Particle Analyzer

[Wu, Yuhong]

Dear All,

I am a new user of ImageJ. We have some images of a grey background
(base material) and dark grey specks (contaminants) that we want to do
image analysis. The first step I have been doing is to use the Finding
maxima function to count the number of specks, which works great. Now
one step further, in addition to the numbers, I'd like to analyze the
size of the particles and ideally get a distribution of the size.
However, the particle analyzer only deals with a binary image. And I
don't really know how you could use finding maxima to find the particles
first and then binarize the image so each local pixel group of maxima is
counted as one speck.

Another question is, I tried to binarize the image directly (which picks
up many unwanted background noise as black specks), however, the
particle analyzer gives a measurement of E-5 pixel^2 as the average
particle size. This does not seem right. What did I do wrong?

The image is a scanned image with 600 or 1200 dpi resolution. The link
to a sample pix is:
http://picasaweb.google.com/yuhongwuspecial/BlackSpeck

Thanks a lot for any information or input!

Yuhong


---------------------------------------------------------------------------
This e-mail is confidential.
If you are not the addressee or an authorized recipient of this message,
any distribution, copying, publication or use of this information for any
purpose is prohibited.
Please notify the sender immediately by e-mail and then delete this message.
Ce message est confidentiel.
Si vous n'etes pas le destinataire designe de ce message ou une personne
autorisee a l'utiliser, toute distribution, copie, publication ou usage a
quelques fins que ce soit des informations contenues dans ce message sont
interdits.
Merci d'informer immediatement l'expediteur par messagerie electronique et
d'ensuite detruire ce message.
---------------------------------------------------------------------------

[David Hovis]

I greatly appreciate that when you change units in the Image-
 >Properties... window, the unit conversion is handled automatically  
for certain types, but I find that there are two units that it doesn't  
handle that I would find useful.

1) mil or thou (1e-3 inches)
2) microinches (1e-6 inches)

(I occasionally do imaging work for people in industry in America, and  
they tend to want things in english units. )

--David


----------------------------------------
David Hovis
Senior Research Associate
Department of Materials Science
Case Western Reserve University
[hidden email]

[Wilfred L. Guerin]

Historical imaging, like the little artifact/museum archive browser
made a decade ago and still in use apparently, would prefer
calibration measures far more archaic, put them all in with easy
selection and conversion... There are likely standard conversion libs
(for plugins) if IJ will handle it.

That tool both helped research purposes importing imaging as well as
provided data on the item, most museums have trouble with artifact
measures and placement for things like kids' arm length and ADA,
especially with mobile or special interest exhibits set up at
libraries or schools. Just to name an obvious use where this need is
demonstrated. (yes ancient art is recorded in its real measure, no
infinite resolution decimal inaccuracies)


On 7/22/08, David Hovis <[hidden email]> wrote:

> I greatly appreciate that when you change units in the Image-
>  >Properties... window, the unit conversion is handled automatically
> for certain types, but I find that there are two units that it doesn't
> handle that I would find useful.
>
> 1) mil or thou (1e-3 inches)
> 2) microinches (1e-6 inches)
>
> (I occasionally do imaging work for people in industry in America, and
> they tend to want things in english units. )
>
> --David
>
>
> ----------------------------------------
> David Hovis
> Senior Research Associate
> Department of Materials Science
> Case Western Reserve University
> [hidden email]
>

[Sami Badawi-2]

Hi Yuhong,

You could try the Particle Counter or the Color Particle Analyzer in
the ShapeLogic 1.2 plugin.
I tried it on your image and it seems to give a reasonable result.
Set the Max distance parameter to 9 or 10. This control how big the
color bins are, and consequently how many particles you will find.
The default value is 50 but your image has very low contrast.

