DAEs for Linear Inverse Problems: Improved Recovery with Provable Guarantees

Jasjeet Dhaliwal, Kyle Hambrook

2022

Abstract

Generative priors have been shown to provide improved results over sparsity priors in linear inverse problems. However, current state of the art methods suffer from one or more of the following drawbacks: (a) speed of recovery is slow; (b) reconstruction quality is deficient; (c) reconstruction quality is contingent on a computationally expensive process of tuning hyperparameters. In this work, we address these issues by utilizing Denoising Auto Encoders (DAEs) as priors and a projected gradient descent algorithm for recovering the original signal. We provide rigorous theoretical guarantees for our method and experimentally demonstrate its superiority over existing state of the art methods in compressive sensing, inpainting, and super-resolution. We find that our algorithm speeds up recovery by two orders of magnitude (over 100x), improves quality of reconstruction by an order of magnitude (over 10x), and does not require tuning hyperparameters.

Paper Citation

in Harvard Style

Dhaliwal J. and Hambrook K. (2022). DAEs for Linear Inverse Problems: Improved Recovery with Provable Guarantees. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 4: VISAPP; ISBN 978-989-758-555-5, SciTePress, pages 97-105. DOI: 10.5220/0010804500003124

in Bibtex Style

@conference{visapp22,
author={Jasjeet Dhaliwal and Kyle Hambrook},
title={DAEs for Linear Inverse Problems: Improved Recovery with Provable Guarantees},
booktitle={Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 4: VISAPP},
year={2022},
pages={97-105},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010804500003124},
isbn={978-989-758-555-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 4: VISAPP
TI - DAEs for Linear Inverse Problems: Improved Recovery with Provable Guarantees
SN - 978-989-758-555-5
AU - Dhaliwal J.
AU - Hambrook K.
PY - 2022
SP - 97
EP - 105
DO - 10.5220/0010804500003124
PB - SciTePress