Evolutionary Lines for Flow Visualization

Abstract

In this work we explore evolutionary algorithms for selected a visualization application. We demonstrate its potential using an example from flow visualization showing promising first results. Evolutionary algorithms, as guided search approach, find close-to-optimal solutions with respect to some fitness function in an iterative process using biologically motivated mechanisms like selection, mutation and recombination. As such, they provide a powerful alternative to filtering methods commonly used in visualization where the space of possible candidates is densely sampled in a pre-processing step from which the best candidates are selected and visualized. This approach however tends to be increasingly inefficient with growing data size or expensive candidate computations resulting in large pre-processing times. We present an evolutionary algorithm for the problem of streamline selection to highlight features of interest in flow data. Our approach directly optimizes the solution candidates with respect to a user selected fitness function requiring significantly less computations. At the same time the problem of possible under-sampling is solved since we are not tied to a preset resolution. We demonstrate our approach on the well-known flow around an obstacle as case with a two-dimensional search space. The blood flow in an aneurysm serves as an example with a three-dimensional search space. For both, the achieved results are comparable to line filtering approaches with much less line computations.