DOCSLIB.ORG

Arxiv:2007.14139V2 [Physics.Optics] 1 Apr 2021 Plane and the Detector Plane

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Arxiv:2007.14139V2 [Physics.Optics] 1 Apr 2021 Plane and the Detector Plane Accelerating ptychographic reconstructions using spectral initializations Lorenzo Valzania,1, 2, ∗ Jonathan Dong,1, 3, ∗ and Sylvain Gigan1 1Laboratoire Kastler Brossel, Ecole´ Normale Sup´erieure - Paris Sciences et Lettres (PSL) Research University, Sorbonne Universit´e,Centre National de la Recherche Scientifique (CNRS) UMR 8552, Coll`egede France, 24 rue Lhomond, 75005 Paris, France 2Laboratory for Transport at Nanoscale Interfaces, Empa, Swiss Federal Laboratories for Materials Science and Technology, 129 Uberlandstrasse, Dubendorf 8600, Switzerland 3Laboratoire de Physique de l'Ecole´ Normale Sup´erieure - Paris Sciences et Lettres (PSL) Research University, Sorbonne Universit´e,CNRS, Universit´ede Paris, 24 rue Lhomond, 75005 Paris, France (Dated: April 2, 2021) Ptychography is a promising phase retrieval technique for label-free quantitative phase imaging. Recent advances in phase retrieval algorithms witnessed the development of spectral methods, in order to accelerate gradient descent algorithms. Using spectral initializations on experimental data, for the first time we report three times faster ptychographic reconstructions than with a standard gradient descent algorithm and improved resilience to noise. Coming at no additional computational cost compared to gradient-descent-based algorithms, spectral methods have the potential to be implemented in large-scale iterative ptychographic algorithms. I. INTRODUCTION and the acquisition parameters. The quest for improved reconstruction performance culminated in the derivation Ptychography is a computational imaging technique of optimal spectral methods for the random measure- that enables label-free, quantitative phase imaging [1]. ments setting [9, 10]. Due to these breakthroughs, spec- It is based on a simple principle: scan a probe across tral methods for phase retrieval are rather well under- a sample, collect the corresponding intensity diffraction stood theoretically. patterns (also known as 'ptychograms'), and reconstruct Although spectral initializations to solve ptychography an image of the object of interest. Because it does not have been showcased on simulated measurements [11], no require complex optical elements, it has been adapted to gain was ever reported when employed on experimental a variety of settings and spectral ranges, from electron data. Here, we provide the first experimental demon- microscopy [2], for which it was originally conceived in stration of ptychographic reconstructions improved and the late 60s [3], to X-rays [4] and extreme ultra-violet accelerated by spectral initializations. light [5], all the way down to the terahertz (THz) range [6]. The computational reconstruction in ptychography re- II. METHODS quires to solve a phase retrieval problem, where the phase of the diffracted electric field has to be recovered Let us begin by describing a ptychogram y(l)(x) with from intensity-only measurements. Such a problem is the following forward model tractable if a minimum overlap ratio (empirically esti- (l) (l) 2 mated around 60% [7]) is ensured between subsequent y (x) = jPdfa(x − x ) (x)gj ; (1) probe positions. However, as a non-linear and non- where x is a two-dimensional spatial coordinate, (x) is convex optimization problem, it is still not completely the complex transmission function of the object, scanned understood and convergence towards its global minimum with the probe a(x) at the positions x(l) for l = 1;:::;L, is not guaranteed. with L being the total number of images. Pd is a known One way to avoid local minima is to provide an ini- linear operator describing the transmission through the tial estimate already close to the solution, and further optical system of optical length d between the object refine it with iterative algorithms. To this end, spec- arXiv:2007.14139v2 [physics.optics] 1 Apr 2021 plane and the detector plane. We hereby point out that tral methods have recently been proposed for the general the model equally applies to Fourier ptychography [12], phase retrieval problem with independent and identically if one regards (x) and a(x) as the Fourier transform of distributed (i.i.d.) random measurements, solved with a the object transmission and the probe functions respec- gradient descent (GD) approach [8]. The initial estimate tively, and Pd is an inverse Fourier transform. In what is defined as the leading eigenvector of a covariance ma- follows the case of ptychography will be considered, with trix, constructed from the experimental measurements a setup sketched in the top row of Fig. 1. To recover the object, ptychography solves a phase re- trieval problem that can be formulated in the vectorized form ∗ These authors equally contributed to this work. Corresponding author: [email protected] y = jS j2 : (2) 2 the bottom row of Fig. 1. Probe Object Propagator Ptychograms In order to apply spectral methods to ptychography, the forward model in the continuous domain of Eq. (1) needs to be brought into the discrete vectorized form of Eq. (2). Let a(l) 2 Cm×n be the discrete and vector- (l) m×m ized form of a(x − x ), and Pd 2 C be the matrix computing the linear transform Pd. Now we can adopt the matrix factorization used in [11] and finally write y = jS j2, with S ≡ PA and 2 3 2 (1) 3 Pd ··· 0 a 6 . .. 