[Insight-users] MR Object Extraction and Registering with SPECT.

[Lydia Ng]

If you are registering the brain of the same person, have you tried =
using a
rigid transform instead of affine. Optimizing for rigid parameters has =
less
degree of freedom so it should be a bit more robust. Even if you do need
affine, you can still try starting with rigid and use the results to
initialize the affine registration.

- Lydia

> -----Original Message-----
> From: Luis Ibanez [mailto:luis.ibanez at kitware.com]
> Sent: Friday, March 19, 2004 5:32 AM
> To: valli gummadi
> Cc: insight-users at itk.org
> Subject: Re: [Insight-users] MR Object Extraction and Registering with
> SPECT.
>=20
>=20
> Hi Srivalli,
>=20
> Did you verified if the coarse levels of the
> pyramid are converging ?
>=20
> If you find that they are using all the alloted
> iterations, that means that they have not
> achieved conversion.
>=20
> Please tune one level of the pyramid at a time,
> starting of course with the coarse levels.
>=20
> BTW: The translation scale value that you are
> using seems to be extremely large for a Brain MR
> image. I would assume that your image has an extent
> of about 300 millimeters (for a typical human skull),
> your scaling seems to be set for an image of 100.000
> millimeters in linear extent.
>=20
>=20
>    Regards,
>=20
>=20
>=20
>        Luis
>=20
>=20
> -----------------------
> valli gummadi wrote:
>=20
> >  Hi,
> >
> >       I am working on Registration of Single and Multi modality =
images.
> I am facing problem in registering multi modality images.
> >
> > Let me explain the procedure and steps I followed to process multi
> modlaity images.
> >
> > Registration of MR and SPECT images:
> >
> > 1.MRimage will be in CORONAL view and SPECT will be in Transverse =
view.
> > I converted the SPECT to CORONAL before giving it for registration.
> >
> > 2.Actual requirement is to extract the Brain object from the MR and =
to
> register that object with the spect. I extracted Object from the MR =
using
> itkNeighbourhoodConnectionImageFilter and used the exctracted brain as
> Fixed image and Registered MROBject with SPECT.
> >
> > 3.The registered spect is transformed and fuse it with the original =
MR
> so that spect fuses with the brain and not with the other tissues.
> >
> > I am  using following itk classes for registration and transform
> > proces:
> > 1. MultiResolutionImageRegistrationMethod
> > 2.MattesMutualInformationImageToImageMetric
> > 3.LinearInterpolateImageFunction
> > 4.RecursiveMultiResolutionPyramidImageFilter
> > 5.CenteredAffineTransform & centered Affine Transform Initializer
> > 6.RegularStepGradientDescentOptimizer
> > 7.ResampleImageFilter
> >
> > no of iterations =3D450;
> > TranslateScale =3D1000000;
> > optimizer Scales =3D1;
> > Optimizer maximum step length =3D16;
> > Optimizer minimum step length =3D2.5;
> > SpatialSamples =3D1000;
> > Bins =3D50;
> > No of Levels =3D3;
> > SetMinimize =3DTRUE;
> >
> >
> > After fusion, I observed that SPECT is not exactly alighned with =
brain
> object of the MR Volume.
> > Please,Anybody help me to implement the registration to acieve the
> expected output.
> >
> > Thanking You,
> > Regards,
> > Srivalli.
> >
> >
> >
> >
> >
>=20
>=20
>=20
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users