Interpolating matrices of different sizes to the same size

Hello, I apologize if this has been asked before, but my search has been futile so far. I am working with several models, and they all have different resolution. Therefore my data is in different size arrays. To continue with my analysis, I simply need the to be the same size. For example:

Model_1 is 60X128

Model_2 is 96X192

Model_3 is 145X192

etc.

Ideally, I would like each array to be 60X128. Although this seems simple, I am unable to come up with some "elegant" way of coding this. I apologize if this post is confusing, because I am still learning jargon of these methods. Any help will be greatly appreciated and thanks in advance.

~Eliott

[Swaminath VENKATESWARAN]

Hello,

Can you share the code for interp2? I have the following data and I am unable to fit them.

Model 1- Each Size of 1020x1

Model 2- Each Size of 1220x1

Model 3- Each Size of 1750x1

Thanks in advance.

Answer 1

[Sven]

Hi Eliott,

Do you have the image processing toolbox? If so, this will be quite easy:

targetSize = [60 128];
Model_2_resized = imresize(Model_2, targetSize);
Model_3_resized = imresize(Model_3, targetSize);

If not, you can instead use interp2. It will let you resample the images at a given set of locations. Let me know if the above works or not, and I can help you out with the code for the interp2 command.

-----==== Answer edited below ====----

Hi Eliott, this code should resize a source image (ie, any X-by-Y matrix) to a target size. It's essentially what imresize() does:

targetSize = [60 128];
sourceIm = double(imread('rice.png')); % you can of course use your own source
sourceSize = size(sourceIm);
[X_samples,Y_samples] = meshgrid(linspace(1,sourceSize(2),targetSize(2)), linspace(1,sourceSize(1),targetSize(1)));
source_resized_to_target_size = interp2(sourceIm, X_samples, Y_samples);
figure, imagesc(sourceIm)
figure, imagesc(source_resized_to_target_size) % note the change in pixels

Did it work for you?

Thanks, Sven.

[wefoust]

Hi Sven, Unfortunately, I don't have access to the image processing toolbox. However, I've been playing with the interp commands with Maltab for a while. I understand the examples, but I am having trouble applying them to my data. Thanks for your help, and I welcome any suggestions/ideas you have.

~Eliott

[Sven]

Hi Eliott, I've updated the answer with interp2 and no image processing toolbox.

[wefoust]

This works well! I originally started playing with the code, and gave up with interp2. I then used interp1, but the code was awfully slow (unusable for the large datasets I have). However, this is exactly what I need. Thanks for following up.

~Eliott