This PDF file contains the front matter associated with SPIE Proceedings Volume 10889, including the Title Page, Copyright Information, Table of Contents, Author and Conference Committee lists.

The depth resolution achieved by a continuous wave time-of-flight (C-ToF) imaging system is determined by the coding (modulation and demodulation) functions that it uses. We present a mathematical framework for exploring and characterizing the space of C-ToF coding functions in a geometrically intuitive space. Using this framework, we design families of novel coding functions that are based on Hamiltonian cycles on hypercube graphs. The proposed Hamiltonian coding functions achieve up to an order of magnitude higher resolution as compared to the current state-of-the-art. Using simulations and a hardware prototype, we demonstrate the performance advantages of Hamiltonian coding in a wide range of imaging settings.

This paper presents the last leg of the evolution of the Master Slave (MS) optical coherence tomography (OCT) technology, towards complex master slave (CMS), where phase information is also delivered. We will show how matrix manipulation of signals can lead to real time display. We have demonstrated that this can be executed on central processing units (CPU)s with no need for graphic processing units (GPU)s, yielding simultaneous display of multiple en-face OCT images (C-scans), two cross-section OCT images (B-scans) and an aggregated image, equivalent to a scanning laser ophthalmoscopy (SLO) image when imaging the retina, which is similar to a confocal microscopy image. The same protocol can obviously be applied employing GPUs when using faster acquisition engines, such as multi MHz swept optical sources.

Probing vibrational modes of molecules by Coherent Raman Scattering (CRS) provides a non-invasive, label free and chemically sensitive imaging modality that is of a large interest for the study of biological systems. Stimulated Raman Scattering and Coherent Anti-Stokes Raman Scattering provide quantitative imaging of the density of resonant vibrational modes within a focal volume. However, those modes are also sensitive to the incident polarization of light, which can give additional insight into the structural order and orientation of molecular assemblies. Accessing such sub-diffraction molecular order information opens ways to understand the interplay between organization and function in biological systems, as well as its alteration in the progression of diseases. A limiting factor of polarized-CRS imaging so far was the long acquisition times due to a slow step wise rotation of the incident polarization state, leading to image acquisition times of 100x100 μm field of views in the minute time scales. In this work, we use electro-optic modulation to fully rotate the incident polarization state at rates of 100 kHz. Coupled with lock-in detection, we show that one can obtain order and orientation parameters of molecular vibrations down to sub-seconds time scales per image. Additional incident intensity modulation permits to access simultaneously molecular density information, with no compromise in signal to noise conditions. Using this method, we demonstrate real-time deformation and dynamics of model membrane systems (multilamellar vesicles). Moreover, we show the sensitivity of the technique down to one single lipid bilayer as present in red blood cell ghosts.

Detection of circulating tumor cells with image cytometry is limited by the sensitivity and specificity of the biomarker panel. We collected confocal images of ~100,000 cells labeled for DNA, lipids, CD45, and Cytokeratin on a model system of MCF7 and WBCs representing disease positive, D+ and disease negative, D- populations. We computed spatial image metrics and performed multivariable regression and feature selection, increasing the separation of the D+ and D- populations to 7 standard deviations with detection limit of ~1 in 480. In conclusion, simple regression analysis holds promise to improve the separation of rare cells in cytometry applications.

Methods for imaging of excised tissue specimens that obviate manual and time-consuming histology processing steps including wax-embedding, sectioning, and separate staining hold considerable promise for improving our ability to render time- and cost-efficient diagnoses. They also could potentiate the clinical adoption of machine learning tools dependent on routine production of digital histology data. A known limitation of point scanning-based ex vivo imaging technologies that precludes ready adoption for clinical use is imaging speed. We have recently described the development of an approach for multiphoton microscopy capable of imaging un-sectioned and un-embedded human tissue samples at millimeter depths with sufficient quality for primary diagnostic interpretation and at speeds more than 30-times faster than a traditional galvanometer-based system. Incorporation of a polygonal mirror combined with stage scanning in an optically efficient geometry yielded significant speed gains, beyond those possible with resonant galvanometers, and without sacrificing image quality. We hypothesized that further gains in speed, with maintenance of image quality, would be attainable by incorporating pulsed laser excitation with a repetition rate beyond the typical 80 MHz speed of standard Ti-Sapphire lasers. In this analysis we describe the use of a newly produced 250 MHz ultra-fast laser in a polygon-based microscope with stage scanning and demonstrate multiphoton human tissue imaging at speeds on par with the fastest whole slide imaging systems and with resolution, contrast, and coloration that matches physical slides but without any of the principal artifacts associated with slide scanning.

