Skip to content

The Wavelet Decompose Retouching Method!

February 2, 2009

I stumbled across this great filter for the GIMP, called wavelet decompose.

This is an insanely powerful tool! Breaking down the image into its frequency components means, as the author of the filter states, you can selectively retouch the problem areas while leaving other important details intact. For example, you can get rid of blemishes while leaving details such as skin texture or wrinkles intact!

So here goes, this is my attempt at retouching.

We start with this photo:

And start the Wavelet Decompose filter (or plugin, or whatever you want to call it). You have a choice of the number of wavelet scales (from my limited experience, 8 to 9 works quite well; the maximum number is limited by the size of the image I believe). Then you click OK, and now you have a bunch of new layers with the frequency components of the image. Lower numbers = higher frequency (the details of the image) and higher numbers = lower frequencies of the image (global characteristics).

So here are wavelets #2, 4, and 6, just to demonstrate what each shows. I skipped the odd numbers just to save space, but what you see is what you expect; something in between the other wavelets.

Now comes the fun part. Let’s try to decrease the dark shadows under the eyes, while maintaining the small wrinkles!

What you do is find the wavelet that contains the “data” of the dark shadows, in this case it was a mix of 4, 5, and 6 and basically just get rid of them using the clone tool, smudge tool, paintbrush tool, etc, just as you would to get rid of moles and such in traditional retouching. Here is an example of wavelet #4 after I got rid of the shadows, and a bunch of the red spots on the nose and cheeks; compare this with the original wavelet #4 layer above.

The result? Here you go!

Now we work on the other elements, for example finding the wavelets that contain data for the red spots, and again cloning/smuding/painting over them, and you get:

Notice the details are all preserved? That’s because I left the high frequency data in wavelets #1 and 2 intact!

After a few more things (getting rid of the scattered white spots, decreasing the redness in the eyes), you have this:

BTW I left the upper half of the image unretouched just to have an internal comparison.

Here is an animated GIF with a before and after of a crop:


Summary? I’m totally loving wavelet decompose!!!!! The author mentions some other application for this, which hopefully I will able to try out soon.

No comments yet

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: