Machine learning holography for 3D particle field imaging

Siyao Shao,Kevin Mallery,S Santosh Kumar,Jiarong Hong,santosh kumar sankar

DOI: https://doi.org/10.1364/OE.379480

IF: 3.8

2020-01-23

Optics Express

Abstract:Siyao Shao, Kevin Mallery, S. Santosh Kumar, Jiarong Hong We propose a new learning-based approach for 3D particle field imaging using holography. Our approach uses a U-net architecture ... [Opt. Express 28, 2987-2999 (2020)]

optics

What problem does this paper attempt to address?

Generalizable Deep Learning Approach for 3D Particle Imaging using Holographic Microscopy

Shyam Kumar,Jiarong Hong

2024-11-25

Abstract:Despite its potential for label-free particle diagnostics, holographic microscopy is limited by specialized processing methods that struggle to generalize across diverse settings. We introduce a deep learning architecture leveraging human perception of longitudinal variation of diffracted patterns of particles, which enables highly generalizable analysis of 3D particle information with orders of magnitude improvement in processing speed. Trained with minimal synthetic and real holograms of simple particles, our method demonstrates exceptional performance on various challenging cases including those with high particle concentrations and noises and a wide range of particle sizes, complex shapes, and optical properties exceeding the diversity of the training datasets.

Optics
Hybrid Digital Holographic Imaging System for Three-Dimensional Dense Particle Field Measurement

Lujie Cao,Gang Pan,Jeremy de Jong,Scott Woodward,Hui Meng

DOI: https://doi.org/10.1364/ao.47.004501

IF: 1.9

2008-01-01

Applied Optics

Abstract:To apply digital holography to the measurement of three-dimensional dense particle fields in large facilities, we have developed a hybrid digital holographic particle-imaging system. The technique combines the advantages of off-axis (side) scattering in suppressing speckle noise and on-axis (in-line) recording in lowering the digital sensor resolution requirement. A camera lens is attached to the digital sensor to compensate for the weak object wave from side scattering over a large recording distance. A simple numerical reconstruction algorithm is developed for holograms recorded with a lens without requiring complex and impractical mathematical corrections. We analyze the effect of image sensor resolution and off-axis angle on system performance and quantify the particle positioning accuracy of the system. The holographic system is successfully applied to the study of inertial particle clustering in isotropic turbulence.
Adaptive In-Focus Particle Detection and Segmentation in Holographic 3d Image with Mechanism-Guided Machine Learning

Hang Zhang,Boyi Wang,Letian Zhang,Yue Zhao,Yu Wang,Jianhan Feng,Wei Xiao,Gaofeng Wang,Yingchun Wu,Xuecheng Wu

DOI: https://doi.org/10.1016/j.optlaseng.2024.108400

IF: 5.666

2024-01-01

Optics and Lasers in Engineering

Abstract:In the domain of three-dimensional (3D) particle holography, diverse imaging characteristics across particles of varying scales and positions pose a notable impediment to efficient information extraction. This research introduces an adaptive detection and segmentation approach employing mechanism-guided machine learning techniques. Not only the carefully chosen detection network but also the cropping pretreatment of input images are used to obtain the precise regions of interest (ROIs) of particles, inspired with the estimated particle size distribution under specific experimental conditions. Restricted by the blurred particle boundary and aiming for a robust method for hyperparameter, cluster-based segmentation is proposed with an extra feature map, which takes the unique change of object signals across different slices compared with background into account. Rigorous evaluation of different detection network configurations, training data, and segmentation methods are conducted on experimental data from an icing wind tunnel and calibration dot board. Concrete ablation experiments and analysis of the mechanism-guided operations (cropping and the introduction of the extra feature map) are provided. The influence of the hyperparameter in segmentation is also illustrated. To showcase its robustness, the proposed method is applied to segment a swirl spray particle field, demonstrating its versatility and applicability across varied particle field scenarios. The code and data for the paper are available at https://github.com/Adiazhang0426/HOLOAI/tree/master and https://drive.google.com/drive/folders/1C9QlaklXpApIpE8AukCNMTKUnZw_YcGJ?usp=drive_link.
Research on 3D Particle Field Reconstruction Method Based on Deep Learning