Recent advances in imaging cytometry enable high-resolution analysis of single-cell phenotypes (both physical and biochemical) at high throughput with the overall aim of revealing the phenotypic variability within an enormous and heterogeneous population of cells. However, analysis of large-scale high dimensional single-cell image data is computationally intensive and soon becomes unscaleable from both a memory and run time perspective. To address this challenge, we develop Accelerated Pheno-Tree (APT) – an unsupervised clustering algorithm tailored for analyzing large-scale high dimensional single-cell image-based data. As a proof-of-concept demonstration, we adopt APT in time-stretch quantitative phase imaging (TS-QPI) – an ultrahigh-throughput label-free imaging technique that allows large-scale single-cell biophysical phenotyping. APT allows fast unbiased clustering and visualization of high-dimensional datasets of above 1 million single cell TS-QPI - bypassing the need for prior knowledge of the data as well as data down-sampling which are common in the existing clustering methods. Integrating two key computational steps, i.e. accelerated non-linear dimension reduction (LargeVis) of the TS-QPI data followed by the graph-based and data-driven agglomerative clustering (based on accelerated minimum spanning tree construction), APT successfully distinguishes multiple cell types (e.g. 7 lung cancer cell lines, and sub-types of PBMC cells) entirely based on their intrinsic biophysical phenotypes (up to 30 dimensions) quantified from label-free TS-QPI (total cell count: 1.1 million cells). We anticipate that APT could be particularly useful in ultralarge-scale single-cell analysis and facilitates exploration of the heterogeneity within cell populations based on single-cell biophysical features with high accuracy.

The real-time characterization of fast-evolving processes with sub-cellular resolution is crucial to unveil biological mechanisms as relevant as brain function or intracellular diffusion. Optical microscopes are optimal tools for 2D imaging of living organisms, but the need for z-focus translation significantly constraints volumetric imaging speed. In this talk, I will show how an acoustic optofluidic lens enables z-focus scanning at rates as high as 1 MHz. Such unsurpassed speed opens the door to novel strategies in 3D confocal and light-sheet systems in which the temporal and spatial resolution can be user-selected based on application, well-beyond the limits of traditional microscopes.

Light-sheet fluorescence microscopy (LSFM) enables both sensitive and rapid volumetric imaging, from single cells to multicellular organisms. Nevertheless, LSFM faces technological and physical limitations on the volumetric acquisition speed. We have realized that there are certain geometrical arrangements that allow lossless parallelization of the LSFM acquisition process. Thereby the volumetric acquisition rate can be increased without inducing higher amounts of phototoxicity or requiring faster detector technologies. We present two implementations of parallelized light-sheet microscopy and discuss applications that either benefit from the increased acquisition speed or improved sensitivity of these systems.

Studying neuronal connections and activities in vivo is critical for understanding the brain. Optical microscopy, with the capability of specific fluorescent labeling and sub-cellular spatial resolution, has become an indispensable tool in neuroscience. However, the major limitation of optical imaging is penetration depth and imaging speed to capture neural signal dynamics in deep brain regions. Recently, by applying adaptive optics, high-energy laser, or long wavelength lasers for nonlinear imaging, penetration depth around 1mm has been achieved in living mouse brains. Nevertheless, this depth barely pierces through the mouse cortex and is far from reaching the bottom of the centimeter-thick mouse brain. For studying deeper regions of the brain, brain slice is one possible approach, yet it is invasive and cut away many neuron connections. In this study, a home-built two-photon microscope is integrated with both a gradient refractive index (GRIN) lens and a tunable acoustic gradient (TAG) lens. The GRIN lens serves as a micro-endoscope which extends the imaging depth to a centimeter while minimizing the invasiveness, and the TAG lens provides ~100kHz axial scanning which enables high-speed volumetric imaging of neuronal response. This novel high-speed volumetric endoscopy system offers an unprecedented opportunity towards studying three-dimensional neuronal dynamics in deep brains regions of a living mouse.

