[Insight-developers] Problems with a couple auto-threshold filters

[Bradley Lowekamp]

Hello,

The mask feature would be VERY useful  to me as well. We tend to use 0 value pixels for boarders and outside ROI, so the mask paradigm would work very easily.

About the design of this new feature, are you thinking about a new set of filters, or just adding an optional input to the base-class, and  choosing the correct image to histogram filter at run-time. While I am usually, an advocate for a filter just doing one thing well, I find the idea of an optional mask image input intuitive.

Brad


On Nov 17, 2011, at 5:59 PM, Richard Beare wrote:

> I'm hoping to have a look at a couple of issues with the
> autothresholding. Not sure when I'll get to it though. I'm also hoping
> to add some masked versions that exploit the masked histogram
> operation.
> 
> 2011/11/18 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>:
>> 
>> Hi Brad,
>> 
>> Sorry I did have a chance to reply earlier.
>> I don't have much time currently unfortunately. I'll try to fix those bug
>> anyway.
>> 
>> For the record, you said in a previous mail this is with the image
>> RA-Short.nrrd and short as pixel type.
>> 
>> Thanks,
>> 
>> Gaëtan
>> 
>> 
>> Le 17 nov. 11 à 15:29, Bradley Lowekamp a écrit :
>> 
>>> Hello,
>>> 
>>> I began wrapping the new histogram based auto-theshold image filters, and
>>> ran into problems with a couple of filters.
>>> 
>>> I am getting segfaults with the LiThresholdFilter. And then with the
>>> HuangThresholdFilter am not always getting the same results ( the md5 of the
>>> mask image changes between a 3 or 4 values on my laptop ).
>>> 
>>> Thanks,
>>> Brad
>>> 
>>> 
>>> ========================================================
>>> Bradley Lowekamp
>>> Medical Science and Computing for
>>> Office of High Performance Computing and Communications
>>> National Library of Medicine
>>> blowekamp at mail.nih.gov
>>> 
>>> 
>>> 
>> 
>> --
>> Gaëtan Lehmann
>> Biologie du Développement et de la Reproduction
>> INRA de Jouy-en-Josas (France)
>> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
>> http://mima2.jouy.inra.fr  http://www.itk.org
>> http://www.bepo.fr
>> 
>> 

========================================================
Bradley Lowekamp  
Medical Science and Computing for
Office of High Performance Computing and Communications
National Library of Medicine 
blowekamp at mail.nih.gov



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20111123/2eaa482e/attachment.htm>