[vtkusers] Precision of 3D measurements

[Kenneth Sloan]

On Jul 1, 2008, at 10:40 AM, Boettcher, Dr. Peter wrote:

> When reconstructing bones out of CT data the in-plane resolution of  
> the images is most of the time much better than the out-of plane  
> resolution. Therefore one could argue that every 3D-measurement  
> made on a 3D-opbject could not be "better" than the slice thickness  
> of the original CT images.
>
> I do not believe this is true - but I have no arguments to defend  
> my feeling.
>
> The problem I am referring to is the following;
> We did axial CT scans on canine elbows. Then we did 3D-measurements  
> (in millimetre) on those 3D-reconstructed bones. The slice  
> thickness used was 1mm and the increment 0.5mm. We expressed the  
> measurements at 0.01mm precision.
>
> Do you support this approach or what precision do you think is  
> appropriate?
>

You would have to tell us more about the method used to "reconstruct"  
and the purpose of the "measurements".

If you make (perhaps reasonable) assumptions about how smooth the  
original material was, and take sufficient care in interpolation,  
then *perhaps* 0.01mm precision is justified.  But, if you are  
examining the bones for pathologies then reporting measurements at  
this precision would be very misleading.  It all depends on your  
model of the object being reconstructed - and how well you have  
validated that model.

So...what was the *purpose* of the measurements, and how were these  
measurements actually used?


--
Kenneth Sloan                                           
KennethRSloan at gmail.com
Computer and Information Sciences                        +1-205-934-2213
University of Alabama at Birmingham              FAX +1-205-934-5473
Birmingham, AL 35294-1170                http://KennethRSloan.com/