Hi,

the problem seems to be connected to a different color intensity of the original image.

You can try to avoid the issue by:

1) adjusting the threshold of grey-scale image

2) adjusting the analyze particles feature, by imposing a lower size limit of the objects to be measured, to remove noise from the image

Thanks Marcello,

I've tried adjusting the threshold but then it makes the analysis uneven, since I don't follow the same steps to achieve the same output.

Does it matter?

The problem is that I can't get consistent results for the same image twice.

I'm looking for a pipeline that process different images the same way (if there's any).

Marcello Alinovi

Hi Daniel,

If the threshold adjustment does not work, what about trying to imposing a lower object size? (that of course is smaller than the smallest observed object).

From the image that you have attached there is a clear distinction between noise and objects, in terms of dimensions.

You can also try to use the command Close- from the binary menu after threshold and then 2/3 times the command dilate followed by erode ( you have to use it the same number of times).

Attached, there is an example of the above mentioned workflow (8-bit->threshold->Close- -> 2x/3x Dilate -> 2x/3x Erode -> Analyze particles

Thanks a lot for the help and attention,

I think I got the idea.

My follow up question for that, is if you think that with this method I can trust the results in independent images?

Sorry for the late reply,

Yes, I think so. As you said, you have to apply the same method to the whole photoset, but if you do so, I think that you'll gel consistent results.

Cheers,

Many thanks for your help.