Re: Scanogram lines in MR
- From: Colby Dillion <colby.dillion@xxxxxxxxx>
- Date: Mon, 25 Jun 2007 06:12:34 -0700
On Jun 25, 4:40 am, anjali <anjali.cha...@xxxxxxxxx> wrote:
On Jun 1, 6:50 pm, Colby Dillion <colby.dill...@xxxxxxxxx> wrote:
It sounds like you are trying to render your own image withlocalizer
lines based on the information by David Clunie at the link below.
http://dclunie.com/medical-image-faq/html/part2.html#DICOMLocalizers
1) unable to findlocalizerimage
If none of images have an image type of "LOCALIZER" but they all
contain position and orientation information you could mark all of the
images as axial, sagittal, or coronal and then run through them to
find a 1 or 2 slice series that was scanned on a different axis.
There are other methods that you could try but this is the simplest
method that I have found.
2) some scanogram lines draw are not resulting in perfect straight
line, some what trapezoid is drawn onlocalizerimage.
Is there any solution to this, or any other way (assumptions) to
convert this trapezoid in line.
If your code is modeled after the code by David Clunie then that is
the intended result. What you are seeing are the boundaries of your
selected slice. If thelocalizeris not perpendicular to the selected
slice then you will see the trapezoid shape. In my code I used
different hues of color to give the perception of depth and to
indicate the foreground and background of the slice lines.
Colby
i did not make any changes in basic code of David Clunie. I know the
reason behind the trapezoid. i am using some other software for
benchmarking, in that straight lines are shown where for same images i
am getting trrapezoid. is there any assumption or way to convert
trapezoid into straight lines (how to show the lines when trapezoid).
If you understand how David's code is working then this should be
fairly trivial, but I will offer a hint since I have seen this
question asked several times.
You have the normal vectors and physical dimensions of both planes.
Find the points where the left and right sides (lines) of your slice
intersect with the localizer plane. David's code could be modified to
project the points onto the localizer image and then draw a line.
Nearly all of the math you need can be found on these two pages:
http://en.wikipedia.org/wiki/Plane_(mathematics)
http://en.wikipedia.org/wiki/Line-plane_intersection
Good luck!
Colby
.
- References:
- Scanogram lines in MR
- From: anjali
- Re: Scanogram lines in MR
- From: Colby Dillion
- Re: Scanogram lines in MR
- From: anjali
- Scanogram lines in MR
- Prev by Date: Re: Scanogram lines in MR
- Next by Date: Should an ADT^A18 (patient merge) update patient demographics as well?
- Previous by thread: Re: Scanogram lines in MR
- Next by thread: I have a question about reading a series dicom files
- Index(es):
Relevant Pages
|