Article
Engineering, Electrical & Electronic
Samia Kazemi, Bariscan Yonel, Birsen Yazici
Summary: We propose a deep learning based network and recovery theory for imaging from intensity-only measurements. The network uses a recurrent structure and a deep decoding prior to perform algorithm updates in a lower dimensional encoded image space. We introduce a separate encoding network for transforming the spectral initialization to an appropriate initial value. The model unrolls a fixed number of optimization iterations into a recurrent neural network, enabling simultaneous learning of the parameters of the decoding and encoding networks and the RNN. We establish a sufficient condition for guaranteed exact recovery under deterministic forward models, and demonstrate the relation between the Lipschitz constants of the trained decoding prior and encoding networks with the convergence rate of the algorithm. We validate the practical applicability of our method in synthetic aperture imaging using high fidelity simulation data.
IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING
(2022)
Article
Engineering, Electrical & Electronic
Dongdong Chen, Mike Davies, Matthias J. Ehrhardt, Carola-Bibiane Schonlieb, Ferdia Sherry, Julian Tachella
Summary: From early image processing to modern computational imaging, successful models and algorithms have relied on the fundamental property of natural signals: symmetry. Symmetry, in the form of equivariance, can also be incorporated into deep neural networks (DNNs) for more data-efficient learning. However, computational imaging presents unique challenges for equivariant network solutions due to the observation of images through noisy and ill-conditioned operators. The emerging field of equivariant imaging (EI) provides improved generalization and new imaging opportunities, with links to acquisition physics, iterative reconstruction, blind compressed sensing, and self-supervised learning.
IEEE SIGNAL PROCESSING MAGAZINE
(2023)
Article
Optics
Alaleh Aminzadeh, Lindon Roberts, Benjamin Young, Cheng- Chiang, Imants D. Svalbe, David I. Paganin, Andrew M. Kingston
Summary: A set of non-configurable transversely-displaced masks has been designed and fabricated to generate high-quality X-ray illumination patterns for imaging techniques. The designs include random binary and orthogonal patterns, fabricated through photolithography and gold electroplating techniques. Experimental results show that a single wafer can be used as an illumination mask, producing high-quality X-ray ghost images.
Article
Optics
Yuheng Wang, Huiyang Wang, Chengxin Zhou, Xianxin Han, Shengde Liu, Jianglei Di, Liyun Zhong
Summary: Learning-based phase imaging maintains a balance between high fidelity and speed. However, obtaining clear and large-scale datasets for supervised training is often difficult or impossible. In this study, we propose a real-time phase imaging architecture, PEPI, based on a physics-enhanced network and equivariance. By leveraging the measurement consistency and equivariant consistency of physical diffraction images, we optimize the network parameters and invert the process from a single diffraction pattern. We also introduce a regularization method, TV-K, based on total variation kernel function constraint, to enhance texture details and high-frequency information in the output.
Article
Optics
Xingchen Zhao, Xiaoyu Nie, Zhenhuan Yi, Tao Peng, Marlan O. Scully
Summary: A non-invasive optical imaging technique using spatial-temporal encoded patterns (STEPs) is introduced, allowing for fast imaging and insensitivity to the motion of scattering media. The technique overcomes the limitations of traditional methods and offers efficient image reconstruction.
PHOTONICS RESEARCH
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Lorenzo Massimi, Tom Partridge, Alberto Astolfo, Marco Endrizzi, Charlotte K. Hagen, Peter R. T. Munro, David Bate, Alessandro Olivo
Summary: This work presents an analytical model describing the noise in the various contrast channels retrieved in edge illumination as a function of the illumination curve sampling. An optimal sampling scheme leading to minimum noise has been determined for the case where 3 or 5 sampling points are used, which represent two of the most common acquisition schemes. Additionally, the correlation between noise in the different channels and the role of the number of points and exposure time have been investigated.
Article
Computer Science, Interdisciplinary Applications
Zhuo-Xu Cui, Sen Jia, Jing Cheng, Qingyong Zhu, Yuanyuan Liu, Kankan Zhao, Ziwen Ke, Wenqi Huang, Haifeng Wang, Yanjie Zhu, Leslie Ying, Dong Liang
Summary: This study proposes a safeguarded methodology for network unrolling, which utilizes network modules as substitutes for regularizers in parallel MR imaging, offering improved explainability and predictability. The study also demonstrates the robustness of the network against noisy interferences.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Chemistry, Analytical
Chaoran Wen, Ting Nie, Mingxuan Li, Xiaofeng Wang, Liang Huang
Summary: This paper proposes an effective network based on Retinex for low-illumination image enhancement, which decomposes, reconstructs, and enhances images to improve image quality under low-illumination conditions. Experimental results show that the method is superior to existing methods both quantitatively and visually.
