Overview and State-of-the-Art of Uncertainty Visualization - Scientific Visualization: Uncertainty, Multifield, Biomedical, and Scalable Visualization - page 11

•

Task oriented user study: a cognitive, task-based user study is conducted to assess
the efficiency or the usability of the method.

1.4.1 Theoretical Evaluation

General guidelines and rules regarding visual depiction of data have been established,
that have proven their efficiency. Bertin in [ 5 ], later translated in [ 6 ], has introduced
the concept of visual variables. These include among others the location, size, ori-
entation, shape, focus and realism. Furthermore he defined four visual properties,
natural ordering, the ability to quantify, the ability to focus user attention (selectiv-
ity) and the ability to associate similar elements (associativity). He studied which of
these properties are verified by the visual variables. Tufte in [ 102 ], through his con-
cepts of graphical excellence and integrity, has proposed a number of guidelines to
enhance the precision and the usability of graphical depiction. Chambers et al. in [ 14 ]
have studied the relative influence of specific patterns on the visual perception, for
example straight lines versus curves, dark versus light objects or small versus large
patterns. This study leads the authors to define general rules for plot construction.
These graphical design principles may be used to conduct a theoretical evalua-
tion of new uncertainty visualization techniques. As already mentioned in Sect. 1.2 ,
Zuk and Carpendale in [ 111 ] have done such an evaluation for eight uncertainty
visualization techniques. The same type of theoretical evaluation was followed by
Riveiroin[ 87 ] to evaluate three uncertainty visualization techniques in the context
of information fusion and decision making.

1.4.2 Low-Level Visual Evaluation

Barthelmé and Mamassian in [ 2 ] studied the influence on noise uncertainty in a
decision-making task. Based on psychometric and simple task experiments, he
proved that users can reliably measure the visual uncertainty and use this information
in their decision-making. Coninx et al. in [ 18 ] conducted psychometric experiments
to measure the impact of contrast sensitivity on the visibility of uncertain noisy pat-
terns. He used this information in order to control the visibility of uncertainty data
in a visualization technique based on the perturbation of colormaps by Perlin noise.

1.4.3 Task-Oriented User Study

Task oriented cognitive user studies are by far the most common way of assessing
the efficiency and usability of uncertainty visualization techniques. In this type of
evaluation a panel of users is typically asked to perform a task that requires not