Time- and Space-efficient Error Calculation for Multiresolution Direct Volume Rendering

Multiresolution data representations crucial for key to viewing large volumetric datasets interactively. There are several stages in the visualization pipeline to render large data sets. When a data set is too large to fit into texture memory, or even into main memory, a “cut” must be made through the multiresolution data hierarchy to attain a coarser subset of the data. Ideally, the subset is chosen such that the error made when visualizing the subset (compared to a visualization of the full data set) is smaller than that of any other subset. For real-time visualization, however, it is computationally too expensive to calculate the exact error during runtime. For example, when coupling a multiresolution representation with a direct volume rendering technique, the screen-space error depends on the chosen transfer function. We present an algorithm that stores an intermediate form of the error, which allows us to efficiently compute a good approximation of screenspace error. The input for our algorithm is any spatially subdivided multiresolution representation of grid-aligned multi-variate volume data. In this study, we focus on octreeand wavelet-based multiresolution techniques. For each level in the multiresolution hierarchy, the algorithm estimates the screen-space error “on the fly,” with respect to transfer function and viewing direction. The error is approximated by means of a two-dimensional histogram of error pairs. The presented method extends previous work in the sense that it balances computational and memory costs with approximation quality of the error estimates.