Identifying high level features of texture perception
CVGIP: Graphical Models and Image Processing
Towards a texture naming system: identifying relevant dimensions of texture
VIS '93 Proceedings of the 4th conference on Visualization '93
Fast Global Illumination for Visualizing Isosurfaces with a 3D Illumination Grid
Computing in Science and Engineering
Numerical Simulation in Molecular Dynamics: Numerics, Algorithms, Parallelization, Applications
Numerical Simulation in Molecular Dynamics: Numerics, Algorithms, Parallelization, Applications
Hi-index | 0.00 |
Large-scale molecular dynamics (GD) simulations are some of the primary applications running on today's supercomputers. These simulations frequently compute the interactions of millions of atoms over millions of time steps, and petascale simulations will target billion atom simulations in the near future. Visualizing the state of such large-scale simulations poses a significant challenge for both data reduction and perception. Standard tools for 3D visualization of MD simulations -- such as direct visualization of the molecules in their simulated physical arrangement -- eventually grow so dense (with more atoms than screen pixels) that only the volumetric shape of the simulation domain is apparent, rather than its detailed content. In this paper, we describe a method for visualizing simulations of a compound in bulk which we refer to as the ensemble display. An ensemble display is produced by superimposing molecular neighborhoods onto fixed reference molecules. This results in a 3D visualization which preserves inter-molecular distances and angles between pairs of molecules while greatly reducing the visual complexity. We performed human-subjects experiments to test the hypothesis that subjects could deduce a bulk physical property of the simulation (temperature) from the ensemble display more accurately than from other common visualizations of GD simulation. Our results show that temperature estimates under the ensemble display had the least error of the tested visualization techniques.