<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Hi,</p>
<p>just to bring some closure to that problem..</p>
<p>I didn't manage to adapt the ramp filter, so I instead modified
rtkProjectionsReader to include an Orientation parameter that sets
the right coefficients for an itkPermuteAxis filter.</p>
<p>I don't think it is worthy to be included in the official source,
but I am happy to share if someone requires it.</p>
<p>Best regards,</p>
<p>Vincent<br>
</p>
<div class="moz-cite-prefix">On 15.04.20 20:13, Simon Rit wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CAF0oig1H=yihYbGNKCLxjJPx0GtCXH5sW-9GzQiSbROos50Bhg@mail.gmail.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<div dir="ltr">
<div>Hi Vincent,</div>
<div>Everything you did is correct but rtkfdk always applies the
ramp filter along the first dimension and does not check if
this corresponds to the direction orthogonal to the rotation
axis. For your case, it is probably not so difficult to adapt
the code. When there is a small rotation, it is debatable
whether one should ignore the rotation for the ramp filter,
resample the image, or recalculate the filter in 2D (see our
recent publication <a
href="https://doi.org/10.1088/1361-6560/ab7875"
moz-do-not-send="true">here</a> which has not been
implemented in RTK). I have opened an <a
href="https://github.com/SimonRit/RTK/issues/335"
moz-do-not-send="true">issue on github</a>, hopefully we'll
find time to fix this in the future (at least raise an
exception above a parametrable rotation threshold).<br>
</div>
<div>Best regards,</div>
<div>Simon<br>
</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Wed, Apr 15, 2020 at 7:00
PM vincent <<a href="mailto:vl@xris.eu"
moz-do-not-send="true">vl@xris.eu</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px
0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Hi
RTK community,<br>
<br>
I hope everyone is doing well in those hard times.<br>
<br>
I have a panel that is rotated by 90° on the CT system. I
used to <br>
rotate the images back in the right orientation before
building the <br>
sinogram with rtkprojections, but then I thought I might give
a try at <br>
the "newdirection" option. Using the matrix [0 1 0 ; 1 0 0 ;
0 0 1] <br>
gives the expected results. So far so good. But when I try to
<br>
reconstruct the latter with the same geometric parameters than
for the <br>
pre-rotated projections, I obtain a very blurry result (vs a
sharp one <br>
in the first case). I am obviously missing something, and I
would <br>
suspect I am not giving the right information to the rtkfdk <br>
application. So far, here is what I have tried:<br>
<br>
-keep the projections in the initial orientation, but give the
<br>
"newdirection" option to the rtkfdk application,<br>
<br>
-specify the "newdirection" in both rtkprojections and rtkfdk,<br>
<br>
-turn to you for advice !<br>
<br>
I thank you very much in advance for any help you might be
able to give.<br>
<br>
Best regards,<br>
<br>
Vincent<br>
<br>
_______________________________________________<br>
Rtk-users mailing list<br>
<a href="mailto:Rtk-users@public.kitware.com" target="_blank"
moz-do-not-send="true">Rtk-users@public.kitware.com</a><br>
<a
href="https://public.kitware.com/mailman/listinfo/rtk-users"
rel="noreferrer" target="_blank" moz-do-not-send="true">https://public.kitware.com/mailman/listinfo/rtk-users</a><br>
<br>
</blockquote>
</div>
</blockquote>
</body>
</html>