[Insight-developers] SimpleITK Community Survey Draft

[Bradley Lowekamp]

Hello all

Just a quick glance through the questions I am concerned about the following: "Pipeline of image filters (eg Gaussian than Sigmoid filter)"

If someone does not know what a pipeline means, in the ITK sense,  then this may be misleading. It may be better to ask if any of the benefits and negatives of a pipeline are need, so that the pros and cons of this design pattern can be reevaluated for the target group. For example ask for feed back on the importance of the following:

tracking filter dependencies
automatic execution of modified dependencies on update
automatic streaming ( streaming is possible with out a pipeline)

minimal memory usage
immediate execution, and direct access to data


Just a quick thought,
Brad


On Sep 16, 2010, at 3:56 PM, Gabe Hart wrote:

> Hi Wes and Alex (and everyone else),
> 
> This is an interesting question indeed.  I went ahead and removed the "Medical Image Analysis" option from the survey.  Wes, I also made a number of the other changes you suggested earlier.  Please let me know if you see anything else that should be changed.
> 
> Thanks for all the feedback,
> -Gabe
> 
> On 09/16/2010 03:48 PM, Wes Turner wrote:
>> 
>> Man not my day to write ... Yes, I meant Alex!  Thanks for being kind! :-)
>> 
>> - Wes
>> 
>> On Thu, Sep 16, 2010 at 3:45 PM, Alexandre GOUAILLARD <agouaillard at gmail.com> wrote:
>> he he,
>> 
>> you meant alex right, not andre?
>> 
>> anyway, I love this question, because I spent few years wit med.
>> doctors (in hospitals) and then few more years with biologists, and
>> seeing the gap (that I only poorly illustrate in my e-mail before) in
>> their answers always puzzled me.
>> 
>> I agree with you that it should not be that complicated and just be
>> medical imaging, or bioimaging, or whatever they want it to be as long
>> as they leave the processing to us.
>> 
>> :)
>> 
>> alex.
>> 
>> 
>> On Thu, Sep 16, 2010 at 8:05 PM, Wes Turner <wes.turner at kitware.com> wrote:
>> > Andre,
>> > First, very nice note.  I think it lays out the differences nicely.  My
>> > question was more narrowly focused.  What I was saying (poorly) was that we
>> > should jettison Medical Image since both radiology and microscopy can be
>> > medical images and it does not add much additional information to the mix.
>> >  Either that or we could remove both  Medical Imaging and Radiology entirely
>> > and replace them with CT/MRI, PET/SPECT/Nuclear Imaging, X-Ray/Fluoroscopy,
>> > and Ultrasound while leaving microscopy.  Or we can just leave it all the
>> > way it is.
>> > Anyway, thanks again for the response, I promise not to use it in a
>> > religious war.
>> > - Wes
>> >
>> > On Thu, Sep 16, 2010 at 2:45 PM, Alexandre GOUAILLARD
>> > <agouaillard at gmail.com> wrote:
>> >>
>> >> hi wes,
>> >>
>> >> On Thu, Sep 16, 2010 at 6:27 PM, Wes Turner <wes.turner at kitware.com>
>> >> wrote:
>> >> > Gabe,
>> >> > Background Questinos:
>> >> > Question 1:
>> >> > I'm not sure how Microscopy and Radiology differ from Medical Image
>> >> > Analysis
>> >> > ...
>> >>
>> >> Here some point of views on what the answer can be. Be carefull some
>> >> of those answer can start religious wars :-)
>> >>
>> >> - microscopy would mainly be used by biologists / radiology by Med. Dr
>> >> (not the same community of users).  "BioImaging" would englobe both
>> >> fields, while Medical Imaging would only refer to the latest.
>> >>
>> >> - microscopy images are made by microscopes | radiology use scanners
>> >> (different hardware)
>> >>
>> >> - medical images are usually relatively small images, greyscale
>> >> (pixeltype), and in majority do not have time (exception here: cardiac
>> >> CT, ultrasound, ...) whereas microscopy can be 2D, 3D, 2D+t, 3D+t, and
>> >> each image can be an image of vectors (pixeltype). up to terabyte per
>> >> experiment. that makes the question of the types very relevant, as
>> >> well as the question of the images fitting into ram.
>> >>
>> >> - objects contained in medical images tends to come in small number
>> >> (usually one organ) and are relatively big compared to the size of the
>> >> image. Microscopy images usually contains numerous targets, small
>> >> compared to the size of the images (cells, colonies of e-coli, ...).
>> >> Algorithms are/should be different.
>> >
>> >
>> >
>> > --
>> > Wesley D. Turner, Ph.D.
>> > Kitware, Inc.
>> > Technical Leader
>> > 28 Corporate Drive
>> > Clifton Park, NY 12065-8662
>> > Phone: 518-881-4920
>> >
>> 
>> 
>> 
>> -- 
>> Wesley D. Turner, Ph.D.
>> Kitware, Inc.
>> Technical Leader
>> 28 Corporate Drive
>> Clifton Park, NY 12065-8662
>> Phone: 518-881-4920
> 
> <ATT00001..txt>

========================================================
Bradley Lowekamp  
Lockheed Martin Contractor 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/20100916/37517d9d/attachment.htm>