[Mrtrix-discussion] empty ROI
Donald Tournier
jdtournier at gmail.com
Wed Jun 4 12:05:29 PDT 2014
Hi Manuel,
If mrtreshold doesn't work, and the values outside the ROI are exactly
zero, you could also use:
mrcalc roi_CC1.mif 0 -neq new_roi_CC1.mif
Cheers,
Donald
--
Dr J-Donald Tournier (PhD)
Senior Lecturer, Biomedical Engineering
Division of Imaging Sciences & Biomedical Engineering
King's College London
A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St
Thomas' Hospital, London. SE1 7EH
T: +44 (0)20 7188 7118 ext 53613
W:
http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering
On 4 Jun 2014 19:00, "Donald Tournier" <jdtournier at gmail.com> wrote:
> Hi Manuel,
>
> OK, make sense then. The threshold is 0.5, I think. You can use
> mrthreshold to sort this out.
>
> Cheers,
> Donald
>
> --
> Dr J-Donald Tournier (PhD)
>
> Senior Lecturer, Biomedical Engineering
> Division of Imaging Sciences & Biomedical Engineering
> King's College London
>
> A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St
> Thomas' Hospital, London. SE1 7EH
> T: +44 (0)20 7188 7118 ext 53613
> W:
> http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering
>
> On 4 Jun 2014 18:28, "Manuel Blesa Cábez" <mblesac at gmail.com> wrote:
>
>> Hi Donald,
>>
>> I checked it for all the ROIs and the values are not 0 and 1. Are 0
>> outside the roi and between 0 and 1 in the ROI. Do you think this is the
>> motive? How can I make it binary?
>>
>> Regards,
>>
>> Manuel Blesa
>>
>>
>> 2014-06-04 19:13 GMT+02:00 Donald Tournier <jdtournier at gmail.com>:
>>
>>> Hi Manuel,
>>>
>>> OK, I guess the transform wasn't the issue... The only other possibility
>>> I can think of is that the values in the ROI image aren't the expected 0 &
>>> 1. Can you load the ROI image in MRView and check that the intensity values
>>> outside the ROI are zero and those inside the ROI are one?
>>>
>>> Cheers,
>>> Donald
>>>
>>> --
>>> Dr J-Donald Tournier (PhD)
>>>
>>> Senior Lecturer, Biomedical Engineering
>>> Division of Imaging Sciences & Biomedical Engineering
>>> King's College London
>>>
>>> A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St
>>> Thomas' Hospital, London. SE1 7EH
>>> T: +44 (0)20 7188 7118 ext 53613
>>> W:
>>> http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering
>>>
>>> On 4 Jun 2014 18:06, "Manuel Blesa Cábez" <mblesac at gmail.com> wrote:
>>>
>>>> Hi Donald,
>>>>
>>>> This is the mrinfo for the fod:
>>>>
>>>> ************************************************
>>>> Image: "fod.mif"
>>>> ************************************************
>>>> Format: MRtrix
>>>> Dimensions: 120 x 110 x 54 x 45
>>>> Voxel size: 1.4 x 1.4 x 2.8 x 1
>>>> Data type: 32 bit float (little endian)
>>>> Data strides: [ 2 3 4 1 ]
>>>> Intensity scaling: offset = 0, multiplier = 1
>>>> Comments: FSL5.0
>>>> Transform: 1 0 0 2.788e+04
>>>> -0 1 0 -3.487e+04
>>>> -0 0 1 -5.623e+04
>>>> 0 0 0 1
>>>>
>>>>
>>>> The rest of the ROIs, have the same mrinfo that the problematic ROI. I
>>>> check it with both programs, fslview and mrview and with both looks ok.
>>>>
>>>> To calculate this ROIs I did it with FSL, because i calculated on the
>>>> template after do the TBSS, to see differences betwen groups, and after
>>>> this I did the inverse transform to the subject space. Can some of this
>>>> steps affect to the transform of the ROI?
>>>>
>>>> Regards,
>>>>
>>>> Manuel Blesa
>>>>
>>>>
>>>> 2014-06-04 18:54 GMT+02:00 Donald Tournier <jdtournier at gmail.com>:
>>>>
>>>>> Hi Manuel,
>>>>>
>>>>> The translation column of the transform seems way out - about 30
>>>>> metres off isocentre... Does the ROI overlay properly onto your fod.mif
>>>>> image within MRView? Does your fod.mif image have the same translation -
>>>>> what does mrinfo report for that image? I expect whatever you used to
>>>>> generate the ROI has corrupted the transform. When you said the ROI looks
>>>>> OK on visual inspection, was that with fslview or MRView? I think fslview
>>>>> just overlays images voxel-wise, with no regard for any differences in the
>>>>> transforms. Basically, if the two images don't overlap in scanner
>>>>> coordinates, then that would explain your issue...
