AW: Re: [Camino-users] question re interpolated tractography

[Philip A Cook]

Hi,

At each step, we calculate the total displacement since the last curvature check. If that displacement exceeds the slice thickness, then curvature is evaluated. The number of steps that lie between curvature checks will vary if you are using FACT, because each step is placed at a voxel boundary.

If you are using a fixed step size, then the number of steps between curvature checks is fixed.


Phil

On Mar 15, 2012, at 5:15 AM, Fögen, Zacharias wrote:

> Hi all,
> 
> Does this mean that curvature is evaluated on every step (with every step being compared to the step <Voxel Size/Slize Thickness> behind it) ; or only every <Voxel Size/Slize Thickness> steps?
> 
> Greetings,
> 
> Zacharias Fögen
> 
> ________________________________
> Von: Florian Fischer [flofisch at students.uni-mainz.de]
> Gesendet: Mittwoch, 14. März 2012 16:20
> An: Fögen, Zacharias
> Betreff: Fwd: Re: [Camino-users] question re interpolated tractography
> 
> 
> 
> -------- Original Message --------
> Subject:        Re: [Camino-users] question re interpolated tractography
> Date:   Wed, 14 Mar 2012 11:18:45 -0400
> From:   Philip A Cook <cookpa at mail.med.upenn.edu><mailto:cookpa at mail.med.upenn.edu>
> To:     Michael Zeineh <mzeineh at stanford.edu><mailto:mzeineh at stanford.edu>
> CC:     Camino-users at www.nitrc.org<mailto:Camino-users at www.nitrc.org> <camino-users at www.nitrc.org><mailto:camino-users at www.nitrc.org>
> 
> 
> 
> Hi Michael,
> 
> In the algorithms currently implemented, the curvature threshold is related to the slice thickness of the data (the z dimension). So if your slice thickness is 2mm, the curvature is evaluated every 2mm, approximately. If you are tracking with a step size of .2 mm, then your curvature will be tested once every 10 steps. With FACT, the displacement per step can vary according to how much of the voxel you traverse, so it is less precise but it should be approximately the same if you have isotropic voxels.
> 
> 
> Phil
> 
> On Mar 13, 2012, at 1:27 PM, Michael Zeineh wrote:
> 
>> Hi all,
>> 
>> A question re interpolation on with tractography and curvature thresholds. Let us say voxel size is 2mm isotropic.
>> 
>> If interpolation is OFF, then the I assume the curvature is measured voxel to voxel along the course of the track, and if the curvature from one voxel to the next (separated by the step size of 2mm) exceeds the threshold, then the track is stopped.
>> 
>> If interpolation is ON (and the step size is the default of 0.2mm), then the same things happens, but across a much shorter distance (which means that the track is much less likely to curve anyways).
>> 
>> So a net effect of this is that for the same track, one obtained with interpolation on, and the other with interpolation off, the measured curvature will be different? So one should have a more strict threshold with interpolation on?
>> 
>> Or is my thought process totally off, and curvature is measured independently from the step size in some way ...
>> 
>> Thank you for any comments,
>> 
>> Michael Zeineh, M.D.-Ph.D.
>> Assistant Professor, Dept of Radiology
>> 300 Pasteur Drive, Room S-047
>> Stanford, CA  94305-5105
>> t:  650-721-1419
>> c: 650-722-2235
>> f:  650-498-5374
>> mzeineh at stanford.edu<mailto:mzeineh at stanford.edu>
>> 
>> 
>> _______________________________________________
>> Camino-users mailing list
>> Camino-users at www.nitrc.org<mailto:Camino-users at www.nitrc.org>
>> http://www.nitrc.org/mailman/listinfo/camino-users
> 
> _______________________________________________
> Camino-users mailing list
> Camino-users at www.nitrc.org<mailto:Camino-users at www.nitrc.org>
> http://www.nitrc.org/mailman/listinfo/camino-users
> 
> _______________________________________________
> Camino-users mailing list
> Camino-users at www.nitrc.org
> http://www.nitrc.org/mailman/listinfo/camino-users