[ITK] Regarding Feature based image Registration in ITK

[Dženan Zukić]

Yes. But I doubt any of the existing similarity metrics will work well, so
I suggest you implement your own.

Regards

On Sat, Jan 23, 2016 at 9:08 AM, Raghav Mehta <raghav.11393 at gmail.com>
wrote:

> Hi,
>
> So you are suggesting that I create an image in which each voxel is
> N-D(number of feature) and than use different existing registration methods
> to do registration??
>
> Raghav Mehta
> MS by Research (ECE)
> Center for Visual Information Technology
> International Institute of Information Technology - Hyderabad
> +91 9712990393
>
> On Sat, Jan 23, 2016 at 7:18 PM, Dženan Zukić <dzenanz at gmail.com> wrote:
>
>> Hi Raghav,
>>
>> have you tried just creating new images with features instead of original
>> intensities and then registering those images? You would then simply use
>> the resulting transform and apply them to original images. You would have
>> to implement your own metric which compares those features and supply the
>> metric class to the registration class.
>>
>> HTH,
>> Dženan
>>
>> On Sat, Jan 23, 2016 at 8:44 AM, Raghav Mehta <raghav.11393 at gmail.com>
>> wrote:
>>
>>> Hi,
>>>
>>> Can anyone provide me information about feature based registration in
>>> ITK.
>>>
>>> I am extracting features at every voxel level and want to use them for
>>> image registration.
>>>
>>> Can anyone point out any examples or projects where this type of methods
>>> is used in ITK??
>>>
>>> Regards,
>>> Raghav Mehta
>>> MS by Research (ECE)
>>> Center for Visual Information Technology
>>> International Institute of Information Technology - Hyderabad
>>> +91 9712990393
>>>
>>> _______________________________________________
>>> Community mailing list
>>> Community at itk.org
>>> http://public.kitware.com/mailman/listinfo/community
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/community/attachments/20160123/2a751f87/attachment.html>