7 Lm×Lm 6 . 7 Lm×n P ≡ 4 . 5 2 C ; A ≡ 4 . 5 2 C ; (L) 0 ··· Pd a (4) from which, using Eq. (3) and the unitarity of P (i.e., P y = P −1), one obtains Leading eigenvector of 풁 Iterative optimization (GD, PIE, etc.) Z = AyP −1 diagfT (y)gPA : (5) FIG. 1. Sketch of a ptychography setup and algorithmic Our choice of the preprocessing function, reading pipeline with our spectral method. Top: an aperture gen- 0 0 erates the probe a(x), that is scanned with shifts x(l) across T (y ) ≡ max(−β; 1 − 1=y ) ; (6) an object of interest (x), for l = 1;:::;L. For each probe 0 position, a camera records a ptychogram y(l)(x) after free where y collects the ptychograms, each normalized to space propagation by a distance d. Middle: the ptychograms its own mean intensity, and 0 < β < 1, was inspired 0 are preprocessed and used to construct a weighted covariance by the theoretical works in [9, 10, 14]. For y & 1, matrix Z. Bottom: the spectral estimate, obtained as the the function corresponds to the optimal preprocessings leading eigenvector of Z, is already close to the solution and of [9, 10]. The lower bound set at low intensity values is further refined using an iterative optimization (dark blue prevents large negative eigenvalues of Z, that will de- point). Initial estimates further away from the solution (light grade the performance of the power iteration algorithm. blue point) may be stuck in local minima. GD: gradient de- It is indeed known that power iterations return the eigen- scent; PIE: ptychographic iterative engine. vector corresponding to the largest eigenvalue in absolute value. In addition, it reminds of the truncated amplitude flow algorithm, whose resilience to noise has been proved Here, 2 n represents the vector of the unknown C [14]. After plugging Eq. (6) into Eq. (3), the power object transmission coefficients at each of its n pixels, method resembles a GD update, proving that no com- S 2 p×n is the sensing matrix, with p being the num- C putational complexity is added when replacing one GD ber of measured intensity values (i.e., p = Lm, where m iteration with one power iteration (see Appendix B). is the number of camera pixels), y 2 p is the vector R In the box Algorithm 1 we summarize the overall al- collecting the ptychograms, and j · j denotes the element- gorithm, consisting of the combination of a spectral ini- wise modulus operation. tialization and a GD optimization with amplitude loss When the entries of S are i.i.d. random variables, p function L( ) ≡ k y −jS jk2, where k·k indicates the is initially estimated as the leading eigenvector of the 2 2 L2-norm. The corresponding code is available at Ref. following n × n matrix (Fig. 1, middle row): [15], and more details on a memory-efficient implemen- Z ≡ Sy diagfT (y)gS ; (3) tation of power iterations are outlined in Appendix C. where Sy denotes the Hermitian conjugate of the ma- trix S, T is a preprocessing function acting element-wise III. EXPERIMENTS on the measurements, and diagfug denotes a diagonal matrix with a general vector u on its main diagonal. A Experiments were performed using the THz imaging widely used method to compute the leading eigenvector is setup at Empa, the Swiss Federal Laboratories for Ma- power iterations, that converge exponentially towards the terials Science and Technology, equipped with a far- eigenvector corresponding to the largest eigenvalue in ab- infrared gas laser (FIRL 100, Edinburgh Instruments, solute value [13]. This estimate is finally used as the first Livingston, Scotland) emitting several tens of mW of estimate of an iterative optimization algorithm like gra- continuous-wave power at the wavelength λ = 96.5 µm. dient descent or, in the case of ptychography, dedicated An uncooled microbolometer array detector featuring m routines like the ptychographic iterative engine (PIE) [1]. = 480 × 640 pixels on a pitch of 17 µm (Gobi-640-GigE, The reconstruction pipeline is graphically summarized in Xenics, Leuven, Belgium), was used as a THz camera 3 Algorithm 1: Solve ptychography with a of the power method, which can be explained by com- spectral initialization paring one iteration of the power method with one GD Input: Measurements y, sensing matrix S, iteration (see Appendix B). When such a decrease came to a halt, power iterations were stopped, and the ob- preprocessing function T , initial estimate 0, number of power iterations M, number of GD tained spectral estimate was used as the initial guess for iterations N, step size γ GD, run with an exponentially decreasing step size, com- y 1 Z = S diagfT (y)gS patible with a backtracking line search, to ensure stable /* Spectral initialization */ convergence.
