These are the results of the A.I. model I trained on my own halftone analog video texture. This is just a prototype and its application is pretty limited. I've experimented with it enough to know its limitations and how I could make it better on the next round of training - I mostly just wanted to prove to myself that this was even possible.
See the 3-hour preparation process (livestream archive) here: ru-vid.comOsUHxgm3c2s
Audio: "Giving In" by Douglas Calii: open.spotify.com/track/5WfOap...
I'm really excited about this because if I continue to train A.I. models on my own stills, I can create whole new never-before-seen textures through the combination of them, and/or by applying the trained models at low weights on top of video edits or when generating new material. I feel that I'm just barely scratching the surface with this.
#glitchart #aiart
23 мар 2024