[Tubetk-developers] Collecting testing data using our ultrasound image acquisition pipeline and abdominal phantom

Andinet Enquobahrie andinet.enqu at kitware.com
Tue Sep 11 09:45:13 EDT 2012


This is a good idea. We will keep it in mind when we start generating
the datasets.

Andinet

> Would you be interested in posting some of this as a mini-grand
> challenge?   That is, pre-identify some as training data and others as
> testing data.  Or specify a good experimental design (including
> metrics) for others to follow.   So, people using the data will then
> be more likely to present studies that could be meaningfully compared.
>
> Note that I am not suggesting that we sequester data and such, but by
> pre-identifying the data and asking people to be honest in their use
> of the training/testing sets, we can get more from this data.
>
> I've also CC'd Patrick.   I know he is interested in hosting grand
> challenges at challenge.kitware.com - perhaps this would fit.
>
> s
>
> On Mon, Sep 10, 2012 at 11:23 AM, Matt McCormick
> <matt.mccormick at kitware.com> wrote:
>> Hi Andinet,
>>
>> Excellent.  It will be great to have some good testing data for
>> algorithm development.  If you can record the pose of the ultrasound
>> transducer well enough relative to the CT acquisition, a "known truth"
>> will be very useful... i.e. what Danielle said before I hit Send ;-)
>>
>> Thanks,
>> Matt
>>
>> On Mon, Sep 10, 2012 at 3:00 PM, Andinet Enquobahrie
>> <andinet.enqu at kitware.com> wrote:
>>> Hi Matt
>>>
>>> My next plan is to generate testing data for our new ultrasound
>>> simulation filters using the ultrasound image acquisition pipeline and
>>> abdominal phantom we have here at KRS.  Here is what I have in mind.
>>>
>>> 1) Record pose of the ultrasound probe for every frame
>>> 2) Generate several datasets with different sweeping pattern
>>>
>>> The idea is we will "compare" these "baseline" ultrasound scans with
>>> the ones that we will simulate from the CT scan of the abdominal
>>> phantom using the same sweeping pattern. Sounds good?
>>>
>>> I need to add some functionality to the PLUS library to make this
>>> possible. I am going to focus on that in the next few days.
>>>
>>> Let me know your thoughts and if you need a different set of testing dataset.
>>>
>>> Thanks,
>>> Andinet
>>>
>>>
>>>
>>>
>>>
>>> --
>>> Andinet Enquobahrie, Ph.D.
>>> Technical Lead
>>> Kitware, Inc.
>>>
>>> http://www.kitware.com
>>> (919) 969-6990 x311
>> _______________________________________________
>> Tubetk-developers mailing list
>> Tubetk-developers at tubetk.org
>> http://public.kitware.com/cgi-bin/mailman/listinfo/tubetk-developers
>
>
>
> --
>
> ==============================
> Stephen R. Aylward, Ph.D.
> Senior Director of Operations - North Carolina
> Kitware, Inc. - North Carolina Office
> http://www.kitware.com
> stephen.aylward (Skype)
> (919) 969-6990 x300



-- 
Andinet Enquobahrie, Ph.D.
Technical Lead
Kitware, Inc.

http://www.kitware.com
(919) 969-6990 x311



More information about the Tubetk-developers mailing list