Particle analysis - strange behaviour

[cherkasovn]

Hi. I used ImageJ occasionally over the years. Always Particle analysis and always it was a breeze. Today I got it again and cannot analyse particles - the outlines look totally strange and are mostly concentrated on edges. I tried Linux and Windows version; re-checked a youtube guide - all seems right.
I get to black dots on white image, press analyse particles and it shows outlines only for noise, not the actual particles; mostly at the edges somehow. The Browser version behaves well.
The image below shows an example - right is my image to analysed.
Left - the black&white image, analysis settings (I adjusted all settings to no avail), the right are the outlines with absence of most particles at the centre.

Am I getting crazy?

[rasband]

Before running the particle analyzer, try setting a threshold using the Image>Adjust>Threshold tool. If that doesn’t work, upload an original image so we can reproduce the problem.

[cherkasovn]

The files are attached. The exact details do not matter. I process the color image, convert to 16 bit grey, then threshold (attached). The outlines do not match particles even remotely - they are only 1-2-3 pixel pieces somehow.

](url)

[rasband]

The particle analyzer works as expected when I convert
the image to 8-bit, subtract the background and set a threshold before running it.

Here’s a macro that does this:

run("8-bit");
run("Subtract Background...", "rolling=50 light");
setAutoThreshold("Otsu");
run("Analyze Particles...", "size=50-Infinity display clear overlay");

[cherkasovn]

Thanks.... but it only somewhat works.
Scenario 1 - I run the macro as you write and it is perfect.
Below is the screenshot with ~1500 particles and reasonable outlines.

I do all the same steps manually. (Note that the checkbox on AutoThreshold says "white objects on black" which likely means black on white; but does not matter). The result black-white looks as before and is perfect.
I click Analyse particles as before with the same settings... and it looks strange somehow - see below. Only 3 particles. Do I completely miss something?
Thanks again!

[mutterer]

Hi
The manual step equivalent to
setAutoThreshold("Otsu");
is choosing Image>Adjust>Threshold... and choosing "Otsu" as thresholding method.
If I do that, I get the expected result.

[cherkasovn]

The Threshold works perfectly, it is the "Analyze Particles" in the screenshots that does not work via the dialog window somehow.