[Insight-users] intensitywindowing before using mutul information metrics

Luis Ibanez luis.ibanez at kitware.com
Sun May 1 11:20:56 EDT 2005



Hi Jef,

Thanks for the additional details on the study that you are performing.

About your questions:



1) Yes, the intensity range that you select *can* be different
    for both images. In fact, given that they are from different
    modalities, it is almost certain that you will have to select
    different intensity ranges for each image.



2) Karthik Krishnan recently added the perfect example that can help
    you with your study. This example computes and saves the Joint
    histogram of the images at every iteration of the registration
    process. By looking at the way the histogram changes over time
    you will be able to determine which regions of the intensity
    distribution of both images are really making a contribution to
    the metric.

    The example in question is:

        Insight/Examples/Registration/
                   ImageRegistrationHistogramPlotter.cxx

     Since this example is brand-new you will have to update your
     CVS checkout of ITK in order to find this file.



3) In order to identify what anatomical structures are inside
    this intensity range you can use a viewer such as ParaView

                http://www.paraview.org

    that allows you to load images and draw iso-contours and/or
    iso-surfaces at specific levels.




   Please let us know if you have any further questions,



      Thanks



           Luis




-------------------------

Jef Vandemeulebroucke wrote:

> Hi,
> 
> Thank you for your advice on my work Luis. Believe me, I do know that it is
> impossible to "rank" the metrics, or even to "characterize" them correctly
> (or completely). Eventhough I am only including 4 metrics in my research,
> the possiblilities seem endless. But in fact, my goal is not to make a
> theoretical study of the metrics, but to find ONE that works for my
> particular US-MR 2D-3D registration problem. From that point of view, it
> doesn't bother me to perform preprocessing on some metrics and not on
> others, as long as it improves the behavior of the metric in question.
> 
> Your advice on using rescaling or windowing  is new to me, it seems
> applicable for all metrics, no?
> 
> Just to make sure I understand what your saying:
> 
> The idea is to only retain that range of intensities in the image that will
> contribute in a good way to the calculation of the metric? For both images
> this range can be different?
> 
> Is there anyway of easily veryfying what an adequate range would (by
> checking the contributions on the metric), or do I have to visually inspect
> images that have been "windowed" to see if the desired anatomical 
> structures
> are well presented?
> 
> 
> Thanks again for all your advice, ITK is an excelllent toolkit!!!!! I would
> call it my bible but I don't want to piss off the new pope;-)
> 
> 
> Jef
> 
>  >From: Luis Ibanez <luis.ibanez at kitware.com>
>  >To: Jef Vandemeulebroucke <jvdmb at hotmail.com>
>  >CC: insight-users at itk.org
>  >Subject: Re: [Insight-users] Normalisation of images necessary?
>  >Date: Sat, 23 Apr 2005 17:27:25 -0400
>  >
>  >
>  >Hi Jef,
>  >
>  >Normalization of the image intensities is not required for
>  >the Image Metrics:
>  >
>  > > MutualInformationHistogramImageToImageMetric
>  > > NormalizedMutualInformationHistogramImageToImageMetric
>  >
>  >However, what is *VERY* important is to make sure that you use
>  >the range of intensities that is relevant to the anatomical
>  >structures that you care to register.
>  >
>  >In other words, your image will have section of the dynamic
>  >range of intensities that are not contributing (and may even
>  >disturb) the evaluation of the Metric. You should then apply
>  >a filter such as
>  >
>  >
>  >      RescaleIntensityImageFilter
>  >
>  >or
>  >
>  >       IntensityWindowingImageFilter
>  >
>  >for preprocessing the images.
>  >
>  >Note that these filters (and its parameters) bring uncertainty
>  >to your comparision of Image metrics. For the sake of fairness
>  >you probably want to apply *exactly* the same preprocessing to
>  >the image that are fed into all your registration metrics.
>  >
>  >
>  >Note that at the end, any comparision of Algorithms is pointless
>  >and useless if you dont' provide the entire set of material that
>  >you used for your comparision. That includes:
>  >
>  >       - Source code
>  >       - Input images
>  >       - Full sets of parameters
>  >
>  >
>  >Only in this way, other people will be able to repeat your
>  >evaluations and tweak them in different ways. The fact that
>  >each metric has many parameters makes very difficult (if not
>  >impossible) to define a "fair" comparison. For example, you
>  >are selecting for Viola Wells parameters such as :
>  >
>  >     - Number of Bins
>  >     - Number of Samples
>  >     - Standard Deviations
>  >
>  >Changes in any of those parameters will result in dramatic
>  >changes on the outcome of the Metric, and therefore will chage
>  >how this metric perform face to other metrics.
>  >
>  >
>  >Conclusions of the sort:
>  >
>  >               "Metric A is better than Metric B"
>  >
>  >are useless and worst of all: misleading.
>  >
>  >
