Authors:
Stavros Papadopoulos
;
Anastasios Drosou
and
Dimitrios Tzovaras
Affiliation:
Centre for Research and Technology Hellas, Greece
Keyword(s):
Data Visualization, Hierarchical, Magnification.
Related
Ontology
Subjects/Areas/Topics:
Abstract Data Visualization
;
Computer Vision, Visualization and Computer Graphics
;
General Data Visualization
;
Information and Scientific Visualization
;
Perception and Cognition in Visualization
;
Visual Representation and Interaction
Abstract:
Non-linear deformations are useful for applications where users face highly cluttered visual displays, either
due to large datasets, or visualizations on small screens, or a combination of both, that increases the density
of the data and makes the perception of patterns difficult. Non-linear deformations have been used to magnify
significant/cluttered regions in data visualization, for the purpose of reducing clutter and enhancing the perception
of patterns. General deformation methods (e.g. logarithmic scaling and fish-eye views) suffer from
several drawbacks, since they do not consider the prominent features that must be preserved in the visualization.
This work introduces a hierarchical approach for non-linear deformation that aims to reduce visual clutter
by magnifying significant regions, and lead to enhanced visualizations of two/three-dimensional datasets on
highly cluttered displays. The proposed approach utilizes an energy function, which aims to determine the
optimal deform
ation for every local region in the data, taking the information from multiple single-layer significance
maps into account. The problem is subsequently transformed into an optimization problem for the
minimization of the energy function under specific spatial constraints. The proposed hierarchical approach
for the generation of the significance map, surpasses current methods, and manages to efficiently identify
significant regions and achieve better results.
(More)