Performance Evaluation of Multiresolution Isosurface Rendering

Abstract

Our work involves the establishment of a standard multiresolution data hierarchy of large scientific datasets. In additional to a conventional level-of-detail design, our model provides a meaningful localized error estimation for each level of the representation. This paper describes a computational study that examines our multiresolution data representation model and evaluates its performance using real life volume datasets. A C++ wavelet library implements the multidimensional transformations and recursive data projections. We explore the space/time tradeoffs of approximation construction within a multiresolution hierarchy for volume data.