SCI Seg3D Mailing List

Text archives Help


[Seg3D] Re: Interpreting the result of the Point Set Registration


Chronological Thread 
  • From: Ayla Khan <ayla@sci.utah.edu>
  • To: seg3d@sci.utah.edu
  • Subject: [Seg3D] Re: Interpreting the result of the Point Set Registration
  • Date: Thu, 19 Dec 2013 22:05:03 -0700
Hi Sergio,

I attempted to reproduce your problem with a very simple tooth dataset, and also increasing the iterations. The translation parameters in X is approximately -8. The Y parameter gets reasonably close to 0. Z is a bit larger, but still fairly close to 0.

Can you send me your project, or give me more information about the data you were working with?

Thanks,

Ayla



On Dec 19, 2013, at 5:24 AM, Sergio Vera wrote:

Dear Seg3D2 staff

I have a binary mask loaded in Seg3D2. With the Transform operator, I generate a copy of that same image but I change its origin to be +8 units in the X axis.

So I have a mask and the same mask but displaced in the X axis.

If I apply the Point set registration algorithm to these masks, after 30 iterations I obtain a new mask that is close to the target mask, but the translation parameters that appear on the ITK interface are very confusing. 

I would expect to obtain a translation in X close to 8 (or -8) and Y and Z translation to be near 0.
Instead I obtain translations on X, Y and Z between 0.5 and 0.9. However, the newly generated mask is really close to the target mask.

Is there a problem with the transformation values? I'm interpreting the results wrongly?

Many thanks

--
Sergio Vera

 Alma IT Systems
 C/ Vilana, 4B, 4º 1ª
 08022 Barcelona
 T. (+34) 932 380 592
 www.alma3d.com


Archive powered by MHonArc 2.6.16.

Top of page