[Vv] [Elastix] Registering two MRI volumes

Pelt, R.F.P. van R.F.P.v.Pelt at tue.nl
Wed Nov 7 10:48:06 CET 2012


Thanks a lot for looking into this so quickly.

I should have realized this, but thanks for sorting this out.
The punishment for taking the intuitive route with Euler angles huh.
Luckily no space stations harmed this time ;-)

Best,

-Roy

-----Original Message-----
From: simon.rit at gmail.com [mailto:simon.rit at gmail.com] On Behalf Of Simon Rit
Sent: dinsdag 6 november 2012 15:54
To: Pelt, R.F.P. van
Cc: vv
Subject: Re: [Elastix] Registering two MRI volumes

Thanks for the data (cc vv mailing list but I have removed your data link).

Congratulations! You have hit the gimbal lock:
http://en.wikipedia.org/wiki/Gimbal_lock. I hope we did not jeopardize a spatial station this time, as with Apollon 11.

Our manual registration tool is not robust to that... The incriminated lines are here:
https://git.creatis.insa-lyon.fr/git/?p=clitk.git;a=blob;f=vv/vvToolRigidReg.cxx;h=28591ba05ed1b6c6cc8af79af440ef5eaea6eb37;hb=f9a4bbf9b3b1c2d8b2ae6b7fe568808a41f48c1e
after line 367.
I have added a bug report:
http://kingkong.grid.creatis.insa-lyon.fr:9002/issues/1741

In the mean time, you can:
- swap the images you're registering:
vv fixed_ssfp.mhd --overlay moving_ssfp.mhd
- change in the code the convention from ZXY to ZYX. That may help.

Good luck!
Simon


On Tue, Nov 6, 2012 at 3:16 PM, Pelt, R.F.P. van <R.F.P.v.Pelt at tue.nl> wrote:
> Dear Simon,
>
> Thank you for your prompt reply!
>
> I prefer to send the data to you by e-mail.
> This data should -not- be distributed elsewhere, so I'd rather not 
> post it on the bug tracker.
> I trust you will treat the data confidentially.
>
> If you prefer to have a bug report in the system, I can of course add 
> an item without the data.
>
> I've also attached a screenshot with the situation.
> As you can see, the sliders have an initial offset, while the angles are 0.
> When dragging the slider, the initial placement 'flips' and the manual registration is impossible.
>
> I have no clue what is going wrong here.
> Perhaps something with the direction cosines / patient orientation?
>
> Looking forward to hear from you, and thanks again for looking into this.
>
> Best,
>
> -Roy
>
>
> -----Original Message-----
> From: simon.rit at gmail.com [mailto:simon.rit at gmail.com] On Behalf Of 
> Simon Rit
> Sent: dinsdag 6 november 2012 14:41
> To: Pelt, R.F.P. van
> Cc: vv
> Subject: Re: [Elastix] Registering two MRI volumes
>
> Hoi Roy,
> Thanks for the report. Would you be able to share your data so that we 
> can have a look? You can either attach them to a bug report
> (http://kingkong.grid.creatis.insa-lyon.fr:9002/projects/vv/issues)
> that can be anonymous or your can send them by email to me.
> Thanks !
> Simon
>
>> ---------- Forwarded message ----------
>> From: Pelt, R.F.P. van <R.F.P.v.Pelt at tue.nl>
>> Date: 6 November 2012 13:09
>> Subject: [Elastix] Registering two MRI volumes
>> To: "elastix at bigr.nl" <elastix at bigr.nl>
>>
>>
>> Dear all,
>>
>>
>>
>> I recently started to look into Elastix.
>>
>> After experimenting with the examples, I'm trying to register two 
>> MRI-acquired volumes (SSFP).
>>
>>
>>
>> In contrast to the example data, these data have their direction 
>> cosines set, and a different origin set in world coordinates.
>>
>> I can use vv to overlay the two data sets, but somehow the sliders 
>> are set at an offset, while the values are zero?
>>
>> As soon as I drag the slider, the bit of initial alignment flicks 
>> away, and there is no means to do a manual registration.
>>
>> Is this some issue with vv?
>>
>>
>>
>> Also my first attempt to register these two volumes did not succeed, 
>> not even with the most basis TranslateTransform setting.
>>
>> I've played around with the scales, but all to no avail.
>>
>>
>>
>> Could you provide me some hint on what I might be doing wrong?
>>
>>
>>
>> Thanks.
>>
>>
>>
>> -Roy
>>
>>
>>
>> Fixed:
>>
>> HeaderSize = -1
>>
>> ObjectType = Image
>>
>> NDims = 3
>>
>> BinaryData = True
>>
>> BinaryDataByteOrderMSB = False
>>
>> CompressedData = False
>>
>> TransformMatrix = 0.0 1.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0
>>
>> Offset = 12.543 -16.259 19.749
>>
>> AnatomicalOrientation = RAI
>>
>> ElementSpacing = 1.118 1.118 0.9
>>
>> DimSize = 224 224 130
>>
>> ElementType = MET_USHORT
>>
>> ElementDataFile = fixed_ssfp.vtk
>>
>>
>>
>> Moving:
>>
>> HeaderSize = -1
>>
>> ObjectType = Image
>>
>> NDims = 3
>>
>> BinaryData = True
>>
>> BinaryDataByteOrderMSB = False
>>
>> CompressedData = False
>>
>> TransformMatrix = 0.0 1.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0
>>
>> Offset = 9.086 -12.779 48.65
>>
>> AnatomicalOrientation = RAI
>>
>> ElementSpacing = 1.09 1.09 0.9
>>
>> DimSize = 224 224 105
>>
>> ElementType = MET_USHORT
>>
>> ElementDataFile = moving_ssfp.vtk
>>
>>
>>
>>
>> _______________________________________________
>> Elastix mailing list
>> Elastix at bigr.nl
>> http://lists.bigr.nl/cgi-bin/mailman/listinfo/elastix
>>
>>



More information about the vv mailing list