[Mristudio-users] crash when saving ROI's

Fri Jul 16 21:12:14 EDT 2010

Hi all,

Thanks SC to follow the question. I just want to add several points;

1) We are looking into the way RoiEditor draws ROI and hope we can improve
the interface.
2) Please note that one of the reason of jagged ROIs in RoiEditor is due to
pixel resolution. The ROIs are drawn based on monitor resolution, which is
very high and you can draw smooth ROIs. However, when the drawn ROIs are
translated into actual ROIs in the images, they have to be converted to
image pixel-dimension. This could be one of the reasons why the final ROIs
do not look as good.
3) If you have two images, such as anatomical T1/T2 and DTI, which are not
co-registered and you want to register them, you can try DiffeoMap. There
are two approaches. For animal studies in which you can assume that the head
position is stable, you can register them theoretically based on scan
parameters such as the locations and sizes of FOV. DiffeoMap has functions
to resize and relocate FOV and matrix. If you can't use this approach for
some reason, you can use the AIR/MUT options.
4) DtiStudio also has image registration functions, but they are for
intra-DTI image registration. If you want cross-modality (e.g. T1 vs DTI)
registration, you need to use DiffeoMap.
5) MUT tends to be more accurate and less sensitive to parameters. However,
it takes much longer time. Typical 30-orientation DTI may take overnight. If
you are using it for cross-modality registration just for 1 image, you
should get results within 30 min or so.

On Thu, Jul 15, 2010 at 8:27 PM, [Seongjin] <choisj70 at gmail.com> wrote:

