[Insight-users] Apply template matching with ITK

Luis Ibanez luis.ibanez at kitware.com
Wed Jul 1 15:58:13 EDT 2009


HI Reinaldo,

Is your goal to segment and quantify shape and size of the droplets ?

If so, you may find useful the methods used for analyzing cells in
microscopy images.

In particular the ones described in the following Insight Journal papers
(whose source code is available).

"Blob Segmentation":
http://www.insight-journal.org/browse/publication/321

"Watersheds"
http://www.insight-journal.org/browse/publication/92

"Consolidated Morphology"
http://www.insight-journal.org/browse/publication/124

and related papers:
http://www.insight-journal.org/browse/publication/642
http://www.insight-journal.org/browse/publication/322


If your droplets are similar enough, and rotationally
symmetric, you could do pattern matching in the Fourier
domain, by computing the Fourier transform of your
large image, and the template, multiplying them in
the Fourier domain, and then computing the inverse
Fourier transform of the product.

This will be the equivalent of compute correlation in
the spatial domain of the image.

You could do this with each one of the templates...
and then count peaks in the output image by using
the LocalMaxima filter.


     Regards,


           Luis


------------------------------------------------------------------------
On Wed, Jul 1, 2009 at 3:40 PM, Reinaldo <rheineo at gmail.com> wrote:

> Hi, Luis
>
> thanks ..
>
> I am analyzing images obtained on emulsions of an electron microscope, is
> composed of an emulsion droplets.
>
> I need to consider the drops as I have in some small images (templates)
> shows that the emulsion, the drops are for different kinds and sizes, the
> intention is to know how many of each type are in the picture.
>
> Templates are greater than 10x10 pixel
> The original image has a pixel size of 576x720
>
> Thanks
> --
> REINALDO
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20090701/d52c7aed/attachment.htm>


More information about the Insight-users mailing list