[Mristudio-users] about effect of masking

[susumu]

There are three ways for skull stripping in RoiEditor.

 

1) Simple thresholding: This produce the same results as the DtiStudio's
masking function.

2) Region growing: if you use this function together with b0 or average(all
DWIs), you should be able to achieve very decent skull stripping. We use
this routinely. We use average(DWIs) often.

3) Skull-stripping button: The region growing may leak out to some tissues
outside the brain if high-intensity areas of the brain and outside tissues
are connected. This skull-stripping function poses some restrictions in the
shape of the object. Namely, the brain surface should be smooth without
anything budding out in an extreme manner. If the region growing function
consistently leaks out and defines tissues outside the brain, you can try
this function. This function also allows you to load "second image", such as
an anatomical T2. The brain contour defined by b0 or average(DWIs) will be
superimposed on the second image (please note, the brain contour will not
align perfectly because of subject motion and DTI (EPI)-related brain
distortion). This b0-defined contour then will be used as an initial guess
of the brain contour of the second image and do the skull strip. We use this
function for skull-strip of co-registered T2 (and T1). It works decently,
but not perfect (I've never seen any perfect automated skull-strip that
works for all images anyway).

 

susumu

  _____  

From: mristudio-users-bounces at mristudio.org
[mailto:mristudio-users-bounces at mristudio.org] On Behalf Of Jim Li
Sent: Monday, December 08, 2008 1:53 PM
To: DTI Studio, ROI Editor, Landmarker Questions/Support
Subject: Re: [Mristudio-users] about effect of masking

 

Hi Susumu,

 

Thanks a lot for your information. I tried the skull-stripping of ROIEditor
using the default setting. The result is not bad, but I hope to add/remove
some areas manually on top of the outcoming mask file to define a better
mask. Also I hope to better understand the default parameters in the
skull-stripping algorithm so I can tailor them to my need. I wish there's an
instruction manual for ROIEditor that can tell me all of these...

 

Cheers,

 

Jim

 

-----------------------------------------

Hi Jim,
 
When you do tensor calculation, you want to mask out the areas outside the
brain. The easiest way is to use the threshold value to mask out the noise
floor. This approach however leaves a lot of extra-brain tissues. RoiEditor
(and many other programs) offers much more tools to define the brain as a
binary map and remove other tissues using tools such as region growing and
skull-strip tools. If you have any binary masking generated by another
program you can load such a file through this function. 
 
Susumu
 
------Original Message------
From: Jim Li
Sender: mristudio-users-bounces at mristudio.org
<http://lists.mristudio.org/mailman/listinfo/mristudio-users> 
To: DTI Studio, ROI Editor, Landmarker Questions/Support
ReplyTo: DTI Studio, ROI Editor,      Landmarker Questions/Support
Subject: [Mristudio-users] about effect of masking
Sent: Dec 5, 2008 17:57
 
Hi all,   Can anybody give me a hint?   In the  DTIstudio software I clicked
the "Tensor, Color map, etc" button in the "Calculation" section of the
"DtiMap" tab. Then, from the pop-out window called "DTI-Map Options", I saw
a choice called "Select a mask-image". Under what situation should we use
that choice? Should it be a default choice when the brain has some lesions,
given that the lesional area can be at about the same level as the
background noise level and thus can be easily wiped out even if we choose a
low background noise level from which to create a mask-image?     Thanks a
lot,   Jim _______________________________________________ Mristudio-users
mailing list Mristudio-users at mristudio.org
<http://lists.mristudio.org/mailman/listinfo/mristudio-users>
http://lists.mristudio.org/mailman/listinfo/mristudio-users 
Sent via BlackBerry by AT&T
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mristudio.org/pipermail/mristudio-users/attachments/20081209/52677cc3/attachment-0001.html