No subject
Thu May 27 11:52:52 EDT 2010
difference is that the Mattes MI implemented in ITK can be run
in multiple-threads. That is, Mattes takes advantage of multiple
cores, while the ITK implementation of ViolaWells does not.
(this is in part due, to their different ways of picking up the
samples).
Note that these comments are very specific to the ITK
implementation of these metrics, and probably should not
be generalized as a comparison of the algorithm themselves.
--
For full details, you may want to look at:
* Viola, P. and Wells III, W. (1997).
* "Alignment by Maximization of Mutual Information"
* International Journal of Computer Vision, 24(2):137-154
and for Mattes:
* [1] "Nonrigid multimodality image registration"
* D. Mattes, D. R. Haynor, H. Vesselle, T. Lewellen and W. Eubank
* Medical Imaging 2001: Image Processing, 2001, pp. 1609-1620.
* [2] "PET-CT Image Registration in the Chest Using Free-form Deformations"
* D. Mattes, D. R. Haynor, H. Vesselle, T. Lewellen and W. Eubank
* IEEE Transactions in Medical Imaging. Vol.22, No.1,
January 2003. pp.120-128.
* [3] "Optimization of Mutual Information for MultiResolution Image
* Registration"
* P. Thevenaz and M. Unser
* IEEE Transactions in Image Processing, 9(12) December 2000.
Regards,
Luis
---------------------------------------------------------------------------------
On Fri, Oct 29, 2010 at 6:08 PM, Dav <masslawi at gmail.com> wrote:
> Dear Ken,
>
> The most important difference between Mattes mutual information and regular
>
> mutual information metric in ITK is their sampling. While the mutual
> information
>
> metric take samples from all of the transformation space, the Mattes
> mutual
>
> information only samples pixels from number of predefined histogram bins
>
> which extremely decreases the computation time and improves the
> performance.
>
> Best regards,
>
> Masslawi.
>
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Kitware offers ITK Training Courses, for more information visit:
> http://www.kitware.com/products/protraining.html
>
> Please keep messages on-topic and check the ITK FAQ at:
> http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>
>
--000e0ce0f792f9f6950493ea8c73
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
<br>Yes,<br><br>The sampling process is probably the main difference.<br><b=
r>Mattes MI picks all its sample points at the first iteration <br>and reus=
e them for subsequent computation.<br><br>ViolaWells MI picks samples at ev=
ery iteration.<br>
<br>--<br><br>From the recent refactoring of the Metrics, another important=
<br>difference is that the Mattes MI implemented in ITK can be run<br>in mu=
ltiple-threads. That is, Mattes takes advantage of multiple<br>cores, while=
the ITK implementation of ViolaWells does not.<br>
<br>(this is in part due, to their different ways of picking up the<br>samp=
les).<br><br><br>Note that these comments are very specific to the ITK<br>i=
mplementation of these metrics, and probably should not<br>be generalized a=
s a comparison of the algorithm themselves.<br>
<br><br>--<br><br>For full details, you may want to look at:<br><br>=C2=A0*=
Viola, P. and Wells III, W. (1997).<br>=C2=A0* "Alignment by Maximiza=
tion of Mutual Information"<br>=C2=A0* International Journal of Comput=
er Vision, 24(2):137-154<br>
<br><br>and for Mattes:<br><br>=C2=A0* [1] "Nonrigid multimodality ima=
ge registration"<br>=C2=A0*=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 D. Mattes, D=
. R. Haynor, H. Vesselle, T. Lewellen and W. Eubank<br>=C2=A0*=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0 Medical Imaging 2001: Image Processing, 2001, pp. 1609-1=
620.<br>
=C2=A0* [2] "PET-CT Image Registration in the Chest Using Free-form De=
formations"<br>=C2=A0*=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 D. Mattes, D. R. =
Haynor, H. Vesselle, T. Lewellen and W. Eubank<br>=C2=A0*=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0 IEEE Transactions in Medical Imaging. Vol.22, No.1,<br>
=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 January 2003. pp.120-128.<br>=C2=
=A0* [3] "Optimization of Mutual Information for MultiResolution Image=
<br>=C2=A0*=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Registration"<br>=C2=A0*=C2=
=A0=C2=A0=C2=A0=C2=A0=C2=A0 P. Thevenaz and M. Unser<br>=C2=A0*=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0 IEEE Transactions in Image Processing, 9(12) December 20=
00.<br>
<br><br><br>=C2=A0=C2=A0=C2=A0 Regards,<br><br><br>=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0 Luis<br><br><br>----------------------------------------=
-----------------------------------------<br><div class=3D"gmail_quote">On =
Fri, Oct 29, 2010 at 6:08 PM, Dav <span dir=3D"ltr"><<a href=3D"mailto:d=
avoud_zzz at yahoo.com">masslawi at gmail.com</a>></span> wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; borde=
r-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><table border=3D"=
0" cellpadding=3D"0" cellspacing=3D"0"><tbody><tr><td style=3D"font: inheri=
t;" valign=3D"top">
Dear Ken,<div><br></div><div>The most important difference between Mattes m=
utual information and regular</div><div><br></div><div>mutual=C2=A0informat=
ion=C2=A0metric in ITK is their sampling. While the mutual information</div=
><div>
<br></div><div>metric take samples from all of=C2=A0the=C2=A0transformation=
space, the Mattes mutual=C2=A0</div><div><br></div><div>information only s=
amples pixels from=C2=A0number of=C2=A0predefined histogram bins=C2=A0</div=
><div><br></div><div>which extremely decreases the computation time and imp=
roves the performance.</div>
<div><br></div><div>Best regards,=C2=A0</div><div><br></div><div>Masslawi.<=
/div></td></tr></tbody></table><br>
<br>_____________________________________<br>
Powered by <a href=3D"http://www.kitware.com" target=3D"_blank">www.kitware=
.com</a><br>
<br>
Visit other Kitware open-source projects at<br>
<a href=3D"http://www.kitware.com/opensource/opensource.html" target=3D"_bl=
ank">http://www.kitware.com/opensource/opensource.html</a><br>
<br>
Kitware offers ITK Training Courses, for more information visit:<br>
<a href=3D"http://www.kitware.com/products/protraining.html" target=3D"_bla=
nk">http://www.kitware.com/products/protraining.html</a><br>
<br>
Please keep messages on-topic and check the ITK FAQ at:<br>
<a href=3D"http://www.itk.org/Wiki/ITK_FAQ" target=3D"_blank">http://www.it=
k.org/Wiki/ITK_FAQ</a><br>
<br>
Follow this link to subscribe/unsubscribe:<br>
<a href=3D"http://www.itk.org/mailman/listinfo/insight-users" target=3D"_bl=
ank">http://www.itk.org/mailman/listinfo/insight-users</a><br>
<br></blockquote></div><br>
--000e0ce0f792f9f6950493ea8c73--
More information about the Insight-users
mailing list