Article
Optics
Changchun Lu, You Zhou, Yanxun Guo, Shaowei Jiang, Zibang Zhang, Guoan Zheng, Jingang Zhong
Summary: The study presents a mask-modulated lensless imaging platform based on translated structured illumination for high-resolution complex-valued object image recovery without involving mechanical scanning. By using an LED array to provide angle-varied illumination, regularized ptychographic iterative engine is employed in the reconstruction process, demonstrating higher resolution and better quality recovery compared to previous implementations.
Article
Optics
Kai Wen, ZhaoLin Gao, Xiang Fang, Min Liu, JuanJuan Zheng, Ying Ma, Zeev Zalevsky, Peng Gao
Summary: This study introduces a partially coherent illumination-based SIM apparatus for dual-modality microscopic imaging, capable of obtaining phase and fluorescent images of the same sample. The system demonstrates superiority in phase and fluorescent imaging, with potential applications in various fields.
Article
Optics
Mostafa Agour, Claas Fallorf, Ralf B. Bergmann
Summary: The method presents a fast and accurate way to measure the 3D shape of micro-parts using LED illumination and depth discrimination. It is robust, fast, and safe, with extended depth of focus in the range of a few millimeters.
Article
Radiology, Nuclear Medicine & Medical Imaging
Yuxin Hu, Yunyingying Xu, Qiyuan Tian, Feiyu Chen, Xinwei Shi, Catherine J. Moran, Bruce L. Daniel, Brian A. Hargreaves
Summary: This study aims to accelerate and improve multishot diffusion-weighted MRI reconstruction using deep learning, achieving a reconstruction time of 0.1 second per image. Results show that the proposed method outperforms traditional locally low-rank reconstruction, with good generalization to different imaging scenarios, enhancing the feasibility of multishot DWI in various clinical and neuroscientific studies.
MAGNETIC RESONANCE IN MEDICINE
(2021)
Article
Optics
Hong-Yun Hou, Ya-Nan Zhao, Jia-Cheng Han, Sheng-Wei Cui, De-Zhong Cao, Hong-Chao Liu, Su-Heng Zhang, Bao-Lai Liang
Summary: This research presents a coherent structured illumination single-pixel imaging scheme that efficiently generates complex-amplitude images of objects using phase modulation and inverse Hadamard transform. The scheme demonstrates high efficiency, image quality, spatial resolution, and stability, showing promising applications in optical metrology and biomedical science.
Article
Materials Science, Multidisciplinary
Kang Fu, Xumin Gao, Feifei Qin, Yongyuan Song, Linning Wang, Ziqi Ye, Yulong Su, Hongbo Zhu, Xiangyang Ji, Yongjin Wang
Summary: This research confirms that III-nitride diodes can simultaneously emit and detect light, and multiple MQW III-nitride diodes are integrated on a chip to achieve different functions.
ADVANCED MATERIALS TECHNOLOGIES
(2021)
Article
Computer Science, Artificial Intelligence
Truong Thanh Nhat Mai, Edmund Y. Lam, Chul Lee
Summary: In this paper, an algorithm unrolling approach to ghost-free HDR image synthesis algorithm is proposed. It unrolls an iterative low-rank tensor completion algorithm into deep neural networks to combine the advantages of both learning- and model-based approaches while overcoming their weaknesses. Experimental results show that the proposed algorithm outperforms state-of-the-art algorithms in terms of HDR image synthesis performance and robustness, using significantly fewer training samples.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Optics
Kyrollos Yanny, Nick Antipa, William Liberti, Sam Dehaeck, Kristina Monakhova, Fanglin Linda Liu, Konlin Shen, Ren Ng, Laura Waller
LIGHT-SCIENCE & APPLICATIONS
(2020)
Article
Optics
Kristina Monakhova, Vi Tran, Grace Kuo, Laura Waller
Summary: Compressive lensless imagers enable novel applications in an extremely compact device, recovering 3D data from 2D measurements through compressive-sensing-based inverse problems. This work proposes an unsupervised approach based on untrained networks for compressive image recovery, demonstrating improved image quality without the need for labeled training data.
