"Analyse Particles" Outline Thickness for Grain Boundary Analysis

[Subscribe Imagej Rkaur]

Hi All!

I'm working on analyzing grains from specific micrographs but when I use the "Analyse Particles" function to look at the area, it takes into account the thickness of the grain boundaries in the image and so gives a reduced grain area.

I've tried to use the Minimum/Maximum filters to change the thickness, but if I reduce the thickness of the boundaries too much, the software doesn't pick up the boundaries at all when using the Analyze tool.

I'm currently analyzing hand traced images (using a fine tip marker) but still can't seem to find any other way to reduce the boundary thickness so the software can give the most accurate reading for the grain area.

Any suggestions would be greatly appreciated!

Thank-you!

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Michael Schmid]

Hi Rkaur (?),

if there are no voids between grains in your images, it seems you can do
with a "dilate - no merge" step.  There is no built-in command with this
name; it is done by Process>Binary>Voronoi. You just have to threshold
the result with lower and upper limits of 0.

Here is an example macro, reducing the very large distance between the
blobs (which would be your grains) to single-pixel width:

run("Blobs (25K)");
setAutoThreshold("Default");
setOption("BlackBackground", false);
run("Convert to Mask"); //the input image, with huge gaps
run("Voronoi");
setThreshold(0, 0);
run("Convert to Mask");

Michael
________________________________________________________________
On 2016-06-07 17:54, Subscribe Imagej Rkaur wrote:

> Hi All!
>
> I'm working on analyzing grains from specific micrographs but when I
> use the "Analyse Particles" function to look at the area, it takes
> into account the thickness of the grain boundaries in the image and
> so gives a reduced grain area.
>
> I've tried to use the Minimum/Maximum filters to change the
> thickness, but if I reduce the thickness of the boundaries too much,
> the software doesn't pick up the boundaries at all when using the
> Analyze tool.
>
> I'm currently analyzing hand traced images (using a fine tip marker)
> but still can't seem to find any other way to reduce the boundary
> thickness so the software can give the most accurate reading for the
> grain area.
>
> Any suggestions would be greatly appreciated!
>
> Thank-you!

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Subscribe Imagej Rkaur]

Hi Michael!

     Thank you for the suggestion. I tried a version of this with one of my simpler images and it seems like the software has been able to detect 98% of the image area (as being a part of the grain area). This was a much better result than what I got before (with only 94% of the image area being detected). I will give this a shot with a few other images and hopefully the outcome is the same.

Thank you so much once again!

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html