Shuai Wang,Fucheng You,Xuewei Wang

DOI: https://doi.org/10.1088/1742-6596/2562/1/012049

2023-08-18

Journal of Physics: Conference Series

Abstract:Reconstructing 3D particle fields from single holograms is an important topic in the computer vision field. To increase network accuracy, we incorporate a channel attention mechanism based on the U-Net architecture in this research; meanwhile, we use the LeakyReLU activation function to accelerate the network convergence. Deep learning is used to extract information from a single hologram that can recreate the 3D particle field. The network receives the digital hologram as input, and the radius and 3D locations of the particles are converted into 2D grayscale images as real labels. Simulation and experimental results show that 2D grayscale images with clear edge textures can be rapidly encoded using the neural network, and the average SSIM and PSNR of the network output results with the real target on the test data set can reach 0.989 and 32.56.
Digital Holography of Particle Fields: Reconstruction by Use of Complex Amplitude.

G Pan,H Meng

DOI: https://doi.org/10.1364/ao.42.000827

IF: 1.9

2003-01-01

Applied Optics

Abstract:Digital holography appears to be a strong contender as the next-generation technology for holographic diagnostics of particle fields and holographic particle image velocimetry for flow field measurement. With the digital holographic approach, holograms are directly recorded by a digital camera and reconstructed numerically. This not only eliminates wet chemical processing and mechanical scanning, but also enables the use of complex amplitude information inaccessible by optical reconstruction, thereby allowing flexible reconstruction algorithms to achieve optimization of specific information. However, owing to the inherently low pixel resolution of solid-state imaging sensors, digital holography gives poor depth resolution for images, a problem that severely impairs the usefulness of digital holography especially in densely populated particle fields. This paper describes a technique that significantly improves particle axial-location accuracy by exploring the reconstructed complex amplitude information, compared with other numerical reconstruction schemes that merely mimic traditional optical reconstruction. This novel method allows accurate extraction of particle locations from forward-scattering particle holograms even at high particle loadings.
A Calibration-Informed Deep Learning Model for Three-Dimensional Particle Reconstruction of Volumetric Particle Image Velocimetry

Hongtao Lin,Qi Gao

DOI: https://doi.org/10.1063/5.0237067

IF: 4.6

2024-01-01

Physics of Fluids

Abstract:Accurately reconstructing three-dimensional particle fields is essential in fluid velocity measurement research. This study addresses the limitations of current three-dimensional (3D) particle reconstruction methods, such as computational efficiency, precision at high particle density, and particle morphology issues, by introducing a calibration-informed deep learning model named the calibrated pixel to voxel convolutional neural network (CPV-CNN) for 3D Particle Reconstruction. This innovative neural network framework employs a unique Gaussian attention mechanism that bridges pixels and voxels, enabling the mapping of pixel features from two-dimensional (2D) particle images to 3D voxel features. This approach eliminates the need for an initial particle field for particle reconstruction, while significantly enhancing reconstruction efficiency. Additionally, the neural network incorporates camera calibration parameters and the physical coordinates of the reconstructed domain, thereby improving the model's generalization capability and flexibility. Numerical experiments demonstrate that CPV-CNN delivers superior results in terms of accuracy, generalization, and robustness in 3D particle reconstruction. The reconstructed particles exhibit favorable morphology, without the elongation issues commonly observed with conventional methods. This achievement illustrates a practical particle reconstruction algorithm based on artificial intelligence (AI) techniques and represents an important step toward developing an end-to-end AI-based particle reconstruction method in the future.
Three-dimensional Vectorial Holography Based on Machine Learning Inverse Design

