Improving Information Visibility in Displayed Data

One method to improve information visibility is the use of pseudocolor, i.e., the mapping of one or more attributes of the information to color. So each value, or range of values corresponds to some color. A common method is to use the rainbow scale, i.e., the assignment is based on the position of the color in the visible spectrum (violet - blue for low values up to red for high values). This scale is based on hue contrast. Some studies have indicated that a scale based on brightness contrast is more effective. The brightness is based on the perceived brightness rather than the physical brightness.

Here is an example of two sets of data images, with the upper pair based on brightness contrast and the lower pair based on hue contrast.

Studies also show that, for colored objects, bright objects on a dark background look bigger than the same objects depicted with dark colors on a bright background. In the above image notice that the green spots in the lower left hand image look larger than the orange spots in the upper left image. There are two implications: care must be taken when making quantitative estimates from the visual data, and small objects can be made more visible by make them brightly colored against a dark background.

Brightness can also be used for increased 3D perception, since most depth cues are sensitive to brightness rather than hue contrast. The assignment of a pseudocolor scale that makes a region brighter enhances the depth perception of this region. This is another reason to choose a scale based on brightness contrast.

One method used to increase the brightness contrast is to use a technique known as mathematical morphology to "shave" the surfaces. The effect is to create a very thin region of data values equal to zero around each surface. This causes a larger gradation in light density, which increases the contrast. Here are two examples of a 2D projection of a 3D image of cerebral glucose metabolism as shown by PET.

Here is an image without brightness contrast enhancement. r_pg132.jpg (13033 bytes)
Here is the same image with brightness contrast enhancement. r_pg133.jpg (15422 bytes)

Different choices of color mapping can also be used to show or hide different aspects of a data set. Here are two images of the same data set, with the upper image showing only the general outline of the data while the lower image shows the internal structure of the data.

Previous Page
HyperVis Table of Contents

Last modified on February 11, 1999 G. Scott Owen,