[Insight-users] Speed of Mutual Information calculation

[Markus Weigert]

Hi Luis,

regarding my last email:

<<<<
Hi Luis,

thanks a lot for your help.
I uploaded a CT - dataset (lungs, Analyze - format) and a MR - dataset.
The CT is the target, the MR is the moving image.

When I try to do automatic rigid registration (translation,
mutual information (0.4 sigma both, 100000 spatial samples,
moving/fixed Image Gaussfilter variance 0.16)
GradientDescent optimizer (200 iters, 200 leraning rate),
it seems to be veeeery slow or doesn't start at all.

With Mattes MI (80 bins, 100000 samples), same transformation and optimizer,
it starts and is reasonably fast.

But the bigger problem is deformable registration afterwards.
Here, I don't get a result by using BSpline (with LBFGSB and MattesMI)
or FEM.

I'm very interested in the FEM - registration filter, but find it very hard
to use.
>>>>

Now, I get better results with BSpline.
I am also interested in getting reasonable results with the FEM - 
registration
filter. Can you perhaps help to find a parameter - setting for this filter, 
because I
tried for long but can't manage to get a good result with it.
It would be very helpful, because I may learn new things about setting these 
parameters
and the filter at all, if somebody shows me a good setting / composition for 
this
specific problem and I may transfer it to other registration tasks.
Till now, I could not solve any problem with FEM - registration.

Thanks again,
Markus



----- Original Message ----- 
From: "Luis Ibanez" <luis.ibanez at kitware.com>
To: "Markus Weigert" <m.weigert at fz-juelich.de>
Sent: Thursday, May 18, 2006 3:45 PM
Subject: Re: [Insight-users] Speed of Mutual Information calculation


>
> Hi Markus,
>
> If you send us your CT and MR images we will
> give it a try to register them and will let you
> know what settings of Metric and Optimizer are
> appropriate for them.
>
> As we mention in a previous email. This size of
> image should be registered in about 2 minutes
> in a standard modern computer.
>
> You can upload your images at:
>
> http://www.kitware.com/KitwareScripts/uploadfile.cgi
>
>
> Please remove any patient and doctor information
> from the images before you upload them.
>
>
> Thanks
>
>
>    Luis
>
>
> =======================
> Markus Weigert wrote:
>> Hi Luis,
>>
>> thanks for your response.
>> I use the Viola - Wells implementation.
>> Strangely, the Mattes implementation is much faster.
>> I compiled for release with dbg. information on VC6
>> and used GradientDescentOptimizer, not 
>> RegularStepGradientDescentOptimizer.
>> I plot the progress from a Command Observer and currently don't use 
>> multiresolution
>> (only on the original resolution).
>>
>> Cheers,
>> Markus
>>
>>
>> ----- Original Message ----- From: "Luis Ibanez" 
>> <luis.ibanez at kitware.com>
>> To: "Markus Weigert" <m.weigert at fz-juelich.de>
>> Cc: <insight-users at itk.org>
>> Sent: Monday, May 15, 2006 4:40 PM
>> Subject: Re: [Insight-users] Speed of Mutual Information calculation
>>
>>
>>>
>>> Hi Markus,
>>>
>>> Nope, this is not the common time for this size of images.
>>>
>>> This type of registration should take about 2 minutes in
>>> a modern standard computer.
>>>
>>>
>>> Some questions:
>>>
>>>
>>> 1) Are you compiling your application for "Release" ?
>>>
>>> 2) Are you using multi-resolution ?
>>>
>>> 3) Are you using the GradientDescent or
>>>    the RegularGradientDescent optimizer ?
>>>
>>> 4) Are you plotting the progress of the optimizer ?
>>>    from a connected Command Observer ?
>>>
>>> 5) Which one of the 5 ITK implementations of
>>>    Mutual Information Metric  are you using  ?
>>>
>>>
>>>
>>> It is very likely that you are letting the optimizer run
>>> for a lot of uncessary iterations.
>>>
>>> Have you measure the time needed for performing One iteration ?
>>> This will indicate if the problem is to have too many iterations,
>>> or to have metric evaluations that are too slow.
>>>
>>>
>>>
>>> The best way to figure out the problem is to analyze the
>>> trace provided by the Command Observer.
>>>
>>> Given that you are testing with a 3D translation transform,
>>> you are in the lucky situation were you can actually plot
>>> the path of the optimizer in the parametric space.
>>>
>>> You could use a tool such as GNUplot, in order to see this
>>> path in 3D.  Other easy options are a VTK script, or saving
>>> the trace in a .vtk file and loading it into ParaView.
>>>
>>> Whi
>>>
>>>
>>>
>>>
>>> =====================
>>> Markus Weigert wrote:
>>>
>>>> Dear insight users,
>>>>  I currently try to register two 3D images (CT and MR)
>>>> by using mutual information as metric.
>>>> The images have a size of approx. 255 * 290 * 75 slices each (MR 
>>>> perhaps even more).
>>>> Although I use a very simple transformation (translation) and a 
>>>> graddescent
>>>> optimizer, one iteration of the optimizer takes more than 1.5h.
>>>>  Is this a common time for images of this size???
>>>> The metric uses 60000 spatial samples.
>>>> I thougt about using BSpline transform in the next step of the 
>>>> registration
>>>> with MI metric too, but I think I can forget to do this, if I have to 
>>>> deal with 3000 Parameters
>>>> to be optimized!
>>>>  Regards,
>>>> Markus
>>>>  ------------------------------------------------------------------------
>>>>
>>>>
>>>> _______________________________________________
>>>> Insight-users mailing list
>>>> Insight-users at itk.org
>>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>>
>>>
>>
>>
>>
>
>