Article
Computer Science, Artificial Intelligence
Changyeop Shin, Hyun Ryu, Eun-Seo Cho, Seungjae Han, Kang-Han Lee, Cheol-Hee Kim, Young-Gyu Yoon
Summary: In this study, we propose a computational microscopy technique called recursive light propagation network (RLP-Net) for volumetric imaging. RLP-Net utilizes a recursive inference scheme and a self-supervised denoising method to enable accurate virtual light propagation.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Acoustics
Daewoon Seong, Euimin Lee, Yoonseok Kim, Sangyeob Han, Jaeyul Lee, Mansik Jeon, Jeehyun Kim
Summary: Spatial sampling density and data size are important factors for the imaging speed of photoacoustic microscopy (PAM). Undersampling methods are commonly used to increase PAM imaging speed by reducing the number of scanning points. In this study, a deep learning-based method for fully reconstructing undersampled 3D PAM data was proposed, which exhibited robustness and outperformed interpolation-based methods at various undersampling ratios. The proposed method significantly improved PAM system performance with 80-times faster imaging speed and 800-times lower data size.
Article
Nanoscience & Nanotechnology
Yilin Luo, Luzhe Huang, Yair Rivenson, Aydogan Ozcan
Summary: Autofocusing is critical for high-quality microscopic imaging, with hardware-based and algorithmic methods being the main approaches. The deep learning-based offline autofocusing method Deep-R is significantly faster than standard online algorithmic autofocusing methods, allowing for rapid and blind autofocusing of microscope images.
Article
Nanoscience & Nanotechnology
Xilin Yang, Luzhe Huang, Yilin Luo, Yichen Wu, Hongda Wang, Yair Rivenson, Aydogan Ozcan
Summary: This paper presents a virtual refocusing method over an extended depth of field using cascaded neural networks and a double-helix point-spread function (DH-PSF). By combining the W-Net model with DH-PSF engineering, the depth of field of a fluorescence microscope was extended experimentally.
Article
Engineering, Electrical & Electronic
Josue Page Vizcaino, Federico Saltarin, Yury Belyaev, Ruth Lyck, Tobias Lasser, Paolo Favaro
Summary: A novel deep learning approach, LFMNet, is introduced for reconstructing confocal microscopy stacks from single light field images rapidly and accurately, enabling imaging of highly dynamic and light-sensitive events, as well as in closed-loop systems where reducing latency is crucial.
IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING
(2021)
Article
Engineering, Multidisciplinary
Cunqiang Liu, Juan Li, Jie Gao, Dongdong Yuan, Ziqiang Gao, Zhongjie Chen
Summary: This study proposes a method based on deep learning to measure and reconstruct macro-texture using any number of pavement views, and demonstrates its effectiveness and stability through experiments. The 3D model allows for the assessment of pavement performance, with average errors of 7.62% for mean texture depth and 6.32% for dynamic fraction coefficient.
Article
Chemistry, Analytical
Cory Juntunen, Isabel M. Woller, Yongjin Sung
Summary: Hyperspectral 3D imaging integrates SPOT and FTS technologies for acquiring projection images from different viewing angles instantly and performing high-spectral-resolution imaging, which is validated through imaging performance using fluorescent beads and sunflower pollens.
Article
Chemistry, Physical
Takashi Sumikama, Filippo Federici Canova, David Z. Gao, Marcos Penedo, Keisuke Miyazawa, Adam S. Foster, Takeshi Fukuma
Summary: In this study, the 3D-AFM method was used to resolve the three-dimensional structures of biopolymers. A computational method using the Jarzynski equality was developed to simulate 3D-AFM images, and it was found that some parts of the fiber structure of biopolymers were resolved in the images. The dependency of 3D-AFM images on scanning velocity was investigated, and optimum scanning velocities were determined. It was also clarified that forces in nonequilibrium processes could be measured in 3D-AFM measurements when the dynamics of polymers were slower than the scanning of the probe.
