AVATAR -- Adaptive Visualization Aid for Touring And Recovery
Lawrence Hall, Kevin W. Bowyer, Nitesh V. Chawla, Jr. T. Moore, W. Philip Kegelmeyer · 2000
This document provides a report on the initial development of software which uses a standard visualization tool to determine, label and display salient regions in large 3D physics simulation datasets. This software uses parallel pattern recognition behind the scenes to handle the huge volume of data. This software is called AVATAR (Adaptive Visualization Aid for Touring and Recovery). It integrates approaches to gathering labeled training data, learning from large training sets utilizing parallelism and the final display of salient data in unseen visualization data sets. The paper uses vorticity fields for a large-eddy simulation to illustrate the method.