Haoran Ren,Wei Shao,Yi Li,Flora Salim,Min Gu

DOI: https://doi.org/10.1126/sciadv.aaz4261

IF: 13.6

2020-01-01

Science Advances

Abstract:The three-dimensional (3D) vectorial nature of electromagnetic waves of light has not only played a fundamental role in science but also driven disruptive applications in optical display, microscopy, and manipulation. However, conventional optical holography can address only the amplitude and phase information of an optical beam, leaving the 3D vectorial feature of light completely inaccessible. We demonstrate 3D vectorial holography where an arbitrary 3D vectorial field distribution on a wavefront can be precisely reconstructed using the machine learning inverse design based on multilayer perceptron artificial neural networks. This 3D vectorial holography allows the lensless reconstruction of a 3D vectorial holographic image with an ultrawide viewing angle of 94° and a high diffraction efficiency of 78%, necessary for floating displays. The results provide an artificial intelligence-enabled holographic paradigm for harnessing the vectorial nature of light, enabling new machine learning strategies for holographic 3D vectorial fields multiplexing in display and encryption.
High-speed Three-Dimensional Digital Holographic Reconstruction of Particles Based on a YOLO Network

Qiran Shi,Liangcai Cao,Yufeng Wu

DOI: https://doi.org/10.1117/12.2593597

2021-01-01

Abstract:A high-speed three-dimensional digital holographic reconstruction algorithm is proposed based on the YOLO architecture, which is able to significantly accelerate the training process. Supervised learning is used to train the network using both simulated and experimental holograms. With the aid of transfer learning, a small set of 2D holograms is sufficient to train the network. The trained network can also be used to label new holograms. These holograms in turn can help train the networks to improve the robustness. It takes hours for the training process, which is more efficient than the previously proposed networks with several days for the same dataset. The network has great potential for high-dynamic scenes and is robust to background noise in the particle field reconstruction.
Dense-U-net: Dense Encoder–decoder Network for Holographic Imaging of 3D Particle Fields

Yufeng Wu,Jiachen Wu,Shangzhong Jin,Liangcai Cao,Guofan Jin

DOI: https://doi.org/10.1016/j.optcom.2021.126970

IF: 2.4

2021-01-01

Optics Communications

Abstract:Digital holographic imaging is able to reconstruct phase and three-dimensional (3D) information of an object from a one-shot two-dimensional (2D) lensless hologram. A dense encoder–decoder network, called Dense-U-net, is proposed to realize the reconstruction of a 3D particle field. The radii and 3D coordinates of the particles are encoded into a 2D grayscale image as the ground truth. The 2D hologram was served as the input of the network. A Dense_Block is designed and added to the encoder and decoder of the traditional U-net model. Each layer takes the inputs from all previous layers and passes its own feature map to all subsequent layers. Thus, as the diffraction is distributed throughout the entire hologram, both the features with high signal-to-noise ratio at the center of the particle and the diffraction features with low signal-to-noise ratio around the particle can be learned, thereby facilitating the full characterization of the particles. Using the specially designed dense connection module, network training can be accomplished with fewer parameters and faster speeds.
Machine-learning techniques for fast and accurate feature localization in holograms of colloidal particles

Mark D. Hannel,Aidan Abdulali,Michael O'Brien,David G. Grier

DOI: https://doi.org/10.1364/OE.26.015221

2018-05-12

Abstract:Holograms of colloidal particles can be analyzed with the Lorenz-Mie theory of light scattering to measure individual particles' three-dimensional positions with nanometer precision while simultaneously estimating their sizes and refractive indexes. Extracting this wealth of information begins by detecting and localizing features of interest within individual holograms. Conventionally approached with heuristic algorithms, this image analysis problem can be solved faster and more generally with machine-learning techniques. We demonstrate that two popular machine-learning algorithms, cascade classifiers and deep convolutional neural networks (CNN), can solve the feature-localization problem orders of magnitude faster than current state-of-the-art techniques. Our CNN implementation localizes holographic features precisely enough to bootstrap more detailed analyses based on the Lorenz-Mie theory of light scattering. The wavelet-based Haar cascade proves to be less precise, but is so computationally efficient that it creates new opportunities for applications that emphasize speed and low cost. We demonstrate its use as a real-time targeting system for holographic optical trapping.

