The VQGAN programs all started with this first image,
but none produced anything I liked after that
HOLOGRAM
UPDATE: (October 6): Adric brought a hologram display unit to Quelab makerspace last night -- the LOOKING GLASS FACTORY PortraitHOLOGRAM display -- and we tried to upload some of my artwork to it, to see my art in holographic 3D. The company said that this unit would accept 3D GLB files; however it looks like one needs to convert those files using the Looking Glass Factory software, and that did not work for us (yet).
We did manage to upload my portrait drawing as a JPG, and view it in holographic 3D. The face broke up into 3D pieces -- that is, the hologram display did not show the face as a coherent unit -- however this fragmented artifact effect is more interesting, and more what I aim to express.
The treatment in Photoshop Elements 2021 (Enhance/Moving Photos.../Counterclockwise) gives the more coherent 3D effect which I expected to see in the hologram display:
Again, the exploded 3D pieces in the face are more interesting, and more in sync with how I created the portrait. I explicitly applied the "style" of a 3D fragmented figure (below) in Deep Dream Generator, to give volume to my line drawing. I created the "style" figure image in Paint 3D:
Artist Jens Haaning delivered two blank canvases instead of his commission, after the Kunsten Museum of Modern Art in Denmark fronted him $84,000. He entitled the new work "Take the Money and Run."
Maybe this is the NFT version of "Take the Money and Run," as it is zero pixels.
previously converted into the style of an Aubrey Beardsley
The above figure was converted into the watercolor-like image below using the VQGAN+CLIP text-to-image program, entering "Sitting woman figure." Of course the result looks nothing like my drawing.
This drawing below is rather vague and sketchy, but I can still feel the weight and attitude of the model. As an afterthought I felt it was blocky, and reminiscent of Mondrian. So I used two text-to-image programs to add some Mondrian:
A) In AI Art Machine I entered "In the style of Mondrian" and used the URL of the above image as the "style_URL." Then I took the last image result and applied that fake Mondrian style to the original drawing using Deep Dream Generator.
B) I wondered how another text-to-image program would handle the input "In the style of Mondrian," and VQGAN+CLIP created something entirely different:
I am trying to carve a relief print of the last pose of the session:
Before I finished carving I ran the above image through a VQGAN+CLIP text-to-image program, using the text "Print in the style of Mexican printmakers."
I used the SAME TEXT -- "A Giacometti figure walking into infinite space" -- in Abraham.ai, a kind of crowdsourcing "text-to-image" generator, but the result came out very differently:
Same prompt "p": "A Giacometti figure walking into infinite space"
Note: Neither of the above two different VQGAN programs allowed me to input an initial image -- both of them were completely "text-to-image."
I then combined TWO of my drawings and the final result was better (note that this source image is a square, the dimension that the VQGAN program uses):
If I simplify this result above, I might be able to make a serigraphy edition of the AI artwork. I used three Paint by Numbers programs to reduce the AI portrait into 4 colors:
There seems to be many "text-to-image" variations on the "VQGAN approach. The CLIP Guided Diffusion HQ is posted on Hugging Face.
I uploaded both the "initial image" and "image prompt" -- the top two images in the collage below -- to the program on Hugging Face, and typed in the text "Walking through the street aimlessly," to get the resulting ambiguous image on the bottom:
I drew the last post on a 7 x 7 inch square of expanded PVC, and cut and printed the plate immediately after the drawing session, on the Tortilla Press.
Ethan Moses brought his homemade camera/dark room to Quelab on September 8th, to take and develop photographs. The ingenious system worked quite well, and Ethan took flattering photos of everyone.