[Mristudio-users] Adjusting Gradients for Oblique Angles
susumu mori
susumu at mri.jhu.edu
Tue Dec 14 17:01:20 EST 2010
Thanks Jeff,
Your observation about VB17 (and other versions) agree with mine. There are
mixed information about the gradient rotation out there; some say it's
rotated and some say not. What is more confusing is that there are two level
of "rotation" questions;
1) Is it rotated at the point of imaging?
2) Is it rotated in the DICOM file?
I have heard mixed information from multiple sources, but so far my
experience says the gradient tables are not rotated at any levels, as you
confirmed. This is interesting.
The small difference you found between the two files (e.g. 0.863, -0.364,
0.352 vs 0.861, -0.366, 0.354) for with and without oblique is NOT caused by
the oblique rotation (the effect of 45 degree rotation can't be this small)
but caused by the different interaction with the small imaging gradients, I
assume.
On Tue, Dec 14, 2010 at 4:00 PM, Jeff Sadino <jsadino.queens at gmail.com>wrote:
> Hi Susumu,
>
> When I was doing my testing during the last couple weeks, I was using the
> VB17 software. I tried large rotations (30-45 degrees) on the sagittal
> orientation, and my gradient, as extracted using DTIStudio 3.0.2, was not
> rotated:
>
> Using no oblique angles:
> 0: 0.000, 0.000, 0.000
> 1: 0.863, -0.364, 0.352
> 2: 0.863, 0.364, 0.352
> 3: 0.863, 0.364, -0.352
> 4: 0.863, -0.364, -0.352
> 5: 0.357, -0.364, -0.860
> 6: 0.357, -0.866, -0.350
> 7: 0.357, -0.866, 0.350
> 8: 0.357, -0.364, 0.860
> 9: 0.357, 0.364, 0.860
> 10: 0.357, 0.866, 0.350
> 11: 0.357, 0.866, -0.3509
> 12: 0.357, 0.364, -0.860
> ImageOrientationPatient: [ 1 0 0 0 1 0 ]
>
> Using an oblique angle:
> 0: 0.000, 0.000, 0.000
> 1: 0.861, -0.366, 0.354
> 2: 0.864, 0.361, 0.350
> 3: 0.861, 0.366, -0.354
> 4: 0.864, -0.361, -0.350
> 5: 0.358, -0.352, -0.865
> 6: 0.357, -0.866, -0.350
> 7: 0.355, -0.861, 0.365
> 8: 0.357, -0.365, 0.860
> 9: 0.358, 0.352, 0.865
> 10: 0.357, 0.866, 0.350
> 11: 0.356, 0.861, -0.364
> 12: 0.357, 0.365, -0.860
> ImageOrientationPatient: [ 1.0000 0 0 0 0.8846
> -0.4664 ]
>
> However, when I tick on the "rotate gradients if applicable" button, then
> the oblique angle is properly taken into account. The case was the same
> when I used oblique angles on the axial orientation.
>
> Although I can not do these same tests for our older software (VB13 and
> VA25A), just by looking at the extracted gradients, I can tell that they are
> not automatically rotated.
>
> I hope that helps and thank you for all of your support!
> Jeff
>
>
> On Tue, Dec 14, 2010 at 10:31 AM, susumu mori <susumu at mri.jhu.edu> wrote:
>
>> Hi Jeff,
>>
>> Thank you for the follow up. Are you using VB17 and did you still need the
>> rotation?
>>
>> Susumu
>>
>>
>> On Tue, Dec 14, 2010 at 3:20 PM, Jeff Sadino <jsadino.queens at gmail.com>wrote:
>>
>>> Thank you everybody! That did the trick :)
>>>
>>> As a heads up to other people, I have heard that 3T Siemens VB17 Trio Tim
>>> is supposed to rotate the gradients automatically. I tried rotating
>>> the imaging plane by 45 degrees in both the sagittal and then the axial
>>> views to test this. On our system, the gradients were not rotated
>>> automatically. The assistance of DTIStudio is still needed in order to
>>> correct for rotated gradients in this scenario. The case is the same using
>>> the VB13 software.