>  >They are only of interest for writing papers in the Dark Side
>  >of the current publishing system where reproducibility is not
>  >supported or even welcomed, and where conclusions are not derived
>  >from ones' own experience but from subjective judgement, such as
>  >the ones provided by the decadent peer-review system.
>  >
>  >
>  >Unfortunately, those practices still percolate the entire community
>  >of medical image processing.
>  >
>  >
>  >
>  >    Regards,
>  >
>  >
>  >
>  >
>  >        Luis
>  >
>  >
>  >
>  >
>  >-----------------------------
>  >Jef Vandemeulebroucke wrote:
>  >
>  >>Hi,
>  >>  I am testing several mutual information metrics, plotting their
>  >> behavior.
>  >>Among the metrics are the two based on histograms:
>  >>  MutualInformationHistogramImageToImageMetric
>  >>NormalizedMutualInformationHistogramImageToImageMetric
>  >>  Do these metrics give better results when the images have been
>  >> normalised, as it is for the Viola-Wells implementation of MI, or is 
> this
>  >> of no importance?
>  >>  Thank you,
>  >>  Jef
>  >>
>  >>
>  >>------------------------------------------------------------------------
>  >>
>  >>_______________________________________________
>  >>Insight-users mailing list
>  >>Insight-users at itk.org
>  >>http://www.itk.org/mailman/listinfo/insight-users
>  >
>  >
>  >
> 
> 
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
> 
> ----- Original Message -----
> From: "Luis Ibanez" <luis.ibanez at kitware.com>
> To: "Jef Vandemeulebroucke" <jvdmb at hotmail.com>
> Cc: <insight-users at itk.org>
> Sent: Saturday, April 23, 2005 11:27 PM
> Subject: Re: [Insight-users] Normalisation of images necessary?
> 
> 
>  >
>  > Hi Jef,
>  >
>  > Normalization of the image intensities is not required for
>  > the Image Metrics:
>  >
>  > > MutualInformationHistogramImageToImageMetric
>  > > NormalizedMutualInformationHistogramImageToImageMetric
>  >
>  > However, what is *VERY* important is to make sure that you use
>  > the range of intensities that is relevant to the anatomical
>  > structures that you care to register.
>  >
>  > In other words, your image will have section of the dynamic
>  > range of intensities that are not contributing (and may even
>  > disturb) the evaluation of the Metric. You should then apply
>  > a filter such as
>  >
>  >
>  >      RescaleIntensityImageFilter
>  >
>  > or
>  >
>  >       IntensityWindowingImageFilter
>  >
>  > for preprocessing the images.
>  >
>  > Note that these filters (and its parameters) bring uncertainty
>  > to your comparision of Image metrics. For the sake of fairness
>  > you probably want to apply *exactly* the same preprocessing to
>  > the image that are fed into all your registration metrics.
>  >
>  >
>  > Note that at the end, any comparision of Algorithms is pointless
>  > and useless if you dont' provide the entire set of material that
>  > you used for your comparision. That includes:
>  >
>  >       - Source code
>  >       - Input images
>  >       - Full sets of parameters
>  >
>  >
>  > Only in this way, other people will be able to repeat your
>  > evaluations and tweak them in different ways. The fact that
>  > each metric has many parameters makes very difficult (if not
>  > impossible) to define a "fair" comparison. For example, you
>  > are selecting for Viola Wells parameters such as :
>  >
>  >     - Number of Bins
>  >     - Number of Samples
>  >     - Standard Deviations
>  >
>  > Changes in any of those parameters will result in dramatic
>  > changes on the outcome of the Metric, and therefore will chage
>  > how this metric perform face to other metrics.
>  >
>  >
>  > Conclusions of the sort:
>  >
>  >               "Metric A is better than Metric B"
>  >
>  > are useless and worst of all: misleading.
>  >
>  >
>  > They are only of interest for writing papers in the Dark Side
>  > of the current publishing system where reproducibility is not
>  > supported or even welcomed, and where conclusions are not derived
>  > from ones' own experience but from subjective judgement, such as
>  > the ones provided by the decadent peer-review system.
>  >
>  >
>  > Unfortunately, those practices still percolate the entire community
>  > of medical image processing.
>  >
>  >
>  >
>  >    Regards,
>  >
>  >
>  >
>  >
>  >        Luis
>  >
>  >
>  >
>  >
>  > -----------------------------
>  > Jef Vandemeulebroucke wrote:
>  >
>  >> Hi,
>  >>  I am testing several mutual information metrics, plotting their
>  >> behavior.
>  >> Among the metrics are the two based on histograms:
>  >>  MutualInformationHistogramImageToImageMetric
>  >> NormalizedMutualInformationHistogramImageToImageMetric
>  >>  Do these metrics give better results when the images have been
>  >> normalised, as it is for the Viola-Wells implementation of MI, or is 
> this
>  >> of no importance?
>  >>  Thank you,
>  >>  Jef
>  >>
>  >>
>  >> ------------------------------------------------------------------------
>  >>
>  >> _______________________________________________
>  >> Insight-users mailing list
>  >> Insight-users at itk.org
>  >> http://www.itk.org/mailman/listinfo/insight-users
>  >
>  >
>  >
>  >
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> 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