Segmentation suggestions

trial.tif (358.1 KB)

hello everyone,

I am new at Fiji and I am looking for the best way to segment this picture. It is a porous structure with pores of different sizes. I would like to segment in a way I can see these pores like grains with simple watershed lines. My idea is to make analysis of porosity and pore size distribution.

i have been trying some options like morphological segmentation with very little success. I have zero skills on this.
could some give me some inputs to find the best approach to do it? i would be very thankful.

best regards

trial.tif (358.1 KB)

HI @aranpedro

I get this

with this macro:

requires( "1.53i" );
setBackgroundColor(0, 0, 0);
setOption("BlackBackground", true);
img=getImageID();
//-------------------------------------------
speedup=true;
setBatchMode(speedup);
selectImage(img);
run("Duplicate...", "title=1");
selectImage("1");
run("Enhance Contrast...", "saturated=50 equalize");
run("Bandpass Filter...", "filter_large=40 filter_small=3 suppress=None tolerance=5");
setAutoThreshold("Moments");
//run("Threshold...");
//setThreshold(0, 137);
run("Convert to Mask");
run("Set Measurements...", "area display scientific add redirect=None decimal=9");
run("Analyze Particles...", "  show=Overlay exclude overlay add");
selectImage(img);
roiManager("Show All without labels");
setBatchMode(!speedup);
exit();

A note: a well-centered image with a single slice of bread would have been preferable.

1 Like

Hi Mathew,

thanks so much for your response. I like a lot your result and I really appreciate your help.

Playing around with morphological segmentation I obtained something also enough accurate for my purpose:
trial-overlaid-dams.tif (1.0 MB)

On my next step to perform the “cell” analysis I am working with the watershed lines image. However, I am struggling to find out how to define a black background to differenciate from the cells:
trial-watershed-lines.tif (358.1 KB)

would you have an idea? sorry for abusing of your knowledge…I could definetely eat the whole loaf way before finishing this task…

many thanks!

best