[Insight-users] question about GradientDifferenceImageToImageMetric metric.

Luis Ibanez luis.ibanez at kitware.com
Wed Nov 23 16:06:50 EST 2005


Hi Yervand,


When you say:

  "where I can take information about
   GradientDifferenceImageToImageMetric"


are you asking for references to papers  ?

If so, you may want to look at the book:

  "Medical Image Registration"

   by  Joseph Hajnal (Editor),
       D. J. Hawkes (Editor),
       Derek Hill (Editor).

   The Biomedical Engineering Series.

http://www.amazon.com/exec/obidos/tg/detail/-/0849300649/qid=1101168544/sr=8-1/ref=pd_csp_1/002-3796138-6356058?v=glance&s=books&n=507846


----

About "Testing".


      Yes, we test all the classes.


Note that there is a difference between

        Testing

and

        Validation


* Testing: is a Software Engineering practice.

* Validation: is an Image Processing *Illusion*.


The term Validation is the preferred question
of reviewers because we all know that validation
is never performed right, simply because it is
an ill-posed problem. In other words...Validation
can never be performed.

ITK Developers stopped using the obsolete term
"Validation" and only use "Evaluation" instead.
Evaluation is a more realistic technical term,
because it is something than can actually be
performed.


Evaluation is a *comparative* measure of how well an
algorithm is solving a problem. It requires therefore
to have other algorithms as a baseline, a quality
criteria, and a problem specification.

Evaluation of Image Metrics for registration can only
be done in the context of a well defined problem
that MUST include the specification of:

1) Anatomical Structure
2) Image modality
3) Image acquisition protocol (not all CTs are equal)
4) Goal of the processing:( diagnosis?, follow-up?
      surgical planning ? surgical guidance ? )
5) Constraints: (acceptable computing time, acceptable
     computing resources,...)


Without those specific settings, anyone who brings up the
term "Validation" is either confused, or it is trying to
confuse you.


The obsession with "Validation" is another detrimental
vestige of the "Publish or Perish" obtuse policy to
which our field has been subject for years.

Once you start applying image processing methods with the
goal of improving health care delivered to patients, instead
of having the goal of adding publications to a CV, then
suddenly the difference between Validation and Evaluation
becomes pristine, and the elements become quite logic.



    Regards,



       Luis



-------------------------
Yervand Karagulyan wrote:
>  I want to know where I can take information about 
> GradientDifferenceImageToImageMetric  metric's algoritm and did you 
> tested  metric's classes?
> 
> ------------------------------------------------------------------------
> Yahoo! FareChase - Search multiple travel sites in one click. 
> <http://us.lrd.yahoo.com/_ylc=X3oDMTFqODRtdXQ4BF9TAzMyOTc1MDIEX3MDOTY2ODgxNjkEcG9zAzEEc2VjA21haWwtZm9vdGVyBHNsawNmYw--/SIG=110oav78o/**http%3a//farechase.yahoo.com/> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users



More information about the Insight-users mailing list