Soft Condensed Matter,Disordered Systems and Neural Networks,Optics
Single-Shot High-Density Volumetric Particle Imaging Enabled by Differentiable Holography

Yang Wu,Jun Wang,Sigurdur T. Thoroddsen,Ni Chen

DOI: https://doi.org/10.1109/tii.2024.3431072

IF: 12.3

2024-01-01

IEEE Transactions on Industrial Informatics

Abstract:Digital holography (DH) stands as a robust imaging method extensively utilized for 3-D imaging. Despite its versatility, when employed in 3-D particle imaging, DH encounters limitations pertaining to resolution and particle concentration. Conventional holographic particle imaging relies on a simplified linearized model and iterative optimization techniques to address pseudoinverse problems. However, this linear assumption inadequately mirrors the complexities of the actual imaging system, resulting in inaccurate particle reconstruction and constrained density. To address these challenges, we introduce an innovative technique named 3-D differentiable holography, which combines a precise theoretical holography model with real-world factors using differentiable programming. This allows us to solve particle imaging problems more effectively using automatic differentiation technology. The effectiveness of our approach is confirmed through numerical simulations and optical experiments, which clearly demonstrate improved imaging accuracy and particle density.
Holographic 3D Particle Reconstruction Using a One-Stage Network

Yunping Zhang,Yanmin Zhu,Edmund Y. Lam

DOI: https://doi.org/10.1364/ao.444856

IF: 1.9

2021-01-01

Applied Optics

Abstract:Volumetric reconstruction of a three-dimensional (3D) particle field with high resolution and low latency is an ambitious and valuable task. As a compact and high-throughput imaging system, digital holography (DH) encodes the 3D information of a particle volume into a two-dimensional (2D) interference pattern. In this work, we propose a one-stage network (OSNet) for 3D particle volumetric reconstruction. Specifically, by a single feed-forward process, OSNet can retrieve the 3D coordinates of the particles directly from the holograms without high-fidelity image reconstruction at each depth slice. Evaluation results from both synthetic and experimental data confirm the feasibility and robustness of our method under different particle concentrations and noise levels in terms of detection rate and position accuracy, with improved processing speed. The additional applications of 3D particle tracking are also investigated, facilitating the analysis of the dynamic displacements and motions for micro-objects or cells. It can be further extended to various types of computational imaging problems sharing similar traits.
Moment-Based Shape-Learning Holography for Fast Classification of Microparticles

Haixin Luo,Jie Xu,Jiannan Jiao,Liyun Zhong,Xiaoxu Lu,Jindong Tian

DOI: https://doi.org/10.1002/adpr.202300120

2023-01-01

Advanced Photonics Research

Abstract:Compact computational in-line holography based on deep learning is an attractive single-shot approach to image microparticles dispersed in 3D volume. The particle shape contains valuable information for species classification, but the dataset acquisition process for network training suffers from labor consuming and low efficiency due to the complex and varied 2D shapes of different particle species. Herein, a moment-based shape-learning holography (MSLH) is proposed where the shape of a microparticle is mathematically characterized using varying weights of Zernike moments. By decomposing and recombining feature shapes of pollen particles into numerous new characteristic shapes as incremental data, MSLH improves the efficiency of dataset preparation. The depth-encoded shape-learning is achieved using a U-net with self-attention mechanism, which enables fast axial depth determination. The shape reconstruction uses a wavelet-based method with more explicit physical meanings, making MSLH a hybrid data-and-model driven approach that requires fewer primary data. Validation results show that MSLH achieves high accuracy in axial position and shape reconstruction, while maintaining good classification effectiveness. It is believed that MSLH is an easy-to-setup, efficient-to-construct, and fast-to-output approach for shape-based classifications of 3D distributed microparticles in dynamic fluid.
Unsupervised Deep Learning Enables 3D Imaging for Single‐Shot Incoherent Holography

