While I was hunting for royalty free face images for vjing purpose found this archive:
https://archive.org/details/1mFakeFaces uploaded by Alexander Reben with a CC BY-NC 4.0 – non commercial licence. I did not wish any better, since I don’t want to use actual human portraits (I did not want to deal with copyrights issues), and these images were generated by a styleGAN algorithm. These images were my source.
One of my long-term interest as a painter is to not just imitate but put the painting process into an animation (animate itself), showing the flow, energy and time-scale of a borning painting.
Literally generating some motionvectors from the images itself/themselves to drive its creating from within itself (re-creating itself).
During the last decade I slowed down many-many footages, morphed with itself and so on, tried out many techniques to generate the moving force from the image itself, so the current test images I show here is just a result of a slowdown process with a sloppy motinvectors and a little bit of a time-shift and displace within the generated footage.
My aim was that the result image should be accurate and punctial as it could be as a handmade picture, I mean like/as in a traditional way – colours, consistency, luminosity, viscousity, and overall balance of the composition.
Following the European painting tradition – I look for a boundary of a real/ thruth and the gitch/kitch or fake images. This process involves some decunstruction as well, and resulting this glitch-like gitch images, also making a mockery of the tradition as well.
‘Misantrophy’ – random images of a 1145 fake/glitch portraits