[Insight-users] ITK measurement units (was: Re: Level Set Deformable Registration Vector Components and Their Units)

[Michael Halle]

Luis Ibanez wrote:

> 
> Hi Kevin,
> 
> The units will be the same as the ones used for the pixel spacing
> in the input images.
> 
> If your images were using spacing in millimeters, then the units
> of the deformation field computed by the deformable registration
> will also be in millimeters.
> 
> If your images were using spacing in microns, then the units
> of the deformation field computed by the deformable registration
> will also be in microns.

Related to this topic, has there ever been any discussion about 
retaining measure and unit names for different data dimensions through 
standardized fields in the metadata dictionary?

I'd suggest two fields for each dimension: the first describing the 
quantity being measured, the second the unit of measure.  For example:

"spatial extent" "mm"
"time" "second"

In the simplest case, ITK wouldn't interpret these strings at all, just 
store them in a standard field for the pipeline to use if it wanted.

One immediate benefit would come for labeling axes in interactive 
applications.  The second would be trivial checks for dimension or unit 
mismatch based on string equality (i.e., a simple check of correctness 
might require the axis and unit to be string-equal, but know nothing 
about the actual units of measure).

Basically, all it would take is a standardized naming convention.  Has 
anyone done something like this?

Thanks.

--Mike