The Effective Representation of Large Graphs - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

The Effective Representation of Large Graphs

Description:

layout tree, DAG, force-directed, ring. metrics highlight structural properties ... Schumann, Strothotte, Raab & Laser: Assessing the affect of non-photorealistic ... – PowerPoint PPT presentation

Number of Views:24
Avg rating:3.0/5.0
Slides: 12
Provided by: ivanh3
Category:

less

Transcript and Presenter's Notes

Title: The Effective Representation of Large Graphs


1
The Effective Representation of Large Graphs
  • D.J. Duke
  • Department of Mathematical Sciences
  • University of Bath

2
Graph Visualization
  • applications include
  • social structures
  • biochemical pathways
  • web access patterns
  • visualization tools
  • typically lt 104 nodes
  • limited control of filtering and interaction

3
Royere
  • modular graph visualization tool
  • developed at CWI Amsterdam
  • modules from Bath and LaBRI
  • www.cwi.nl/InfoVisu/
  • components provide
  • layout tree, DAG, force-directed, ring
  • metrics highlight structural properties
  • clustering metric/neighbourhood
  • interaction pan/zoom/fisheye etc.

4
Metrics
  • significance
  • based on
  • structural properties
  • application weights
  • Strahler metric
  • S(v) max S(ki) C
  • C (p ? 1), if S(ki) S(kj)
  • C (p ? 2), otherwise
  • Other metrics, e.g. Fission (Marshall, 2001)

5
Skeletons
  • metric determines degree of interest
  • elide nodes with DOI below threshold
  • but provide an overview of unexplored regions.

6
Ongoing work in graph viz.
  • visualization techniques
  • interaction between clustering and layout
  • efficient algorithms
  • generalized skeletons
  • human aspects
  • role of task and domain knowledge
  • interaction
  • cognitive demands

7
Cognitive foundations
8
Non-photorealistic rendering
  • traditional rendering in CG
  • aim for level of photo realism
  • simulate light-material interaction
  • recent growth in NPR
  • enhance perception of detail
  • artistic techniques
  • sketching, shading, painting
  • image or object-space techniques

9
NPR in visualization
  • motivations
  • characterise structure of surfaces
  • differentiate regions (through texture)
  • represent higher-dimensional data, e.g. tensors
  • capture impression of motion in flow
  • challenges
  • real-time interactive rendering
  • dealing with abstraction / simplification

10
The value of sketchiness
Schumann, Strothotte, Raab Laser Assessing the
affect of non-photorealistic rendered images in
CAD. In Proc. CHI96, ACM Press
11
NPR and graph visualization
Write a Comment
User Comments (0)
About PowerShow.com