I had some CT scans taken this year due to medical issues
and found the data could be provided on DVD on request ;)

Images were in DICOM format and ImageMagick didn't like them,
but they contained uncompressed 16 bit pixel data so conversion
was mostly an issue of discarding the header and guessing width.

This is a stack of 91 horizontal slices that were turned into
a 3d pigment simply as a gradient of image_maps. This allows to
render an arbitrary slice or texture an arbitrary shape, which
is endless fun but doesn't provide too impressive images.

So I turned the pigment into a function and rendered it as
volumetric media, using a density color_map to enhance certain
brightness ranges. This last one looks quite promising.

Post a reply to this message

Attachments:
Download 'inner_values.png' (357 KB)

Preview of image 'inner_values.png'

On 04/11/13 22:31, Christian Froeschlin wrote:
> I had some CT scans taken this year due to medical issues
> and found the data could be provided on DVD on request ;)
> 

I see your problem here; your head and arms have fallen off ;-)

Nice work. A brilliant example of putting a positive spin on a possible
problem.

John
-- 
Protect the Earth
It was not given to you by your parents
You hold it in trust for your children

Post a reply to this message

>Christian Froeschlin  on date 04/11/2013 23.31 wrote:
> I had some CT scans taken this year due to medical issues
> and found the data could be provided on DVD on request ;)
>
> Images were in DICOM format and ImageMagick didn't like them,
> but they contained uncompressed 16 bit pixel data so conversion
> was mostly an issue of discarding the header and guessing width.
>
> This is a stack of 91 horizontal slices that were turned into
> a 3d pigment simply as a gradient of image_maps. This allows to
> render an arbitrary slice or texture an arbitrary shape, which
> is endless fun but doesn't provide too impressive images.
>
> So I turned the pigment into a function and rendered it as
> volumetric media, using a density color_map to enhance certain
> brightness ranges. This last one looks quite promising.
>
>
>
>
>
>
Very nice render.
I suggest you to manage DICOM images with ImageJ instead of ImageMagick, 
your next export will be simpler.
;-)
Paolo

Post a reply to this message

Christian Froeschlin <chr### [at] chrfrde> wrote:
> I had some CT scans taken this year due to medical issues
> and found the data could be provided on DVD on request ;)
>
> Images were in DICOM format and ImageMagick didn't like them,
> but they contained uncompressed 16 bit pixel data so conversion
> was mostly an issue of discarding the header and guessing width.
>
> This is a stack of 91 horizontal slices that were turned into
> a 3d pigment simply as a gradient of image_maps. This allows to
> render an arbitrary slice or texture an arbitrary shape, which
> is endless fun but doesn't provide too impressive images.
>
> So I turned the pigment into a function and rendered it as
> volumetric media, using a density color_map to enhance certain
> brightness ranges. This last one looks quite promising.

First, I really hope that the "possible problems" John mentioned are not to
serious. "Some CT" indicates a certain seriousness, I hope the best for you.

My first thought was, if it is more easy to convert the data to df3. I got the
impression that DICOM is not so far away from df3, after ripping off the
metadata. Seems to be big-endian too. May be one can exchange the headers of the
voxel bulk and get a df3 file. I don't know, only an idea.

The picture is fine but hard to interpret in some areas. If I look at the bottom
of the image I think I'm looking at your back in an upward direction. If I look
at the scapula and the clavicula only, I cannot tell apart what is in front and
what is back. Both are dominated by the green lung in large parts.

As you speak from publishing your data I can only remind you that DICOM files
contain medical metadata, which you will strip off first most likely. And
second, even the voxel data itself can reveal your disease to someone who is
able to read and visualise them properly. From your given image I cannot read
your disease but from detailed CT data may be... I only want to remind you of
that... If you will still do so: 1GB df3 files can be compressed to 3MB using
the highest compression rate with WinZip, so a DVD is not needed most likely.

Best regards,
Michael

Post a reply to this message

> First, I really hope that the "possible problems" John mentioned are not to
> serious. "Some CT" indicates a certain seriousness, I hope the best for you.

Thanks!

It looked serious for a while but currently everything seems fine.

> My first thought was, if it is more easy to convert the data to df3. I got the
> impression that DICOM is not so far away from df3, after ripping off the
> metadata. Seems to be big-endian too. May be one can exchange the headers of the
> voxel bulk and get a df3 file. I don't know, only an idea.

Yes df3 would have been an option. Using the image stack was simply
the way of least resistance (least work for me at the moment) and it
renders quite fast too.

> The picture is fine but hard to interpret in some areas. If I look at the bottom
> of the image I think I'm looking at your back in an upward direction. If I look
> at the scapula and the clavicula only, I cannot tell apart what is in front and
> what is back. Both are dominated by the green lung in large parts.

It is a front view but difficult to tell with this "transparent
data. In fact the back view can be recognized by the scan bed that
is also included in the data.

> As you speak from publishing your data I can only remind you that DICOM files
> contain medical metadata, which you will strip off first most likely

I think that was a misunderstanding, I did not intend to publish
the DICOM files or derived data sets. I think the render itself is
not so problematic and unsuitable for medical diagnostics. Also it
is a clean scan and should only serve to show off my health ;)

Post a reply to this message

> I had some CT scans taken this year due to medical issues and found
> the data could be provided on DVD on request ;)
>

Very nice, and a good candidate for a new "ykybrtlw" sentence:

"You know you've been Raytracing too long when... you ask your doctor
for the CT scan data". ;)

--
Jaime

Post a reply to this message

Christian Froeschlin <chr### [at] chrfrde> wrote:
> > First, I really hope that the "possible problems" John mentioned are not to
> > serious. "Some CT" indicates a certain seriousness, I hope the best for you.
>
> Thanks!
>
> It looked serious for a while but currently everything seems fine.
>
> > My first thought was, if it is more easy to convert the data to df3. I got the
> > impression that DICOM is not so far away from df3, after ripping off the
> > metadata. Seems to be big-endian too. May be one can exchange the headers of the
> > voxel bulk and get a df3 file. I don't know, only an idea.
>
> Yes df3 would have been an option. Using the image stack was simply
> the way of least resistance (least work for me at the moment) and it
> renders quite fast too.
>
> > The picture is fine but hard to interpret in some areas. If I look at the bottom
> > of the image I think I'm looking at your back in an upward direction. If I look
> > at the scapula and the clavicula only, I cannot tell apart what is in front and
> > what is back. Both are dominated by the green lung in large parts.
>
> It is a front view but difficult to tell with this "transparent
> data. In fact the back view can be recognized by the scan bed that
> is also included in the data.
>
> > As you speak from publishing your data I can only remind you that DICOM files
> > contain medical metadata, which you will strip off first most likely
>
> I think that was a misunderstanding, I did not intend to publish
> the DICOM files or derived data sets. I think the render itself is
> not so problematic and unsuitable for medical diagnostics. Also it
> is a clean scan and should only serve to show off my health ;)

I really like to read this. If I misunderstood your sentence about a DVD
distribution please forgive me. In a way Jaime is right and not. In Germany a
lot of radiologists are giving away the data together with an application to
view the data to their patients. They ask their patients if they will have it,
you must not ask them to can render the data. I had to judge several scans from
my mother's severe osteoporosis. For some reason she relies more on my amateur
opinion than on the practitioners. I have worked for some 15 years as a male
nurse mainly in oncology so I have the one or other experience with CT data but
I'm far away from being an expert.

Best regards,
Michael

Post a reply to this message