>>>>>
>>>>> Cheers,
>>>>> Donald
>>>>>
>>>>> --
>>>>> Dr J-Donald Tournier (PhD)
>>>>>
>>>>> Senior Lecturer, Biomedical Engineering
>>>>> Division of Imaging Sciences & Biomedical Engineering
>>>>> King's College London
>>>>>
>>>>> A: Department of Perinatal Imaging & Health, 1st Floor South Wing, St
>>>>> Thomas' Hospital, London. SE1 7EH
>>>>> T: +44 (0)20 7188 7118 ext 53613
>>>>> W:
>>>>> http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering
>>>>>
>>>>> On 4 Jun 2014 17:39, "Manuel Blesa Cábez" <mblesac at gmail.com> wrote:
>>>>>
>>>>>> Hi Donald,
>>>>>>
>>>>>> This is the 'mrinfo roi_CC1.mif'
>>>>>>
>>>>>> ************************************************
>>>>>> Image: "roi_CC1.mif"
>>>>>> ************************************************
>>>>>> Format: MRtrix
>>>>>> Dimensions: 120 x 110 x 54
>>>>>> Voxel size: 1.4 x 1.4 x 2.8
>>>>>> Data type: 32 bit float (little endian)
>>>>>> Data strides: [ -1 2 3 ]
>>>>>> Intensity scaling: offset = 0, multiplier = 1
>>>>>> Comments: FSL5.0
>>>>>> Transform: 1 0 0 2.788e+04
>>>>>> -0 1 0 -3.487e+04
>>>>>> -0 0 1 -5.623e+04
>>>>>> 0 0 0 1
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> Manuel Blesa
>>>>>>
>>>>>>
>>>>>> 2014-06-04 18:33 GMT+02:00 Donald Tournier <jdtournier at gmail.com>:
>>>>>>
>>>>>>> Hi Manuel,
>>>>>>>
>>>>>>> Can you post the output of 'mrinfo roi_CC1.mif' ? Would be good to
>>>>>>> also have some info as to how the image was generated. Hopefully that'll
>>>>>>> help to narrow down the problem...
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Donald
>>>>>>>
>>>>>>> --
>>>>>>> Dr J-Donald Tournier (PhD)
>>>>>>>
>>>>>>> Senior Lecturer, Biomedical Engineering
>>>>>>> Division of Imaging Sciences & Biomedical Engineering
>>>>>>> King's College London
>>>>>>>
>>>>>>> A: Department of Perinatal Imaging & Health, 1st Floor South Wing,
>>>>>>> St Thomas' Hospital, London. SE1 7EH
>>>>>>> T: +44 (0)20 7188 7118 ext 53613
>>>>>>> W:
>>>>>>> http://www.kcl.ac.uk/medicine/research/divisions/imaging/departments/biomedengineering
>>>>>>>
>>>>>>> On 4 Jun 2014 17:01, "Manuel Blesa Cábez" <mblesac at gmail.com> wrote:
>>>>>>>
>>>>>>>> Hi all,
>>>>>>>>
>>>>>>>> I have a strange problem, I'm sure is easy to solve but I'm stuck
>>>>>>>> on this and I don't find the solution.
>>>>>>>>
>>>>>>>> I want to calculate the tracks for a ROI, and when I run it I
>>>>>>>> obtain the following message:
>>>>>>>>
>>>>>>>> tckgen fod.mif tracks_CC1.tck -algorithm SD_STREAM -grad
>>>>>>>> ag140128a_6_HARDI_WT_ref.b -seed_image roi_CC1.mif -mask mask.mif -number
>>>>>>>> 1000
>>>>>>>>
>>>>>>>> tckgen [ERROR]: Cannot use image roi_CC1.mif as ROI - image is empty
>>>>>>>>
>>>>>>>> I opened the ROI and it looks ok, and is not "outside" of the fod
>>>>>>>> map. I did the same, for another ROI and it works correctly. I don't know
>>>>>>>> way this ROI has this problem, Somebody can help me? Maybe is due to the
>>>>>>>> location of the ROI and the default parameters? Thanks in advance.
>>>>>>>>
>>>>>>>> Best regards,
>>>>>>>>
>>>>>>>> Manuel Blesa
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Mrtrix-discussion mailing list
>>>>>>>> Mrtrix-discussion at www.nitrc.org
>>>>>>>> http://www.nitrc.org/mailman/listinfo/mrtrix-discussion
>>>>>>>>
>>>>>>>>
>>>>>>
>>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.nitrc.org/pipermail/mrtrix-discussion/attachments/20140604/df06e96d/attachment.html>
More information about the Mrtrix-discussion
mailing list