# Image Approximation (imshow) trouble -- I believe I'm doing Image Compression??

14 views (last 30 days)
Jessica on 23 Feb 2024
Commented: Jessica on 26 Feb 2024
I have been asked to do the following (which I've done successfully):
Here is the beginning of what I entered, and I'm attaching a PDF for the rest.
I get confused when it gets to the second set of directions where I am to show an "image approximation". After all I've researched, again never have coded before, I cannot seem to enter anything correct because all I get is a white image. I used this help page to try to get me on the correct path: https://www.mathworks.com/help/matlab/math/image-compression-with-low-rank-svd.html but have had no luck. Here are the 2nd set of directions specifically that I'm struggling with. The math part of all the coding, I'm golden! It's the image part, and displaying that image where I'm getting the 1st image to display properly, but not the compressed/approximated images to come up as anything other than white.
Thank you!
DGM on 24 Feb 2024
As to why the image is white, it's probably because your floating-point images are improperly scaled for their class because of the manner in which you handled them or cast them.

Karl on 23 Feb 2024
Edited: Karl on 24 Feb 2024
The code below builds on code given by Torsten in relation to a related question: https://mathworks.com/matlabcentral/answers/2084643-how-to-do-a-rank-1-approximation.
% Read test image, which has data type uint8.
imshow(im1)
title(['original image']);
% Set rank of approximation.
k = 10;
% Convert to and from type double.
% The svd() function used in rank_k_approximation()
% requires data of type single or double.
im2 = uint8(rank_k_approximation(double(im1), k));
imshow(im2)
title(['rank-',num2str(k),' approximation']);
function Ak = rank_k_approximation(A, k)
%RANK_K_APPROXIMATION Calculate rank-k approximation of matrix A.
% The matrix A must be of type single or double
% (required by the function svd).
[U, S, V] = svd(A);
Ak = zeros(size(A));
for i=[1:k]
Ak = Ak + S(i,i)*U(:,i)*V(:,i).';
end
end
##### 3 CommentsShow 1 older commentHide 1 older comment
Karl on 25 Feb 2024
Edited: Karl on 25 Feb 2024
If you post the syntax errors that you're seeing, and the code that produced them, someone might be able to suggest changes for getting things to work with a MATLAB release from 2006.
In case it could be useful, MathWorks offers a free version of MATLAB Online:
Jessica on 26 Feb 2024
Thank you, my Professor helped me enough where I'm not going to fail anyways.

### Categories

Find more on Image Preview and Device Configuration in Help Center and File Exchange

R2006b

### Community Treasure Hunt

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

Start Hunting!