Direct Volume Rendering with Nonparametric Models of Uncertainty

We present a nonparametric statistical framework for the quantification, analysis, and propagation of data uncertainty in direct volume rendering (DVR). The state-of-the-art DVR allows preserving transfer function (TF) ground truth when visualizing uncertain data; however, existing is restricted to parametric models uncertainty. In this paper, we address limitations by extending distributions. exploit quantile interpolation technique derive probability distributions representing viewing-ray sample intensities closed form, which accurate efficient computation. evaluate our proposed through qualitative quantitative comparisons with mean-field models, such as uniform Gaussian, well Gaussian mixtures. addition, an extension 2D TFs improved classifications. show applicability quantification ensemble, downsampled, bivariate versions scalar field datasets.