JOURNAL OF PHYSICAL CHEMISTRY LETTERS
(2022)
Article
Computer Science, Interdisciplinary Applications
Xuan Wang, Min Liu, Yaonan Wang, Jiawang Fan, Erik Meijering
Summary: This paper proposes a neuron centerline extraction method based on a 3D tubular flux model using a two-stage CNN framework. The method learns flux features from neuron images and extracts the centerline with a spatial weighted average strategy. The experiments show that the proposed method outperforms other state-of-the-art methods and the extracted centerline improves neuron reconstruction performance.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2022)
Article
Multidisciplinary Sciences
Marco Leonetti, Lorenzo Pattelli, Simone De Panfilis, Diederik S. Wiersma, Giancarlo Ruocco
Summary: Through directly measuring spatially resolved intensity correlations of light inside a disordered medium using DNA strings decorated with emitters, researchers uncovered deviations in the size and polarization degrees of freedom of bulk speckles from theoretical predictions. These deviations are explained by correlations among polarization components and non-universal near-field contributions at the nanoscale.
NATURE COMMUNICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Francisco Gomez-Donoso, Julio Castano-Amoros, Felix Escalona, Miguel Cazorla
Summary: The popularity of intelligent and autonomous vehicles has grown, but 2D object detection methods have issues in urban environments. To overcome this, the use of structure from motion is proposed to reconstruct 3D information from images and merge 2D and 3D data to differentiate actual objects from depictions.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Thomas Wollmann, Karl Rohr
Summary: The article presents a novel deep neural network for object detection in microscopy images, which includes a feature extractor, a centroid proposal network, and a layer for ensembling detection hypotheses over all image scales and anchors. Utilizing anchor regularization and a new loss function to address class imbalance, along with an improved non-maximum suppression algorithm, experiments demonstrate the method's outstanding performance on challenging data.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Biochemical Research Methods
Xu Mei, Qiyin Fang, P. ravi Selvaganapathy
Summary: Measurement of oxygen concentration in 3D hydrogels is crucial for 3D cell culture and tissue engineering. This study presents a fast and low-cost phosphorescence lifetime imaging design that can accurately measure oxygen concentration in 3D. By combining light-sheet illumination and frequency-domain lifetime measurement, along with a commercial camera, this design is customizable and cost-effective for specific research needs.
BIOMEDICAL OPTICS EXPRESS
(2023)
Article
Environmental Sciences
Zuobang Zhou, Xiangguo Jin, Lei Liu, Feng Zhou
Summary: This study proposes an extended factorization framework (EFF) that combines deep learning-based instance segmentation and an improved factorization method to achieve 3D geometry reconstruction of space targets. Experiments on simulated and measured data demonstrate the feasibility of the proposed framework.
Article
Biotechnology & Applied Microbiology
Andreas Boden, Francesca Pennacchietti, Giovanna Coceano, Martina Damenti, Michael Ratz, Ilaria Testa
Summary: A new 3D pRESOLFT microscope method with sub-80-nm resolution has been developed to visualize the volumetric architecture of organelles and molecules inside whole living cells, allowing for targeted 3D confinement of fluorescence and observation of dynamic structural alterations in cells.
NATURE BIOTECHNOLOGY
(2021)
Article
Engineering, Biomedical
Rui Cao, Scott D. Nelson, Samuel Davis, Yu Liang, Yilin Luo, Yide Zhang, Brooke Crawford, Lihong Wang
Summary: Ultraviolet photoacoustic microscopy can be used to evaluate thick bone specimens without sectioning them during surgery. This technique allows for rapid diagnosis of bone pathology and determination of tumor margins.
NATURE BIOMEDICAL ENGINEERING
(2023)
Article
Optics
Rui Cao, Jingjing Zhao, Lei Li, Lin Du, Yide Zhang, Yilin Luo, Laiming Jiang, Samuel Davis, Qifa Zhou, Adam de la Zerda, Lihong Wang
Summary: Optical-resolution photoacoustic microscopy can visualize wavelength-dependent optical absorption at the cellular level. However, it has a limited depth of field. To overcome this limitation, needle-shaped beam photoacoustic microscopy is proposed. This approach extends the depth of field and provides new perspectives for slide-free intraoperative pathological imaging and in vivo organ-level imaging.
