Preserve the label information when extracing boundaries from label matrix

Hello, I would like to get the boundary labels when extracting boundaries from label matrix.

Right now, I understand that I can generate boundary masks from the following function:

bm = boundarymask(labelmatrix), however, the generated masks are binary, that is to say, I wouldn't know which mask correspond to what label. I am wondering if there is a way to preserve such information, so that I can know which boundary is for label 1, 2, 3..., etc.

Thanks in advance!

Answer 1

[DGM]

You can do something like this:

A = imread('blobs.png');

L = bwlabel(A);
B = L.*boundarymask(L); % multiply

[min(L(:)) max(L(:))] % label array has 28 blobs

ans = 1×2

0 28

[min(B(:)) max(B(:))] % boundary array also has 28 labeled blob boundaries

ans = 1×2

0 28

Answer 2

[yanqi liu]

clc; clear all; close all;
A = imread('kobi.png');
L = superpixels(A,100);
% label id
inds = sort(unique(L(:)));
% for every label id
mask = boundarymask(L);
figure
imshow(mask)
hold on;
for i = 1:length(inds)
    maski = boundarymask(L==inds(i));
    % display every label by different color
    [ri,ci] = find(maski);
    plot(ci,ri,'.','Color',rand(1,3))
end