[Insight-users] Tune Parameters
valli gummadi
valli gummadi" <valli_gummadi at rediffmail.com
28 Feb 2004 11:38:40 -0000
This is a multipart mime message
--Next_1077968320---0-202.54.124.178-3540
Content-type: text/html;
charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
<P>=0A<BR>=0A Hi All,<BR>=0A<BR>=0A Our objective is to=
develop a registration tool which would work on single / muliti modality i=
mages.<BR>=0AWe used <BR>=0AMattesMutualInformationImageToImageMetric<BR>=
=0ALinearInterpolateImageFunction<BR>=0AMultiResolutionImageRegistrationMet=
hod<BR>=0ARecursiveMultiResolutionPyramidImageFilter<BR>=0ACenteredAffineTr=
ansform<BR>=0ARegularStepGradientDescentOptimizer<BR>=0Aclasses in achievin=
g the above tasks.<BR>=0A <BR>=0AWe had to fine tune the parameters such as=
TranslateScale,StepLength,Iterations,Levels etc. for the classes. Th=
e parameters had to be modified / fine tuned for each and every dataset. I =
worked with SPECT SPECT volumes, MR SPECT, and CT PET volumes. We could fin=
e tune the parameters by trial and error method to improve the registration=
of these datasets. <BR>=0A <BR>=0AThe problem is that if I need to fine tu=
ne these parameters for every dataset by trial and error method it is real =
tough. Instead We would like to know more about these parameters so that th=
e fine tuning work becomes simpler for us.<BR>=0AI would like to know the r=
elation between images and the parameters. <BR>=0AFor example, I know the r=
otation included in the moving image. If i know the parameter which is rela=
ted with rotation then it will be easy for me to make the application to wo=
rk for various datasets..<BR>=0A<BR>=0A<BR>=0A<BR>=0AThankig you,<BR>=0AReg=
ards,<BR>=0ASrivalli.<BR>=0A=0A</P>=0A<br><br>=0A<A target=3D"_blank" HREF=
=3D"http://clients.rediff.com/signature/track_sig.asp"><IMG SRC=3D"http://a=
ds.rediff.com/RealMedia/ads/adstream_nx.cgi/www.rediffmail.com/inbox.htm at Bo=
ttom" BORDER=3D0 VSPACE=3D0 HSPACE=3D0 HEIGHT=3D74 WIDTH=3D496></a>=0A
--Next_1077968320---0-202.54.124.178-3540
Content-type: text/plain;
charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
=0A Hi All,=0A=0A Our objective is to develop a registration tool whic=
h would work on single / muliti modality images.=0AWe used =0AMattesMutualI=
nformationImageToImageMetric=0ALinearInterpolateImageFunction=0AMultiResolu=
tionImageRegistrationMethod=0ARecursiveMultiResolutionPyramidImageFilter=0A=
CenteredAffineTransform=0ARegularStepGradientDescentOptimizer=0Aclasses in =
achieving the above tasks.=0A =0AWe had to fine tune the parameters such as=
TranslateScale,StepLength,Iterations,Levels etc. for the classes. The par=
ameters had to be modified / fine tuned for each and every dataset. I worke=
d with SPECT SPECT volumes, MR SPECT, and CT PET volumes. We could fine tun=
e the parameters by trial and error method to improve the registration of t=
hese datasets. =0A =0AThe problem is that if I need to fine tune these para=
meters for every dataset by trial and error method it is real tough. Instea=
d We would like to know more about these parameters so that the fine tuning=
work becomes simpler for us.=0AI would like to know the relation between i=
mages and the parameters. =0AFor example, I know the rotation included in t=
he moving image. If i know the parameter which is related with rotation the=
n it will be easy for me to make the application to work for various datase=
ts..=0A=0A=0A=0AThankig you,=0ARegards,=0ASrivalli.=0A
--Next_1077968320---0-202.54.124.178-3540--