Article
Chemistry, Multidisciplinary
Artem Goncharov, Hyou-Arm Joung, Rajesh Ghosh, Gyeo-Re Han, Zachary S. Ballard, Quinn Maloney, Alexandra Bell, Chew Tin Zar Aung, Omai B. Garner, Dino Di Carlo, Aydogan Ozcan
Summary: This study demonstrates a point-of-care serodiagnosis assay that can simultaneously quantify three biomarkers of acute cardiac injury. Using a paper-based fluorescence vertical flow assay (fxVFA) processed by a low-cost mobile reader, the platform achieves fast testing time and high sensitivity for the target biomarkers. Validation using human serum samples shows minimal cross-reactivity and accurate quantification of the three biomarkers. Blind testing with individually activated cartridges further confirms the accuracy and reliability of this multiplexed computational fxVFA. The inexpensive paper-based design and handheld footprint make this platform a promising point-of-care sensor for diagnostics in resource-limited settings.
Article
Optics
Deniz Mengu, Anika Tabassum, Mona Jarrahi, Aydogan Ozcan
Summary: Multispectral imaging has applications in various fields and a diffractive optical network-based system is proposed to create a virtual spectral filter array. This system performs coherent imaging over a large spectrum and routes specific spectral channels onto an array of pixels without filters or recovery algorithms. The system's compact design and polarization insensitivity make it transformative for imaging applications where high-density multispectral pixel arrays are not readily available.
LIGHT-SCIENCE & APPLICATIONS
(2023)
Article
Optics
Bijie Bai, Yuhang Li, Yi Luo, Xurong Li, Ege Cetintas, Mona Jarrahi, Aydogan Ozcan
Summary: Researchers have proposed an all-optical processor that can directly classify unknown objects using broadband illumination detected with a single pixel, achieving accurate classification of handwritten digits through random diffusers. This system is based on passive diffractive layers and can operate at any part of the electromagnetic spectrum.
LIGHT-SCIENCE & APPLICATIONS
(2023)
Review
Optics
Bijie Bai, Xilin Yang, Yuzhu Li, Yijie Zhang, Nir Pillar, Aydogan Ozcan
Summary: Histological staining is an important technique in clinical pathology and research, but it is expensive, time-consuming, and limited in resource-limited settings. Deep learning techniques have provided a solution by digitally generating histological stains, which are rapid, cost-effective, and accurate alternatives to chemical staining methods. This review provides an overview of recent advances in deep learning-enabled virtual histological staining and discusses its future potential.
LIGHT-SCIENCE & APPLICATIONS
(2023)
Article
Multidisciplinary Sciences
Jingxi Li, Tianyi Gan, Yifan Zhao, Bijie Bai, Che-Yung Shen, Songyu Sun, Mona Jarrahi, Aydogan Ozcan
Summary: The first demonstration of unidirectional imagers is reported, presenting polarization-insensitive and broadband uni-directional imaging based on successive diffractive layers that are linear and isotropic. The diffractive unidirectional imager maintains its functionality over a large spectral band and works under broadband illumination. The diffractive unidirectional imaging using structured materials will have applications in security, defense, telecommunications, and privacy protection.
Article
Engineering, Biomedical
Tairan Liu, Yuzhu Li, Hatice Ceylan Koydemir, Yijie Zhang, Ethan Yang, Merve Eryilmaz, Hongda Wang, Jingxi Li, Bijie Bai, Guangdong Ma, Aydogan Ozcan
Summary: An automated plaque assay leveraging lens-free holographic imaging and deep learning rapidly and accurately detects the cell-lysing events caused by viral replication.
