Making Generative Portraits Utilizing Processing

Advertisements

[ad_1]

Hello everybody! ? Haven’t talked to you guys for some time. Hope issues are going effectively in your finish. Final semester I used to be learning overseas in Budapest however now I’m again at Colgate. I had the time of my life touring throughout Europe and assembly/staying with fascinating individuals. I’ll attempt to write an article about that quickly. Until then sit tight and luxuriate in this text about generative artwork.

I’m taking a sophisticated studio tasks class at Colgate. For this class, I had the chance to choose no matter medium I needed to concentrate on. I used to be barely confused between selecting up drawing (scribble artwork) and digital (generative) artwork. I’ve been specializing in each for some time now so it was arduous to determine which one I needed to spend my entire semester on. The anxiousness was actual! You’ll be able to see a few of my scribble artwork and generative artwork beneath.

Scribble Art

As an alternative of deciding to choose up one in every of these I made a decision to problem myself by selecting each! I requested myself, is it potential to provide you with an algorithm that may generate scribble artwork from a supply picture? Early analysis advised that it was doable and for the primary project I settled on working with generative scribble artwork for this class.

I searched round and located some assets. I haven’t labored with processing an excessive amount of. I’ve spent solely a month or two at max exploring this system so it was a pleasant alternative to get to know this system higher as effectively.

A helpful useful resource was this. Jason shared high-level particulars for the algorithm he was utilizing to make portraits like this:

Stirman Portrait

The opposite evening I sat down and determined to provide you with an algorithm for this (impressed by Jason’s algorithm) and that is the output I bought:

Source Image

Output Image

In case you re-run the algorithm it’s going to produce a barely totally different output (that is the place the time period generative is available in):

My algorithm varies barely from Jason’s algorithm in the best way it generates scribbles exterior the principle portrait. I like these sorts of scribbles so I explicitly tried so as to add them in at random areas.

That is simply the very primary model of the algorithm. There are a bunch of issues I wish to change:

  1. Be sure that there are fewer scribbles over “white” areas within the supply picture
  2. Range the burden of the scribbles
  3. Perhaps strive totally different colours
  4. Scale back the entire variety of scribbles

By the tip of this semester, I additionally plan on making a plotter in order that I can add extra “bodily” contact to my generative artwork items. At the moment, it feels very indifferent from me as an artist and I believe making a plotter will assist change that and also will indulge the “engineer” facet of me.

When you have any concepts/enhancements/options I’m all ears ? I’ll see you within the subsequent article! ❤️

[ad_2]