I write about ChatGPT and language AI tools a lot lately. Not so much about ai image generation. But this one was fun to try out and play with.
Illusion Diffusion is a technology related to optical illusions using stable diffusion.
The tool created by AP123 creates optical illusions that can transform one image into another, for example, changing the perception of “a duck” to “a rabbit” or a “cartoon drawing of a bird” to a “cartoon drawing of a frog”. Or lets say a logo of your company that gets perfectly blent into a city, garden, beach or something else.
The technology is based on a Stable Diffusion process that allows for these optical transformations. The primary source of information on this topic is a GitHub repository by a user @tancik.
Other resources and demonstrations are available online, such as the Diffusion Illusions website which showcases how optical illusions and steganography can be created using this method.
Here are some examples where I played with my company’s logo.
