Glitch in 'include holes' option of Analyse particles?

[Teresa W]

I have an irregular foam (readily thresholded). When I apply Analyze particles, ImageJ can readily differentiate and size any pores that are formed by the pore wall that is in continuous contact with the edge of the ROI. However, any pores formed from internal fragments of pore wall not attached directly to  the ROI are invisible, even if I choose the 'include holes' option. Is this a fundamental limitation of the Analyze particles function, or is it a glitch in the 'include holes' function? I tried this months ago, and again recently. Thanks Teresa

[Michael Schmid]

Hi Theresa,

maybe I don't understand you correctly, but it seems to me that you want
to keep inner holes as they are (as holes, not becoming part of the
measurement). So, 'include holes' should be off.
Of course, if you want to measure the pores, you should threshold the
pores, not the walls (you should see red pores and gray walls) - your mail
is unclear to me in this respect.

Michael

______________________________________________________________________
On Sun, January 31, 2010 08:30, Teresa W wrote:
> I have an irregular foam (readily thresholded). When I apply Analyze
> particles, ImageJ can readily differentiate and size any pores that are
> formed by the pore wall that is in continuous contact with the edge of the
> ROI. However, any pores formed from internal fragments of pore wall not
> attached directly to  the ROI are invisible, even if I choose the 'include
> holes' option. Is this a fundamental limitation of the Analyze particles
> function, or is it a glitch in the 'include holes' function? I tried this
> months ago, and again recently. Thanks Teresa

[Teresa W]

Hello Michael. See my earlier entry a few posts ago which includes a raw data slice of my irregularly shaped foam.

Since my post I have been reading back a bit more (Russ, J. e-book CRC Press "Image analysis of food structure"). From his information in Ch 4. Russ desceribes a Particle recognition function as inherently limited in its ability to identify internal structures, as it operates at a pixel level. Feature, as opposed to particle, recognition is required for my analysis.

The reason for my question was really whether the "include holes" option expands ImageJ Particle recognition to Feature recognition in some way. And whether it is working as expected or not, as when I use it on my slices, there is no difference between my slice outputs whether I choose "include holes" or not.

I see there is an ImageJ Plug-in, FeatureJ (http://www.imagescience.org/meijering/software/featurej/) from the Biomedical Imaging Group of the Swiss Federal Institute of Technology. Next stage for me, try this out.

To make some general comments about ImageJ, I have found it very hard to apply at the entry level. I have spent some months wandering in every now and then and trying something out with no luck. I recommend the Russ book as an excellent readable and practical introduction for those with zilch experience to digital analysis methods for material analysis.  I also recommend the Particle recognition video tutorial at the http://imagejdocu.tudor.lu/ site. The little tricks in this e.g. tweaking the thresholding to reveal data on what looks like an empty output view (Ahah! That explains a lot!) helped me immensely.