working with multiple matrix avoiding "for"

I have 253 vertically pasted matrix of size 221x431 (final dimensión of 55913x431). I would like to calculate the coefficient of variation (standard deviation/mean) of each element of the matrix, obtaining a final matrix of size 221x431. I have the following code which works but I would like to know if I could do it in a more simple and efficient way without using "for". Could you give me any suggestion? thank you for your help.

% b -> start matrix, size (55913,431)
for col = 1 : 431
   for r = 1 : 221   
      m =b(r: 221 : 55913,col);
      variation(r,col)= std(m)/mean(m);
   end
end

Answer 1

[Azzi Abdelmalek]

b=rand(55913,431);% b is your matrix
a=permute(reshape(b',431,221,[]),[2 1 3]);
variation=std(a,0,3)./mean(a,3);

I tested the two methods speed

With the for loop : Elapsed time is 7.797265 seconds.
Without loop :      Elapsed time is 0.479849 seconds.

Answer 2

[Andrei Bobrov]

k = 221;

s = size(b,1);
b2 = reshape(b,k,s/k,[]);
variation = squeeze(std(b2,0,2)./mean(b2,2));

Answer 3

[Jose Luis]

Thank you for the answers, they were very helpful! Jose