aiqbench analysis
This paper presents PSI3D, a scalable inference framework that leverages 2D latent diffusion models to reconstruct massive 3D volumes via slice-wise Markov chain Monte Carlo sampling. By integrating stochastic total variation regularization across slices, the method effectively addresses inter-slice consistency challenges in high-dimensional scientific imaging tasks like OCT super-resolution. The approach demonstrates significant performance gains over traditional and learning-based baselines for handling large-scale data where standard diffusion models typically fail due to computational constraints.
PSI3D: Plug-and-Play 3D Stochastic Inference with Slice-wise Latent Diffusion Prior
arXiv:2512.18367v2 Announce Type: replace-cross Abstract: Diffusion models are highly expressive image priors for Bayesian inverse problems. However, most diffusion models cannot operate on large-scale, high-dimensional data due to high training and inference costs. In this work, we introduce a Plug-and-play algorithm for 3D stochastic inference with latent diffusion prior (PSI3D) to address massive ($1024\times 1024\times 128$) volumes. Specifically, we formulate a Markov chain Monte Carlo approach to reconstruct each two-dimensional (2D) slice by sampling from a 2D latent diffusion model. To enhance inter-slice consistency, we also incorporate total variation (TV) regularization stochastically along the concatenation axis. We evaluate our performance on optical coherence tomography (OCT) super-resolution. Our method significantly improves reconstruction quality for large-scale scientific imaging compared to traditional and learning-based baselines, while providing robust and credible reconstructions.
Abstract:Diffusion models are highly expressive image priors for Bayesian inverse problems. However, most diffusion models cannot operate on large-scale, high-dimensional data due to high training and inference costs. In this work, we introduce a Plug-and-play algorithm for 3D stochastic inference with latent diffusion prior (PSI3D) to address massive ($1024\times 1024\times 128$) volumes. Specifically, we formulate a Markov chain Monte Carlo approach to reconstruct each two-dimensional (2D) slice by sampling from a 2D latent diffusion model. To enhance inter-slice consistency, we also incorporate total variation (TV) regularization stochastically along the concatenation axis. We evaluate our performance on optical coherence tomography (OCT) super-resolution. Our method significantly improves reconstruction quality for large-scale scientific imaging compared to traditional and learning-based baselines, while providing robust and credible reconstructions.
Image and Video Processing (eess.IV); Machine Learning (cs.LG)
arXiv:2512.18367 [eess.IV]
(or arXiv:2512.18367v2 [eess.IV] for this version)
https://doi.org/10.48550/arXiv.2512.18367 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 220 words
Related coverage
Presentation: Exploding GPUs
InfoQ · Oct 16, 2026
Jev: System One models for Prod, not God — with Diogo Almeida, CEO, TypeSafe AI
Latent Space · Sep 21, 2026
PLAN: Parallel Liquid-Inspired Approximation Network for Efficient Representation Learning in Flexible Job Shop Scheduling
arXiv · Aug 5, 2026
Text-Vision Synergistic Token Caching: A Training-Free Framework for Efficient Vision-Language-Action Inference
arXiv · Sep 29, 2026