Due to computational limitations, I need to downsample the 3D (1500x1500x1700) matrix by a factor of 2. The binning process should merg the 2x2x2 voxels and assign the mean intensity value of the group to the corresponding voxel in the reduced image.

 Akzeptierte Antwort

Jan
Jan am 24 Jul. 2022

1 Stimme

X = rand(1500, 1500, 1700); % 28.5 GB - does not run in the forum!
Y = reshape(X, 2, 750, 2, 750, 2, 850);
Z = squeeze(sum(Y, [1,3,5])) / 8;
size(Z)

4 Kommentare

Rizwan Khan
Rizwan Khan am 24 Jul. 2022
thanks Jan for your quick reply. I am getting an error if the dimension is odd.
One of my matrix is 1212x1212x1561 and the error is: "Size arguments must be real integers"
Jan
Jan am 25 Jul. 2022
Bearbeitet: Jan am 25 Jul. 2022
@Rizwan Khan: What do you wnt to do, if the dimension is odd? Omit the last vector or include it in the output without averaging? I cannot guess, what you need.
X = rand(1501, 1501, 1701);
s = size(X);
s = s - rem(s, 2);
XX = X(1:s(1), 1:s(2), 1:s(3)); % Crop, Needs a lot of memory...
Y = reshape(X, 2, s(1) / 2, 2, s(2) / 2, 2, s(3) / 2);
Z = squeeze(sum(Y, [1,3,5])) / 8;
Rizwan Khan
Rizwan Khan am 25 Jul. 2022
Yes, I omitted the last vector before downsampling.
thanks Jan
Brilliant Jan. Thank you.

Melden Sie sich an, um zu kommentieren.

Weitere Antworten (0)

Community Treasure Hunt

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

Start Hunting!

Translated by