Data Driven Computational Methods In Optical Imaging And Characterization

A comprehensive and up-to-date textbook and reference for computational imaging, which combines vision, graphics, signal processing, and optics. Computational imaging involves the joint design of imaging hardware and computer algorithms to create novel imaging systems with unprecedented capabilities. In recent years such capabilities include cameras that operate at a trillion frames per second, microscopes that can see small viruses long thought to be optically irresolvable, and telescopes that capture images of black holes. This text offers a comprehensive and up-to-date introduction to this rapidly growing field, a convergence of vision, graphics, signal processing, and optics. It can be used as an instructional resource for computer imaging courses and as a reference for professionals. It covers the fundamentals of the field, current research and applications, and light transport techniques. The text first presents an imaging toolkit, including optics, image sensors, and illumination, and a computational toolkit, introducing modeling, mathematical tools, model-based inversion, data-driven inversion techniques, and hybrid inversion techniques. It then examines different modalities of light, focusing on the plenoptic function, which describes degrees of freedom of a light ray. Finally, the text outlines light transport techniques, describing imaging systems that obtain micron-scale 3D shape or optimize for noise-free imaging, optical computing, and non-line-of-sight imaging. Throughout, it discusses the use of computational imaging methods in a range of application areas, including smart phone photography, autonomous driving, and medical imaging. End-of-chapter exercises help put the material in context.

This dissertation presents my work in application of computational techniques and the resulting enhancements to several non-linear and ultrafast optical imaging and spectroscopy modalities. The importance of novel computational optical imaging schemes which aim to overcome the limitations of conventional imaging techniques by leveraging the availability of computational resources and the vast body of literature in computational signal processing is emphasized. In particular the computational techniques of compressive sensing and two dimensional phase retrieval are introduced in the broad context as inversion techniques suitable for optical applications including coherent anti-Stokes Raman holography, non-linear spectroscopy, and ultrashort pulse characterization. It is shown that both computational techniques seek to improve key signal metrics such as higher signal to noise ratio (SNR) and better resolution than can be obtained traditionally within the specific imaging modality.Coherent anti-Stokes Raman scattering (CARS) holography is a novel imaging modality which combines the principles of coherent anti-Stokes Raman scattering and holography to provide label-free, chemical selective, scanning-free, and single shot 3D imaging modality. Compressive CARS holography is introduced as a sparsity constrained holographic image reconstruction technique to enhance the optical sectioning capability of CARS holography by suppressing out-of-focus background noise inherent in 3D images processed from a typical single 2D hologram. The advantages of compressive sensing guided signal acquisition strategy in optical spectroscopy is presented by proposing compressive multi-heterodyne optical spectroscopy as a novel technique for ultra-high resolution frequency comb spectroscopy. Using numerical simulations, our proposed compressive frequency comb spectroscopy technique is shown to be well-suited for recording narrow line spectra at ultra-high sampling over broad spectral range by leveraging sparsity inherent in such spectra.We next present applications of phase retrieval in optical imaging and spectroscopy. In particular we use 2D phase retrieval technique to enhance the resolution of sum frequency generation vibrational spectroscopy (SFG-VS) whose unique surface selectivity enables qualitative and quantitative study of chemical species at surfaces/interfaces. Specifically, our key contribution is that we show that our 2D phase retrieval based inversion algorithm enables measurement of characteristic molecular vibrational spectra of air/dimethyl sulfoxide interface at resolutions significantly better than that achievable in conventional SFG-VS acquisition system. Lastly, we address the limitation of the commonly used pulse characterization technique: frequency resolved optical gating (FROG) to spatio-temporally characterize the ultrafast pulse. Using a simple spectral holographic recording technique, we present a modified 2D phase retrieval based algorithm to measure the spectral phase at every spatial location in the vicinity of focus of an objective and thereby track the spatio-temporal evolution of the pulse along its optical axis.