NATURE BIOMEDICAL ENGINEERING
(2023)
Review
Optics
Xurong Li, Jingxi Li, Yuhang Li, Aydogan Ozcan, Mona Jarrahi
Summary: This article reviews the development of terahertz imaging technologies and discusses different types of hardware and computational imaging algorithms. It explores opportunities for capturing various image data and briefly introduces the prospects and challenges for future high-throughput terahertz imaging systems.
LIGHT-SCIENCE & APPLICATIONS
(2023)
Article
Acoustics
Karteekeya Sastry, Yang Zhang, Peng Hu, Yilin Luo, Xin Tong, Shuai Na, Lihong V. Wang
Summary: This article introduces a geometric calibration method applicable to various photoacoustic computed tomography (PACT) systems. By using surrogate methods to obtain the speed of sound and point source locations, the linear problem in the transducer coordinates is solved. The estimation error is characterized, informing the arrangement of point sources. The method is demonstrated in a three-dimensional PACT system, showing improvements in contrast-to-noise ratio, size, and spread of point source reconstructions by (80 ± 19)%, (19 ± 3)%, and (7 ± 1)%, respectively. The calibrated image of a healthy human breast reveals previously invisible vasculatures. This work introduces a method for geometric calibration in PACT and paves the way for improving PACT image quality.
Article
Computer Science, Artificial Intelligence
Luzhe Huang, Hanlong Chen, Tairan Liu, Aydogan Ozcan
Summary: GedankenNet is a self-supervised learning model that achieves image reconstruction without the need for labelled or experimental training data, demonstrating superior generalization on hologram reconstruction tasks.
NATURE MACHINE INTELLIGENCE
(2023)
Article
Automation & Control Systems
Che-Yung Shen, Jingxi Li, Deniz Mengu, Aydogan Ozcan
Summary: This study presents a diffractive processor for all-optical multispectral quantitative phase imaging of transparent objects. The processor encodes the phase profile of the input object at predetermined wavelengths into spatial intensity variations at the output plane using diffractive layers optimized through deep learning. Numerical simulations demonstrate its capability to perform quantitative phase imaging at multiple spectral bands and the generalization of the design is validated through tests on unseen objects. This diffractive multispectral processor offers a compact and power-efficient solution for high-throughput quantitative phase microscopy and spectroscopy due to its all-optical processing capability.
ADVANCED INTELLIGENT SYSTEMS
(2023)
Article
Optics
Jingxi Li, Tianyi Gan, Bijie Bai, Yi Luo, Mona Jarrahi, Aydogan Ozcan
Summary: Large-scale linear operations are crucial for complex computational tasks, and optical computing offers advantages in terms of speed, parallelism, and scalability. The deep-learning-based design of a broadband diffractive neural network enables the performance of a large group of complex-valued linear transformations. By assigning different illumination wavelengths to each transformation, a single diffractive network can execute multiple linear transformations simultaneously or sequentially. This technology allows for the approximation of unique linear transforms with negligible errors, and the spectral multiplexing capability can be increased by increasing the number of diffractive neurons.
ADVANCED PHOTONICS
(2023)
Article
Automation & Control Systems
Md Sadman Sakib Rahman, Aydogan Ozcan
Summary: This study demonstrates for the first time a time-lapse image classification scheme using a diffractive network, improving classification accuracy and generalization performance by leveraging the lateral movements of the input objects and/or the diffractive network. Numerical exploration reveals a blind testing accuracy of 62.03% on the optical classification of objects from the CIFAR-10 dataset using time-lapse diffractive networks, achieving the highest inference accuracy so far. Time-lapse diffractive networks will be widely beneficial for spatiotemporal analysis of input signals using all-optical processors.
ADVANCED INTELLIGENT SYSTEMS
(2023)
Article
Mathematical & Computational Biology
Michael John Fanous, Nir Pillar, Aydogan Ozcan
Summary: Traditional staining methods have drawbacks, while computational virtual staining using deep learning techniques has emerged as a powerful solution. Virtual staining can be combined with neural networks to correct microscopy aberrations and enhance resolution, significantly improving sample preparation and imaging in biomedical microscopy.
FRONTIERS IN BIOINFORMATICS
(2023)
