Hi,

i'm developing a software for image reconstruction from few x-ray
projection images (<10) from different arbitrary directions. The images
are DICOM.  Until now i applied reconstruction algorithms like FBP, ART
or CG to the gray values of the projections.

The algorithms work fine,  but what is the real meaning of the gray
values in x-ray projections? Is it the line integral of the attenuation
coefficients? That would mean i can apply the standard algorithms. Or
is there a log relation between the gray values and  the value of the
line integrals caused by Beer's Law?

Here the DICOM-Header Information:

(0028,0002) US 1                                        #   2, 1
SamplesPerPixel
(0028,0004) CS [MONOCHROME2]                            #  12, 1
PhotometricInterpretatio
(0028,0010) US 1346                                     #   2, 1 Rows
(0028,0011) US 2026                                     #   2, 1
Columns
(0028,0030) DS [0.143\0.143]                            #  12, 2
PixelSpacing
(0028,0100) US 16                                       #   2, 1
BitsAllocated
(0028,0101) US 10                                       #   2, 1
BitsStored
(0028,0102) US 9                                        #   2, 1
HighBit
(0028,0103) US 0                                        #   2, 1
PixelRepresentation
(0028,1050) DS [511.500]                                #   8, 1
WindowCenter
(0028,1051) DS [1023.000]                               #   8, 1
WindowWidth

Thanks
Oliver Weinheimer

Report this post to a moderator

[  Post Follow-up to this message ]

On 6/28/06 9:17 AM, in article
1151511442.257369.165450@x69g2000cwx.googlegroups.com, "oliwe"
<mail@oliwe.com> wrote:

> The algorithms work fine,  but what is the real meaning of the gray
> values in x-ray projections? Is it the line integral of the attenuation
> coefficients? That would mean i can apply the standard algorithms. Or
> is there a log relation between the gray values and  the value of the
> line integrals caused by Beer's Law?

I am not in the medical imaging field so I may not understand the jargon.
"what is the real meaning of the gray values in x-ray projections?" does not
make syntactical sense. Thy formulating your question formally using
mathematical (algegraic) description.

Bill
-- Ferme le Bush

Report this post to a moderator

[  Post Follow-up to this message ]

oliwe wrote:
> Hi,
>
> i'm developing a software for image reconstruction from few x-ray
> projection images (<10) from different arbitrary directions. The images
> are DICOM.  Until now i applied reconstruction algorithms like FBP, ART
> or CG to the gray values of the projections.
>
> The algorithms work fine,  but what is the real meaning of the gray
> values in x-ray projections? Is it the line integral of the attenuation
> coefficients? That would mean i can apply the standard algorithms. Or
> is there a log relation between the gray values and  the value of the
> line integrals caused by Beer's Law?

The meaning of the image pixel values varies depending on the source of
the images, but generally for CR or DR images, the value is related to
the radiation exposure to the image receptor.  The equation mapping
receptor exposure (receptor response to be more precise) to pixel value
depends on the manufacturer of the CR or DR unit.  Various
manufacturers use different equations for the mapping.  It would also
be very difficult to relate pixel value to a precise dose since this
would require assuming that the CR or DR system was properly
calibrated.  Converting to a line integral of the linear attenuation
coefficients would also be equally problematic because this would
require knowledge of the incident beam intensity.  Theoretically given
the manufacturer's response/pixel value mapping this could be
determined from a non-attenuated region of the image, but these areas
usually tend to be saturated and useless for this kind of
determination.

For an image that was scanned in from film, there would be no such
relationship and no way to relate it back to attenuation coefficients.

Report this post to a moderator

[  Post Follow-up to this message ]