diffused reality resynthesis

 

I ran an experiment out of curiosity where i started with a screen capture of an image from Daniel Marcolina's recent Diffused Reality exhibition as the one shot input to a resynthesis algorithm.  Above is the first resynthesized image, which does an interesting take of capturing the feel of his imagery (at least the one i used as the input, minus the blue tints in the original).  I then slowly modulated the latent representation used as the input to the resynthesis until i arrived at below.  Animation below follows that complete progression.  Daniel was an early Studio Artist user and has a long time innovative graphic design business on the east coast.  Interesting to see he is getting into ai art now.