<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Thanks Simon,<br>
<br>
I have implemented both of the techniques you have mentioned. Nui
scatter correction does do a reasonable job, but I have found it
leaves significant artefacts when there is a substantial change in
organ shape or location between scans, and I hope that MC will allow
me to investigate the true effect of this technique when I have the
scatter ground truth.<br>
<br>
WLS also has been outperforming LS in phantom experiments, but only
when photon flux is being starved, and has been difficult to get a
reasonable estimate of I0 from Varian data (also have yet to set up
the scanner to even produce such a low current).<br>
<br>
I am doing my developments in MATLAB rather than RTK at the moment
as I find it quicker to test things out, and utilise Toolboxes, but
it is my hope to eventually move over once I have found more
stability. I would love to contribute to what is now a very
comprehensive and powerful software package.<br>
<br>
Cheers,<br>
<br>
Jonathan<br>
<br>
<div class="moz-cite-prefix">On 01/12/15 08:58, Simon Rit wrote:<br>
</div>
<blockquote
cite="mid:CAF0oig2LgKGM9jq7x+iKK95wp=9w-6tGuP1EtJUHEoUGnqnU3A@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>
<div>Hi,<br>
</div>
For scatter and assuming you have the planning CT, you could
consider the <a moz-do-not-send="true"
href="http://www.openrtk.org/RTK/news/201507_press.php">solution
of Yang et al</a> (that he implemented from Niu et al) that
is much simpler than Monte Carlo. I know two teams that have
already implemented it using RTK and we will try to do our own
in the future.<br>
</div>
<div>For photon statistics, Cyril has recently been working on a
weighted least square solution, see, e.g., <a
moz-do-not-send="true"
href="https://github.com/SimonRit/RTK/commit/b133c0bb96bff6805dd28c7820a89310629ff031">this
commit</a>. But I don't think it's mature yet.<br>
</div>
<div>Note that we discussed pre-processing in July in Lyon, the
minutes are <a moz-do-not-send="true"
href="http://wiki.openrtk.org/index.php/RTK/Meetings/TrainingNov15#Pre-processing">here</a>.<br>
</div>
<div>Good luck, it's not an easy task,<br>
</div>
Simon</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Tue, Dec 1, 2015 at 9:50 AM,
Jonathan Mason <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:s1015431@staffmail.ed.ac.uk" target="_blank">s1015431@staffmail.ed.ac.uk</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div style="word-wrap:break-word">Thank you Simon and Chao,
<div><br>
</div>
<div>I apologise for the confusion—I can see that my
message was not very clear. I am essentially trying to
compensate for low photon flux and scatter, but am
finding it difficult with pre corrected data. I will
certainly take time to look through your suggested
publications, so thanks for that.</div>
<div><br>
</div>
<div>Without having access to low level physical data, I
am considering simulating a scanner with MC, where I
will have the ground truth for scatter field and photon
statistics, and try to produce reconstructions on this.</div>
<div><br>
</div>
<div>Many thanks,</div>
<div><br>
</div>
<div>Jonathan</div>
<div>
<div class="h5">
<div><br>
<div>
<blockquote type="cite">
<div>On 1 Dec 2015, at 07:27, Simon Rit <<a
moz-do-not-send="true"
href="mailto:simon.rit@creatis.insa-lyon.fr"
target="_blank">simon.rit@creatis.insa-lyon.fr</a>>
wrote:</div>
<br>
<div>
<div dir="ltr">
<div>
<div>
<div>
<div>Hi Jonathan,<br>
</div>
I'm not sure I fully get it either.
