基于深度学习的非视域成像

Tingyi Yu; Mu Qiao; Honglin Liu; Shensheng Han

doi:10.3788/AOS201939.0711002

基于深度学习的非视域成像

Translated title of the contribution: Non-Line-of-Sight Imaging Through Deep Learning

Tingyi Yu, Mu Qiao, Honglin Liu, Shensheng Han

Research output: Contribution to journal › Article › peer-review

9 Scopus citations

Abstract

Aiming at the problem of non-line-of-sight imaging under incoherent illumination, we propose a solution based on deep learning. Combining the classical semantic segmentation and residual model in the field of computer vision, a URNet network structure is constructed and the classical bottleneck layer structure is improved. The experimental results show that the improved model has more details of recovery images and generalization ability. Compared with speckle autocorrelation imaging method under incoherent illumination, the recovery performance of this method is greatly improved.

Translated title of the contribution	Non-Line-of-Sight Imaging Through Deep Learning
Original language	Chinese (Traditional)
Article number	0711002
Journal	Guangxue Xuebao/Acta Optica Sinica
Volume	39
Issue number	7
DOIs	https://doi.org/10.3788/AOS201939.0711002
State	Published - Jul 10 2019
Externally published	Yes

All Science Journal Classification (ASJC) codes

Electronic, Optical and Magnetic Materials
Atomic and Molecular Physics, and Optics

Keywords

Deep learning
Imaging systems
Non-line-of-sight imaging
Residual model
Semantic segmentation

Access to Document

10.3788/AOS201939.0711002

Cite this

@article{1a452dbcbc0242a1bd996c01ae6fc817,

title = "基于深度学习的非视域成像",

abstract = "Aiming at the problem of non-line-of-sight imaging under incoherent illumination, we propose a solution based on deep learning. Combining the classical semantic segmentation and residual model in the field of computer vision, a URNet network structure is constructed and the classical bottleneck layer structure is improved. The experimental results show that the improved model has more details of recovery images and generalization ability. Compared with speckle autocorrelation imaging method under incoherent illumination, the recovery performance of this method is greatly improved.",

keywords = "Deep learning, Imaging systems, Non-line-of-sight imaging, Residual model, Semantic segmentation",

author = "Tingyi Yu and Mu Qiao and Honglin Liu and Shensheng Han",

year = "2019",

month = jul,

day = "10",

doi = "10.3788/AOS201939.0711002",

language = "Chinese (Traditional)",

volume = "39",

journal = "Guangxue Xuebao/Acta Optica Sinica",

issn = "0253-2239",

publisher = "Chinese Optical Society",

number = "7",

}

TY - JOUR

T1 - 基于深度学习的非视域成像

AU - Yu, Tingyi

AU - Qiao, Mu

AU - Liu, Honglin

AU - Han, Shensheng

PY - 2019/7/10

Y1 - 2019/7/10

N2 - Aiming at the problem of non-line-of-sight imaging under incoherent illumination, we propose a solution based on deep learning. Combining the classical semantic segmentation and residual model in the field of computer vision, a URNet network structure is constructed and the classical bottleneck layer structure is improved. The experimental results show that the improved model has more details of recovery images and generalization ability. Compared with speckle autocorrelation imaging method under incoherent illumination, the recovery performance of this method is greatly improved.

AB - Aiming at the problem of non-line-of-sight imaging under incoherent illumination, we propose a solution based on deep learning. Combining the classical semantic segmentation and residual model in the field of computer vision, a URNet network structure is constructed and the classical bottleneck layer structure is improved. The experimental results show that the improved model has more details of recovery images and generalization ability. Compared with speckle autocorrelation imaging method under incoherent illumination, the recovery performance of this method is greatly improved.

KW - Deep learning

KW - Imaging systems

KW - Non-line-of-sight imaging

KW - Residual model

KW - Semantic segmentation

UR - http://www.scopus.com/inward/record.url?scp=85071610316&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071610316&partnerID=8YFLogxK

U2 - 10.3788/AOS201939.0711002

DO - 10.3788/AOS201939.0711002

M3 - Article

AN - SCOPUS:85071610316

SN - 0253-2239

VL - 39

JO - Guangxue Xuebao/Acta Optica Sinica

JF - Guangxue Xuebao/Acta Optica Sinica

IS - 7

M1 - 0711002

ER -

基于深度学习的非视域成像

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this