{"componentChunkName":"component---src-templates-publications-js","path":"/line-ensembles","result":{"data":{"markdownRemark":{"html":"

\n \n \n teaser density\n \n

\n

\n \n \n teaser density selection\n \n

\n

\n \n \n teaser density map\n \n

\n

\n \n \n teaser friction\n \n

\n

\n \n \n teaser friction selection\n \n

\n

\n \n \n teaser friction map\n \n

\n

Figure 1: Weather observations: density-based selection compared with structural selection.

\n

Abstract (Summary)

\n

This work studies how to expose meaningful patterns in crowded collections of trajectories. It scores each trajectory against a tensor representation of its neighborhood, then maps disagreements into an image-space field. Excluding a trajectory from its own reference reduces bias. Users can query and remove subsets to reveal further patterns without replacing the original paths with a bundled approximation. Experiments use synthetic trajectories, weather observations, aircraft routes, and brain fibers to examine structural differences that density views can hide.

\n

Links

\n\n

Figures

\n

\n \n \n \n

\n

\n \n \n \n

\n

\n \n \n structure tensor 3\n \n

\n

Figure 2: Building the tensor representation at different neighborhood sizes.

\n

\n \n \n integral raw\n \n

\n

\n \n \n \n

\n

\n \n \n integral execute\n \n

\n

Figure 3: Evaluating agreement along complete trajectories.

\n

\n \n \n synthetic1 raw\n \n

\n

\n \n \n synthetic2 raw\n \n

\n

\n \n \n synthetic3 raw conn\n \n

\n

\n \n \n synthetic3 raw disconn\n \n

\n

Figure 4: Synthetic input collections.

\n

\n \n \n synthetic1 density\n \n

\n

\n \n \n \n

\n

\n \n \n \n

\n

\n \n \n synthetic2 density\n \n

\n

\n \n \n \n

\n

\n \n \n \n

\n

\n \n \n synthetic3 density conn\n \n

\n

\n \n \n synthetic3 messiness conn\n \n

\n

\n \n \n \n

\n

\n \n \n synthetic3 density disconn\n \n

\n

\n \n \n synthetic3 messiness disconn\n \n

\n

\n \n \n \n

\n

Figure 5: Comparing density, disagreement, and categorical views.

\n

\n \n \n \n

\n

Figure 6: Runtime measurements as dataset size increases.

\n

\n \n \n france density\n \n

\n

\n \n \n france sif\n \n

\n

\n \n \n france inliers4288\n \n

\n

\n \n \n \n

\n

Figure 7: Separating dominant flight routes from unusual trajectories.

\n

\n \n \n \n

\n

\n \n \n dti sif\n \n

\n

\n \n \n dti bundling\n \n

\n

\n \n \n dti filtered\n \n

\n

Figure 8: Comparing brain-fiber views and a selected coherent subset.

","frontmatter":{"title":"Structuring Line Ensembles with Path-Integrated Fidelity and Structural Inconsistency Fields","year":"2026","authors":["Yumeng Xue","Patrick Paetzold","Bin Chen","Yunhai Wang","Christophe Hurter","Oliver Deussen"],"publication":"IEEE Transactions on Visualization and Computer Graphics (Proc. IEEE VIS 2026), 2026"}}},"pageContext":{"id":"1e3731aa-8dd7-5f4c-9fbe-c78dc0977237"}},"staticQueryHashes":["63159454"]}