how do I count the number of identical rows in a matrix?

8 views (last 30 days)
Jeremie on 25 Oct 2012
Commented: reetu hooda on 24 May 2018
I'm trying understand how to apply the huffman coding function to a matrix of the type:
0 22
1 12
0 4
1 1
which is the result of run length encoding on a bi level image. I need to obtain a vector with all the possible characters, an another with the corresponding probabilities for the huffman coding function to accept.
more simply put, I'm trying to find a way to find all the identical rows in this matrix, divide it by the total number of rows, and write the result in a vector.
Thanks very much for your help, and let me know if I didn't follow the appropriate formatting rules or if this question has been covered before - a search did not give any results.
  1 Comment
Matt J
Matt J on 25 Oct 2012
Your example doesn't have any identical rows in it. Show an example that does and what the output should be.

Sign in to comment.

Answers (1)

Kye Taylor
Kye Taylor on 25 Oct 2012
Let the N-by-2 matrix you're working with be denoted A.
Then, you can get a vector of size M-by-2 containing the unique rows with the command
uRows = unique(A,'rows');
Compute your empirical probability of each unique row with the commands
a = mat2cell(uRows,ones(1,size(uRows,1)),2); % convert to cell for cellfun
% function handle that takes a 1x2 row vector as input
% and returns the number of matching rows in A.
g = @(row)nnz(ismember(A,row,'rows'));
% the probability
puRows = cellfun(g,a)/size(A,1);

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!