how to extract objects from an image after filling in the objects in the binary image

I need to delete the objects I have marked below from the binary image but I don't know how to do it can you help

Answer 1

[KALYAN ACHARJYA]

Original BW image

BW_result=bwareafilt(BW,2); % Get the largest 2 max size objects
result=~(BW_result.*BW);  % Nand Logical Operation or do other as per requirements

[betty]

thank you it was very useful :)

[KALYAN ACHARJYA]

Welcome! Please change the logical operation as per your requirements

Answer 2

[Image Analyst]

How do you want to specify the objects? Like you're going to manually, interactively click on them? Or based on their size, shape, or location? What exactly indicates those two particular blobs in the image out of the 6 in there?

I mean one way is

[labeledImage, numShapes] = bwlabel(binaryImage);
extractedTwoShapes = ismember(labeledImage, [3, 5]);

but I doubt that's what you want to do.

[betty]

Thank you, but what I want to do is delete these two objects that I marked from the image. In the code you wrote, only those two objects remain, but they should be deleted and the other objects should remain in the image.

[Image Analyst]

Try

[labeledImage, numShapes] = bwlabel(binaryImage);
labelsToKeep = 1:numShapes;
labelsToKeep[3] = []; % Let's not keep blob #3.
labelsToKeep[5] = []; % Let's not keep blob #5.
extractedTwoShapes = ismember(labeledImage, labelsToKeep);

But you did not say how we're supposed to know that #3 and #5 are the ones we want to delete. Again, what is that based on? Size, shape, location? What??????

[betty]

I'm sorry I forgot to mention. We need to delete it by location.

[Image Analyst]

OK, you can use regionprops() to get both the centroids and bounding boxes. How are you defining the locations of blobs to delete? Like their centroid is in a certain box in your image? Or the edge of your bounding box is in some range?

Answer 3

[yanqi liu]

sir，may be use mask to filter the block，such as

clc; clear all; close all;
im=imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/808264/image.jpeg');
im=imcrop(im,[82   117   519   518]);
bw = im2bw(im);
bt = bw;
bw1 = bwareafilt(bw, 1);
bw2 = bwareafilt(bw, 3);
bw2(bw1) = 0;
bw(logical(bw2)) = 0;
figure; imshow(bt);

figure; imshow(bw);