New to ImageJ Analysis


#1

Hello,

I am new to ImageJ and I am doing an internship over summer which requires me to analyse some pictures on ImageJ. I am trying to measure Area, Perimeter etc of stained cells using Analyse Particles option . However, I am not sure how to remove the background colour using the same threshold for each picture so that I can obtain only the cells. This will make my results less biased.

I will appreciate your help!

Thank you,
PN


#2

Welcome to the community!

Analyse Particles requires a binary image, so the step that preceeds this is segmentation: the process of separating objects of interest (your cells) from background.

Have a try at segmenting your images and post back if you run into problems. An example image of the type you are trying to work with will always get more responses!

Good luck!


#4

Thank you for your reply.
Here are some pictures of my work.I need to remove the black background.


#5

@pana17

Are you looking to segment the cells based on the ‘red’ channel then? So… it’s a little difficult to really robustly test this using a .png image - would be best if you could share the original image file (you can always link to a file-sharing site to do so if you cannot upload it directly here). But … let’s just see for now …

In general, your signal-to-noise is not so bad from what I can tell - and you should be able to easily segment those cells. If you wanted - you could add a gaussian blur to ‘get rid’ of those tiny non-specific signals before thresholding. But if you just apply a ‘Li’ or ‘Huang’ threshold via auto-thresholding methods - seems to work well even on the raw image… and then use analyze particles to create your ROIs (just make sure to not select cells touching the edges… so use exclude on edges). You can always set a lower threshold for size exclusion of those tiny background signals…

Here are some helpful links too to help you get started with ImageJ and Segmentation:

I hope this helps! And if you have more questions - just ask. We are here to help!

eta :slight_smile: