why did i get this error "Error using == Matrix dimensions must agree " ?? with MATLAB 2013a while not with MATLAB 2010b

when i am using this function: -------------- function [Am Al Amm] = regionadjacency(L)

    % Identify the unique labels in the image, excluding 0 as a label.
    %labels = setdiff(unique(L(:))',0);
    labels=unique(L(:))';
    if isempty(labels)
        warning('There are no objects in the image')
        Am = [];
        Al = {};
        return
    end

    Am = zeros(max(labels));   % Allocate adjacency matrix
    Amm = Am;    
    Al = cell(1, max(labels)); % and adjacency list.

    % Strategy: Dilate each labeled region and use that as a mask on the
    % original labeled image.  This extracts the original region plus a one
    % pixel wide section of any adjacent regions.  This ends up in the image 'r'
    % in the code below. We then find the unique labels in 'r' ensuring the
    % label of the original region and 0 are not included in the label list.
    % This forms a list of region labels that are adjacent to the original
    % region.  We then set the entries in the adjacency list and adjacency
    % matrix accordingly.

    for n = labels
        r = zeros(size(L));
        r = L(imdilate(L==n, ones(3)));
        Al{n} = setdiff(unique(r(:)), [n 0]);
        Am(n, Al{n}) = 1;

        % Experimental: Get strength of 'adjacency' by counting No of pixels in
        % neighbouring region that is adjacent
        for m = Al{n}
            Amm(n,m) = sum(r(:) == m);
        end

    end
------------
i get this error "Error using  ==  Matrix dimensions must agree" for this line " Amm(n,m) = sum(r(:) == m);" when i am using MATLAB 2013a but not with 2010b, so what should i do especially i need to work with 2013a??

Answer 1

[Andy]

The only thing that makes sense is that in 2010b,

>>setdiff(unique(r(:)), [n 0]);

outputs a row vector, whereas in 2013a, it outputs a column vector. Try looking there

[Doaa Adel]

thank u very much, yes, you are right. now, it works