For Loop Indexing of matrices

I have a matrix of dimensions 36x25x355.

I would like to extract 36x25 arrays from this data and index them 1,2,3 and so on, as descrtibed in attached picture. I would like to be able to do this for any size matrix multipe times so I need to create a for loop.

As 355 does not divide wholly into 25, I would like the last array to be smaller than the others.

Any help with this woud be greatly appreciated.

[Jan]

I assume "36x25x355" is a typo. You mean 36x355, don't you?

[Aidan Turner]

Apologies. Yes it's a 3 dimenstional NIfTI file, so in 2 dimensions 36x355

Answer 1

[Matt J]

Using mat2tiles

https://www.mathworks.com/matlabcentral/fileexchange/35085-mat2tiles-divide-array-into-equal-sized-sub-arrays

it's as simple as,

A=mat2tiles(yourMatrix,[36,25])

[Aidan Turner]

Thanks for your help Matt,

How could I then extract each cell as it's own array?

[Steven Lord]

While you can extract each cell to an individual variable, that is not recommended. See this answer for a lengthy discussion of why it is not recommended. Wherever you would have used A1 in your code (for example) use A{1} instead.

[Jan]

@Steven Lord: Thanks for this important advice.

[Aidan Turner]

Thank you both for your contributions. My end goal is to sum the induvidual arrays. How would I go about this?

[Matt J]

How can all the arrays be added together if An may be of a different size than the rest?

To sum the first n-1 tiles, you do not need to split them up at all:

sum(  reshape(   yourMatrix(:,1:350),   36,25,[])  ,3 )

[Aidan Turner]

I was intending to populate the remaining space in the final array with zeroes. Really sorry, I'm fairly new to Matlab and this is a steep learning curve for me.

[Aidan Turner]

@Matt J Would there be a way to calculate the max value for each point in the 36x25? For example, I would want the maximum value of the first point out of alll the reshaped matrices.

[Steven Lord]

From this additional information about what you're trying to do, splitting this matrix into cells in a cell array doesn't seem like the best option anymore. Pad your array with columns of missing data at the end until its width is a multiple of the width of the "blocks" you want to create, then reshape the data into a 3-dimensional array.

n = 7;
A = magic(n)

A = 7×7

30 39 48 1 10 19 28 38 47 7 9 18 27 29 46 6 8 17 26 35 37 5 14 16 25 34 36 45 13 15 24 33 42 44 4 21 23 32 41 43 3 12 22 31 40 49 2 11 20

paddedA = [A, NaN(n, 1)]

paddedA = 7×8

30 39 48 1 10 19 28 NaN 38 47 7 9 18 27 29 NaN 46 6 8 17 26 35 37 NaN 5 14 16 25 34 36 45 NaN 13 15 24 33 42 44 4 NaN 21 23 32 41 43 3 12 NaN 22 31 40 49 2 11 20 NaN

B = reshape(paddedA, [n 4 2])

B = 
B(:,:,1) =

    30    39    48     1
    38    47     7     9
    46     6     8    17
     5    14    16    25
    13    15    24    33
    21    23    32    41
    22    31    40    49


B(:,:,2) =

    10    19    28   NaN
    18    27    29   NaN
    26    35    37   NaN
    34    36    45   NaN
    42    44     4   NaN
    43     3    12   NaN
     2    11    20   NaN

Now you can use the dim input argument to functions like sum and max to compute the sum or maximum in a specific dimension.

C = max(B, [], 3)

C = 7×4

30 39 48 1 38 47 29 9 46 35 37 17 34 36 45 25 42 44 24 33 43 23 32 41 22 31 40 49

[Matt J]

Would there be a way to calculate the max value for each point in the 36x25? For example, I would want the maximum value of the first point out of alll the reshaped matrices.

Yes, max, min, mean, median... Whatever you like:

max(  reshape(   yourMatrix(:,1:350),   36,25,[])  ,[],3 )

Answer 2

[Jan]

X  = rand(36, 355);
w  = 25;              % Width of the tiles

sX = size(X);         
M  = [repmat(w, 1, floor(sX(2) / w)), rem(sX(2), w)];
M  = M(M ~= 0);   % Crop last tile, if it is 0
C  = mat2cell(X, sX(1), M)