RISCAN Surface Extraction to Meshgrid Surface Representation in MPEG-4" />

From TRISCAN Surface Extraction to Meshgrid Surface Representation in MPEG-4

A.I. Salomie, A. Munteanu, R. Deklerck, D. Cernea, P. Schelkens, and J. Cornelis (Belgium)


Surface extraction, meshes, surface representations, view dependent coding, differential chain code, wavelet coding.


TRISCAN is a new method to extract automatically multi resolution surfaces from object representations expressed as scalar fields (e.g. discrete 3D data) or scalar functions (e.g. implicit surfaces); it puts no limitations on the shape of the objects as all surface topologies are handled correctly. TRISCAN aims to obtain a description of the "global connectivity" of the object's surface, expressed by a connectivity-wireframe (CW), a polygonal mesh that implicitly contains all information for the polygonization of the object's surface. Therefore, the approach is different from other automatic surface extraction methods, e.g. Marching Cubes, which directly target the polygonization of the surface. The CW can be encoded efficiently in a hierarchical way using a 3D extension of the Freeman chain-code, and stored as part of MESHGRID. The paper illustrates how to extract the surface using the TRISCAN approach, and how to store the result as a MESHGRID in order to exploit (1) the particular animation capabilities of MESHGRID, which are not available in other surface representation methods, and (2) MESHGRID's compact and scalable view-dependent stream.

Important Links:

Go Back