[Insight-users] Naively/smoothly interpolating a hole in a grayscale image

[Juan Cardelino]

Hello David,
      The problem you are referring to is called "inpainting". Theres
plenty of algorithms available to do that, from the 2000s when the
term was first introduced. They success mostly depends on the type of
image you have. There is a family of algorithms intended for heavily
textured images (See Efros and Leung for a starter), and another one
that works well with images that have some geometric structure. (see
Bertalmio et al for that).
Recently some inpainting code was submitted to the IJ, I think an
implementation of a paper by Criminisi et al.
Apart from ITK, there are plenty of implementations that can be
downloaded from the web.
I hope this helps.
Best regards,
                  Juan

On Tue, Jul 5, 2011 at 8:48 PM, David Doria <daviddoria at gmail.com> wrote:
> Consider a grayscale image with a hole (undefined pixels) in the
> middle. ITK has some nearest-neighbor interpolation classes, and some
> "better" interpolation classes (WindowedSincInterpolateImageFunction),
> but all of these seem to rely on all of the information around a pixel
> being known. Is there any filter that will fill a hole like that
> reasonably? I guess something like an iterative "fill pixels who have
> defined boundary pixels, then repeat now that there are new undefined
> pixels near pixels determined in the previous step, repeat"? Or is
> there a better way to do this in one step?
>
> Thanks,
>
> David
> _____________________________________
> 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.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
>