[ITK-users] k-means on 12 bit data

[Simon Alexander]

Yes, a histogram and a single pass on the data is the way to do this
efficiently for the case in question.

I'll happily write one but wanted to avoid the effort if it already exists.


On Tue, Apr 1, 2014 at 1:30 PM, Bradley Lowekamp <blowekamp at mail.nih.gov>wrote:

> Hello,
>
> If I recall correctly this filter uses a sample list for the computation
> [1]. I Would think that the algorithm may be more efficient with a
> histogram type sample container. I would suspect that you could reuse many
> of the building blocks for the current filter but with the generation of a
> histogram instead of the adaptor.
>
> Hope that helps,
> Brad
>
> [1]
> https://github.com/InsightSoftwareConsortium/ITK/blob/master/Modules/Segmentation/Classifiers/include/itkScalarImageKmeansImageFilter.h#L96
>
>
> On Apr 1, 2014, at 12:59 PM, Simon Alexander <skalexander at gmail.com>
> wrote:
>
> > Hi,
> >
> > I was using itk::ScalarImageKmeansImageFilter on some largish volumetric
> data, all unsigned 12 bit, and found it unusably slow for the particular
> task.
> >
> > Is there an efficient k-means filter for constrained data ranges like
> this?
> >
> > For this use it should be computable a couple of orders of magnitude
> faster than the general method, so I'm just trying to avoid re-implementing
> if possible.
> >
> > - Simon
> > _____________________________________
> > 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/20140401/d8e3c89e/attachment.html>