Load more

Recommended publications
  • Subwavelength Resolution Fourier Ptychography with Hemispherical Digital Condensers
    Subwavelength resolution Fourier ptychography with hemispherical digital condensers AN PAN,1,2 YAN ZHANG,1,2 KAI WEN,1,3 MAOSEN LI,4 MEILING ZHOU,1,2 JUNWEI MIN,1 MING LEI,1 AND BAOLI YAO1,* 1State Key Laboratory of Transient Optics and Photonics, Xi’an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi’an 710119, China 2University of Chinese Academy of Sciences, Beijing 100049, China 3College of Physics and Information Technology, Shaanxi Normal University, Xi’an 710071, China 4Xidian University, Xi’an 710071, China *[email protected] Abstract: Fourier ptychography (FP) is a promising computational imaging technique that overcomes the physical space-bandwidth product (SBP) limit of a conventional microscope by applying angular diversity illuminations. However, to date, the effective imaging numerical aperture (NA) achievable with a commercial LED board is still limited to the range of 0.3−0.7 with a 4×/0.1NA objective due to the constraint of planar geometry with weak illumination brightness and attenuated signal-to-noise ratio (SNR). Thus the highest achievable half-pitch resolution is usually constrained between 500−1000 nm, which cannot fulfill some needs of high-resolution biomedical imaging applications. Although it is possible to improve the resolution by using a higher magnification objective with larger NA instead of enlarging the illumination NA, the SBP is suppressed to some extent, making the FP technique less appealing, since the reduction of field-of-view (FOV) is much larger than the improvement of resolution in this FP platform. Herein, in this paper, we initially present a subwavelength resolution Fourier ptychography (SRFP) platform with a hemispherical digital condenser to provide high-angle programmable plane-wave illuminations of 0.95NA, attaining a 4×/0.1NA objective with the final effective imaging performance of 1.05NA at a half-pitch resolution of 244 nm with a wavelength of 465 nm across a wide FOV of 14.60 mm2, corresponding to an SBP of 245 megapixels.
    [Show full text]
  • Near-Field Fourier Ptychography: Super- Resolution Phase Retrieval Via Speckle Illumination
    Near-field Fourier ptychography: super- resolution phase retrieval via speckle illumination HE ZHANG,1,4,6 SHAOWEI JIANG,1,6 JUN LIAO,1 JUNJING DENG,3 JIAN LIU,4 YONGBING ZHANG,5 AND GUOAN ZHENG1,2,* 1Biomedical Engineering, University of Connecticut, Storrs, CT, 06269, USA 2Electrical and Computer Engineering, University of Connecticut, Storrs, CT, 06269, USA 3Advanced Photon Source, Argonne National Laboratory, Argonne, IL 60439, USA. 4Ultra-Precision Optoelectronic Instrument Engineering Center, Harbin Institute of Technology, Harbin 150001, China 5Shenzhen Key Lab of Broadband Network and Multimedia, Graduate School at Shenzhen, Tsinghua University, Shenzhen, 518055, China 6These authors contributed equally to this work *[email protected] Abstract: Achieving high spatial resolution is the goal of many imaging systems. Designing a high-resolution lens with diffraction-limited performance over a large field of view remains a difficult task in imaging system design. On the other hand, creating a complex speckle pattern with wavelength-limited spatial features is effortless and can be implemented via a simple random diffuser. With this observation and inspired by the concept of near-field ptychography, we report a new imaging modality, termed near-field Fourier ptychography, for tackling high- resolution imaging challenges in both microscopic and macroscopic imaging settings. The meaning of ‘near-field’ is referred to placing the object at a short defocus distance with a large Fresnel number. In our implementations, we project a speckle pattern with fine spatial features on the object instead of directly resolving the spatial features via a high-resolution lens. We then translate the object (or speckle) to different positions and acquire the corresponding images using a low-resolution lens.