Article
Optics
Yina Wang, Henry Pinkard, Emaad Khwaja, Shuqin Zhou, Laura Waller, Bo Huang
Summary: When studying cellular dynamics with fluorescent microscopy, trade-offs need to be made between light exposure and image quality, with image denoising playing a key role. Deep learning based image denoising has shown promising results by incorporating prior knowledge and temporal information. However, challenges arise due to the need for large training datasets, which can be addressed by combining self-supervised learning with transfer learning to maintain denoising performance without task-specific training data.
Article
Biochemical Research Methods
Ruiming Cao, Michael Kellman, David Ren, Regina Eckert, Laura Waller
Summary: We present a practical extension of the 3D differential phase contrast (DPC) method that does not require precise motion stage scanning, but uses a hand-scanning approach with automatic axial position calibration. The algorithm solves for the 3D refractive index of the sample through computational inverse problem. The method also incorporates optimized illumination patterns for improved performance.
BIOMEDICAL OPTICS EXPRESS
(2022)
Article
Optics
Kyrollos Yanny, Kristina Monakhova, Richard W. Shuai, Laura Waller
Summary: In this study, we propose a deep-learning-based approach that leverages knowledge about the system's spatially varying point spread function (PSF) for fast 2D and 3D reconstructions. Our method outperforms traditional iterative methods in terms of speed and performance.
Review
Optics
Vivek Boominathan, Jacob T. Robinson, Laura Waller, Ashok Veeraraghavan
Summary: This article reviews recent advances in lensless imaging technology and describes the design principles to be considered when developing and using lensless imaging systems.
Article
Multidisciplinary Sciences
Michael L. Whittaker, David Ren, Colin Ophus, Yugang Zhang, Laura Waller, Benjamin Gilbert, Jillian F. Banfield
Summary: This study investigates the visualization of hydrated interfaces and proposes a new method for direct imaging the electric double layer (EDL) structure of layered minerals. The presence of hydrated ion waves and their interactions with minerals are revealed. This research is of great significance for understanding geochemical processes and mineral behavior.
NATURE COMMUNICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Cem Yalcin, Nathan Tessema Ersaro, Mohammad Meraj Ghanbari, George Bocchetti, Sina Faraji Alamouti, Nick Antipa, Daniel Lopez, Nicolas C. Pegard, Laura Waller, Rikky Muller
Summary: This article presents a varifocal mirror consisting of an array of piston-motion MEMS micromirrors and a custom ASIC, offering fast operation with dwelling capability and high diffraction efficiency.
IEEE JOURNAL OF SOLID-STATE CIRCUITS
(2022)
Editorial Material
Biochemical Research Methods
Henry Pinkard, Laura Waller
Summary: Advancements in microscopy, computer vision, and open source software are coming together to introduce a new era of self-controlling microscopes.
Article
Biochemical Research Methods
Yi Xue, David Ren, Laura Waller
Summary: Fluorescence microscopy and phase microscopy provide imaging and label-free measurement of refractive index (RI) of biological samples, respectively. This study proposes a new computational multi-modal imaging method based on epi-mode microscopy to reconstruct both 3D fluorescence and 3D RI, enabling the digital correction of multiple-scattering effects in fluorescence images.
BIOMEDICAL OPTICS EXPRESS
(2022)
Article
Optics
Joseph D. Malone, Neerja Aggarwal, Laura Waller, Audrey K. Bowden
Summary: Computational spectroscopy breaks the traditional one-to-one spatial-to-spectral pixel mapping by multiplexing spectral data over a sensor region. DiffuserSpec is a simple computational spectrometer that uses commercially available diffusers to generate unique speckle patterns for each wavelength. Using Scotch tape as a diffuser, we demonstrate narrowband and broadband spectral reconstructions with 2-nm resolution over an 85-nm bandwidth in the near-infrared.
Correction
Optics
Lei Tian, Laura Waller
Summary: This errata corrects typographical errors in the derived transfer functions in [Opt. Express 23, 11394 (2015)].
Correction
Optics
Kyrollos Yanny, Nick Antipa, William Liberti, Sam Dehaeck, Kristina Monakhova, Fanglin Linda Liu, Konlin Shen, Ren Ng, Laura Waller
LIGHT-SCIENCE & APPLICATIONS
(2023)
Article
Optics
Eric Li, Stuartq Sherwin, Gautam Gunjala, Laura Waller
Summary: Fourier ptychographic microscopy is a computational imaging technique that provides high-resolution images with quantitative phase information over a large field-of-view. Model mismatch caused by unknown optical aberrations can limit reconstruction quality, but a separate calibration process can improve accuracy.