G’mic Smooth Abstract Filter Demonstration

A yellow crocus processed with the g'mic smooth abstract filter.

Recently I’ve been exploring the G’mic Smooth Abstract filter. Because I was unable to find documentation for this filter, I created this post.

Controls for the G’mic Smooth Abstract Filter

The five options you can make for this filter are:

  1. Smoothness
  2. Regularization method
  3. Regularization iterations
  4. Geometry
  5. Details

Smoothness

You can adjust smoothness with the slider or type in a value. The range is 0 to 100. As you can see, increasing the smoothness value softens the picture.

In the gallery above I kept all of the settings at their default except for smoothness.

Regularization Method

Select from a drop down menu with three choices: Isometric, Delaunay Guided and Edge Oriented. The Edge Oriented method gives a sharper result. The Delauney Guided gives the softest result, but has emphasis on the bits that radiate. Also, the center of the flower seems crispest.

The gallery shows the different methods for two smoothness values. The number of iterations for all examples is 20.

Regularization Iterations.

Next comes the number of iterations. You can adjust the number of iterations using the slider or type in a value. The range is from 0 to 100, and the image becomes softer with more iterations.

For the examples in this gallery I used the isometric method. At 0 iterations the method doesn matter, because it is not applied.

Geometry

This value appears to control how big the chunks are, and the range is from 0 to 4, with 4 having the biggest chunks.

Details

Unlike the other options, the details slider increases the definition at higher values. As with the other controls, you can either use a slider or type in a value. The range is from 0 to 50, and at 50 you have your original back.

Mixing and matching

I tried to present the controls in a logical way above. With so many controls, and a continuous scale available for most of them, you can get an infinite amount of variations for any photo. Here is a YouTube video that shows how to use the filter:

The best thing to do is of course experiment. When you open the filter the default values are softness of 75, isometric, 20 iterations, geometry = 1 and details =30. Hopefully this post will help you decide where to start depending on the effect you want to achieve.

Getting the tools

Gโ€™mic is open-source software that contains a set of over 500 filters for photo processing. I run it as a plug-in within the GIMP, but I believe that it works with other programs as well. The web site for Gโ€™mic is https://gmic.eu/download.html. If you want to get going with the GIMP check out this post.

3 Replies to “G’mic Smooth Abstract Filter Demonstration”

    1. Good to know. I had trouble figuring out how to present such a complex filter. I’ll go back to the drawing board.