k-means segmentation of an image - matlab

Question

I came across the following code, which segments an image using num no. of clusters, via the k-means clustering algorithm. However, I could not understand the meaning of the second statement in the first for loop. Please help me understand what the statement is doing and also what ~= means(here).
Also, when I run the code, I get the following error:
??? Attempt to grow array along ambiguous dimension.
Error in ==> kmeansseg at 42
color(rgb_label ~= k) = 0;
It seem that I get this error for every value other than num=3. So, does it mean that I cannot cluster an rgb image into more than 3 colors? The input image has 6 colors that I can tell apart. Can someone suggest a solution for this?
Function call:
>> f=imread('peppers.png');
>> kmeansseg(f,6)
Here is the code:
function kmeansseg(im,num)
figure(1),imshow(im), title('original image');
cform = makecform('srgb2lab');
lab_im = applycform(im,cform);
ab = double(lab_im(:,:,2:3));
nrows = size(ab,1);
ncols = size(ab,2);
ab = reshape(ab,nrows*ncols,2);
nColors = num;
[cluster_idx, cluster_center] = kmeans(ab,nColors,'distance','sqEuclidean', ...
'Replicates',3);
pixel_labels = reshape(cluster_idx,nrows,ncols);
figure(2),imshow(pixel_labels,[]), title('image labeled by cluster index');
segmented_images = cell(1,nColors);
rgb_label = repmat(pixel_labels,[1 1 nColors]);
for k = 1:nColors
color = im;
color(rgb_label ~= k) = 0; %meaning?
segmented_images{k} = color;
end
figure(3),imshow(segmented_images{1}), title('objects in cluster 1');
figure(4),imshow(segmented_images{2}), title('objects in cluster 2');
figure(5),imshow(segmented_images{3}), title('objects in cluster 3');
end
end

Answer 1

It is setting to zero any elements of the image that don't correspond to that particular label. That's how you get a series of segmented images. It gets the segregation labels from the rgb_label variable.
What ~= means there is "for every pixel of the segmentation image is NOT equal to the current segmentation number, set the image pixel to zero, leaving the other image pixels unchanged"
In regards to your edit - it looks like the color and rgb_label matrices do not have the same dimensions.