Paper Detail

SP^3: Spherical Priors for Plug-and-Play Restoration

Sean Man, Ron Raphaeli, Matan Kleiner, Or Ronai

huggingface Score 6.5

Published 2026-06-15 · First seen 2026-06-16

General AI

Abstract

In this paper, we introduce SP^3, a novel Plug-and-Play algorithm that accelerates maximum a posteriori image restoration by replacing denoisers with Spherical Encoders (SE) as generative priors. SP^3 approximates the intractable proximal prior step by utilizing the SE tightly structured latent space as a robust projection onto the natural image manifold. Alternating this projection with a closed-form data-consistency step, via Half-Quadratic Splitting, achieves stable convergence without requiring gradient computation during inference. This unique formulation unlocks "anytime" restoration capabilities, producing sharp, plausible images from the first iteration. Evaluations across a variety of image restoration tasks demonstrate that SP^3 achieves perceptual quality comparable to state-of-the-art zero-shot diffusion and flow methods while being 3-630times faster.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{man2026sp,
  title = {SP\textasciicircum{}3: Spherical Priors for Plug-and-Play Restoration},
  author = {Sean Man and Ron Raphaeli and Matan Kleiner and Or Ronai},
  year = {2026},
  abstract = {In this paper, we introduce SP\textasciicircum{}3, a novel Plug-and-Play algorithm that accelerates maximum a posteriori image restoration by replacing denoisers with Spherical Encoders (SE) as generative priors. SP\textasciicircum{}3 approximates the intractable proximal prior step by utilizing the SE tightly structured latent space as a robust projection onto the natural image manifold. Alternating this projection with a closed-form data-consistency step, via Half-Quadratic Splitting, achieves stable convergence without requir},
  url = {https://huggingface.co/papers/2606.16396},
  keywords = {Plug-and-Play, maximum a posteriori, denoisers, Spherical Encoders, generative priors, latent space, natural image manifold, Half-Quadratic Splitting, stable convergence, anytime restoration, perceptual quality, zero-shot diffusion, flow methods, code available, huggingface daily},
  eprint = {2606.16396},
  archiveprefix = {arXiv},
}

Metadata

{}