[Elastix] RAS or LPS coordiante system?

Bilal Tahir b.tahir at sheffield.ac.uk
Fri Apr 19 17:40:21 CEST 2013


Hi Stefan,

I stored the landmarks in world coordinates.
I used the input .txt format and did set the DirectionCosines to "true".

I think that as Slicer uses RAS coordinates as default, its just a matter
of simply multiplying the X and Y coordinates by -1.

For example, if we have the following RAS coordinates in slicer:

5, -4, 10

The alternative to use would be:

-5, 4, 10

Notice how the Z in the sup-inf direction hasn't changed sign.

This seems to have worked and is giving me appropriate TRE results.

Here is the link for how slicer deals with coordinates:

http://www.slicer.org/slicerWiki/index.php/Coordinate_systems

I use slicer for landmarking over VV as it allows me to identify landmarks
which can then be manually moved with the click of a mouse by a
radiologist. However, I am interested to explore some other software for
this. If anyone uses alternative software that they
recommend specifically for landmarking then please let me know.

Best wishes,

Bilal


On 19 April 2013 11:29, S. Klein <s.klein at erasmusmc.nl> wrote:

>  Hi Bilal, elastix takes into account the direction cosines of the images.
> Do you store the landmarks in world coordinates or in voxel coordinates?
> Do you use the .vtk format or the input point .txt format to supply the
> points to transformix?
> In your registration, did you set UseDirectionCosines to "true"?
> Kind regards,
> Stefan
>
>
> On 18/04/2013 02:52, Bilal Tahir wrote:
>
> Dear All,
>
>  I recently moved from VV to Slicer for defining expert landmarks on
> fixed and moving images prior to registration. I then use transformix to
> deform the fixed points so I can compare them with original moving points
> via the target registration error.
>
>  One thing I noticed is that Slicer defines the points (via fudicial
> markers in the annotations module) in the RAS coordinate system as default.
> Putting these values into transformix gives me completely wrong results. Is
> Elastix based on the LPS coordinate system as default or another one? Below
> is part of  my header data for the fixed image:
>
>  ObjectType = Image
> NDims = 3
> BinaryData = True
> BinaryDataByteOrderMSB = False
> CompressedData = False
> TransformMatrix = 1 0 0 0 1 0 0 0 1
> Offset = -215.592 -417.592 -323.5
> CenterOfRotation = 0 0 0
> AnatomicalOrientation = RAI
> ElementSpacing = 0.816406 0.816406 1.78125
> DimSize = 512 512 160
> ElementType = MET_SHORT
> ElementDataFile = fixed.raw
>
>  Here is the header data for the moving image:
>
>  ObjectType = Image
> NDims = 3
> BinaryData = True
> BinaryDataByteOrderMSB = False
> CompressedData = False
> TransformMatrix = 1 0 0 0 1.52656e-015 -1 0 1 5.55112e-017
> Offset = -212.209 -155 200.502
> CenterOfRotation = 0 0 0
> AnatomicalOrientation = RSA
> ElementSpacing = 1.6797 1.6797 10
> DimSize = 256 256 24
> ElementType = MET_SHORT
> ElementDataFile = moving.raw
>
>  I look forward to hearing your replies.
>
>  Cheers,
>
> Bilal Tahir,
> James Morrison Researcher in Radiotherapy Imaging
>
>
> _______________________________________________
> Elastix mailing listElastix at bigr.nlhttp://lists.bigr.nl/cgi-bin/mailman/listinfo/elastix
>
>
> --
> Stefan Klein+31 10 7043049http://www.bigr.nl/people/StefanKlein
>
>


-- 
Bilal Tahir,
James Morrison Researcher in Radiotherapy Imaging
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bigr.nl/pipermail/elastix/attachments/20130419/2bae9cd3/attachment.html>


More information about the Elastix mailing list