FlowGraph: A compound hierarchical graph for flow field exploration

Abstract

Visual exploration of large and complex 3D flow fields is critically important for understanding many aero-and hydro-dynamical systems that dominate various physical and natural phenomena in the world. In this paper, we introduce the FlowGraph, a novel compound graph representation that organizes streamline clusters and spatial regions hierarchically for occlusion-free and controllable visual exploration. Our approach works with any seeding strategies as long as the domain is well covered and important flow features are captured. By transforming a flow field to a graph representation, we enable observation and exploration of the relationships among streamline clusters, spatial regions and their interconnection in the transformed space. The FlowGraph not only provides a visual mapping that abstracts streamline clusters and spatial regions in various levels of detail, but also serves as a navigation tool that guides flow field exploration and understanding. Through brushing and linking in conjunction with the spatial streamline view, we demonstrate the effectiveness of FlowGraph with several visual exploration and comparison tasks that can not be well accomplished using the streamline view alone. As occlusion and clutter are almost ubiquitous in 3D flows, the FlowGraph represents a promising direction for enhancing our ability to understand large and complex flow field data.

Integral Curve Clustering and Simplification for Flow Visualization: A Comparative Evaluation

“…Besides, AHC can be coupled with many customized similarity measures that describe either the spatial distance or shape similarity of integral curves. Examples include the average-linkage AHC with a weighted endcurve-distance [12], average-linkage AHC with weighted form of signature-based similarity and mean distance [13], single-linkage AHC with mean-of-thresholded-closest-distance for fiber bundle clustering [24], Wards-variance AHC with segment matching cost distance [27], penalized-linkage AHC with a DTW-based histogram similarity measure [28], average-linkage with the string matching cost measure [14], self-defined AHC with a graph-based similarity measure [29], average-linkage AHC with a specific spatio-temporal similarity measure for adjacent blood flow pattern classification [25], [26].…”

Graphs in Scientific Visualization: A Survey

“…Rather than only considering streamline clusters or spatial regions as nodes, FlowGraph developed by Ma et al . [MWS13] integrates both streamline clusters or spatial regions as nodes and thus presents a more complete picture. As a compound graph, FlowGraph consists of two kinds of nodes (L‐nodes and R‐nodes) and three kinds of edges (L‐L edges, R‐R edges and L‐R edges) where ‘L’ denotes streamline and ‘R’ denotes spatial region.…”

FLDA: Latent Dirichlet Allocation Based Unsteady Flow Analysis

“…TransGraph [16] is proposed to visualize information transition along time between blocks, where users can select different levels of hierarchy. Ma et al [34] proposed FlowGraph to explore the dual relationships between streamlines and blocks with rich interaction and query techniques. Jänicke et al [26] extract local flow patterns as nodes in graph, and their transitions as edges where users can track features over time.…”