[ITK-users] [ITK] Pointset registration

Wed Nov 11 17:09:01 EST 2015

It seems that you want to do a simple landmark registration (you touch exactly the same points with the stylus, in the same order, as you defined in the image). If this is the case, then you should not use ICP but itk::LandmarkBasedTransformInitializer (with optional preprocessing step to determine the scale), as it is done in this Slicer module:
https://github.com/Slicer/Slicer/blob/master/Modules/CLI/FiducialRegistration/FiducialRegistration.cxx

Andras

From: Community [mailto:community-bounces at itk.org] On Behalf Of Matias Montroull
Sent: Wednesday, November 11, 2015 4:58 PM
To: Matt McCormick <matt.mccormick at kitware.com>
Cc: insight-users at itk.org Users <insight-users at itk.org>
Subject: Re: [ITK] [ITK-users] Pointset registration

Yes, here it goes..

you will notice there are points that I had to repeat, the original dataset was 4 points on each file and I was getting warnings:  Number of unknowns(12) greater than number of data (4)

"puntosfijos.txt" is the fixed pointsets
"puntosflotantes.txt" is the moving pointset..

Thanks for your help.

Matias.

El mié., 11 de nov. de 2015 a la(s) 6:43 p. m., Matt McCormick <matt.mccormick at kitware.com<mailto:matt.mccormick at kitware.com>> escribió:
Hi Matias,

Can the datasets be shared?

Thanks,
Matt

On Wed, Nov 11, 2015 at 4:38 PM, Matias Montroull <matimontg at gmail.com<mailto:matimontg at gmail.com>> wrote:
> Thanks.
>
> Here's another question, it seems I need 12 points to perform the
> registration and obtain the matrix, now, the results I'm getting are not
> very good. We have the algorithm programmed manually (not using ITK) and
> when we compare the results are different (way too different).
> I just pulled the example and modified to 3 dimensions and changed the
> translationtransform to an affine transform. Is there an example using
> Affine that I can check? I'm thinking is not just changing the transform,
> there may be other variables I need to tweak..
> Matias.
>
> El mié., 11 de nov. de 2015 a la(s) 5:51 p. m., Dženan Zukić
> <dzenanz at gmail.com<mailto:dzenanz at gmail.com>> escribió:
>>
>> Nowhere, which means it's the >> operator's standard white-space
>> separators (spaces, tabs, newlines).
>>
>> On Wed, Nov 11, 2015 at 3:03 PM, Matias Montroull <matimontg at gmail.com<mailto:matimontg at gmail.com>>
>> wrote:
>>>
>>> by the way, where in the code is the space separator specified?
>>>
>>> El mié., 11 de nov. de 2015 a la(s) 5:02 p. m., Matias Montroull
>>> <matimontg at gmail.com<mailto:matimontg at gmail.com>> escribió:
>>>>
>>>> that worked! Thanks for the advise, I was under the impression that
>>>> points needed to be separated by ";"
>>>> Thanks again
>>>>
>>>> El mié., 11 de nov. de 2015 a la(s) 5:00 p. m., Dženan Zukić
>>>> <dzenanz at gmail.com<mailto:dzenanz at gmail.com>> escribió:
>>>>>
>>>>> By looking at the code, the numbers should be separated by space, not
>>>>> semicolon. Can you try that?
>>>>>
>>>>> HTH
>>>>>
>>>>> On Wed, Nov 11, 2015 at 2:51 PM, Matias Montroull <matimontg at gmail.com<mailto:matimontg at gmail.com>>
>>>>> wrote:
>>>>>>
>>>>>> Hi Dženan, thanks for the suggestion, I tried that example and I think
>>>>>> I'm missign something as my disk drive became crazy.
>>>>>> I basically created 2 txt files with a set of 4 points each in this
>>>>>> format:
>>>>>> -0.585938;92.406328;38.200
>>>>>> 44.7654;52.73442;61.000
>>>>>> -53.5;49.218792;68.200
>>>>>> -1.757814;81.445382;70.600
>>>>>>
>>>>>> second file has the same format but different numbers of course.
>>>>>>
>>>>>> Is there anything wrong with how I entered the point sets in both
>>>>>> files?
>>>>>>
>>>>>> As you can see it is 3 points per line separated by this character ";"
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Matias.
>>>>>>
>>>>>>
>>>>>> El mar., 10 de nov. de 2015 a la(s) 6:16 p. m., Dženan Zukić
>>>>>> <dzenanz at gmail.com<mailto:dzenanz at gmail.com>> escribió:
>>>>>>>
>>>>>>> Have you seen these two examples:
>>>>>>>
>>>>>>> http://www.itk.org/Doxygen/html/Examples_2RegistrationITKv3_2IterativeClosestPoint3_8cxx-example.html
>>>>>>>
>>>>>>> http://www.itk.org/Doxygen/html/Examples_2RegistrationITKv4_2IterativeClosestPoint3_8cxx-example.html
>>>>>>>
>>>>>>> HTH
>>>>>>>
>>>>>>> On Tue, Nov 10, 2015 at 3:18 PM, Matias Montroull
>>>>>>> <matimontg at gmail.com<mailto:matimontg at gmail.com>> wrote:
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I have 2 sets of points (one is points on an image and another one
>>>>>>>> is points on a NDI tracker) and I need to obtain the transformation matrix.
>>>>>>>> What would be the best method? I'm currently using a Solver but I'd like to
>>>>>>>> implement this using ITK and better if it is SimpleITK.
>>>>>>>>
>>>>>>>> Thank you!
>>>>>>>>
>>>>>>>> Matias.
>>>>>>>> --
>>>>>>>> Matias
>>>>>>>>
>>>>>>>> _____________________________________
>>>>>>>> Powered by www.kitware.com<http://www.kitware.com>
>>>>>>>>
>>>>>>>> Visit other Kitware open-source projects at
>>>>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>>>>
>>>>>>>> Kitware offers ITK Training Courses, for more information visit:
>>>>>>>> http://www.kitware.com/products/protraining.php
>>>>>>>>
>>>>>>>> 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://public.kitware.com/mailman/listinfo/insight-users
>>>>>>>>
>>>>>>>
>>>>>> --
>>>>>> Matias
>>>>>
>>>>>
>>>> --
>>>> Matias
>>>
>>> --
>>> Matias
>>
>>
> --
> Matias
>
> _____________________________________
> Powered by www.kitware.com<http://www.kitware.com>
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.php
>
> 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://public.kitware.com/mailman/listinfo/insight-users
>
--
Matias
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/insight-users/attachments/20151111/2b1b8e9b/attachment-0001.html>