ABOUT THE PROJECT (Beta / more to come):

During the AI reverse diffusion process, an image is subsequentlyconstructed by a certainly trained model through denoising and digital manipulation. Guided by prompts and probability distribution.

Using my 3D modelled palms depth maps I planted into screenshots taken from google street view at the locations Ed Ruscha provided in his book, I generated all palm trees using ControlNet Stablediffusion. In a first step I used Clip Interrogator to describe the palms. I then further developed my prompt, defined the model, pre-processors and necessary parameters.

The machine ‘perceived’, interpreted, created and halluzinated based on the information it has been given. The compositions are rooted in the palms 1971 locations and how I would compose an image (my screenshots from google street view and the fixed camera position on the google car which refers back to Ed Ruscha and REVISITED. Composing the templates I had to keep in mind that they would have to be analyzed by pre-processors to get good results. The basic visual esthetics were definied by my prompt. A lot of them did not work as planned and I retweaked them until I was happy with the outcome that then was the basis for hundreds of further hours of tweaking, editing, upscaling and refining.

The works in this series and the juxtaposition of evolutionary steps reflect upon the quest for authenticity in the digital age and our data-driven world. They aim to create a thought-provoking exploration at the intersection of humanity and technology as well as appropriation and creation in arts with the help of AI.

CON[DIF]FUSION (VERSE III)

Each branch and leaf meticulously placed.
With unseen hands, this world is traced.
In the lattice of neural threads aligned,
Creating vistas where palm trees unwind.

More information about REVISITED: DOWNLOAD HERE