    [Show full text]
  • Sub-Diffraction Imaging Using Fourier Ptychography and Structured Sparsity
    SUB-DIFFRACTION IMAGING USING FOURIER PTYCHOGRAPHY AND STRUCTURED SPARSITY Gauri Jagatap, Zhengyu Chen, Chinmay Hegde, Namrata Vaswani Electrical and Computer Engineering, Iowa State University, Ames, IA 50011 ABSTRACT fewer samples suffice. The assumption of sparsity is natural in sev- eral applications in imaging systems, such as sub-diffraction imag- We consider the problem of super-resolution for sub-diffraction ing, X-ray crystallography, bio-imaging and astronomical imaging imaging. We adapt conventional Fourier ptychographic approaches, [12, 13, 14]; in these applications, the object to be imaged is often for the case where the images to be acquired have an underlying modeled as sparse in the canonical (or wavelet) basis. structured sparsity. We propose some subsampling strategies which Moving beyond sparsity, several algorithms that leverage more can be easily adapted to existing ptychographic setups. We then refined structured sparsity modeling assumptions (such as block use a novel technique called CoPRAM with some modifications, sparsity) achieve considerably improved sample-complexity for re- to recover sparse (and block sparse) images from subsampled pty- constructing images from compressive measurements [15, 16, 17]. chographic measurements. We demonstrate experimentally that this However, analogous algorithms in the context of phase retrieval are algorithm performs better than existing phase retrieval techniques, in not well-studied. In recent previous work [18], we have developed terms of quality of reconstruction, using fewer number of samples. a theoretically-sound algorithmic approach for phase retrieval that Index Terms— Phase retrieval, ptychography, sparsity, non- integrate sparsity (as well as structured sparsity) modeling assump- convex algorithms tions within the reconstruction process. However, that work also assumes certain stringent probabilistic assumptions on the measure- 1.
    [Show full text]
  • This Is a Repository Copy of Ptychography. White Rose
    This is a repository copy of Ptychography. White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/127795/ Version: Accepted Version Book Section: Rodenburg, J.M. orcid.org/0000-0002-1059-8179 and Maiden, A.M. (2019) Ptychography. In: Hawkes, P.W. and Spence, J.C.H., (eds.) Springer Handbook of Microscopy. Springer Handbooks . Springer . ISBN 9783030000684 https://doi.org/10.1007/978-3-030-00069-1_17 This is a post-peer-review, pre-copyedit version of a chapter published in Hawkes P.W., Spence J.C.H. (eds) Springer Handbook of Microscopy. The final authenticated version is available online at: https://doi.org/10.1007/978-3-030-00069-1_17. Reuse Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item. Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request. [email protected] https://eprints.whiterose.ac.uk/ Ptychography John Rodenburg and Andy Maiden Abstract: Ptychography is a computational imaging technique. A detector records an extensive data set consisting of many inference patterns obtained as an object is displaced to various positions relative to an illumination field.