The Particle Counter or the Color Particle Analyzer should both gather
the information you are asking for. The difference between them is:

Particle Counter:
* Faster and more stable
* Prints the average color for each particle

Color Particle Analyzer:
* Gives more geometric information
* Still has at least one bug
* Prints the average color for each particle
* Prints the average gray value for each particle, making it is easier
to distinguish light and dark spots

-Sami Badawi
http://www.shapelogic.org

On Tue, Jul 22, 2008 at 6:07 PM, Wu, Yuhong <[hidden email]> wrote:

> Dear All,
>
> I am a new user of ImageJ. We have some images of a grey background
> (base material) and dark grey specks (contaminants) that we want to do
> image analysis. The first step I have been doing is to use the Finding
> maxima function to count the number of specks, which works great. Now
> one step further, in addition to the numbers, I'd like to analyze the
> size of the particles and ideally get a distribution of the size.
> However, the particle analyzer only deals with a binary image. And I
> don't really know how you could use finding maxima to find the particles
> first and then binarize the image so each local pixel group of maxima is
> counted as one speck.
>
> Another question is, I tried to binarize the image directly (which picks
> up many unwanted background noise as black specks), however, the
> particle analyzer gives a measurement of E-5 pixel^2 as the average
> particle size. This does not seem right. What did I do wrong?
>
> The image is a scanned image with 600 or 1200 dpi resolution. The link
> to a sample pix is:
> http://picasaweb.google.com/yuhongwuspecial/BlackSpeck
>
> Thanks a lot for any information or input!
>
> Yuhong
>
>
> ---------------------------------------------------------------------------
> This e-mail is confidential.
> If you are not the addressee or an authorized recipient of this message,
> any distribution, copying, publication or use of this information for any
> purpose is prohibited.
> Please notify the sender immediately by e-mail and then delete this message.
> Ce message est confidentiel.
> Si vous n'etes pas le destinataire designe de ce message ou une personne
> autorisee a l'utiliser, toute distribution, copie, publication ou usage a
> quelques fins que ce soit des informations contenues dans ce message sont
> interdits.
> Merci d'informer immediatement l'expediteur par messagerie electronique et
> d'ensuite detruire ce message.
> ---------------------------------------------------------------------------
>

[Michael Schmid]

Hi Yuhong,

the Process>Binary>Find Maxima function has an option to create
different output types.
You can try "Maxima within tolerance" or "segmented particles".
These create a binary output that can be used for "Analyze Particles".

When selecting "segmented particles": you should set a threshold so
that the particles are above the lower threshold (you need to convert
the image to grayscale for this).
It may be helpful to run Process>Subtract Background before
thresholding.
The "above lower threshold" option of "Find Maxima" works only if
the particles are maxima (not minima) of the pixel values; otherwise
you have to invert the image first.
(inverting the LUT does not help since it does not modify the pixel
values).

After "Analyze Particles", you can get the size distribution by
Analyze>Distribution.

Concerning your second question: possibly the image read by ImageJ
is spatially calibrated in inches (or cm) and the area displayed is
in square inches (or other units). See Image>Show Info for any possible
calibration.


Michael
________________________________________________________________


On 23 Jul 2008, at 00:07, Wu, Yuhong wrote:

> Dear All,
>
> I am a new user of ImageJ. We have some images of a grey background
> (base material) and dark grey specks (contaminants) that we want to do
> image analysis. The first step I have been doing is to use the Finding
> maxima function to count the number of specks, which works great. Now
> one step further, in addition to the numbers, I'd like to analyze the
> size of the particles and ideally get a distribution of the size.
> However, the particle analyzer only deals with a binary image. And I
> don't really know how you could use finding maxima to find the  
> particles
> first and then binarize the image so each local pixel group of  
> maxima is
> counted as one speck.
>
> Another question is, I tried to binarize the image directly (which  
> picks
> up many unwanted background noise as black specks), however, the
> particle analyzer gives a measurement of E-5 pixel^2 as the average
> particle size. This does not seem right. What did I do wrong?
>
> The image is a scanned image with 600 or 1200 dpi resolution. The link
> to a sample pix is:
> http://picasaweb.google.com/yuhongwuspecial/BlackSpeck
>
> Thanks a lot for any information or input!
>
> Yuhong
>
>
> ----------------------------------------------------------------------
> -----
> This e-mail is confidential.
> If you are not the addressee or an authorized recipient of this  
> message,
> any distribution, copying, publication or use of this information  
> for any
> purpose is prohibited.
> Please notify the sender immediately by e-mail and then delete this  
> message.
> Ce message est confidentiel.
> Si vous n'etes pas le destinataire designe de ce message ou une  
> personne
> autorisee a l'utiliser, toute distribution, copie, publication ou  
> usage a
> quelques fins que ce soit des informations contenues dans ce  
> message sont
> interdits.
> Merci d'informer immediatement l'expediteur par messagerie  
> electronique et
> d'ensuite detruire ce message.
> ----------------------------------------------------------------------
> -----