# Little issue whith image analysis

Hi everyone,

I am looking for help,
I will be gratefull if this forum can help me ^^

I have several images like that:

I want Image J to calculate the percentage of brown in my image one. Then I will do the same whith the image 2. To comparate my two images.

Can you help me ?

Thank you

Léa

1 Like

I have to add that I am a Numb in Image J ^^

you can try this:

• Split the image into its three colour channels (Menu Image > Color > Split Channels)
• Then, subtract the green channel from the blue channel (Menu Process > Image Calculator )
• Apply a threshold to segment areas with high signal (Menu Image > Adjust Threshold)
• Measure the area above threshold (Menu Analyze > Set Measurements, select “Area”; menu Analyze > Measure)

Cheers,
Robert

1 Like

Ok, thanks you haesl !
If I substract the blue from the green, it look ok. ( if I but the image in 8 bit and the I threshold, do you think it can be ok too ?)

To see the area, i have to clik on Analyse> measure every time I want to calculate the area from a new image ?

And finally, Can you tell me want mean the “Set measurement” > Area ? It mean than Image J calculate the area of everything black on the image ?

Thank you a lot

if the thresholded image resulting from the 8-bit image is ok as well, you can decide on your own. Why not subracting blue from green channel? I mean, you try different ways of the analysis on a handfull of images. Then you make the decision which method is the best one. Finally, in the future you will only apply this method to all your images. So, give it a try and look at your results sceptically. There is no general method which solves all issues best. Try some, then decide for one. And try to use the same method for all images. Otherwise the results may not be usable for scientific research.
To measure something (like the area) on an image, you need to click “Measure”, of course. Alternatively, you can use the “M” key on your keyboard.
The “Set measurements” thingy is needed to configure what to measure. You only need to do that once. And yes, it measures all pixels above the threshold, in the binary image it may be black.
Finally: If you have to process a huge number of images, you may want to automate the whole procedure. Therefore, activate the macro recorder in the menu Plugins > Macro > Record before doing your analysis. You will see that it records all your actions. Later you can rerun the same steps on another image. This requires some programming skills. If you want to go in this direction, this website may be a good starting point:

Have a nice weekend!

Best,
Robert

2 Likes

@haesleinhuepf

Best,

Léa

1 Like

Hi everyone,

I have a request quite similar to the one above, and I would like to confirm whether my approach with ImageJ is correct.
It is for my professional job, it is very important, and I do not know ImageJ or even macro very well.

I will be very grateful for your help…

I have several images like that :
(before) (after)

I am looking for the area of all elements in orange, and the percentage of the total surface it covers.

I went through different tutorials, and forum topics, and I just want to confirm if my procedure for analyzing the area is correct:

1. Set a scale
2. Select the region to analyze by drawing a rectanle comprising the total surface
3. Do a color threshold
4. Analyze > Analyze particules
5. Results > Area and summary

Here is what I get :
(before) (after)

• Are 1, 2, 3…19, …the numeroted particles in black appearing on previous picture ?
• how can I visualize the numeroted particles on my original image ?
• how can I visualize the numeroted particles on my black and white image ?

Here is the summary I get:

Before: count = 718 ; % Area : 35.747
After: count = 872 ; % Area : 21.285

I do not understand:

1. Count
For “after” picture, there is less orange than for “before”, however the summary counts more particules for “after” than for “before”… does ImageJ analyze black particles ?

2. % Area
Is it the % area of my orange particles compared to the entire surface of my sample ? It is in contradiction with my question 1 though…

Am I having the correct result with my procedure: is “% area”, the % of surface covered by orange particules ?

Many thanks,

Alexandra

This looks to be the correct result. You have more particles in the “after”, probably because larger areas in the “before” image were split up. However, on average the particles in the “after” are smaller, leading to a lower percentage coverage. The long list is indeed the area of the individual particles resulting from the particle analyser.
To visualize the particles on ypou original image, select “add to manager” in the particle analyzer. Than after running the particle analyzer, go to image->Overlay->From ROI manager

Hi @aelFR,

just guessing: Is the number of objects in the images really decisive? If not, why not measuring the area of pixels above a threshold in the red channel of the image?

Cheers,
Robert

Ok thank you @rvanthof !
It is all clear to me now.
I will try the way you describe for visualizing the particles.

@haesleinhuepf : thanks ! The number of objects is not decisive but I prefer not to consider pixel, since I have a lot of image to analyze, not having the same colored background (I want to be most precise in my result and just select my particles).

1 Like