home   |<<   <--   -->   >>|

Application area

We are concerned with medical datasets, like CT or MRI, comprised of scalar values on a regular grid.

We make the simplifying assumption that the surfaces we wish to visualize, are isosurfaces of the field. This implies a few things:

• When I say curvature, I mean implicit surface curvature.
• Also, the surface normal is defined by the gradient vector. Actually, its the negative gradient vector, normalized, based on the assumption that higher values are more "inside" the object.
• And this means that curvature measurement will be based on analyzing how the gradient vector changes in the volume.