These pictures, fragments of work-in-progress images, let you take a look at the upcoming series of ultra large-scale artworks titled Future ants, meant to be included in a future Game Backgrounds & Assets digital product. Yes, it’s about ants - AI-empowered ants of the future and their incredible habitat! Rendered in the spirit of MC Escher’s geometry, an ingredient I like adding to the spatial arrangement of my creations these days which makes the whole thing even more mind-boggling.
Technically, each image in the series is a result of applying a method of my own design during img2img refinement of the base image with a high denoise value and a particularly crafted prompt, using a model that is known to produce best results in such conditions (DreamshaperXL, in this case). The output is a very different picture saturated with all sorts of extremely detailed content but still retaining a recognisable shape of the base image. (I call this technique the CanvasInfusion style.) The latter in this case is a picture of a sad dog I generated a long time ago (as it seems now); it comes attached as the last image in the picture pairs below.
Attached below are 3 pairs of demo pictures. The main picture of each pair is a teaser fragment (what I call ‘pixel morsel’) representing only a small part (between 6-13% of the pixel space) of a much larger, 8x8K (64 MP) image that provded an interesting content, at least in one of its areas (I have produced many images of this kind as raw material). The second attached picture in a pair is a scaled down version of the raw image, it shows which part of it was used to produce the fragment. Eventually, all processed images will be merged into a single coherent master image of the HyperPixel standard, to be included in the future product.
It’s a work in progress, which means I am still experimenting (and having great fun along the way) with the structure and components of the future ant civilization. Any suggestions are very welcome!
fragment area in the image
fragment area in the image
original (base) image
pixel morsel fragment 3
fragment area in the image
original (base) image