    [Show full text]
  • Low-Rank Fourier Ptychography
    LOW RANK FOURIER PTYCHOGRAPHY Zhengyu Chen, Gauri Jagatap, Seyedehsara Nayer, Chinmay Hegde, Namrata Vaswani Iowa State University ABSTRACT reconstruction problem is nascent, and several important questions remain unanswered. Rigorous guarantees of correctness of the meth- In this paper, we introduce a principled algorithmic approach for ods proposed in [8, 10] are not yet available. More worrisome is the Fourier ptychographic imaging of dynamic, time-varying targets. To issue of sample complexity of image reconstruction, since a basic the best of our knowledge, this setting has not been explicitly ad- requirement in all existing methods is that the number of measure- dressed in the ptychography literature. We argue that such a setting ments must exceed the resolution of the image. However, this re- is very natural, and that our methods provide an important first step quirement can be particularly challenging when imaging a dynamic towards helping reduce the sample complexity (and hence acquisi- scene involving a moving target; for a video sequence with q images tion time) of imaging dynamic scenes to managaeble levels. With each with resolution n, without using any structural prior assump- significantly reduced acquisition times per image, it is conceivable tions, the number of observations must be at least Ω(nq), which can that dynamic ptychographic imaging of fast changing scenes indeeed quickly become prohibitive. becomes practical in the near future. Index Terms— Phase retrieval, ptychography, low rank 1.2. Our Contributions 1. INTRODUCTION In this paper, we introduce a principled algorithmic approach for Fourier ptychographic imaging of dynamic, time-varying targets. To 1.1. Motivation the best of our knowledge, this setting has not been explicitly ad- dressed in the ptychography literature.
    [Show full text]
  • Phase Retrieval for Fourier Ptychography Under Varying Amount of Measurements
    BOOMINATHAN ET AL.: PHASE RETRIEVAL FOR FOURIER PTYCHOGRAPHY 1 Phase retrieval for Fourier Ptychography under varying amount of measurements Lokesh Boominathan12 1 Computational Imaging Lab, [email protected] Indian Institute of Technology Madras, Mayug Maniparambil1 Chennai, India [email protected] 2 Aindra Systems Honey Gupta1 Bengaluru, India [email protected] Rahul Baburajan1 [email protected] Kaushik Mitra1 [email protected] Abstract Fourier Ptychography is a recently proposed imaging technique that yields high- resolution images by computationally transcending the diffraction blur of an optical system. At the crux of this method is the phase retrieval algorithm, which is used for computationally stitching together low-resolution images taken under varying illumina- tion angles of a coherent light source. However, the traditional iterative phase retrieval technique relies heavily on the initialization and also needs a good amount of overlap in the Fourier domain for the successively captured low-resolution images, thus increas- ing the acquisition time and data. We show that an auto-encoder based architecture can be adaptively trained for phase retrieval under both low overlap, where traditional tech- niques completely fail, and at higher levels of overlap. For the low overlap case we show that a supervised deep learning technique using an autoencoder generator is a good choice for solving the Fourier ptychography problem. And for the high overlap case, we show that optimizing the generator for reducing the forward model error is an appropriate choice. Using simulations for the challenging case of uncorrelated phase and amplitude, we show that our method outperforms many of the previously proposed Fourier ptychog- raphy phase retrieval techniques.
    [Show full text]
  • Sample Efficient Fourier Ptychography for Structured Data
    1 Sample Efficient Fourier Ptychography for Structured Data Gauri Jagatap, Zhengyu Chen, Seyedehsara Nayer, Chinmay Hegde, Senior Member, IEEE and Namrata Vaswani, Fellow, IEEE Abstract—We study the problem of recovering structured data shifts induced by the optical lens setup. However, the sensing from Fourier ptychography measurements. Fourier ptychogra- apparatus is incapable of estimating the phase of the complex phy is an image acquisition scheme that uses an array of images values, and only the magnitudes can be measured. to produce high-resolution images in microscopy as well as long- distance imaging, to mitigate the effects of diffraction blurring. This setup can be molded to that of the classical problem The number of measurements is typically much larger than the of phase retrieval [7], [8], [9], which is a non-linear, ill- size of the signal (image or video) to be reconstructed, which posed inverse problem. In phase retrieval, the goal is to recon- translates to high storage and computational requirements. struct a discretized image (or video) of size n (or nq) from The issue of high sample complexity can be alleviated by utiliz- noisy, magnitude-only observations of the image’s discrete ing structural properties of the image (or video). In this paper, we first discuss a range of sub-sampling schemes which can reduce Fourier transform (DFT) coefficients. A generalized version the amount of measurements in Fourier ptychography setups; of this problem replaces the DFT coefficients with a generic however, this makes the problem ill-posed. Correspondingly, we linear operator constructed by sampling certain families of impose structural constraints on the signals to be recovered, probability distributions.
