Split 4000x10 Matrix into 400 10x10 Matrices and calculate stats for each matrix?

Hi

Im trying to split a 4000x10 Matrix into 400 10x10 Matrices? Once the matrix is split up I need to calcualte the mean, median and standard deviation for each of these matrices.

Regards Akshay

[Stephen23]

@Akshay J: learn to use the dimensions of arrays (see Geoff Hayes' answer) or cell arrays. Do not try to create 400 individual matrices. Read this to know why magically creating lots of variables is a bad way to write code:

https://www.mathworks.com/matlabcentral/answers/196952-how-to-create-new-variables-in-batches-with-strcat

Answer 1

[Geoff Hayes]

Akshay - try using reshape to split your 4000x10 matrix into a 10x10x400 matrix. For example,

 A = randi(255,4000,10);
 B = reshape(A,10,10,[]);

Then you can iterate over each of the 400 matrices to calculate your statistics

 for k=1:size(B,3)
     mtx = B(:,:,k); % the kth 10x10 matrix
     % calculate the mean and standard deviation
 end

[gooniyath]

Hi Geoff thanks for the help. How do I get it to split up the the data row by row instead of column by column? Cheers

[Stephen23]

@Akshay J: use permute beforehand to swap the columns and rows:

B = permute(A,[2,1,3]);
B = reshape(B,...)

you might want to use permute again after that too.

Answer 2

[Image Analyst]

Akshay, you can use blockproc to process a block of your array (or image) by moving the block, which can be of any size, in "jumps" that you specify. So you have a 4000 row by 10 column matrix. You can move in jumps of 10 rows and then compute the functions you want. Here is the code (requires the Image Processing Toolbox for the blockproc() function):

% Create sample data.
yourMatrix = randi(9, 4000, 10);
% Define how we will move in jumps over the array.
blockSize = [10, 1];

% Define the function that we will apply to each block.
% First in this demo we will take the median value in the block
% and create an equal size block where all elements have the median value.
medianFilterFunction = @(theBlockStructure) median(theBlockStructure.data(:));

% Block process the array to replace every pixel in the 
% 10 pixel by 10 pixel block by the median of the pixels in the block.
blockyImageMedian = blockproc(single(yourMatrix), blockSize, medianFilterFunction); % Works.
[rows, columns] = size(blockyImageMedian)

% Block process the image to replace every pixel in the 
% 10 pixel by 10 pixel block by the mean of the pixels in the block.
% The matrix is 4000 rows by 10 pixels across which will give 4000/10 = 400 blocks.
meanFilterFunction = @(theBlockStructure) mean2(theBlockStructure.data(:));
blockyImageMean = blockproc(yourMatrix, blockSize, meanFilterFunction);
[rows, columns] = size(blockyImageMean)

% Block process the image to replace every pixel in the 
% 10 pixel by 10 pixel block by the standard deviation of the pixels in the block.
StDevFilterFunction = @(theBlockStructure) std(double(theBlockStructure.data(:)));
blockyImageSD = blockproc(yourMatrix, blockSize, StDevFilterFunction);
[rows, columns] = size(blockyImageSD)