optical density calibration

[Emily Jordan]

Hi Everyone,
I'm trying to analyze some autoradiography images for optical density and
I'm having trouble calibrating the image- namely because when I open the
tiff file and input my standard values everything is opposite, so what
should be the highest value on my standard selection on the image comes up
as the lowest.  Somehow the image settings are inverted so low optical
densities are coming out as high instead but I can't figure out how to
change it.  Any ideas?

Thanks,
Emily

[Randy Polson]

Hi-

If the source image is a black and white TIFF image you can just invert the
Look Up Table (LUT).  This can be accessed from the menu item
Image>LUT>Invert.
If the image is color and the above doesn't work, then you need to separate
the image to component colors, (red,green,blue), invert each and then
reassemble.

image>color>split channels
for each (red,green, blue)
  image>lut>invert image
image>color>merge channels
  (uncheck make composite, check ignore source LUT)

Best regards,
Randy

On Sat, Mar 3, 2012 at 4:23 AM, Emily Jordan <[hidden email]>wrote:

> Hi Everyone,
> I'm trying to analyze some autoradiography images for optical density and
> I'm having trouble calibrating the image- namely because when I open the
> tiff file and input my standard values everything is opposite, so what
> should be the highest value on my standard selection on the image comes up
> as the lowest.  Somehow the image settings are inverted so low optical
> densities are coming out as high instead but I can't figure out how to
> change it.  Any ideas?
>
> Thanks,
> Emily
>

[Terry Shippee]

I have recently run into the same problem.  Changing the LUT does not change the values of the pixels.  For 8 bit images ...Plugins>Morphology>Extened Minima.  This work qround flips all of my values, as shown by the histogram.
        I may be way off base but it works for me.

also,  Plugins>image inverter  seems to do the same thing.

Terry
________________________________________
From: ImageJ Interest Group [[hidden email]] on behalf of Randy Polson [[hidden email]]
Sent: Saturday, March 03, 2012 3:12 PM
To: [hidden email]
Subject: Re: optical density calibration

Hi-

If the source image is a black and white TIFF image you can just invert the
Look Up Table (LUT).  This can be accessed from the menu item
Image>LUT>Invert.
If the image is color and the above doesn't work, then you need to separate
the image to component colors, (red,green,blue), invert each and then
reassemble.

image>color>split channels
for each (red,green, blue)
  image>lut>invert image
image>color>merge channels
  (uncheck make composite, check ignore source LUT)

Best regards,
Randy

On Sat, Mar 3, 2012 at 4:23 AM, Emily Jordan <[hidden email]>wrote:

> Hi Everyone,
> I'm trying to analyze some autoradiography images for optical density and
> I'm having trouble calibrating the image- namely because when I open the
> tiff file and input my standard values everything is opposite, so what
> should be the highest value on my standard selection on the image comes up
> as the lowest.  Somehow the image settings are inverted so low optical
> densities are coming out as high instead but I can't figure out how to
> change it.  Any ideas?
>
> Thanks,
> Emily
>

[Gabriel Landini]

On Monday 05 Mar 2012 04:01:54 Shippee, Terry wrote:
> I have recently run into the same problem.  Changing the LUT does not change
> the values of the pixels.  For 8 bit images ...Plugins>Morphology>Extened
> Minima.  This work qround flips all of my values, as shown by the
> histogram. I may be way off base but it works for me.

I am afraid that yes, it is way off base and I suggest that you do not do it
like that.
The extended minima is a morphological procedure for gresycale images that
does something completely different. Please read about this in: Soille P.
Morphological Image Analysis. 2nd Corrected Edition, Springer, 2004.

In 3 sentences:
The extended-minima transform is the regional minima of the H-minima
transform.
The H-minima transform suppresses all minima whose depth is less than H (same
as preflood).
The regional minima are connected sets of pixels with a given value H (plateau
at depth H)  such that every pixel in the neighbourhood of the regional  
minima has a strictly higher value.

It might be that it just happens to result in an inverted image when applied
to a binary one (but this is *not* intended to be applied to binary images).
Applying this instead of inverting is plainly wrong. Even if it happens to
work, it would be impossible to justify in image processing terms.

> also,  Plugins>image inverter  seems to do the same thing.

No, they do not do "the same thing". In you case, just use "Invert". It would
be much quicker too.

Regards

Gabriel