Analyze particles question..

[NatashaW]

Dear all,

I have a fluorescent objects (green focis) in my image which are grouped together (every group belong to a cell) so I applied a mask from ROI manager to cluster them and I succeeded in this step.
Now my fluorescent objects are clustered in groups, and I need to count the particles in every group and give some details about them.
I should apply the Particle_Analysis plugin for every group to count them, but I should convert the image to a binary mask before applying it and in this way I'll lose the important details (size, intensity).
Maybe there are some features or methods that I don't know in imageJ to save these details?
Anyone can help me or give me an advice?

Thank you,

[Michael Schmid]

Hi Natasha,

two remarks on this:

(1) Analyze particles is restricted to the current selection (region  
of interest, roi) so if you have a roi encompassing one group you  
restrict the measurements to this group.

(2) If you can't define the particles by simple thresholding (which  
does not affect the image data, so you can still measure min, max,  
mean intensity, etc.): Did you have a look at the 'redirect to'  
feature in 'Set Measurements'? It allows you to do the measurements  
on an image that is different from the one that determines the  
particles. Set 'redirect to' to the image with the gray values and  
run 'Anlayze particles' on the mask.

Michael
________________________________________________________________

On 6 May 2010, at 18:35, NatashaW wrote:

> Dear all,
>
> I have a fluorescent objects (green focis) in my image which are  
> grouped
> together (every group belong to a cell) so I applied a mask from  
> ROI manager
> to cluster them and I succeeded in this step.
> Now my fluorescent objects are clustered in groups, and I need to  
> count the
> particles in every group and give some details about them.
> I should apply the Particle_Analysis plugin for every group to  
> count them,
> but I should convert the image to a binary mask before applying it  
> and in
> this way I'll lose the important details (size, intensity).
> Maybe there are some features or methods that I don't know in  
> imageJ to save
> these details?
> Anyone can help me or give me an advice?
>
> Thank you,

[NatashaW]

Thank you Michael !! That's what I was looking for..
This feature works only on gray images or also on RGB images?

another question: I didn't understand if I can define the particles by simple thresholding, how this will help me to count them easier?

[Michael Schmid]

Hi Natasha,

there is an experimental version of 'Color Threshold' in ImageJ. In  
contrast to grayscale thresholding it is destructive - you can't use  
it like grayscale thresholding for analyzing particles and measuring  
the original pixel values. Anyhow, measurements of min, mix, mean  
intensity for RGB images use the gray value, so you can convert you  
image to grayscale before doing any measurements and then have  
'standard' grayscale thresholding.

'Analyze>Analyze Particles' counts thresholded particles and does  
measurements on them (depending on what you set in Analyze>Set  
Measurements. In some cases, 'Process>Find Maxima' is also useful for  
counting; use preview for finding a suitable value of the tolerance.

Michael
________________________________________________________________

On 7 May 2010, at 11:21, NatashaW wrote:

> Thank you Michael !! That's what I was looking for..
> This feature works only on gray images or also on RGB images?
>
> another question: I didn't understand if I can define the particles by
> simple thresholding, how this will help me to count them easier?