.47
%@doi 10.1109/SIBGRAPI.2006.15
%T Estimation of Multiple Orientations and Multiple Motions in Multi-Dimensional Signals
%D 2006
%A Stuke, Ingo,
%A Barth, Erhardt,
%A Mota, Cicero,
%@affiliation Institute for Signal Processing, University of Luebeck
%@affiliation Institute for Neuro- and Bioinformatics, University of L uebeck
%@affiliation Departamento de Matemática, Universidade Federal do Amazonas
%E Oliveira Neto, Manuel Menezes de,
%E Carceroni, Rodrigo Lima,
%B Brazilian Symposium on Computer Graphics and Image Processing, 19 (SIBGRAPI)
%C Manaus, AM, Brazil
%8 8-11 Oct. 2006
%I IEEE Computer Society
%J Los Alamitos
%S Proceedings
%K multiple orientations, multiple motions, transparency, occlusion.
%X Estimation of multiple orientations in multi-dimensional signals is a strong non-linear problem. A solution form this problem is presented in two steps. First, it is linearized by introducing the so-called "mixed orientations parameters" as an unique, albeit implicit, descriptor of the orientations. Next, the non-linearities are solved in order to find the individual orientations. For two-dimensional signals, e.g., images, this decomposition step is solved by seeking for the roots of polynomials. For multi-dimensional signals, the decomposition problem is reduced to a cascade of decompositions problems in two dimensional signals and solved. Therefore, a full solution for the estimation of any numbers of orientations in any dimension is achieved.
%@language en
%3 MotaC_EstimationOrientationMotion.pdf