seed0028

This work is a result of StyleGAN2-ADA.

The notebook was forked by Doug Rosman from Derrick Schultz.

The model was trained on Google Colab, with a dataset of about 1000 images of my paintings.

The idea of this work was to experiment with a possible emotional connection between a human and AI.

I used to have a strong interest in painting for more than ten years. At that time, I was painting in the kitchen every day; my paintings were everywhere in that small apartment, on the floor, on the table, there were my paintings in progress everywhere. I‘ve liked oil painting, drawing, tempera. I’ve tried almost all types of supports: different types of paper, wood panels made by different techniques, foundations of various formulas, and various color techniques, from pure handmade pigments to the cheapest children's watercolor- I've experimented with everything. I paint with these colors and expose them to the sun to observe how the color changed. In order to obtain the perfect natural-like texture, I tried to place my painting in the shade of the garden and let it mildew naturally throughout the entire summer. Year after years, my interest in painting suddenly disappeared at a point—perhaps as people say, everything has an end. Then, I began to have a strong interest in ATS arts.
I felt so moved when I saw the work that AI produced based on a dataset of my paintings from previous years -in which I can clearly see my personal art language, the very delicate brush touch. At that moment, years and years of artist life seem to be compressed into a tiny flash moment, a sort of meaningful feeling which probably only I can perceive has appeared from this interstice.

You may also like