Hi James,
I'm not completely sure what you're trying to do, but it seems like
you want to check the two calibrations to make sure that the people
who felt like doing it a different way are getting the same numbers.
Inversion for 8 bits is "inverted_value = 255-current_value", so by
subtracting inverted grays from the originals you're going to get
"255-current_value-inverted_value". I think you want to do
"255-inverted_value" which would be the same as the original. A
bitwise NOT should also invert i.e. "!original = inverted" and
likewise "!inverted = original"
Justin
On 10/11/06, James Beals <
[hidden email]> wrote:
> Can some one explain what is happening when a gray level images is inverted?
>
> Background.
> We measure the density of in situ auto radiograms.
> Some folks here invert the gray level values.
> So that the darker areas have higher values than the brighter areas.
> black=255, white=0, and measure density like that.
>
> If I calibrate to optical density numbers, using a Stuoffers 21 step wedge (1=0.06 to 21=3.00 OD)
> and the (Rodbard) function, I get very good numbers, from different cameras and scanners.
> So I know how to do it right.
>
> But folks here have been inverting the gray levels and making measurements of that for some
> time.
>
> So, I imaged the standard and plot the values I get a prety good log curve.
> If I copy and inverted gray levels, and plot them agents the original, the plots cross over at about
> 125, the 4th step in the standard or about 0.50 OD.
> If I subtract the inverted gray levels from the gray levels, I get a very strange
> curve, that starts at about 250 crosses the gray levels step 7 just bellow 100, goes into the
> negative at step 4-5 and down to -255. The difference should be 0.
>
> So i have to think that there is a difference between Inversion and difference.
> any clues ??
> thanks
> James
>
Locked
