1
Divisions by zero
18,136 readers652 users here now
Communities about Anarchism, Generative AI, Copylefts, Neurodivergence, Filesharing, and Free Software. (And Math!)
Follow the Anarchist Code of Conduct and honor the Disengage Rule.
Don't be shitty to each other. Keep it SFW. Obey the spirit of The Golden Rules. Fuck around and find out.
👶 New to lemmy? Start here and here
Please help cover server costs
 |
 |
|---|---|
| Ko-fi | Liberapay |
This instance provides a wiki service which only users of this lemmy can use. If you want a wiki for your community or your account or whatever, feel free to use it
This instance currently allows new community creation, however the following subjects are explicitly not allowed as communities.
- Nothing positive or promoting cryptocurrency, blockchain or NFTs
- Nothing extreme-right-wing. This includes conspiracy theories, SovCit, Pro-Police, AnCaps etc. We'll know it when we see it, don't test us!
- No anti-science topics.
- No Tankie shit.
- No TERFs.
- No Porn.
Preferably you'll stay within the topics endorsed by this instance (see first line)
Note that you are expected to attempt to create an active community and not just squat on a name. Inactive communities will be purged after receiving a warning.
Find us on Matrix and regale us with your tales of adventure!
When going to other communities, respect their rules AND our rules when they are more restrictive. Do not give cause for others to de-federate our instance please.
Alternative Frontends
founded 2 years ago
ADMINS
2
4
Visual Anagrams: Generating Multi-View Optical Illusions with Diffusion Models
(dangeng.github.io)
tl;dr: We use pretrained diffusion models to make optical illusions
Abstract
We address the problem of synthesizing multi-view optical illusions: images that change appearance upon a transformation, such as a flip or rotation. We propose a simple, zero-shot method for obtaining these illusions from off-the-shelf text-to-image diffusion models. During the reverse diffusion process, we estimate the noise from different views of a noisy image. We then combine these noise estimates together and denoise the image. A theoretical analysis suggests that this method works precisely for views that can be written as orthogonal transformations, of which permutations are a subset. This leads to the idea of a visual anagram--an image that changes appearance under some rearrangement of pixels. This includes rotations and flips, but also more exotic pixel permutations such as a jigsaw rearrangement. Our approach also naturally extends to illusions with more than two views. We provide both qualitative and quantitative results demonstrating the effectiveness and flexibility of our method. Please see our project webpage for additional visualizations and results: this https URL
Paper: https://arxiv.org/abs/2311.17919
Code: https://github.com/dangeng/visual_anagrams
Progect Page: https://dangeng.github.io/visual_anagrams/
3
view more: next ›