[Insight-users] Question on cost function of registration

Luis Ibanez luis.ibanez at kitware.com
Fri Dec 22 08:19:50 EST 2006


Hi Alex,

You may want to look at the introductory tutorials:

  http://www.itk.org/HTML/Tutorials.htm

as well as to the advanced ITK course:

http://www.na-mic.org/Wiki/index.php/Dissemination:EPFL_Workshop_2005

in this advanced course you will find the session:

"Writing a new filter"
http://www.na-mic.org/Wiki/images/b/bc/Insight-Writing-a-New-Filter.ppt


Note that an image metric is not implemented as a filter,
but still, most of the advice in that presentation will
be relevant. In particular what refers to defining the traits
of the class.



  Regards,


     Luis


========================
Alexander.Wang wrote:
>  
> Dan,
>  
> Have you ever written your own class. It's my first time to do this 
> work. Could you give me some advice, especially on the build step. A 
> cmake file is needed, I think, but I find few information about this. Or 
> is there any tutorial available that I have missed? Thanks.
>  
> Regards.
> Alex
>  
> 在2006-12-21,"Daniel Mace" <dlm19 at duke.edu> 写道:
> 
> Alex,
> 
> The class you probably want to extend is the itk::ImageToImageMetric
> (for pixel metrics) class. There are also other classes (point to Image,
> point to point metrics), that might be more appropriate. Check out here
> for the full list of metric classes:
> http://smsc.cnes.fr/PLEIADES/OTB/Doxygen/html/classitk_1_1CostFunction.html
> 
> The easiest class to look at to understand the code would probably be
> the itk::MeanSquaresImageToImageMetric. Depending on your metric, it's
> usually not too difficult to write your own metric class; The only
> method I tend to change is the GetValue() method. Since the similarity
> metric is already in ITK, the simplest approach would be to cut and
> paste that class and modify the GetValue() method to add in your
> deformation term.
> 
> Cheers,
> Dan
> 
> Alexander.Wang wrote:
>> Hi all,
>> I am working on deformable registration, and want to try a new cost
>> function when implementing registration. The cost funtion includes a
>> similarity term and a deformation term. I only find the Metric
>> representing similarity provided in ITK, maybe I missed something.
>> Could anyone tell me how to implement this type of cost function? Do I
>> need to write a derived class of metric? Any information or reference
>> is appreciated.
>> Thanks in advance!
>> Alexander Wang
>>
>>
>>
>> 昨 天 和 女 友 b u y 疯 了 ! ! !
>> 年 末 万 种 精 品大 促 销 , 8 0 万 大 奖 天 天 送 ( 绝 对 保 真 )
>> <http://adtaobao.allyes.com/main/adfclick?db=adtaobao&bid=600 <http://adtaobao.allyes.com/main/adfclick?db=adtaobao&bid=600>,597,58&cid=29985,198,1&sid=32501&show=ignore&url=http://www.taobao.com/theme/lucky_new_year/index.php>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Insight-users mailing list
>> Insight-users at itk.org <mailto:Insight-users at itk.org>
>> http://www.itk.org/mailman/listinfo/insight-users
>>   
> 
> 
> 
> 
> 
> 昨 天 和 女 友 b u y 疯 了 ! ! !
> 年 末 万 种 精 品 大 促 销 , 8 0 万 大 奖 天 天 送 ( 绝 对 保 真 ) 
> <http://adtaobao.allyes.com/main/adfclick?db=adtaobao&bid=600,597,58&cid=29985,198,1&sid=32501&show=ignore&url=http://www.taobao.com/theme/lucky_new_year/index.php> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users


More information about the Insight-users mailing list