English

HyperbolicDiffusion: Sharp & Scalable Tiled Generation on the Hyperbolic Plane

Computer Vision and Pattern Recognition 2026-08-04 v1

Abstract

Planar tiled diffusion denoises overlapping windows of one rectangular canvas. The hyperbolic plane has no such canvas, and its area grows exponentially with radius. We introduce HyperbolicDiffusion, a training-free method for generating finite visual fields directly on the hyperbolic plane H2. Our Hyperbolic Blooming Cover reduces window placement to a compact dynamic program that runs in seconds while providing strong theoretical guarantees. Permanent surface IDs form a shared latent canvas: a standard diffusion model denoises local windows, whose predictions are fused back onto H2. Because curvature causes residual disagreement and blur at multi-window junctions, a geometry-derived second stage re-noises and repairs precisely those regions. The resulting fields are sharp, reprojectable, and consistent across viewpoints, providing a prompt-driven generative counterpart to Escher's Circle Limit series.

Cite

@article{arxiv.2608.03422,
  title  = {HyperbolicDiffusion: Sharp & Scalable Tiled Generation on the Hyperbolic Plane},
  author = {Hugo Caselles-Dupré},
  journal= {arXiv preprint arXiv:2608.03422},
  year   = {2026}
}

Comments

Work in progress. Updated version incoming