Yuheng Wang,Huiyang Wang,Shengde Liu,Tao Huang,Weina Zhang,Jianglei Di,Joseph Rosen,Xiaoxu Lu,Liyun Zhong

DOI: https://doi.org/10.1002/lpor.202301091

2024-02-23

LASER & PHOTONICS REVIEWS

Abstract:This paper reports a novel 3D imaging strategy for single‐shot incoherent holography based on unsupervised deep learning. The proposed method achieves single‐shot high‐fidelity reconstruction of complex objects with intensity variations for incoherent holography for the first time. The proposed PSF calibration method is universally adaptable to common optical imaging systems and compatible with any subsequent reconstruction algorithm. Incoherent digital holography no longer requires spatial coherence of the light field, breaking through the imaging resolution of coherent digital holography. However, traditional reconstruction methods cannot avoid the inherent contradiction between temporal resolution and signal‐to‐noise ratio, which is mitigated by deep learning methods, and there are problems such as dataset labeling and insufficient generalization ability. Here, a self‐calibrating reconstruction approach with an untrained network is proposed by fusing the plug‐and‐play nonlinear reconstruction block, the forward physics imaging model, and a physically enhanced neural network. Measurement consistency and total variation kernel function regularization are used to optimize the network parameters and invert the potential process. The results show that the proposed method can achieve high fidelity, high signal‐to‐noise ratio, dynamic, and artifact‐free 3D reconstruction using a single hologram without the need for datasets or labels. In addition, the peak signal‐to‐noise ratio of the reconstructed image with the proposed method is improved by a factor of 4.6 compared to the previous methods. The proposed method leads to considerable performance improvement on the imaging inverse problem, bringing new enlightenment for high‐precision unsupervised incoherent digital holographic 3D imaging.
An Approach For Sensing Underwater Particles Using Digital Holographic Imaging

Chao Liu,Xiaoping Wang,Yingtian Hu,Ming Tang,Zheng Shi

DOI: https://doi.org/10.1117/12.2505602

2018-01-01

Abstract:The purpose of this paper is to use digital holographic imaging to capture the hologram of the underwater particles. Digital holography includes two steps: wavefront recording and wavefront reconstruction. In the experiments of the underwater particles, some technique issues, such as the principle of digital holography, zero-order item suppressing and reconstruction algorithm are discussed. The complex amplitude of the underwater particles is reconstructed using numerical methods based on the Kirchhoff-Fresnel integral formula. In this case, it enables reconstruction of the object wavefield in a plane at a distance d from hologram. This novel method for imaging the underwater particles using digital holography is an extremely valuable technique for the study of underwater particles fields as it has many advantages such as large depth of field, no focusing, high resolution and has the ability to record the wavefront information of all particles in the entire volume. Finally, this paper demonstrates it is possible for digital holographic imaging setup with less complexity to image underwater particles according to our laboratory results.
End-to-end learning of 3D phase-only holograms for holographic display

Liang Shi,Beichen Li,Wojciech Matusik

DOI: https://doi.org/10.1038/s41377-022-00894-6

2022-08-03

Abstract:Computer-generated holography (CGH) provides volumetric control of coherent wavefront and is fundamental to applications such as volumetric 3D displays, lithography, neural photostimulation, and optical/acoustic trapping. Recently, deep learning-based methods emerged as promising computational paradigms for CGH synthesis that overcome the quality-runtime tradeoff in conventional simulation/optimization-based methods. Yet, the quality of the predicted hologram is intrinsically bounded by the dataset's quality. Here we introduce a new hologram dataset, MIT-CGH-4K-V2, that uses a layered depth image as a data-efficient volumetric 3D input and a two-stage supervised+unsupervised training protocol for direct synthesis of high-quality 3D phase-only holograms. The proposed system also corrects vision aberration, allowing customization for end-users. We experimentally show photorealistic 3D holographic projections and discuss relevant spatial light modulator calibration procedures. Our method runs in real-time on a consumer GPU and 5 FPS on an iPhone 13 Pro, promising drastically enhanced performance for the applications above.
Towards real-time photorealistic 3D holography with deep neural networks