    [Show full text]
  • REFLECTIVE MODE FOURIER PTYCHOGRAPHY by Basel
    REFLECTIVE MODE FOURIER PTYCHOGRAPHY by Basel Salahieh Committee Members Rongguang Liang (Chair), Jeffrey Rodriguez (Member), Tom Milster (Member) __________________________ Copyright © Basel Salahieh 2015 A Master’s Report Submitted to the College of the OPTICAL SCIENCES In Partial Fulfillment of the Requirements For the Degree of MASTER OF SCIENCE In the Graduate College THE UNIVERSITY OF ARIZONA 2015 Table of Contents List of Figures ................................................................................................................................. 3 Abstract ........................................................................................................................................... 4 Overview of Reflective Imaging Systems ...................................................................................... 5 Background in Optics ................................................................................................................. 5 Limitations of Conventional Imaging Systems........................................................................... 6 Introduction to Fourier Ptychography ............................................................................................. 7 Algorithm .................................................................................................................................... 7 Simulating the Iterative FP Algorithm ........................................................................................ 9 Case Studies on Selected FP Parameters .................................................................................
    [Show full text]
  • Data Preprocessing Methods for Robust Fourier Ptychographic Microscopy
    Data preprocessing methods for robust Fourier ptychographic microscopy 1,2,3 1,2,3 1 1,* YAN ZHANG , AN PAN , MING LEI , AND BAOLI YAO 1 State Key Laboratory of Transient Optics and Photonics, Xi’an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi’an 710119, China 2University of Chinese Academy of Sciences, Beijing 100049, China 3These authors contribute equally to this paper *Corresponding author: [email protected] Fourier ptychographic microscopy (FPM) is a recently proposed computational imaging technique with both high resolution and wide field-of-view. In current FP experimental setup, the dark-field images with high-angle illuminations are easily submerged by stray light and background noise due to the low signal-to-noise ratio, thus significantly degrading the reconstruction quality and also imposing a major restriction on the synthetic numerical aperture (NA) of the FP approach. To this end, an overall and systematic data preprocessing scheme for noise removal from FP’s raw dataset is provided, which involves sampling analysis as well as underexposed/overexposed treatments, then followed by the elimination of unknown stray light and suppression of inevitable background noise, especially Gaussian noise and CCD dark current in our experiments. The reported non-parametric scheme facilitates great enhancements of the FP’s performance, which has been demonstrated experimentally that the benefits of noise removal by these methods far outweigh its defects of concomitant signal loss. In addition, it could be flexibly cooperated with the existing state-of-the- art algorithms, producing a stronger robustness of the FP approach in various applications.
    [Show full text]
  • PTYCHNET : CNN BASED FOURIER PTYCHOGRAPHY Armin Kappeler
    PTYCHNET : CNN BASED FOURIER PTYCHOGRAPHY Armin Kappeler, Sushobhan Ghosh, Jason Holloway, Oliver Cossairt, Aggelos Katsaggelos Department of Electrical Engineering and Computer Science, Northwestern University, Evanston, IL 60208, USA ABSTRACT Fraunhofer diffraction Fourier ptychography is an imaging technique that overcomes ψ(x, y) ψˆ(u, v) A(u, v) I(x, y) the diffraction limit of conventional cameras with applications ψ(x, y) ψˆ(u, v) A(u, v) I(x, y) in microscopy and long range imaging. Diffraction blur causes ψ(x, y) ψˆ(u, v) A(u, v) I(x, y) resolution loss in both cases. In Fourier Ptychography, a co- Image herent light source illuminates an object, which is then im- sensor aged from multiple viewpoints. The reconstruction of the ob- Coherent Transmissiveψ(x, y) ψˆ(u, v) A(u, v) I(x, y) ject from these set of recordings can be obtained by an itera- source object tive phase retrieval algorithm. However, the retrieval process is Lens slow and does not work well under certain conditions. In this (aperture) paper, we propose a new reconstruction algorithm that is based on convolutional neural networks and demonstrate its advan- Fig. 1: Example setup for Fourier Ptychography (FP). Coherent tages in terms of speed and performance. light diffracts through a translucent medium into the far-field. A Index Terms— Fourier Ptychography, Convolutional Neu- lens samples a portion of the Fourier domain which is recorded ral Network, CNN as intensity images at the sensor. See Section 2.1 for details. 1. INTRODUCTION on the algorithm for retrieving the high resolution image.
