2021
DOI: 10.1109/tci.2021.3063870
|View full text |Cite
|
Sign up to set email alerts
|

Holographic 3D Particle Imaging With Model-Based Deep Network

Abstract: Gabor holography is an amazingly simple and effective approach for three-dimensional (3D) imaging. However, it suffers from a DC term, twin-image entanglement, and defocus noise. The conventional approach for solving this problem is either using an off-axis setup, or compressive holography. The former sacrifices simplicity, and the latter is computationally demanding and timeconsuming. To cope with this problem, we propose a model-based holographic network (MB-HoloNet) for threedimensional particle imaging. Th… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
13
0
1

Year Published

2021
2021
2024
2024

Publication Types

Select...
6
1

Relationship

2
5

Authors

Journals

citations
Cited by 34 publications
(14 citation statements)
references
References 35 publications
0
13
0
1
Order By: Relevance
“…In software, reconstruction quality can be potentially improved by varies basis functions or learned ones. [ 22–24 ] The framework is eligible in principle for different hardware configurations as well, resulting in a change of the forward image formation model (matrix boldA) in Equation ().…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…In software, reconstruction quality can be potentially improved by varies basis functions or learned ones. [ 22–24 ] The framework is eligible in principle for different hardware configurations as well, resulting in a change of the forward image formation model (matrix boldA) in Equation ().…”
Section: Discussionmentioning
confidence: 99%
“…First, conventional algorithms are separate, in that particle and fluid flow reconstructions are sequential, with possible error accumulations throughout the computation pipeline. Second, there are works pushing forward for a better holographic particle/volume reconstruction by hand‐crafted or trained priors, including depth‐of‐field extension with wavelet transform, [ 17 ] depth‐resolved reconstruction with 3D deconvolution, [ 18 ] compressing sensing approach with fused lasso regularization [ 19 ] or sparsity, [ 20 ] digital filtering followed by one‐pass 3D deconvolution, [ 21 ] and the deep learning approaches, [ 22–24 ] but little has been done in improving the fluid flow reconstruction.…”
Section: Introductionmentioning
confidence: 99%
“…Additionally, by developing a loss function, this study successfully obtained 3D particle images with a particle density 300 times higher than that of Shimobaba et al (2019b). Chen et al (2021) incorporated compressive sensing into DNN and trained it using end-to-end learning. The input of the DNN were 3D particle holograms, whereas the output was 3D volume data of the particles.…”
Section: Supervised Learningmentioning
confidence: 99%
“…SSIM, with the best possible score 1, reflects the similarity of two images, and higher SSIM value indicates better performance. Inspired by Chen et al [41], the number of scatterer per pixel (nspp) is introduced to measure the scatterer density. It can be expressed as…”
Section: A Numerical Investigationmentioning
confidence: 99%
“…However, the research activity in applications of this technique in 3-D mmW imaging is still quite rare. A most related work released recently in [41] proposes a model-based holographic network (MB-HoloNet), while this work mainly focuses on the Gabor holographic imaging system. And the measurement model built in this work is still a linear equation, which cannot be directly applied to near-field mmW 3-D holography because of the large-scale matrix computation.…”
mentioning
confidence: 99%