Fast alternating minimization algorithm for coded aperture snapshot spectral imaging based on sparsity and deep image priors

Qile Zhao, Xianhong Zhao, Xu Ma*, Xudong Chen, Gonzalo R. Arce

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Coded aperture snapshot spectral imaging is a computational imaging technique used to reconstruct three-dimensional hyperspectral images from one or several two-dimensional projection measurements. However, fewer projection measurements or more spectral channels lead to a severely ill-posed problem, in which case regularization methods have to be applied. In order to significantly improve the accuracy of reconstruction, this paper proposes a fast alternating minimization algorithm based on the sparsity and deep image priors of natural images. By integrating deep image prior into the principle of compressive sensing reconstruction, the proposed algorithm can achieve state-of-the-art results without any training dataset. Extensive experiments show that the Fama-SDIP method significantly outperforms prevailing methods on simulation and real HSI datasets.

Original languageEnglish
Pages (from-to)4121-4131
Number of pages11
JournalApplied Optics
Volume64
Issue number14
DOIs
Publication statusPublished - 10 May 2025
Externally publishedYes

Fingerprint

Dive into the research topics of 'Fast alternating minimization algorithm for coded aperture snapshot spectral imaging based on sparsity and deep image priors'. Together they form a unique fingerprint.

Cite this