    [Show full text]
  • Parallelized Aperture Synthesis Using Multi-Aperture Fourier Ptychographic Microscopy
    Parallelized aperture synthesis using multi-aperture Fourier ptychographic microscopy PAVAN CHANDRA KONDA, JONATHAN M. TAYLOR, ANDREW R. HARVEY* Imaging Concepts Group, School of Physics and Astronomy, University of Glasgow, Scotland, G12 8QQ, UK *Corresponding author: [email protected] We report a novel microscopy platform, termed Multi-Aperture Fourier ptychographic microscopy (MA-FPM), capable of realizing gigapixel complex field images with large data acquisition bandwidths – in gigapixels per second. MA-FPM is a synthetic aperture technique: an array of objectives together with tilt-shift illumination are used to synthesize high-resolution, wide field-of-view images. Here, the phase is recovered using Fourier ptychography (FP) algorithms, unlike conventional optical synthetic aperture techniques where holographic measurements are used. The parallel data-acquisition capability due to multiple objectives provides unprecedented bandwidth enabling imaging of high-speed in vitro processes over extended depth-of-field. Here, we present a proof-of-concept experiment demonstrating high-quality imaging performance despite using nine- fold fewer illumination angles compared to an equivalent FP setup. Calibration procedures and reconstruction algorithm were developed to address the challenges of multiple imaging systems. Our technique provides a new imaging tool to study cell cultures, offering new insights into these processes. space-bandwidth product (SBP) [7,9] is achieved with the quid pro quo 1. INTRODUCTION of increased recording time. We report here the first demonstration of multi-aperture Fourier ptychographic microscopy (MA-FPM), in which Perhaps the most important advantage of the emerging technique of multiple objectives, each forming an image on an independent detector Fourier ptychography (FP) [1] is the ability to record gigapixel images array, capture this diffracted light and hence increase the space- with sub-micron resolution using low numerical-aperture (NA) optics.
    [Show full text]
  • Fast Digital Refocusing and Depth of Field Extended Fourier Ptychography Microscopy
    Fast digital refocusing and depth of field extended Fourier ptychography microscopy GUOCHENG ZHOU,1,2 SHAOHUI ZHANG,1,2,* CHUANJIAN ZHENG,1 TONG LI,1 YAO HU,1 AND QUN HAO1,* 1School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China 2These authors contributed equally to this work *[email protected] *[email protected] Abstract: Fourier ptychography microscopy (FPM), sharing its roots with synthetic aperture technique and phase retrieval method, is a recently developed computational microscopic super-resolution technique. By turning on the light-emitting diode (LED) elements sequentially and acquiring the corresponding images that contain different spatial frequencies, FPM can achieve a wide field-of-view (FOV), high-spatial-resolution imaging and phase recovery simultaneously. Conventional FPM assumes that the sample is sufficiently thin and strictly in focus. Nevertheless, even for a relatively thin sample, the non-planar distribution characteristics and the non-ideal position/posture of the sample will cause all or part of FOV to be defocused. In this paper, we proposed a fast digital refocusing and depth-of-field (DOF) extended FPM strategy by taking the advantages of image lateral shift caused by sample defocusing and varied- angle illuminations. The lateral shift amount is proportional to the defocus distance and the tangent of the illumination angle. Instead of searching the optimal defocus distance in optimization strategy, which is time consuming, the defocus distance of each subregion of the sample can be precisely and quickly obtained by calculating the relative lateral shift amounts corresponding to different oblique illuminations. And then, digital refocusing strategy rooting in the Fresnel propagator is integrated into FPM framework to achieve the high-resolution and phase information reconstruction for each part of the sample, which means the DOF the FPM is effectively extended.
    [Show full text]