Calibrating density and creating calibrated image

[Diana-3]

I am working with autoradiography images - these are grey scale and need to
be calibrated to uCi using images of radioactive standards. I know how to
make and apply calibrations using the calibrate function and a rodbard curve.

My question: is it possible to CONVERT the original image (grey scale) into
a calibrated image where each pixel is converted into the new (calibrated)
value. In other words, i want to create and save this new calibrated image
as a map of uCi.

I read that if one saves calibrate image as tiff, the calibration curve will
be saved with it - this means next time i open this image in ImageJ and
perform measurements, it will measure uCi and not grey scale. But..., i want
to do region of interest analysis in another software which is why i need
saved images with calibrated pixels. I would be very keen to find out if
this is possible - perhaps someone out there has made an add-on that can do
this??

To explain - the reason i want to perform ROI analysis in another software
is because i want to save the ROI's - not just numbers but the actual
coordinates, which can be loaded into the image later, so that we can
sinspect and edit ROI's if necessary... I'm under impression that ImageJ
cannot do this but if i'm wrong please let me know!


Many thanks,
Diana

[Michael A. Miller]

>>>>> "Diana" == Diana  <[hidden email]> writes:

    > I am working with autoradiography images - these are grey
    > scale and need to be calibrated to uCi using images of
    > radioactive standards. I know how to make and apply
    > calibrations using the calibrate function and a rodbard
    > curve.

    > My question: is it possible to CONVERT the original image
    > (grey scale) into a calibrated image where each pixel is
    > converted into the new (calibrated) value. In other words,
    > i want to create and save this new calibrated image as a
    > map of uCi.


You can apply your calibration with the image math functions
available in the Process-Math->... menu.  If the original image
is integer data, make sure to convert the image to 32 bit
floating point values first (using Image->Type->32 bit).

If you save the image as raw values (or tiff), it can be loaded
into another program.

Mike


--
Michael A. Miller                             [hidden email]
  Department of Radiology, Indiana University School of Medicine

[R. C.]

You can use something like this: (first convert to 32bit image)

// Radiochromic film calibration
// A. Perez Rozos - Enero 2009
// H. Virgen de la Victoria. Malaga. spain
// -------------------------------------------------------
 macro "Calibration" {
// a0...a3 are the coeficients of the calibration 3th degree polyn.
a0=2191.9
a1=-0.1411
a2=0.000003206
a3=-0.0000000000252

w1 = getWidth(); h1 = getHeight();
for (i=0; i<w1; i++) {
 for (j=0;j<h1; j++) {
  vpixel=getPixel(i,j);
  cpixel=a0 + a1 * vpixel + a2 *vpixel*vpixel +a3 *vpixel*vpixel*vpixel;
// Negative values are equaled to 0
  if (cpixel<0) cpixel=0;
  setPixel(i,j,cpixel);

 };
showProgress((i/w1));
};
print("Calibration curve applied - EBT Junio de 2008");
};




Alberto Perez
Medical Physicist
Hospital Universitario Virgen de la Victoria
Malaga. Spain
[hidden email]


2009/6/17 Michael A. Miller <[hidden email]>
- Mostrar texto citado -


>>>>> "Diana" == Diana  <[hidden email]> writes:

   > I am working with autoradiography images - these are grey
   > scale and need to be calibrated to uCi using images of
   > radioactive standards. I know how to make and apply
   > calibrations using the calibrate function and a rodbard
   > curve.

   > My question: is it possible to CONVERT the original image
   > (grey scale) into a calibrated image where each pixel is
   > converted into the new (calibrated) value. In other words,
   > i want to create and save this new calibrated image as a
   > map of uCi.



You can apply your calibration with the image math functions
available in the Process-Math->... menu.  If the original image
is integer data, make sure to convert the image to 32 bit
floating point values first (using Image->Type->32 bit).

If you save the image as raw values (or tiff), it can be loaded
into another program.

Mike


--
Michael A. Miller                             [hidden email]
 Department of Radiology, Indiana University School of Medicine

[Michael Schmid]

Hi Diana,

you could also convert your calibrated image to 32 bit; then make  
sure that Edit>Options>Conversions> 'Scale when converting' is  
deselected and transform it back to 16-bit (or 8 bits, if your  
analysis software cannot read 16-bit tiffs).

Of course, you can have only integer values after the conversion, and  
the values will be between 0 and 65535 (16-bit) or 0-255 (8-bit).

Michael
________________________________________________________________

On 17 Jun 2009, at 19:06, Diana wrote:

> I am working with autoradiography images - these are grey scale and  
> need to
> be calibrated to uCi using images of radioactive standards. I know  
> how to
> make and apply calibrations using the calibrate function and a  
> rodbard curve.
>
> My question: is it possible to CONVERT the original image (grey  
> scale) into
> a calibrated image where each pixel is converted into the new  
> (calibrated)
> value. In other words, i want to create and save this new  
> calibrated image
> as a map of uCi.
>
> I read that if one saves calibrate image as tiff, the calibration  
> curve will
> be saved with it - this means next time i open this image in ImageJ  
> and
> perform measurements, it will measure uCi and not grey scale.  
> But..., i want
> to do region of interest analysis in another software which is why  
> i need
> saved images with calibrated pixels. I would be very keen to find  
> out if
> this is possible - perhaps someone out there has made an add-on  
> that can do
> this??
>
> To explain - the reason i want to perform ROI analysis in another  
> software
> is because i want to save the ROI's - not just numbers but the actual
> coordinates, which can be loaded into the image later, so that we can
> sinspect and edit ROI's if necessary... I'm under impression that  
> ImageJ
> cannot do this but if i'm wrong please let me know!
>
>
> Many thanks,
> Diana