Removing objects in an image in MATLAB using Image Processing Toolbox

How can I remove the seven (7) big cells in the picture below so that I can do an accurate count of the smaller cells in the background?

[Towfeeq Fairooz]

While looking for solution to my problem, I came across similar problem as yours. You may want to have a look at it and try to adapt as per your needs (thats what experts here advice). I am just a newbie here :)

https://uk.mathworks.com/matlabcentral/answers/386250-how-can-i-remove-the-background-this-image-i-do-thresholding-but-i-can-t-remove-the-background-i?s_tid=srchtitle

Good luck with that.

Answer 1

[Tala]

I am not sure removing the 7 big objects would affect your accuracy in counting the smaller ones. I think having connected components makes your problem a bit difficult. If that is not an issue, you could use morphological operations to count the blobs...maybe something lile this:

I=rgb2gray(YourImage);
Ie = imerode(I,se);
Iobr = imreconstruct(Ie,I);
BW = imbinarize(Iobr,'adaptive','Sensitivity',0.7);
Bw2=imcomplement(BW);
BW3 = imfill(BW2,'holes');
BW4 = bwareaopen(BW3, 500);
vislabels(BW4) % this should be available online. I downloaded it several years ago

and you will get this:

If you want to disconnect the blobs (for example object 17 is actually 3 objects), I suggest exploring watershed filtering.