Liang Shi,Beichen Li,Changil Kim,Petr Kellnhofer,Wojciech Matusik

DOI: https://doi.org/10.1038/s41586-020-03152-0

IF: 64.8

2021-03-10

Nature

Abstract:The ability to present three-dimensional (3D) scenes with continuous depth sensation has a profound impact on virtual and augmented reality, human–computer interaction, education and training. Computer-generated holography (CGH) enables high-spatio-angular-resolution 3D projection via numerical simulation of diffraction and interference<a href="/articles/s41586-020-03152-0#ref-CR1">1</a>. Yet, existing physically based methods fail to produce holograms with both per-pixel focal control and accurate occlusion<a href="/articles/s41586-020-03152-0#ref-CR2">2</a>,<a href="/articles/s41586-020-03152-0#ref-CR3">3</a>. The computationally taxing Fresnel diffraction simulation further places an explicit trade-off between image quality and runtime, making dynamic holography impractical<a href="/articles/s41586-020-03152-0#ref-CR4">4</a>. Here we demonstrate a deep-learning-based CGH pipeline capable of synthesizing a photorealistic colour 3D hologram from a single RGB-depth image in real time. Our convolutional neural network (CNN) is extremely memory efficient (below 620 kilobytes) and runs at 60 hertz for a resolution of 1,920 × 1,080 pixels on a single consumer-grade graphics processing unit. Leveraging low-power on-device artificial intelligence acceleration chips, our CNN also runs interactively on mobile (iPhone 11 Pro at 1.1 hertz) and edge (Google Edge TPU at 2.0 hertz) devices, promising real-time performance in future-generation virtual and augmented-reality mobile headsets. We enable this pipeline by introducing a large-scale CGH dataset (MIT-CGH-4K) with 4,000 pairs of RGB-depth images and corresponding 3D holograms. Our CNN is trained with differentiable wave-based loss functions<a href="/articles/s41586-020-03152-0#ref-CR5">5</a> and physically approximates Fresnel diffraction. With an anti-aliasing phase-only encoding method, we experimentally demonstrate speckle-free, natural-looking, high-resolution 3D holograms. Our learning-based approach and the Fresnel hologram dataset will help to unlock the full potential of holography and enable applications in metasurface design<a href="/articles/s41586-020-03152-0#ref-CR6">6</a>,<a href="/articles/s41586-020-03152-0#ref-CR7">7</a>, optical and acoustic tweezer-based microscopic manipulation<a href="#ref-CR8">8</a>,<a href="#ref-CR9">9</a>,<a href="/articles/s41586-020-03152-0#ref-CR10">10</a>, holographic microscopy<a href="/articles/s41586-020-03152-0#ref-CR11">11</a> and single-exposure volumetric 3D printing<a href="/articles/s41586-020-03152-0#ref-CR12">12</a>,<a href="/articles/s41586-020-03152-0#ref-CR13">13</a>.

multidisciplinary sciences
Neural Network Based 3D Tracking with a Graphene Transparent Focal Stack Imaging System

Dehui Zhang,Zhen Xu,Zhengyu Huang,Audrey Rose Gutierrez,Cameron J. Blocker,Che-Hung Liu,Miao-Bin Lien,Gong Cheng,Zhe Liu,Il Yong Chun,Jeffrey A. Fessler,Zhaohui Zhong,Theodore B. Norris

DOI: https://doi.org/10.1038/s41467-021-22696-x

IF: 16.6

2021-01-01

Nature Communications

Abstract:Recent years have seen the rapid growth of new approaches to optical imaging, with an emphasis on extracting three-dimensional (3D) information from what is normally a two-dimensional (2D) image capture. Perhaps most importantly, the rise of computational imaging enables both new physical layouts of optical components and new algorithms to be implemented. This paper concerns the convergence of two advances: the development of a transparent focal stack imaging system using graphene photodetector arrays, and the rapid expansion of the capabilities of machine learning including the development of powerful neural networks. This paper demonstrates 3D tracking of point-like objects with multilayer feedforward neural networks and the extension to tracking positions of multi-point objects. Computer simulations further demonstrate how this optical system can track extended objects in 3D, highlighting the promise of combining nanophotonic devices, new optical system designs, and machine learning for new frontiers in 3D imaging.
Diffraction model-driven neural network with semi-supervised training strategy for real-world 3D holographic photography

