Density = mass/volume. You appear to want cells/nuclei per unit area. Assuming cells are not too confluent, the method below works on DAPI.

1. File > Open \ Open DAPI image, if DAPI is already grey-scale, skip to step 3

2. Image > Color > Split Channel \ Keep Blue == DAPI

3. Image > Duplicate \ Duplicate image, work on the copy

4. Image > Adjust Threshold \ Try Yen or Triangle w/ Dark background

5. Process>Binary>Watershed \optional, it will divide cells/nuclei that overlap

6. Analyze > Set Measurements \ Area, Mean gray value if you want DAPI signal intensity \Redirect to: (DAPI channel) , not the copy. The copy is a mask to measure the nuclei of the DAPI image.

7. Select the area of interest on the copy or select the entire image using the rectangle or oval selection tool.

8. Analyze > Measure \ Measures the area of interest

9. Analyze > Analyze Particles \ 50-Infinity, 0.05-1.00, Outlines \ Display Results

10. Put your results on a spreadsheet. Total number of particles / total area of interest = cells per area. You can use the scale bar tool to convert pixels to microns, if you have a scale bar in your image.

A similar approach can be used on cell area. You will need good dark-field images or the cytoskeleton or membrane labeled to find the cells.

Density = mass/volume. You appear to want cells/nuclei per unit area. Assuming cells are not too confluent, the method below works on DAPI.

1. File > Open \ Open DAPI image, if DAPI is already grey-scale, skip to step 3

2. Image > Color > Split Channel \ Keep Blue == DAPI

3. Image > Duplicate \ Duplicate image, work on the copy

4. Image > Adjust Threshold \ Try Yen or Triangle w/ Dark background

5. Process>Binary>Watershed \optional, it will divide cells/nuclei that overlap

6. Analyze > Set Measurements \ Area, Mean gray value if you want DAPI signal intensity \Redirect to: (DAPI channel) , not the copy. The copy is a mask to measure the nuclei of the DAPI image.

7. Select the area of interest on the copy or select the entire image using the rectangle or oval selection tool.

8. Analyze > Measure \ Measures the area of interest

9. Analyze > Analyze Particles \ 50-Infinity, 0.05-1.00, Outlines \ Display Results

10. Put your results on a spreadsheet. Total number of particles / total area of interest = cells per area. You can use the scale bar tool to convert pixels to microns, if you have a scale bar in your image.

A similar approach can be used on cell area. You will need good dark-field images or the cytoskeleton or membrane labeled to find the cells.

^this is the easiest way. just by segmentation.

Thank you, Robert, for your detailed answer!

However, I would like to have a parameter for each cell that indicates how dense it is packed. Maybe something like a nearest neighbor distance? Was just wondering if such a parameter exists in cell biology that is universally accepted.

If you add the centroid property during set measurements you will obtain X-Y coordinates for each particle.

Analyze > Set Measurements \Centroid

There is a plugin I tried for nearest neighbor distance :

https://icme.hpc.msstate.edu/mediawiki/index.php/Nearest_Neighbor_Distances_Calculation_with_ImageJ

You will have to download, install, and run it after you generate the coordinate data.

wouldnt it be easier to just have a parameter where the area of cells is divided by the whole area. This way you could also cells which are not clearly segmented count in.

And if you have different cell sizes, you can also set different classes and use the same parameter.

Thanks, Robert, I will try this plugin! EDIT: This worked very well, thanks again!

@Tim: you mean I just set a radius around each cell and calculate the number of cells within this radius/area? Yeah, I was planning to try this out as well.

edit: oh, now I got it. you want to do it for each cell ...

yes you are right you can do that to each cell also.

you can use a segmentation algorithm (like watershed or via treshold marking or edge detection) then you get a binary picture (for example: white is the area with no cell and black the area with cell) then you just divide the "black area" by the "white area". I think this will be the density/package of your measure area. you will get results like its shown in this vid: https://youtu.be/D1qBaFwuF4E?t=275.