Time critical isosurface refinement and smoothing

Abstract

Figure 1: (left) Smoothing and (right) refinement with smoothing computed from a coarse representation (center) using the same progressive algorithm. AbstractMulti-resolution data-structures and algorithms are key in Visualization to achieve real-time interaction with large data-sets. Research has been primarily focused on the off-line construction of such representations mostly using decimation schemes. Drawbacks of this class of approaches include: (i) the inability to maintain interactivity when the displayed surface changes frequently, (ii) inability to control the global geometry of the embedding (no selfintersections) of any approximated level of detail of the output surface.In this paper we introduce a technique for on-line construction and smoothing of progressive isosurfaces (see Figure 1). Our hybrid approach combines the flexibility of a progressive multiresolution representation with the advantages of a recursive subdivision scheme. Our main contributions are: (i) a progressive algorithm that builds a multi-resolution surface by successive refinements so that a coarse representation of the output is generated as soon as a coarse representation of the input is provided, (ii) application of the same scheme to smooth the surface by means of a 3D recursive subdivision rule, (iii) a multi-resolution representation where any adaptively selected level of detail surface is guaranteed to be free of self-intersections.¥

3D finite element meshing from imaging data

“…The chain-gang algorithm [33] was presented for isosurface rendering of super adaptive resolution (SAR) and resolves discontinuities in SAR data sets. Progressive multiresolution representation and recursive subdivision are combined effectively, and isosurfaces are constructed and smoothed by applying the edge bisection method [45]. A surface wave-front propagation technique [65] is used to generate multiresolution meshes with good aspect ratio.…”

ViSUS: Visualization Streams for Ultimate Scalability

“…This would make any interactive exploration of the data impossible, because each modification of the isovalue w 0 would require the computation of a new full resolution isosurface, which cannot be done at interactive rates. This problem can be solved in part by using the paradigm of progressive computing that Pascucci et al [P00,PB00] have proven viable for the case of isosurface extraction. Figure 3 shows the conceptual structure of the scheme.…”

A Selective Refinement Approach for Computing the Distance Functions of Curves

“…This forces the distance transform algorithm to subdivide more often in areas where the original scalar field gradients are widely varying. This hierarchy is similar to the data structure used in [8] in that it involves minimal preprocessing and grants access to all of the isocontours in the scalar field.…”