Qiwei Fang,Huadong Zheng,Xinxing Xia,Junchang Peng,Tengfei Zhang,Xingyu Lin,Yingjie Yu

DOI: https://doi.org/10.1364/oe.538649

IF: 3.8

2024-12-05

Optics Express

Abstract:Qiwei Fang, Huadong Zheng, Xinxing Xia, Junchang Peng, Tengfei Zhang, Xingyu Lin, Yingjie Yu Compared to traditional 2D displays, 3D display technology provides richer information to the viewer. Learning-based computer-generated ... [Opt. Express 32, 45406-45420 (2024)]

optics
Application of Deep Learning in Digital Holographic Microscopy

Meng Zhang,Ding Hao,Nie Shouping,Ma Jun,Yuan Caojin

DOI: https://doi.org/10.3788/l0p202158.1811006

2021-01-01

Abstract:Digital holographic microscopy (DHM) has attracted attention in the fields of biological imaging and materials science due to its advantages in high-precision quantitative phase imaging. However, the existence of conjugate images, the problem of phase wrapping, and the limited resolution have always hindered the wide application of DHM. In recent years, deep learning, as a specialized model for data feature extraction in machine learning, has been widely used in the field of optical imaging. In addition to improving imaging efficiency, its potential to solve imaging inverse problems has also been continuously explored by researchers, opening up a new path for the optical imaging. In this paper, we start from the working principle of deep learning applied to DHM, introduces its ideas and important mathematical concepts to solve the inverse problem of optical imaging, and at the same time summarizes the complete implementation process of deep learning. A brief summary of the research progress in recent years of deep learning in holographic reconstruction, auto-focusing and phase recovery, and holographic denoising and super-resolution is given, and summarize the existing problems in this research field and look forward to the development trend of research.

Machine learning holography for 3D particle field imaging

Generalizable Deep Learning Approach for 3D Particle Imaging using Holographic Microscopy

Hybrid Digital Holographic Imaging System for Three-Dimensional Dense Particle Field Measurement

Adaptive In-Focus Particle Detection and Segmentation in Holographic 3d Image with Mechanism-Guided Machine Learning

Research on 3D Particle Field Reconstruction Method Based on Deep Learning

Digital Holography of Particle Fields: Reconstruction by Use of Complex Amplitude.

A Calibration-Informed Deep Learning Model for Three-Dimensional Particle Reconstruction of Volumetric Particle Image Velocimetry

Three-dimensional Vectorial Holography Based on Machine Learning Inverse Design

High-speed Three-Dimensional Digital Holographic Reconstruction of Particles Based on a YOLO Network

Dense-U-net: Dense Encoder–decoder Network for Holographic Imaging of 3D Particle Fields

Machine-learning techniques for fast and accurate feature localization in holograms of colloidal particles

Single-Shot High-Density Volumetric Particle Imaging Enabled by Differentiable Holography

Holographic 3D Particle Reconstruction Using a One-Stage Network

Moment-Based Shape-Learning Holography for Fast Classification of Microparticles

Unsupervised Deep Learning Enables 3D Imaging for Single‐Shot Incoherent Holography

An Approach For Sensing Underwater Particles Using Digital Holographic Imaging

End-to-end learning of 3D phase-only holograms for holographic display

Towards real-time photorealistic 3D holography with deep neural networks

Neural Network Based 3D Tracking with a Graphene Transparent Focal Stack Imaging System

Diffraction model-driven neural network with semi-supervised training strategy for real-world 3D holographic photography

Application of Deep Learning in Digital Holographic Microscopy