Spectral mappings

The spectral mapping transform allows you to compute a weighted average of a data object's range components. Actually, it lets you create three different weighted averages at the same time, into three output range components. By default, VisBio treats the first component as red, the second as green, and the third as blue. Thus, the spectral mapping converts a data object with many range components into one with exactly three (weighted averages).

To create a spectral mapping, first import your dataset as described in the Importing a dataset from disk topic. Next, create a dimensional collapse based on the dataset as described in the Dimensional collapse topic, along your "Channels" dimension. This step converts the spectral dimension into a set of distinct range components that the spectral mapping can use for its weighted average. Finally, select the collapsed data object from the list, then click the "Add >" button and choose "Spectral mapping" from the popup menu.

Choose a name for the spectral mapping, and VisBio will create another data object, a child of the collapsed data object, that represents the spectrally mapped data. To alter the mapping's weights, select the mapped data object and click the "Edit" button. A window similar to this one will appear:

Spectral mapping parameters

Each range component can be assigned a weight between -1.0 and 1.0, which determines its contribution to the first, second and third range components of the mapped data object. By default, the mapping fully weights the first N/3 channels toward the first component, the next N/3 toward the second component, and the last N/3 toward the third component. When you are finished altering the weight sliders, click the "Apply" button to lock in your changes. VisBio will recompute thumbnails as necessary for the new mapping parameters.

For example, say you have imported a 3D dataset with 17 spectral channels (with one range component indicating the intensity of that pixel for the given channel). Collapsing the Channel dimension results in a single-image data object with 17 range components. After creating a spectral mapping for that image, the final data object is a single image with 3 range components (which can be mapped to red, green and blue color components in a display), whose values can be controlled via the spectral weighting sliders.

See also: