D-HAN: An End-to-End Adaptive Neural Network for Snapshot Compressive Imaging

Miguel Marquez; Yingming Lai; Xianglei Liu; Cheng Jiang; Shian Zhang; Henry Arguello; Jinyang Liang

doi:10.1109/PN58661.2023.10222945

D-HAN: An End-to-End Adaptive Neural Network for Snapshot Compressive Imaging

Miguel Marquez
, Yingming Lai
, Xianglei Liu
, Cheng Jiang
, Shian Zhang
, Henry Arguello
, Jinyang Liang^*

^*Corresponding author for this work

Institut national de la recherche scientifique
Universidad Industrial de Santander
East China Normal University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Snapshot compressive imaging (SCI) has surged as a crucial tool for data visualization in applications limited to a single-shot. However, most existing methods shortfall in the reiterated use of random coded apertures and the idealization of the shearing function behavior. To overcome these limitations, we develop a new end-to-end convolutional neural network, termed deep high-dimensional adaptive net (D-HAN), that supplies the SCI systems with multifaceted supervision in the encoding operation, the shearing process, and the reconstruction. D-HAN is implemented in a representative SCI system for hyperspectral and ultrahigh-speed imaging.

Original language	English
Title of host publication	2023 Photonics North, PN 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350326734
DOIs	https://doi.org/10.1109/PN58661.2023.10222945
State	Published - 2023
Externally published	Yes
Event	2023 Photonics North, PN 2023 - Montreal, Canada Duration: 12 Jun 2023 → 15 Jun 2023

Publication series

Name	2023 Photonics North, PN 2023

Conference

Conference	2023 Photonics North, PN 2023
Country/Territory	Canada
City	Montreal
Period	12/06/23 → 15/06/23

Keywords

Snapshot compressive imaging
coded aperture design
end-to-end neural networks
shearing estimation

Access to Document

10.1109/PN58661.2023.10222945

Cite this

@inproceedings{f1e5a14695c74073b7c18a5379331136,

title = "D-HAN: An End-to-End Adaptive Neural Network for Snapshot Compressive Imaging",

abstract = "Snapshot compressive imaging (SCI) has surged as a crucial tool for data visualization in applications limited to a single-shot. However, most existing methods shortfall in the reiterated use of random coded apertures and the idealization of the shearing function behavior. To overcome these limitations, we develop a new end-to-end convolutional neural network, termed deep high-dimensional adaptive net (D-HAN), that supplies the SCI systems with multifaceted supervision in the encoding operation, the shearing process, and the reconstruction. D-HAN is implemented in a representative SCI system for hyperspectral and ultrahigh-speed imaging.",

keywords = "Snapshot compressive imaging, coded aperture design, end-to-end neural networks, shearing estimation",

author = "Miguel Marquez and Yingming Lai and Xianglei Liu and Cheng Jiang and Shian Zhang and Henry Arguello and Jinyang Liang",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 Photonics North, PN 2023 ; Conference date: 12-06-2023 Through 15-06-2023",

year = "2023",

doi = "10.1109/PN58661.2023.10222945",

language = "英语",

series = "2023 Photonics North, PN 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2023 Photonics North, PN 2023",

address = "美国",

}

Marquez, M, Lai, Y, Liu, X, Jiang, C, Zhang, S, Arguello, H & Liang, J 2023, D-HAN: An End-to-End Adaptive Neural Network for Snapshot Compressive Imaging. in 2023 Photonics North, PN 2023. 2023 Photonics North, PN 2023, Institute of Electrical and Electronics Engineers Inc., 2023 Photonics North, PN 2023, Montreal, Canada, 12/06/23. https://doi.org/10.1109/PN58661.2023.10222945

TY - GEN

T1 - D-HAN

T2 - 2023 Photonics North, PN 2023

AU - Marquez, Miguel

AU - Lai, Yingming

AU - Liu, Xianglei

AU - Jiang, Cheng

AU - Zhang, Shian

AU - Arguello, Henry

AU - Liang, Jinyang

PY - 2023

Y1 - 2023

N2 - Snapshot compressive imaging (SCI) has surged as a crucial tool for data visualization in applications limited to a single-shot. However, most existing methods shortfall in the reiterated use of random coded apertures and the idealization of the shearing function behavior. To overcome these limitations, we develop a new end-to-end convolutional neural network, termed deep high-dimensional adaptive net (D-HAN), that supplies the SCI systems with multifaceted supervision in the encoding operation, the shearing process, and the reconstruction. D-HAN is implemented in a representative SCI system for hyperspectral and ultrahigh-speed imaging.

AB - Snapshot compressive imaging (SCI) has surged as a crucial tool for data visualization in applications limited to a single-shot. However, most existing methods shortfall in the reiterated use of random coded apertures and the idealization of the shearing function behavior. To overcome these limitations, we develop a new end-to-end convolutional neural network, termed deep high-dimensional adaptive net (D-HAN), that supplies the SCI systems with multifaceted supervision in the encoding operation, the shearing process, and the reconstruction. D-HAN is implemented in a representative SCI system for hyperspectral and ultrahigh-speed imaging.

KW - Snapshot compressive imaging

KW - coded aperture design

KW - end-to-end neural networks

KW - shearing estimation

UR - https://www.scopus.com/pages/publications/85171580273

U2 - 10.1109/PN58661.2023.10222945

DO - 10.1109/PN58661.2023.10222945

M3 - 会议稿件

AN - SCOPUS:85171580273

T3 - 2023 Photonics North, PN 2023

BT - 2023 Photonics North, PN 2023

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 12 June 2023 through 15 June 2023

ER -

D-HAN: An End-to-End Adaptive Neural Network for Snapshot Compressive Imaging

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this