What I can say is that I was a postdoc
at the NKI until the end of 2009 where
I was working on the Elekta
reconstruction software. The way <a
moz-do-not-send="true"
href="http://www.openrtk.org/Doxygen/classrtk_1_1ProjectionsReader.html"
target="_blank">rtk::ProjectionsReader</a>
works for Elekta projections is close
to what they were doing at the time,
i.e., Boellaard scatter correction,
cropping and simple log to go to a
line integral. I know they have been
improving it since (with lag and
scatter corrections) but I'm not sure
what's commercialized what's not. In
any case, they publish / present what
they do (see publications of M. van
Herk, J-J. Sonke and L. Ploeger).<br>
</div>
I'm not expert of Varian data but for
sure, when you get pre-corrected
projections, it's hard to know what part
you're correcting (uncorrected problems
or miscorrections). Instead of reverse
engineering, which can be tough because
Varian does a lot of stuff on the raw
data I believe (look at the publications
of J. Star-Lack), I would start from the
raw data if you can access them.<br>
</div>
Good luck,<br>
</div>
Simon<br>
</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Mon, Nov 30,
2015 at 7:50 PM, Chao Wu <span dir="ltr"><<a
moz-do-not-send="true"
href="mailto:wuchao04@gmail.com"
target="_blank">wuchao04@gmail.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">
<div dir="ltr">Hi Jonathan,
<div><br>
</div>
<div>I do not quite understand your
problem.</div>
<div>I think most information can be
retrieved by the relationship
between flux (I) and attenuation
(u):</div>
<div>I = I0 * exp (- integral(u * dL)
)</div>
<div>Of course there are other effects
like beam hardening and scattering
involved but this model is the
basic.</div>
<div><br>
</div>
<div>Regards,</div>
<div>Chao</div>
</div>
<div>
<div>
<div class="gmail_extra"><br>
<div class="gmail_quote">2015-11-30
17:29 GMT+01:00 Jonathan Mason <span
dir="ltr"><<a
moz-do-not-send="true"
href="mailto:j.mason@ed.ac.uk"
target="_blank">j.mason@ed.ac.uk</a>></span>:<br>
<blockquote class="gmail_quote"
style="margin:0 0 0
.8ex;border-left:1px #ccc
solid;padding-left:1ex">Good
afternoon RTK users,<br>
<br>
I have access to "raw data"
from a Varian OBI scanner, and
have<br>
experimented with RTK for
performing its reconstruction.
However, it is<br>
not really "raw" in the sense
that the coefficients are
proportional to<br>
the photon flux recorded at a
given sensor, but is instead
proportional<br>
to the Hounsfield attenuation
along that path. With this
data, it means<br>
that one can reconstruct using
FDK and other iterative
techniques, which<br>
assume a linear model, but the
connection to the underlying
physics is<br>
abstract.<br>
<br>
The problem I then have when
trying to develop
reconstruction<br>
techniques, is that I do not
know whether I am really
correcting for<br>
physical distortions or just
correcting for Varian's
preprocessing,<br>
which has been finely tuned
for its own FDK method.<br>
<br>
My question is whether anybody
has dug into the steps that
manufacturers<br>
such as Varian or Elekta
perform to arrive at these
projection images?<br>
And if they think that if
reverse engineered, could
provide a richer set<br>
of information to facilitate
advanced strategies.<br>
<br>
Best wishes,<br>
<br>
Jonathan Mason<br>
<span><font color="#888888"><br>
--<br>
The University of
Edinburgh is a charitable
body, registered in<br>
Scotland, with
registration number
SC005336.<br>
<br>
_______________________________________________<br>
Rtk-users mailing list<br>
<a moz-do-not-send="true"
href="mailto:Rtk-users@public.kitware.com" target="_blank">Rtk-users@public.kitware.com</a><br>
<a moz-do-not-send="true"
href="http://public.kitware.com/mailman/listinfo/rtk-users"
rel="noreferrer"
target="_blank">http://public.kitware.com/mailman/listinfo/rtk-users</a><br>
</font></span></blockquote>
</div>
<br>
</div>
</div>
</div>
<br>
_______________________________________________<br>
Rtk-users mailing list<br>
<a moz-do-not-send="true"
href="mailto:Rtk-users@public.kitware.com"
target="_blank">Rtk-users@public.kitware.com</a><br>
<a moz-do-not-send="true"
href="http://public.kitware.com/mailman/listinfo/rtk-users"
rel="noreferrer" target="_blank">http://public.kitware.com/mailman/listinfo/rtk-users</a><br>
<br>
</blockquote>
</div>
<br>
</div>
</div>
</blockquote>
</div>
<br>
</div>
</div>
</div>
</div>
<br>
The University of Edinburgh is a charitable body, registered
in<br>
Scotland, with registration number SC005336.<br>
<br>
</blockquote>
</div>
<br>
</div>
</blockquote>
<br>
</body>
</html>