Visual Anagrams: Generating optical illusions with diffusion models
beefman Thursday, November 30, 2023
Summary
The linked article is about a simple method to generate multi-view optical illusions. The method supports a broad range of transformations including rotations, flips, color inversions, skews, jigsaw rearrangements, and random permutations. The method uses an off-the-shelf diffusion model to estimate noise in different views or transformations and aligns them by applying the inverse view and averaging together. The article also discusses constraints for the view function and the need for transformations to be orthogonal. The article concludes by referencing previous work in this area and informing readers that the website is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
826
71
Summary
dangeng.github.io