Infusing Learned Priors into Model-Based Multispectral Imaging

TL;DR

This paper presents a novel accelerated gradient method that integrates a learned 3D neural network denoiser for improved multispectral image reconstruction from noisy measurements, demonstrating generalizability across various problems.

Contribution

It introduces a new AGM variant of RED that employs a 3D deep neural network denoiser to leverage spatiationspectral correlations in MS images, enhancing reconstruction quality.

Findings

The MS-RED algorithm effectively reconstructs multispectral images from noisy data.

A single trained DNN can be applied to multiple MS imaging tasks.

The approach outperforms traditional regularization methods in experiments.

Abstract

We introduce a new algorithm for regularized reconstruction of multispectral (MS) images from noisy linear measurements. Unlike traditional approaches, the proposed algorithm regularizes the recovery problem by using a prior specified \emph{only} through a learned denoising function. More specifically, we propose a new accelerated gradient method (AGM) variant of regularization by denoising (RED) for model-based MS image reconstruction. The key ingredient of our approach is the three-dimensional (3D) deep neural net (DNN) denoiser that can fully leverage spationspectral correlations within MS images. Our results suggest the generalizability of our MS-RED algorithm, where a single trained DNN can be used to solve several different MS imaging problems.