Async-RED: A Provably Convergent Asynchronous Block Parallel Stochastic Method using Deep Denoising Priors

TL;DR

The paper introduces ASYNC-RED, an asynchronous parallel algorithm for inverse problems that leverages deep denoisers, offering faster convergence and theoretical guarantees for large-scale image recovery tasks.

Contribution

It develops a novel asynchronous RED algorithm that enables parallel processing, reducing computational complexity and providing convergence analysis.

Findings

ASYNC-RED achieves faster convergence than serial methods.

The algorithm is validated on large-scale image recovery tasks.

Theoretical convergence guarantees are established under explicit assumptions.

Abstract

Regularization by denoising (RED) is a recently developed framework for solving inverse problems by integrating advanced denoisers as image priors. Recent work has shown its state-of-the-art performance when combined with pre-trained deep denoisers. However, current RED algorithms are inadequate for parallel processing on multicore systems. We address this issue by proposing a new asynchronous RED (ASYNC-RED) algorithm that enables asynchronous parallel processing of data, making it significantly faster than its serial counterparts for large-scale inverse problems. The computational complexity of ASYNC-RED is further reduced by using a random subset of measurements at every iteration. We present complete theoretical analysis of the algorithm by establishing its convergence under explicit assumptions on the data-fidelity and the denoiser. We validate ASYNC-RED on image recovery using pre-trained deep denoisers as priors.