Fast, volumetric imaging over large scales has been a long-standing challenge in biological microscopy. To address this issue, we developed a variant of confocal microscopy that provides simultaneous multiplane imaging over large field of view and at video rate. Our apparatus, called multi-Z confocal microscopy, differs from a conventional confocal microscope in both its illumination and detection parts. First, axially elongated illumination is achieved by under filling the back aperture of the microscope objective. The resulting low NA provides an axial extent of the illumination of the order of 100 µm. The light from the sample is then collected by the same objective, now taking advantage of the full NA which ensures high collection efficiency, and send to the detection unit. The latter is comprised by four reflecting pinholes axially distributed in the image plane such that they are conjugated to different depths within the sample. Each detection channel spans a probe volume at a different depth and volumetric imaging is obtained by simply combining the four channels. In our current configuration, each imaging plane covers a field of view of 1.2 mm and the distance between two planes is equal to 25 µm. In other words, we image 1200x1200x100 µm3 at 30Hz. We first demonstrated the applicability of our technique by imaging entire C. elegans in vivo with a cellular resolution. Secondly we applied our technique to image multiple layers of neurons in mouse brain. We were able to record the activity of 550 neurons, with 100-150 neurons present in each imaging plane.

We demonstrate a 512 x 16 CMOS single photon avalanche diode (SPAD) line sensor with per-pixel on-chip histogramming for video rate spectral fluorescence lifetime imaging (sFLIM). On-chip histogramming provides 32-bin histograms per pixel with 11bit/bin dynamic range. In addition, bin widths in time can be programmed from 51.20 ps to 6.55 ns, providing a histogram range from 1.64 ns to 209.72 ns to suit a wide range of fluorescence decays. At the end of a user defined exposure time, the full histogram data (i.e. 32-bins/pixel and 512 pixels) is first transferred to a FPGA in 84.48 μs via 64 data I/O pads at a 33.33 MHz I/O rate. The sensor data is then binned into two user defined spectral bands to provide spectral separation between different fluorophores, before being transferred to a PC via a USB3 connection for further processing. Fluorescence lifetimes for each spectral band are then rapidly estimated in software by applying the Centre-of-Mass Method (CMM), providing two 128 x 128 size spectral lifetime images in 1.384 s (i.e. with a frame rate of 0.72 fps). The frame rate can be increased by reducing the number of bins, reaching a maximum frame rate when only 2 bins are used with the Rapid Lifetime Determination (RLD) algorithm. In this paper we study the lifetime accuracy vs frame rate trade-offs by varying the number of histogram bins while carefully adjusting the bin widths for maximum bin counts. We validate the results using a Rhodamine 110 and Rhodamine B mixture solution which we separate them spectrally by their fluorescence lifetimes.

The Flamingo is a modular, shareable light sheet microscope suited to a new model of scientific collaboration. Each microscope is customized for a given application, equipped to travel from lab to lab and providing widespread access to advanced microscopy. It is a compact selective plane illumination microscope (SPIM) that can be turned upright for multi-view imaging of hanging samples or turned on its side for samples in a dish. Rapid multi-color imaging is achieved via sCMOS cameras and several possible laser lines. With its selective sheet illumination, the Flamingo is well suited for fast and gentle imaging of developing organisms.

