Refine
Document Type
- Conference Proceeding (857) (remove)
Institute
- Professur Informatik im Bauwesen (336)
- In Zusammenarbeit mit der Bauhaus-Universität Weimar (174)
- Professur Baubetrieb und Bauverfahren (104)
- Professur Theorie und Geschichte der modernen Architektur (94)
- Graduiertenkolleg 1462 (32)
- Institut für Strukturmechanik (ISM) (23)
- Professur Angewandte Mathematik (18)
- Institut für Konstruktiven Ingenieurbau (IKI) (12)
- Junior-Professur Computational Architecture (11)
- Bauhaus-Institut für Geschichte und Theorie der Architektur und Planung (10)
Keywords
- Computerunterstütztes Verfahren (286)
- Architektur <Informatik> (200)
- CAD (158)
- Angewandte Informatik (145)
- Angewandte Mathematik (145)
- Computer Science Models in Engineering; Multiscale and Multiphysical Models; Scientific Computing (74)
- Bauhaus (68)
- Arbeitsschutz (61)
- Baustelle (53)
- Architekturtheorie (52)
We present an algebraically extended 2D image representation in this paper. In order to obtain more degrees of freedom, a 2D image is embedded into a certain geometric algebra. Combining methods of differential geometry, tensor algebra, monogenic signal and quadrature filter, the novel 2D image representation can be derived as the monogenic extension of a curvature tensor. The 2D spherical harmonics are employed as basis functions to construct the algebraically extended 2D image representation. From this representation, the monogenic signal and the monogenic curvature signal for modeling intrinsically one and two dimensional (i1D/i2D) structures are obtained as special cases. Local features of amplitude, phase and orientation can be extracted at the same time in this unique framework. Compared with the related work, our approach has the advantage of simultaneous estimation of local phase and orientation. The main contribution is the rotationally invariant phase estimation, which enables phase-based processing in many computer vision tasks.