[Insight-users] Validation of deformable registration

Jonathan Orban jonathan.orban at uclouvain.be
Wed Mar 11 04:40:09 EDT 2009


Hi Marta,

You should take a look at "A Framework for Comparison and Evaluation of 
Nonlinear Intra-Subject Image Registration Algorithms" from Martin 
Urschler, Stefan Kluckner and Horst Bischof 
<http://insight-journal.org/browse/publication/166>. This article 
compares different quantitative field difference measurements.

Moreover, another option you might try is to apply your registration 
strategy on the POPI data 
<http://www.creatis.insa-lyon.fr/rio/popi-model>. For these data, you 
don't have a simulated field that you can consider as ground truth but 
you have several markers pointed out by experts which you can use as 
ground truth.

Regards,

Jonathan

Marta a écrit :
> Hi Karthik, hi Luis,
>
> thank you both a lot! I will try that tomorrow!
>
> Do you also have a suggestion on the best way to compare the 
> deformation fields for validation? Is it a proper way to compare them 
> by SSD in each direction separately?
>
> Regards,
> Marta
>
> Am 10.03.2009 um 21:55 schrieb Luis Ibanez:
>
>>
>> Martha,
>>
>> Along the lines of what Karthik suggests,
>> you can use the examples:
>>
>>
>>    Insight/Examples/Registration/
>>                  BSplineWarping1.cxx
>>                  BSplineWarping2.cxx
>>                  ThinPlateSplineWarp.cxx
>>
>>
>> They will generate deformation fields for you.
>>
>>
>>  regards,
>>
>>
>>      Luis
>>
>>
>> ---------------------
>> Karthik Krishnan wrote:
>>> On Tue, Mar 10, 2009 at 8:48 AM, Marta <t_s_80 at gmx.de 
>>> <mailto:t_s_80 at gmx.de>> wrote:
>>>    Hello once again,
>>>    I still have the questions below, but I also have a new one that
>>>    bothers me to the topic.
>>>    What is a good way to find data for my synthetic deformation field?
>>>    I mean, it isn't enough to deform the image "anyhow", it's 
>>> important to
>>>    have a realistic deformation, right? How can I imitate that? Are
>>>    there any tools to create an synthetic deformation field?
>>> You can use one of the kernel splines to specify landmark pairs at 
>>> corresponding anatomical locations to generate a Kernel transform 
>>> with a realistic deformation.
>>> ------------------------------------------------------------------------ 
>>>
>>> _____________________________________
>>> Powered by www.kitware.com
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>> Please keep messages on-topic and check the ITK FAQ at: 
>>> http://www.itk.org/Wiki/ITK_FAQ
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.itk.org/mailman/listinfo/insight-users
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ITK FAQ at: 
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>



More information about the Insight-users mailing list