Title :
Surface reconstruction and compression using multiresolution arbitrary topology G1 continuous splines
Author :
Stoddart, A.J. ; Baker, M.S.
Author_Institution :
Surrey Univ., Guildford, UK
Abstract :
In previous work (Saminathan et al. (1997)) we have developed a G 1 continuous arbitrary topology deformable surface called `slime´. Surface reconstruction was shown for point clouds and volumetric field functions within a conventional snake optimisation approach. Hoppe (1995) has proposed an efficient alternative based on a local conjugate gradient optimisation and local mesh editing. We have been able to apply this framework to the slime representation. In this paper we present a technique using local editing operations on the arbitrary topology control mesh. In this way we can produce a surface reconstruction at multiple resolutions. In our results we show how a complex surface of 33000 planar faces can be compressed into 451 vertices (or curved patches) for very little loss of visual quality
Keywords :
computer vision; conjugate gradient methods; data compression; image reconstruction; mesh generation; optimisation; splines (mathematics); topology; complex surface; conjugate gradient method; continuous splines; image compression; mesh; multiresolution arbitrary topology; optimisation; point clouds; surface reconstruction; topology deformable surface; volumetric field functions; Application software; Clouds; Computer graphics; Computer vision; Deformable models; Energy resolution; Face detection; Image reconstruction; Surface reconstruction; Topology;
Conference_Titel :
Pattern Recognition, 1998. Proceedings. Fourteenth International Conference on
Conference_Location :
Brisbane, Qld.
Print_ISBN :
0-8186-8512-3
DOI :
10.1109/ICPR.1998.711265