Dynamic Subspace

Category: Computers

  • Cyberpunk Brain in a Box Image Created with Stable Diffusion

    These images are called Brain in a Box 1, 2, and 3. The idea behind them was a electronic-organic computer assemblage that fit into a 19″ server cabinet. The brains are modeled on my fMRI scans, and the background cabling and box perspective come from another controlnet layer using a snapshot of a networking setup with bundles of ethernet cables. The lighting details and computing element details varied based on my prompt.

    I liked how #2 seems like a universe of constellations of lights and wires beneath a transparent brain-shaped cover.

    Number 3 is the brightest of the series. Its brain combines the previous two aspects–transparency and brain folds.

  • Anthropomorphic Cat Wearing a Hoodie and Working on Computers Made with Stable Diffusion

    Anthropomorphic Cat Wearing a Hoodie and Working on Computers Made with Stable Diffusion.

    In this Stable Diffusion made image of an anthropomorphic cat wearing a hoodie and jeans, we find the subject in a workshop surrounded by computer parts and multiple monitors. The main parts that needed repair with inpainting were his hoodie pull strings and what was displayed on the computer monitors (originally, the center one had a ghastly distorted face, but I was able to replace it with computer code). Also, almost all of the anthropomorphic cats need inpainting on the paws to make them look like real paws instead of fur-covered human hands and fingers.

  • Anthropomorphic Cat Astronaut on the International Space Station

    Anthropomorphic Cat Astronaut on the International Space Station. Made with Stable Diffusion.

    One of my favorite kinds of images to create with Stable Diffusion are those involving anthropomorphic cats. Here’s a cat as an astronaut aboard the International Space Station. It required a lot of inpainting for the star field and the Earth’s curvature. SD 1.5 models often have trouble with keeping track of a line, of say a table or in this case a celestial object, bisected by a foreground subject. The final image here isn’t perfect but it was as good enough for me.

  • A Post-Apocalyptic Nightmare Made with Stable Diffusion

    A skeletal being with glowing eyes towers over a man in the foreground. Made with Stable Diffusion.

    I created this nightmare image of a skeletal being with glowing eyes in a post-apocalyptic landscape using Stable Diffusion and A1111 earlier this year. It required very little prompting to come up with this frightening picture.

  • Williams’ Moon Patrol Gameplay Transformed With Stable Diffusion

    Moon Patrol game screenshot transformed with Stable Diffusion.

    For the next few weeks, I’m going to be sharing some of the images that I have created using Stable Diffusion, an opensource generative AI text-to-image model created by Stability.AI. Today begins the series with images based on a classic video game.

    Earlier this year, I used the low-resolution screenshot of Williams’ 1982 Moon Patrol video game from its Wikipedia page and manipulated it with Stable Diffusion and Automatic1111’s stable-diffusion-webui tool. After many, many iterations with img2img and some inpainting, I arrived at the image above. The most difficult part of the image was creating the moon buggy with six wheels, something that most SD 1.5 derived models seem to abhor.

    Later, I took another stab at transforming the gameplay screenshot to a high resolution version. This time, I used controlnet to create the moon buggy based on the Alvis Stalwart. With inpainting, I was able to achieve great detail and lighting on the buildings.

    Wouldn’t it be cool to see a new Moon Patrol game with high resolution graphics and ray tracing? Layer on backstory, mythos, and a brooding protagonist and it could be the next Halo series!