I'm new to image processing and I've been testing out different Gaussian blur filters.

Out of curiosity, I generated an object that is a 2D Gaussian with sigma = 1. enter image description here

I convolved this object with a Gaussian of sigma = 1, and while their individual pixel values are different they look very nearly the same. I then changed the sigma of my Object to 5 (in both x and y directions) and found that convolution with a Gaussian of sigma = 1 gave the same result. Is there something deep going on here, or is a Gaussian of sigma = 1 just very weak?

  • $\begingroup$ Please, add the image file as attachment. Did you tried to code? $\endgroup$ – Jose Enrique Calderon Nov 26 '18 at 3:01
  • 1
    $\begingroup$ I suggest that you try the same thing in one dimension first. Then you may see what's happening more clearly. $\endgroup$ – bill s Nov 26 '18 at 4:28
  • 3
    $\begingroup$ Basically, a Gaussian with sigma=5 convolved with a Gaussian with sigma=1 equals a Gaussian with sigma=Sqrt[5^2 + 1^2] ~= 5.1. So they should look similar. You can find a proof here: en.wikipedia.org/wiki/… $\endgroup$ – Niki Estner Nov 26 '18 at 7:35

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.