Computational Optical Biomedical Spectroscopy and Imaging covers recent discoveries and research in the field by some of the best inventors and researchers in the world. It also presents useful computational methods and applications used in optical biomedical spectroscopy and imaging. Topics covered include: New trends in immunohistochemical, genome, and metabolomics imaging Computer-aided diagnosis of interstitial lung diseases based on CT image analysis Functional near-infrared spectroscopy and its applications in neurosciences Applications of vibrational spectroscopic imaging in personal care studies Induced optical natural fluorescence spectroscopy for Giardia lamblia cysts Nanoimaging and polarimetric exploratory data analysis Fluorescence bioimaging with applications to chemistry Medical imaging instrumentation and techniques The book also discusses future applications, directions, opportunities, and challenges of optical biomedical spectroscopy and imaging in technical industry, academia, and government. This valuable resource introduces key concepts of computational methods used in optical biomedical spectroscopy and imaging in a manner that is easily understandable to beginners and experts alike.

This open access book, edited and authored by a team of world-leading researchers, provides a broad overview of advanced photonic methods for nanoscale visualization, as well as describing a range of fascinating in-depth studies. Introductory chapters cover the most relevant physics and basic methods that young researchers need to master in order to work effectively in the field of nanoscale photonic imaging, from physical first principles, to instrumentation, to mathematical foundations of imaging and data analysis. Subsequent chapters demonstrate how these cutting edge methods are applied to a variety of systems, including complex fluids and biomolecular systems, for visualizing their structure and dynamics, in space and on timescales extending over many orders of magnitude down to the femtosecond range. Progress in nanoscale photonic imaging in Göttingen has been the sum total of more than a decade of work by a wide range of scientists and mathematicians across disciplines, working together in a vibrant collaboration of a kind rarely matched. This volume presents the highlights of their research achievements and serves as a record of the unique and remarkable constellation of contributors, as well as looking ahead at the future prospects in this field. It will serve not only as a useful reference for experienced researchers but also as a valuable point of entry for newcomers.

This comprehensive and self-contained text for researchers and professionals presents a detailed account of optical imaging from the viewpoint of both ray and wave optics.

Computational Methods for Microstructure-Property Relationships introduces state-of-the-art advances in computational modeling approaches for materials structure-property relations. Written with an approach that recognizes the necessity of the engineering computational mechanics framework, this volume provides balanced treatment of heterogeneous materials structures within the microstructural and component scales. Encompassing both computational mechanics and computational materials science disciplines, this volume offers an analysis of the current techniques and selected topics important to industry researchers, such as deformation, creep and fatigue of primarily metallic materials. Researchers, engineers and professionals involved with predicting performance and failure of materials will find Computational Methods for Microstructure-Property Relationships a valuable reference.

One of the greatest challenges in computational imaging is scaling it to work outside the lab. The main reasons for that challenge are the strong dependency on precise calibration, accurate physical models, and long acquisition times. These prevent practical progress towards medical imaging and seeing through occlusions such as fog in the wild. This dissertation demonstrates that with data-driven and probabilistic modeling we can alleviate these dependencies, and pave the way towards real-world time-resolved computational imaging through extreme scattering conditions using visible light. The ability to image through scattering media in the visible part of the electromagnetic spectrum holds many applications in various industries. For example, seeing through fog would enable autonomous robots to operate in challenging weather conditions; augment human driving; and allow airplanes, helicopters, and drones to take off and land in dense fog conditions. In medical imaging, the ability to see into the body with near-infrared light would reduce the exposure to ionizing radiation and provide more clinically meaningful data. In order to image in diverse and extreme scattering conditions, we develop novel algorithms inspired by techniques in signal processing, optimization, statistical analysis, compressive sensing, and machine learning that leverage time-resolved sensing. More specifically, we demonstrate techniques that computationally leverage all of the optical signal, including scattered light, as opposed to locking onto a specific part of the optical signal. Furthermore, we show that by introducing probabilistic formulation to the imaging problem, the resulting system does not require user input for calibration and priors; this makes our systems more practical for real-world scenarios and enables them to operate in a wide range of scattering conditions. We consider four cases of imaging through scattering media with increasing complexity: 1. A theoretical analysis of time-resolved single pixel imaging, which demonstrates scene reconstruction even when the entire scene is measured with a single pixel, an equivalent of simple scattering or a blur that is easy to model. 2. A data-driven calibration invariant technique for imaging through simple scattering (a sheet of paper). 3. Imaging through a thick tissue phantom by utilizing all of the optical signal with minimal assumptions on the tissue properties. 4. Imaging through a wide range of dense, dynamic, and heterogeneous fog conditions. In that case, we introduce a probabilistic model that is able to recover the occluded target reflectance and depth without any assumption about the fog.