Segmentation of overlapped aggregate images without a background

Is it possible to segment the visible portion of the particles in images like this (not taken by me, downloaded from the internet) to get binary images like the attached? The particles have different textures, no definitive background, different illumination and massively overlapped. I tried this Matlab answer (https://www.mathworks.com/matlabcentral/answers/2151-separating-objects-in-an-image) and got the attached. When I tried to separate them using bwlabel(), I got something like the attached, along with many useless separations.

If yes, then is it possible to generalize it to work for many similar images ?

I appreciate the help from all the hardworking people from Matlab answers forum. Thank you

Answer 1

[Image Analyst]

It's probably possible, at least to some accuracy. You might look here for algorithms on rocks and gravel or particles: http://www.visionbib.com/bibliography/contents.html

[sumana]

Are you sure that this is a genuine site ? I do not think all the links are. I just tried.