[Insight-users] spatial filter constrains

[Soren Christensen]

Hi,
 I am an ITK newbie and was looking for a way to smooth - or more generally
- apply any convolution type filter to just the parts of the image defined
by an image mask.
Fx., for a neighbourhood mean filter - in the case where a reliable mask
exists -  I would like to disregard pixels not in the mask so they do not
enter into the mean. I have previously been doing that by filling in NaNs
outside my mask and doing nanmean in Matlab with decent results. I did find
some previous posts on Nan's which appears to not be generally supported,
but I was wondering if there is a similar way of achieving what I want to
do?
I am aware of edge-preserving filters and looking for a direct substitute to
the nan-mean approach.

Thanks
Soren
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20100112/225b7c88/attachment.htm>