> Salvador,
>
> Your mouse pointer position is not the center of your ROI. It is top-left
> corner of your ROI. There is 4-arrow buttons (named move ROI) right-lower
> part of the ROI editor control panel. Adjust your ROI with the buttons. ROI
> will move by one pixel unit.
> The drawn ROI will be pixelized. So if you draw a circular ROI it will have
> pixelized boundary. For the small ROI, rectangular shape will make more
> sense. You can define the ROI dimensions (width and hight) with option
> "fixed size".
> If you fixed the ROI to the object, the arrow button will not work.
> Make your that you move the ROI location before pressing the add button.
>
> I am not aware of ROI importing between DTI studio and ImageJ,
>
> Best,
> -SC
>
> On Thu, Jul 15, 2010 at 4:42 AM, Castaneda, Salvador <
> salvador.castaneda at med.uni-tuebingen.de> wrote:
>
>>  SC
>>
>> So that sounds good... i am using AIR though because the MUT doesnt do
>> anything it gets stuck.... So i am using AIR and 90 for noise reduction
>> because i heard that's what you have to use for siemens machines.... also
>> affine. tri linear interpolation and standard derivation. So im just
>> changing the affine and the 90 noise reduction. The images look ok... bu
>> then again i cant compare it to the MUT because i cant use it. SC, dont you
>> find it troublesome that when you draw the ROIS in ROIeditor, the roi
>> selected isnt really what the selection ends up to be? I mean When you are
>> selecting your pixels and then choose to add as an object, the object you
>> make is not exactly what you had chosen. Is there anyway to fix this? i
>> tried reducing the pencil diameter or doing a fill in the roi and then
>> adding but it always looks kinda crappy... it's still better than picking
>> the ROIS from dtistudio though... I also heard that you can edit your rois
>> importing the images from Dtistudio into ImageJ but i dont know how to do
>> this. Maybe i would try it because in there you can import the anatomy
>> sequences to draw your rois on top of them.
>> anyway, any input in any of these inquiries would be helpful!
>>
>> best regards
>>
>> Salvador
>>
>>  ------------------------------
>> *Von:* mristudio-users-bounces at mristudio.org [mailto:
>> mristudio-users-bounces at mristudio.org] *Im Auftrag von *[Seongjin]
>> *Gesendet:* Mittwoch, 14. Juli 2010 14:47
>>
>> *An:* DTI Studio, ROI Editor, DiffeoMap Questions/Support
>> *Betreff:* Re: [Mristudio-users] crash when saving ROI's
>>
>>  Salvador,
>>
>> What I know about difference between AIR and Mutual info based (MUT)
>> registration, AIR is somewhat sensitive to noise level and has more options
>> to be determined. What I heard from developer group about the noise level in
>> AIR is that the noise level should be determined by empirical way. And it
>> has more options for fitting methods. But which one works better, you should
>> decide after a couple of trial and errors.
>> MUT has less option to be determined by user. MUT is based on information
>> entropy maximization as far as I know. It is one of the good registration
>> methods. But I have no idea about which one is better between AIR or MUT. I
>> prefer to use MUT because less user options, which means it is more
>> automatic. At least, I don't need to play with it to find a proper noise
>> level for AIR.
>>
>> Best,
>> -SC
>>
>> On Wed, Jul 14, 2010 at 7:09 AM, Castaneda, Salvador <
>> salvador.castaneda at med.uni-tuebingen.de> wrote:
>>
>>>  SC
>>>
>>> Sorry again, Do you know if there is such a big difference between the
>>> mutual image registration and the automatic. And what exactly does this do
>>> just improve the quality of the image all together or what?
>>> thanks in advanced
>>>
>>> Salvador Castaneda
>>>
>>>  ------------------------------
>>> *Von:* mristudio-users-bounces at mristudio.org [mailto:
>>> mristudio-users-bounces at mristudio.org] *Im Auftrag von *Castaneda,
>>> Salvador
>>> *Gesendet:* Mittwoch, 14. Juli 2010 12:14
>>>
>>> *An:* DTI Studio, ROI Editor,DiffeoMap Questions/Support
>>> *Betreff:* Re: [Mristudio-users] crash when saving ROI's
>>>
>>>   SC
>>>
>>> Thats exactly what i do, i take the image files from dtistudio and load
>>> them in the roi editor. i wanted to upload an anatomical picture to better
>>> draw y rois because i work with rats. is it possible to dothis even if the
>>> matrix and fov are not the same?
>>> thanks
>>>
>>>
>>>  ------------------------------
>>> *Von:* mristudio-users-bounces at mristudio.org [mailto:
>>> mristudio-users-bounces at mristudio.org] *Im Auftrag von *[Seongjin]
>>> *Gesendet:* Mittwoch, 14. Juli 2010 01:45
>>> *An:* DTI Studio, ROI Editor, DiffeoMap Questions/Support
>>> *Betreff:* Re: [Mristudio-users] crash when saving ROI's
>>>
>>>  Darren,
>>>
>>> I experienced the similar problem. I wanted to use ROI drawn in ROI panel
>>> but it did not work at all. ROI saving and loading functions seem to work in
>>> FIBER panel, but not in the ROI panel.
>>>
>>> If you want to use ROI for DTI metrics such as FA or ADC measurements, I
>>> recommend ROIeditor (a companion tool to DTIstudio). That is more powerful.
>>>
>>> I always process data in DTI studio. Then load all the processed data
>>> into ROI editor.
>>> Finally, do the measurement with ROI editor.
>>>
>>> The saved ROIs can be reused in the DTI studio for fiber tracking.
>>> However, again, the ROIs could not be used in ROI panel of the DTI studio on
>>> my computers (home computer, laptop, and office computer).
>>>
>>> I hope this experience might be a bit of help for your problem in using
>>> DTI studio.
>>>
>>> -SC
>>>
>>> On Tue, Jul 13, 2010 at 5:28 PM, Darren Gitelman <
>>> d-gitelman at northwestern.edu> wrote:
>>>
>>>> Hi
>>>>
>>>> I am having problems with DTI studio crashing when trying to save ROI's.
>>>>
>>>>
>>>> This happens after I draw an ROI and click the save button. Then a
>>>> dialog (ROI File Format) comes up giving a choice of different formats. I
>>>> choose Graphic Compatible (default). After clicking OK, a crash window comes
>>>> up a few seconds later. It is labeled MriView3D MFC application. The only
>>>> technical information I can find is the following.
>>>>
>>>> AppName: dtistudio-latest-x86.exe     AppVer: 1.0.0.1     ModName:
>>>> msvcr80.dll
>>>> ModVer: 8.0.50727.4053     Offset: 00014458
>>>>
>>>> This occurs whether I have drawn 1 or more ROIs. (fixed size ROI, oval).
>>>> The size of the ROI does not matter (4 voxel or 10 voxel)
>>>> It occurs with DTI studio x86 and x64.
>>>> It occurs with versions 3.0.2 and 2.4.1 of DTIstudio.
>>>> It occurs whether I have a complete dataset loaded 35 directions, or
>>>> even if I load just 1 image and draw an ROI on it.
>>>> It occurs on an HP core duo 2 system running Windows XP SP3, and on a
>>>> redhat linux KVM virtual machine running windows XP x64.
>>>> Reinstalling vcredist_x86 or vcredist_x64 from the DTIstudio website
>>>> does not help on any of the systems.
>>>>
>>>> Any suggestions on how to fix?
>>>>
>>>> Darren
>>>>
>>>> _______________________________________________
>>>> 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/20100716/fd2dce48/attachment-0001.html 