>>>
>>> Mahalo again!
>>> Jeff Sadino
>>>
>>> On Tue, Dec 14, 2010 at 8:29 AM, Meredith Reid <mer.reid at gmail.com>wrote:
>>>
>>>> I think the last row of your rotation matrix should be the cross product
>>>> of the two rows from ImageOrientationPatient, resulting in:
>>>>
>>>> 1.0000 0 0
>>>> 0 0.8846 -0.4664
>>>> 0 0.4664 0.8846
>>>>
>>>> Meredith
>>>>
>>>>
>>>> On Mon, Dec 13, 2010 at 5:03 PM, Jeff Sadino <jsadino.queens at gmail.com>wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> Thank you Venkat and Hangyi for you insights. I tried both methods,
>>>>> but I was not able to replicate the values that DTIStudio gave me. The
>>>>> values that DTIStudio gave me, after clicking the "rotate gradients if
>>>>> applicable" button and for the first direction, is:
>>>>> 0.861, -0.489, 0.142
>>>>>
>>>>> Here is my original gradient table:
>>>>> 0 0 0
>>>>> 0.8608 -0.3657 0.3538
>>>>> 0.8643 0.3611 0.3502
>>>>> 0.8608 0.3657 -0.3538
>>>>> 0.8644 -0.3608 -0.3502
>>>>> 0.3582 -0.3523 -0.8646
>>>>> 0.3571 -0.8658 -0.3504
>>>>> 0.3554 -0.8606 0.3648
>>>>> 0.3566 -0.3652 0.8599
>>>>> 0.3579 0.3521 0.8648
>>>>> 0.3571 0.8658 0.3504
>>>>> 0.3557 0.8609 -0.3638
>>>>> 0.3569 0.3655 -0.8597
>>>>>
>>>>> And here is my ImageOrientationPatient info:
>>>>> 1.0000 0 0
>>>>> 0 0.8846 -0.4664
>>>>>
>>>>> If I do it according to Venkat, it becomes:
>>>>> 1.0000 0 0
>>>>> 0 0.8846 -0.4664
>>>>> 0 0 1.0000
>>>>>
>>>>> If I multiply my gradient by this, for the first direction, I get:
>>>>> 0.8608 -0.3235 0.5244
>>>>>
>>>>> If I do it according to Hangyi, the rotation matrix becomes:
>>>>> 1.0000 0 0
>>>>> 0 0.8846 -0.4664
>>>>> 0 0.8846 -0.4664
>>>>>
>>>>> If I multiply my gradient by this, for the first direction, I get:
>>>>> 0.8608 -0.0106 0.0056
>>>>>
>>>>> The only thing that I can think of is there is probably a step in there
>>>>> that normalizes the gradient to one. i.e.: sqrt(grad_x^2 + grad_y^2 +
>>>>> grad_z^2) = 1.
>>>>>
>>>>> Can anybody help point me to the next step? Thank you very much!!
>>>>> Jeff Sadino
>>>>>
>>>>> On Fri, Dec 10, 2010 at 4:07 AM, Hangyi Jiang <hjiang at jhmi.edu> wrote:
>>>>>
>>>>>> hello,
>>>>>>
>>>>>> the 3rd row should be the outer product of the given two vectors,
>>>>>> not [0,0,1]
>>>>>>
>>>>>> regards
>>>>>>
>>>>>> hangyi
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ________________________________
>>>>>> From: mristudio-users-bounces at mristudio.org [
>>>>>> mristudio-users-bounces at mristudio.org] On Behalf Of Rajagopalan,
>>>>>> Venkateswaran [rajagov2 at ccf.org]
>>>>>> Sent: Friday, December 10, 2010 9:01 AM
>>>>>> To: DTI Studio, ROI Editor, DiffeoMap Questions/Support
>>>>>> Subject: Re: [Mristudio-users] Adjusting Gradients for Oblique Angles
>>>>>>
>>>>>> Dear Jeff Sadino,
>>>>>>
>>>>>> This is what I did for manual oblique angle correction
>>>>>>
>>>>>> multiply your gradient matrix (I don't know how many directions you
>>>>>> have) if you have 12 directional data then you will have 3*13 (one b=0
>>>>>> values too) with the "Image Orientation patient" matrix
>>>>>>
>>>>>> [ 1 0 0 0 .8846 -.4664] rewritten as (these are unit vectors with z
>>>>>> direction being 0 0 1 not mentioned in the Image Image orientation patient
>>>>>> from Dicom header
>>>>>>
>>>>>> [ 1 0 0
>>>>>> 0 0.8846 -0.4664
>>>>>> 0 0 1]
>>>>>>
>>>>>> 3*3
>>>>>>
>>>>>> Multiply the ablove Image orientation patient with your gradient
>>>>>> matrix.
