PROPERTIES
Advantages
-
Acceleration
-
Progressive refinement
-
View-independent
-
Data compression (Laplacian Pyramid)
Disadvantages
-
Higher memory consumption
-
Transferfunction-dependent
To talk in general about the properties of a pyramid data structure is
difficult, because the approaches are very different, so also their properties.
So these properties are special or often fulfilled. A common advantage
is the speedup, this is the main reason why pyramids are used. Progressive
refinement which is related to a speedup, but it is associated with a lower
quality and can be easy realized because of the multiresolution representation
in the pyramid. A big advantage is the view-independence, so you can use
the same pyramid for an animation of the data set, without changing the
pyramid, because the construction is a very expensive process. This is
only valid if in the pyramid stored colors are not view-dependent, so you
should only use view-independent shading models for computing the colors.
The storage of colors and opacities in the pyramid are also a problem,
because if the transfer-function changes, the pyramid has to be reconstructed,
so for the sake of performance the transfer-function should not often change.
It is also possible to compress the volume with a Laplacian pyramid, so
you can use higher-resolutional volume data for visualization or you have
decreasing hardware requirements, but this is only a special property and
it is only a approximation of the original volume data stored. Normally
you need additional memory for the pyramid, depending on the data stored
in it. This should be normally no problem, e.g. for image pyramids, but
the volume data needs much memory and restricts sometimes the data sets,
so also the additional memory can be a restriction.
Main Page