| Description |
It is common to extract isosurfaces from simulation eld data to visualize and gain understanding of the underlying physical phenomenon being simulated. As the input parameters of the simulation change, the resulting isosurface varies, and there has been increased interest in quantifying and visualization of these variations as part of the larger interest in uncertainty quantification. In this thesis, we propose an analysis and visualization pipeline for examining the intrinsic variation in isosurfaces caused by simulation parameter perturbation. Drawing inspiration from the shape modeling community, we incorporate the use of heat-kernel signatures (HKS) with a simple nite-difference approach for quantifying the degree to which a region (or even a point) on an isosurface has undergone intrinsic change. Coupled with a clustering technique and the use of color maps, our pipeline allows the user to select the level of fidelity with which they wish to evaluate and visualize the amount of intrinsic change. The pipeline is described with a simple example to walk the reader through the different steps, and experimental validation of parameter choices in the pipeline is provided to justify our design. Then we present canonical and simulation examples to demonstrate the pipeline's use in different applications. |
