[Insight-users] Re: itkNormalizedMutualInformationHistogramImageToImageMetric

Luis Ibanez luis.ibanez at kitware.com
Fri Jul 16 11:37:21 EDT 2004


Hi Tina,

The histogram size will probably have a negligeable
effect on the computation time of this algorithm,
since the histogram size only matters in the later
computation of the Normalized Mutual Information.

My guess is that most of the time for computing the
metric goes into the computation of the histogram
itself.

You can easily verify such hypothesis by inserting
TimeProbes into the source code of the Metric, and
checking how long it takes to perform each one
of the intermediate steps.

For information on TimeProbes please look at:
http://www.itk.org/Insight/Doxygen/html/classitk_1_1TimeProbe.html



The time required for computing a histogram for an
image is usually independent of the number of
histogram bins.

What you probably want to do is to use a Multi-Resolution
registration scheme, since in that way you will actually
reduce the time required for computing the histogram at
every iteration.


Please let us know if you have further questions,



    Thanks,


     Luis



-------------------
Tina Wang wrote:

> Luis,
>  
> Thank you very much for helps. I have updated the CVS code and it works 
> good for 2D data.
>  
> When I switched to 3D, signed short data, it was extremely slow. Could 
> you give me some tips about how to choose the right histogrambin size or 
> others which could speed the operation.
>  
> Thanks,
>  
> Tina
> 
> */Luis Ibanez <luis.ibanez at kitware.com>/* wrote:
> 
> 
>     Hi Tina,
> 
> 
>     Please update your CVS checkout of ITK.
> 
>     We just commited an example on the use of
>     the Normalized Mutual Information metric.
> 
> 
>     You will find this new example under
> 
> 
>     Insight/Examples/Registration/
>     ImageRegistration14.cxx
> 
> 
>     Please let us know if you have further
>     questions,
> 
> 
> 
>     Thanks
> 
> 
> 
>     Luis
> 
> 
> 
>     -----------------
>     Tina Wang wrote:
> 
>      > Luis,
>      > Do you have any sample code to show how to use this metric?
>      > thanks,
>      > Tina
>      >
>      > __________________________________________________
> 
> 
> 
> ------------------------------------------------------------------------
> Do you Yahoo!?
> Yahoo! Mail 
> <http://us.rd.yahoo.com/mail_us/taglines/50x/*http://promotions.yahoo.com/new_mail/static/efficiency.html> 
> - 50x more storage than other providers!





More information about the Insight-users mailing list