Inaccurate circularity measurements

Hi all, I’ve been having a lot of trouble getting accurate circularity measurements with ImageJ/Fiji. I’m trying to measure the circularity of organoids cultured in different gels to compare how effective they are at encouraging acinar formation, but the values I am getting are not reflective of how circular they actually are as organoids in Matrigel (my positive control) should have circularity values greater than 0.035!

When I take the measurements, I open up TIFF files, set the scale (distance in pixels = 1; known distance = 1.9; pixel aspect ratio = 1, unit = micron) and use the freehand line tool to trace the perimeter of the organoid. The measurements I have set include area, perimeter and shape descriptors.

It seems to be an issue with my software, as a colleague kindly analysed my images (with the same scale and measurement settings I use) and got much better results (0.8 - 0.9 for the same organoids). However, I have no idea what exactly could be wrong and none of my colleagues have any clue what the issue is either.

If anyone has any ideas, advice or solutions, please comment below!
Many thanks!!

EDIT: Thank you so much for responding @gabriel! I have uploaded a screenshot of one of my attempts to measure circularity and the raw TIFF file (please see below). Here, I opened the TIFF file, set the scale of the image to 0.5263 pixels/micron and then used the freehand line tool to trace around the boundary of the organoid before taking the measurement. Area, perimeter and shape descriptor measurements were all selected under ‘Analyse > Set Measurements…’. I haven’t used any macro to calculate these measurements as I am not code-literature, although I definitely would be willing to learn how to!

circularity imagej d12|574x500
img_000000000_Off_000.tif (1.3 MB)

Post an image and the code/macro that you are using for others to try.

Hi Gabriel, I’ve just learnt that using the freehand selection tool rather than the freehand line tool to trace around the organoids gives much more accurate circularity measurements! I’m not sure why this is, since both the tools seem to do the same thing on the surface. I am super relieved that there is nothing wrong with my software but a bit puzzled as to why I have gotten such different results using tools that seem interchangeable - do you know what the difference is between the two?

Thank you so much for responding - I hope I haven’t wasted your time! I’ll edit my post ASAP with the solution in the event that anyone else encounters the same problem :smiley: