DeepGhost: real-time computational ghost imaging via deep learning

Abstract The potential of random pattern based computational ghost imaging (CGI) for real-time applications has been offset by its long image reconstruction time and inefficient reconstruction of complex diverse scenes. To overcome these problems, we propose a fast image reconstruction framework for...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Saad Rizvi, Jie Cao, Kaiyu Zhang, Qun Hao
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2020
Materias:
R
Q
Acceso en línea:https://doaj.org/article/0263495455744848bd5b08f1bf5b51bb
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:0263495455744848bd5b08f1bf5b51bb
record_format dspace
spelling oai:doaj.org-article:0263495455744848bd5b08f1bf5b51bb2021-12-02T15:39:49ZDeepGhost: real-time computational ghost imaging via deep learning10.1038/s41598-020-68401-82045-2322https://doaj.org/article/0263495455744848bd5b08f1bf5b51bb2020-07-01T00:00:00Zhttps://doi.org/10.1038/s41598-020-68401-8https://doaj.org/toc/2045-2322Abstract The potential of random pattern based computational ghost imaging (CGI) for real-time applications has been offset by its long image reconstruction time and inefficient reconstruction of complex diverse scenes. To overcome these problems, we propose a fast image reconstruction framework for CGI, called “DeepGhost”, using deep convolutional autoencoder network to achieve real-time imaging at very low sampling rates (10–20%). By transferring prior-knowledge from STL-10 dataset to physical-data driven network, the proposed framework can reconstruct complex unseen targets with high accuracy. The experimental results show that the proposed method outperforms existing deep learning and state-of-the-art compressed sensing methods used for ghost imaging under similar conditions. The proposed method employs deep architecture with fast computation, and tackles the shortcomings of existing schemes i.e., inappropriate architecture, training on limited data under controlled settings, and employing shallow network for fast computation.Saad RizviJie CaoKaiyu ZhangQun HaoNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 10, Iss 1, Pp 1-9 (2020)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Saad Rizvi
Jie Cao
Kaiyu Zhang
Qun Hao
DeepGhost: real-time computational ghost imaging via deep learning
description Abstract The potential of random pattern based computational ghost imaging (CGI) for real-time applications has been offset by its long image reconstruction time and inefficient reconstruction of complex diverse scenes. To overcome these problems, we propose a fast image reconstruction framework for CGI, called “DeepGhost”, using deep convolutional autoencoder network to achieve real-time imaging at very low sampling rates (10–20%). By transferring prior-knowledge from STL-10 dataset to physical-data driven network, the proposed framework can reconstruct complex unseen targets with high accuracy. The experimental results show that the proposed method outperforms existing deep learning and state-of-the-art compressed sensing methods used for ghost imaging under similar conditions. The proposed method employs deep architecture with fast computation, and tackles the shortcomings of existing schemes i.e., inappropriate architecture, training on limited data under controlled settings, and employing shallow network for fast computation.
format article
author Saad Rizvi
Jie Cao
Kaiyu Zhang
Qun Hao
author_facet Saad Rizvi
Jie Cao
Kaiyu Zhang
Qun Hao
author_sort Saad Rizvi
title DeepGhost: real-time computational ghost imaging via deep learning
title_short DeepGhost: real-time computational ghost imaging via deep learning
title_full DeepGhost: real-time computational ghost imaging via deep learning
title_fullStr DeepGhost: real-time computational ghost imaging via deep learning
title_full_unstemmed DeepGhost: real-time computational ghost imaging via deep learning
title_sort deepghost: real-time computational ghost imaging via deep learning
publisher Nature Portfolio
publishDate 2020
url https://doaj.org/article/0263495455744848bd5b08f1bf5b51bb
work_keys_str_mv AT saadrizvi deepghostrealtimecomputationalghostimagingviadeeplearning
AT jiecao deepghostrealtimecomputationalghostimagingviadeeplearning
AT kaiyuzhang deepghostrealtimecomputationalghostimagingviadeeplearning
AT qunhao deepghostrealtimecomputationalghostimagingviadeeplearning
_version_ 1718385850626080768