I have an image I have to threshold before I can create a selection to overlay on another image. I have stained nuclei that are among some other stuff, that also stained, but undesirable. How can I measure the nuclei I want to keep, in order to exclude the other stained particles? and how would I go about it? I have done it manually, by applying a threshold and manually deleting the stuff I don’t want… I am sure there is a way to have them excluded in a set way, based on size? intensity? Thanks!
Sample image…I want to keep only the bright circular spots, and get rid of of the other stained stuff before I create a selection to overlay in another chanel. Any help appreciated it.
H2bMcherryOption2_original(red).zip (2.6 MB)
I played with your image a bit … I found some success just using auto thresholding (using the Threshold tool and clicking ‘Auto’). That is more desirable for you in any case in this analysis - to use threshold automatically.
So - you can do a few things to get you started. Run Auto Threshold and ‘Try All’ and be sure to check the ‘white objects on black background’ option (as for you that is what you have in your image). This will give you a montage of all the method options available for thresholding in ImageJ - so you can compare the resulting masks.
Just by a quick glance… it seems you have a few good options (just depends on exactly what you wan to measure - I’m still not sure what you want in your image): Huang, IsoData, Li, MaxEntropy, Moments, Otsu, RenyiEntropy, or Yen. All of those in auto are looking decent… you can then use that method in auto-mode to generate a mask using Create Mask and then Analyze Particles to exclude those tiny objects you don’t want via size.
Just read up a little bit on Segmentation in ImageJ here. That should lay a good foundation of the available tools to get you started.
Hope this helps!
couldn’t download your example files but anyway…
You might want to have a look at the video from the last ImageJ conference 2015 in Madison. This gives you a stepwise introduction into the basics in image segmentation by thresholding. In most cases it involves some image filtering beforehand. The example chosen specifically deals with nuclei. So this might help you in getting forward with your task.