I've been using a Gaussian Filter and know how to use it but I don't understand the physical significance behind using the filter. We use it but why? What is the significance? I'm just starting out in image processing, so please no high level responses. Thanks!

  • $\begingroup$ What type of filter? LP,HP, BP,HR? $\endgroup$
    – Adel Bibi
    May 31, 2014 at 11:29
  • $\begingroup$ Not sure what your comment means, but I'm told is for a Gaussian Smoothing Filter which tries to mimic the shape of a 2D Gaussian Bell function. Sorry I'm super noob here... I'm just new to it all $\endgroup$
    – user481610
    May 31, 2014 at 11:41
  • $\begingroup$ Possible duplicate of this answer $\endgroup$ May 31, 2014 at 22:14
  • $\begingroup$ dsp.stackexchange.com/questions/40872 $\endgroup$
    – Mark
    Apr 1, 2021 at 8:09

1 Answer 1


As far as your comment suggests, you are using a LP Gaussian Filter. LP = Low pass. This means it passes the low frequency component of the image, and prevents the high frequency component.

In image processing, you can imagine the high frequency component as the places in the image where you have a sudden jump in pixel values. Like moving from a black background and suddenly to white one.

These components are associated with edge detection normally, but at the same type they are associated with noise. It's because noise has a high frequency. How come? There is for example a type of noise known as "Salt and Pepper" where there appears spikes in the image. You see the it as a weird dot in the image. These normally have way different value than the surrounding.

With this sort of a filter (Low Pass Gaussian Filter) you can suppress the noise, but you will lose some sharpness in the image due to killing some parts of the edges. Therefore; this is called smoothing. It's because you tend to make the image smooth in a way that pixel values are not changing rapidly but rather in an elegant way.

You can imagine this as averaging. It's because you are substituting the values in a certain region of the image by it's average. That will help in smoothing the image.

  • $\begingroup$ @user12345 No problem! $\endgroup$
    – Adel Bibi
    May 31, 2014 at 12:04

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.