0
$\begingroup$

What does it mean when an image looks pixelated after a Gaussian filter (Multivariate Gaussian Distribution filter) is applied (in context of mu, Sigma, and the meshgrid created), and what parameters can one alter to minimize this effect and make the image look smoother?

Sample code (MATLAB):

d1 = 20
d2 = 20
d3 = 20
mu = [0 0 0]; 
Sigma = 30*[1 0 0; 0 1 0; 0 0 1];
[X1,X2,X3] = meshgrid(linspace(-10,10,d1)', linspace(-10,10,d2)', linspace(-10,10,d3)');
X4 = [X1(:) X2(:) X3(:)];
p = mvnpdf(X4, mu, Sigma);
F = reshape(p,d1,d2,d3);

I am convolving the filter above with a 3-D matrix, a roughly 90*60*100 matrix.

$\endgroup$
1
$\begingroup$

An image shouldn't look pixelated after filtered by a Gaussian (lowpass) kernel... instead look at the following figure:

enter image description here

| improve this answer | |
$\endgroup$
0
$\begingroup$

Pixelation often occurs when an image's dimensions are distorted. If the image looks pixelated, it is probably not so much a problem with the Gaussian as it is with the way the image is sized or resized.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.