[Insight-users] [ITK Community] Upsampling and Interpolators

[Dženan Zukić]

You can upscale by a non-integer multiplier. In fact, you can change
extent, orientation of axes etc at the same time with a single resampling
step. I don't think that integer multiplier upscaling is optimized, so
there should be no performance penalty by not using that case.


On Fri, Feb 21, 2014 at 3:33 PM, Jose Ignacio Prieto <
joseignacio.prieto at gmail.com> wrote:

> Thanks a lot Luis,
> And does anybody know if you can upsample to a non integer multiplier of
> the previous resolution. If so, does it have a performance penalty over an
> integer multiplier?
>
> Thanks
>
>
> On Wed, Feb 19, 2014 at 2:37 AM, Luis Ibanez <luis.ibanez at kitware.com>wrote:
>
>> Hi Jose,
>>
>> Yes, your intuition is correct in this case.
>>
>> If you keep the images as a 4D image, the interpolator will
>> indeed blend values across time points. Since after all, they
>> are just seen as another spatial dimension.
>>
>>
>> You may want to do the 3D resampling within the original
>> 3D images, before you compose them into a 4D image.
>>
>>
>> Note that since ITK 4.0 there is a set of classes in ITK
>> for managing ND+Time.
>>
>>
>> You will find them under
>>
>>                 /ITK/Modules/Video/Core/include
>>
>>
>> In particular,
>> you may find interesting the class
>>
>>                          itkVideoStream.h
>>
>>
>> where you could put 3D images into a time sequence.
>>
>>
>>     Thanks
>>
>>          Luis
>>
>>
>>
>>
>> On Tue, Feb 18, 2014 at 4:40 PM, Jose Ignacio Prieto <
>> joseignacio.prieto at gmail.com> wrote:
>>
>>> Hi all, I am trying to upsample an image using this example as base
>>> http://www.itk.org/Doxygen/html/Filtering_2ResampleVolumesToBeIsotropic_8cxx-example.html#_a5
>>> I have MRI cine data with spacing like 1:1:8. I mounted all time frames
>>> into a 4D image and I would like to upsample on Z axis to get 1:1:1. I was
>>> wondering if I should better divide the 4D image  into a vector of 3D
>>> images so the interpolator wouldn't mix different times frames or it
>>> doesn't matter. ( also because I am interested in performance). Is it ok to
>>> use the linear interpolator for MRI? Or I should use BSPline? If so, is it
>>> faster to work on 3D  for the splines??
>>>
>>> Thank you very much
>>>
>>> Ignacio Prieto
>>>
>>> Software Developer
>>> Biomedical Imaging Center
>>> Pontificia Universidad Católica de Chile
>>>
>>> _____________________________________
>>> Powered by 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://www.itk.org/mailman/listinfo/insight-users
>>>
>>> _______________________________________________
>>> Community mailing list
>>> Community at itk.org
>>> http://public.kitware.com/cgi-bin/mailman/listinfo/community
>>>
>>>
>>
>
>
> --
> José Ignacio Prieto
> celular(nuevo): 94348182
>
> _____________________________________
> Powered by 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://www.itk.org/mailman/listinfo/insight-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20140221/616b326b/attachment.html>