After converting all the 24 matrices of 32x30 into 24 vectors of 960 elements each, take each vector of 960 elements and sort them.
Suppose,
A = matrix of 32x30x24
for i = 1:24
EA = A(:,:,i);
v1 = EA(:); v2 = reshape(EA,1,[]); v3 = reshape(EA,numel(EA),1); % Convert 2D mat. to vector
u = unique(v3);
n = histc(v3,u);
[n,idx] = sort(n);
freqn = v3(idx); % Most frequenting nos.
myNos(i) = freqn;
end
I am not sure, this would help or not, but surely the logic here will work to find the desired modes from the tensor (3D matrix), stored in myNos.
You may able to compare each no.
