CODEWITHSUNDEEP
matlabmatrixcells
Pedro Marques
Pedro Marques

Reputation: 175

Operations in cells

I have a cell 5x5 and each cell has 100x100 single data. I want to calculate the mode in each matrix 100x100 and then do some operations that depends on the mode. How Can I do this?

The cell: enter image description here

Each cell has 100x100 data like this: enter image description here

I have this code that calculate the mode on each cell, now I want to compare each value on 100x100 matrix with corresponding cell mode.

I = imread('DSM.tif');
c = mat2cell(I, [100,100,100,100,100], [100,100,100,100,100])
for i=1:5
    for j=1:5
        mode_cell = mode(c{i,j}(:))
    end
end

I did this code: 

modes = cellfun(@(x) mode(x(:)), c, 'UniformOutput', false);
modes = cell2mat(cellfun(@(x) mode(x(:)), c, 'UniformOutput', false));
for i = 1 :5
for j =1 :5
    for i2=1 :100
        for j2=1 :100
            cell = c{i,j};
            if cell(i2,j2)<modes(i,j)
                teste(i,j)=0;
            else
                teste(i,j)=1;
            end
        end
    end
end
end

But with this code the matrix teste is only 100x100. And I want to attach all testes matrix, and in the end I want an 500x500 matrix. How can I do this?

Upvotes: 2

Views: 67

Answers (1)

dfrib
dfrib

Reputation: 73166

Edit: Updated answer as per the update of your question

Condensed solution (+):

%// data
I = imread('DSM.tif');
c = mat2cell(I, [100,100,100,100,100], [100,100,100,100,100])
    %// 5x5 cell matrix with 100x100 data entries

%// solution
teste = cell2mat(cellfun(@(x) bsxfun(@ge, x, mode(x(:))), c, 'UniformOutput', false));
    %// 500x500 logical matrix

Explanation:

As previously (see the previous answer below), make use of the cellfun command, but include the 'UniformOutput', false' name-value pair such that the resulting modes is a 5x5 cell matrix (rather than a value matrix). After computing the modes, use a subsequent cellfun call to compute the logicals as per your specifications.

%// data
I = imread('DSM.tif');
c = mat2cell(I, [100,100,100,100,100], [100,100,100,100,100])
    %// 5x5 cell matrix with 100x100 data entries

%// modes: 5x5 cell matrix with single value entries (mode)
modes = cellfun(@(x) mode(x(:)), c, 'UniformOutput', false);

%// teste: 5x5 cell matrix with 100x100 logical (0/1) entries 
teste = cellfun(@ge, c, modes, 'UniformOutput', false);

The logicals in teste describes, for each value in each 100x100 data block, if the value is less than the mode value for the corresponding block:

0 : values in block less than block mode,
1 : values greater or equal to block mode.

Finally, if you want to convert the 5x5 cell array teste (with each cell holding 100x100 logical entries) to a single 500x500 logical matrix, you can make use of the cell2mat command, applied on cell matrix teste:

testeSingleValueMatrix = cell2mat(teste)
    %// 500x500 logical matrix

Now, by including a bsxfun command in a single cellfun, we can boil down the above to one single line; the condensed solution given in (+) above, namely

teste = cell2mat(cellfun(@(x) bsxfun(@ge, x, mode(x(:))), c, 'UniformOutput', false));
    %// 500x500 logical matrix

Answer previous to the edit of the original question (computing modes and assign into a value matrix)

You can make use of cellfun command

myCellMatrix = ... %// 5x5 cell matrix with 100x100 data entries
modes = cellfun(@(x) mode(x(:)), myCellMatrix);

Output will be a 5x5 value matrix containing the mode for each 100x100 data set.

Upvotes: 4

Related Questions