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> Venkat
>>>>>>
>>>>>> ________________________________
>>>>>> From: mristudio-users-bounces at mristudio.org on behalf of Jeff Sadino
>>>>>> Sent: Thu 12/9/2010 6:09 PM
>>>>>> To: DTI Studio, ROI Editor, DiffeoMap Questions/Support
>>>>>> Subject: [Mristudio-users] Adjusting Gradients for Oblique Angles
>>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> Thank you for the DTI training session in Turtle Bay last week. It
>>>>>> was very helpful for us!
>>>>>>
>>>>>> I am wondering how to manually adjust the gradients for an oblique
>>>>>> angle. I am using Siemens 3T Trio Tim VB17.
>>>>>>
>>>>>> My default gradient (for one direction) is:
>>>>>> 0.861, -0.366, 0.354
>>>>>>
>>>>>> After clicking "Rotate gradients if applicable" in DTIStudio 3.0.2
>>>>>> (March2010), the gradient is:
>>>>>> 0.861, -0.489, 0.142
>>>>>>
>>>>>> I couldn't find exactly which tag in the DICOM header contains the
>>>>>> rotation information. The closest I could find is
>>>>>> "ImageOrientationPatient". The value for this is:
>>>>>> [ 1 0 0 0 .8846 -.4664]
>>>>>>
>>>>>> How do I go from the default gradient to the rotated gradient? Also,
>>>>>> can I use this same method on the VB13 and VA25 software?
>>>>>>
>>>>>> Thank you very much for your help!
>>>>>> Jeff Sadino
>>>>>>
>>>>>> _______________________________________________
>>>>>> mristudio-users mailing list
>>>>>> mristudio-users at mristudio.org
>>>>>> http://lists.mristudio.org/mailman/listinfo/
>>>>>> Unsubscribe, send a blank email to:
>>>>>> mristudio-users-unsubscribe at mristudio.org
>>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> mristudio-users mailing list
>>>>> mristudio-users at mristudio.org
>>>>> http://lists.mristudio.org/mailman/listinfo/
>>>>> Unsubscribe, send a blank email to:
>>>>> mristudio-users-unsubscribe at mristudio.org
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> mristudio-users mailing list
>>>> mristudio-users at mristudio.org
>>>> http://lists.mristudio.org/mailman/listinfo/
>>>> Unsubscribe, send a blank email to:
>>>> mristudio-users-unsubscribe at mristudio.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> mristudio-users mailing list
>>> mristudio-users at mristudio.org
>>> http://lists.mristudio.org/mailman/listinfo/
>>> Unsubscribe, send a blank email to:
>>> mristudio-users-unsubscribe at mristudio.org
>>>
>>>
>>
>> _______________________________________________
>> mristudio-users mailing list
>> mristudio-users at mristudio.org
>> http://lists.mristudio.org/mailman/listinfo/
>> Unsubscribe, send a blank email to:
>> mristudio-users-unsubscribe at mristudio.org
>>
>>
>
> _______________________________________________
> mristudio-users mailing list
> mristudio-users at mristudio.org
> http://lists.mristudio.org/mailman/listinfo/
> Unsubscribe, send a blank email to:
> mristudio-users-unsubscribe at mristudio.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mristudio.org/pipermail/mristudio-users/attachments/20101214/fd8e2e8d/attachment-0001.html
More information about the mristudio-users
mailing list