Local contrast enhancement: gaussian vs. bilateral blurring

I’ve been recently playing a bit with different blurring methods for enhancing the local contrast of an image.

One can find in different places over the internet the suggestion to use bilateral (or “surface”) blurring instead of the usual “unsharp mask at large radii”, so I thought it would be nice to show a “live” comparison with some screencast…

So here it comes: Google Workspace Updates: New community features for Google Chat and an update on Currents
(direct link on youtube: https://www.youtube.com/watch?v=Uj4cmXlezVc)

1 Like

Nice tutorial… but i will be honest some explainations why you pick values would have been nice.

Thanks!

Unfortunately, for the moment I have absolutely no way to decently add some speech to the videos… basically I have no quiet moment and/or place during day or evening when I can sit down and speech in the microphone :wink:

Concerning the affect of the filter parameters, in fact I still have to better figure out myself what is the interplay between the spatial variance and the number of iterations of the bilateral filter, at least in terms of final result on the local contrast.

In the gaussian blur case, the “radius” of the filter defines the minimum scale of the details that will be enhanced by the filter: small values will only enhance fine textures. However, the bilateral blur case seems to be more complicated and I’m still experimenting with it…