How can the objects touching the boundary of the ROI polygon in an image be removed in Matlab?
7 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
Is there a way that one can define a new image boundary to concide with the boundary of the polygon obtained using roipoly. I would like to use the imclearborder function to remove objects touching the boundary of the polygon. Can such objects be removed and how in Matlab?
0 comentarios
Respuestas (2)
DGM
el 27 de Oct. de 2024 a las 8:48
You might want to use imclearborder(), but that's not what imclearborder() does. If you have logical masks, then:
% say you have a logical mask
workingmask = imread('cmanpolyblobs.png')>128;
imshow(workingmask)
% and pretend you created a mask
% using roipoly() or drawpolygon()
polygonmask = imread('cmantifmk.png')>128;
imshow(polygonmask)
% get rid of blobs which touch only the border of the polygon mask
L = bwlabel(workingmask);
selectedL = unique(L(bwperim(polygonmask)));
outmask = ~ismember(L,selectedL);
% visualize the three masks as a single image
% only one blob remains in outmask.
% even though it intersects with the polygon,
% it does not intersect with its border.
testpict = double(cat(3,workingmask,polygonmask,outmask));
imshow(testpict)
0 comentarios
Image Analyst
el 27 de Oct. de 2024 a las 13:43
Editada: Image Analyst
el 27 de Oct. de 2024 a las 13:51
"define a new image boundary to concide with the boundary of the polygon obtained using roipoly" <== To get a binary image mask from roipoly you can do
[x, y] = roipoly();
polygonMask = poly2mask(x, y, rows, columns);
This is a new, separate image.
"remove objects touching the boundary of the polygon" <== To remove other objects touching that mask, you can OR the mask into your other image
mask2 = polygonMask | yourOtherMask;
Then you can label the image and use ismember to remove all blobs with the label number of the polygon mask, as @DGM showed you.
Or you could use bwselect to find the polygon and touching blobs and remove then by indexing, something like (untested):
[r, c] = find(polygonMask); % Get all locations of blob pixels
mask2 = polygonMask | yourOtherMask; % Add polygon mask to your other blobs mask.
touchingBlobs = bwselect(mask2, r(1), c(1)); % This is only the polygon and blobs that touch it.
% Erase touching blobs from mask2:
mask2(touchingBlobs) = false; % mask2 is now the final image.
0 comentarios
Ver también
Productos
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!