The capability of Single-Photon Avalanche Diodes (SPADs) to detect photons with picosecond timing precision and shotnoise limited performance has given rise to a range of biological and biomedical applications, from Fluorescence Lifetime Imaging Microscopy (FLIM) to Raman Spectroscopy and Positron Emission Tomography (PET). The use of SPAD sensors has also been successfully demonstrated in Single-Molecule Localisation Microscopy. Traditionally implemented as point detectors, recent advances in SPAD technology, such as compact, binary pixels and back-side illuminated, 3D-stacked architectures, have led to 2-D imaging arrays of increasing resolution and fill factor. Combined with high frame rates (in the kFPS range), and negligible read noise, the sensors offer an exciting prospect for capturing fast temporal dynamics in life science cellular imaging. The work in this paper considers the application of SPAD imaging arrays to widefield fluorescence lifetime imaging of high-speed particles in microscopy. We demonstrate, using a time-gated binary SPAD array, that by tracking particles, and spatially re-assigning the underlying photon counts accordingly, lifetime estimates for fast-moving objects are possible. Moreover, we give the first demonstration of FLIM using a SPAD imaging array with on-chip histogramming of photon arrival time, with potential frame rates of several 100FPS. Both FLIM techniques are illustrated using experimental results based on fluorescent microspheres undergoing Brownian motion. The results pave the way towards applications in live-cell microscopy, such as the monitoring of the fluorescence lifetime of highly mobile cell structures, with a view, for example, to study molecular interactions using Förster Resonance Energy Transfer (FRET) measurements.

We present an achromatic confocal laser scanning system capable of recording spectrally resolved fluorescence lifetime images (sFLIMs) at a rate of >8 frames per second (FPS) for a 128 x 128 image. This frame rate was achieved by optimizing the processing of lifetime calculations from previous results which demonstrated >4 FPS sFLIM imaging. The imaging system is achromatic for a spectral range of 400 - 900 nm, achieved by using reflective optics instead of a transmissive lens system, except for the primary objectives. Two excitation sources have been integrated into the system, 485 nm and 640 nm laser diodes with a pulse width of <70 ps and <90 ps respectively. Imaging is performed via a galvanometric mirror system which scans the laser beam over the sample with the ability to change the Field of View (FOV) on the fly. The collected fluorescence signal is focused into a multimode fiber via a second objective and recollimated onto a transmissive grating for spectral dispersion onto a novel complementary metal–oxide–semiconductor single photon avalanche diode (CMOS SPAD) line array sensor. This sensor can perform lifetime histogram generation on-chip and process over 16.5 Giga events/s enabling fast lifetime data acquisition. High speed sFLIM is demonstrated through imaging of convallaria majalis sections.

Using Time-Correlated Single Photon Counting (TCSPC) for the purpose of fluorescence lifetime measurements is often limited in speed due to pile-up and dead-time artifacts. This is particularly critical in fast imaging applications. With modern instrumentation this limitation can be lifted by reducing the dead-time of the TCSPC electronics to the absolute minimum imposed by the speed of the electrical detector signals. Another, complementing approach to speedy image acquisition is parallelization by means of simultaneous, time tagged readout of many detector channels. This of course puts high demands on the data throughput of the TCSPC system. Here we present a new integrated design, providing up to 8 independent input channels, an extremely short dead-time, very high time tagging throughput over USB 3.0, and a timing resolution of 80 ps. Apart from design features and benchmark results of the instrument as such, we show application results from spectrally resolved (sFLIM) and high speed confocal fluorescence lifetime imaging (rapidFLIM). We put special focus on life science applications, paving the way to monitor sub-second dynamics in live cell imaging, including lifetime based Förster Resonant Energy Transfer (FLIM-FRET) imaging. We furthermore show how the inevitable pulse-pile-up occurring in the detector signals at high photon flux can be corrected for and how this data acquisition scheme excels in terms of photon collection efficiency in comparison to other approaches.

In this talk, I introduce high-throughput intelligence-powered image-activated cell sorting or an “imaging” version of fluorescence-activated cell sorting [Cell 175, 1 (2018)] that realizes real-time image-based intelligent cell sorting at an unprecedented rate of ~100 events per second. This technology integrates high-speed cell microscopy, focusing, and sorting methods on rapid hybrid software-hardware data-management infrastructure, enabling real-time automated operation for data acquisition, data processing, decision-making, and actuation. I present two unique applications in microbiology and hematology enabled by the technology and discuss how technology enables machine-based scientific discovery in diverse biological and medical sciences.

While fluorescence imaging flow cytometry is a promising method for high-throughput single-cell analysis, it has not been suitable for analysis of large populations of cells (e.g., blood samples) due to its low imaging sensitivity at a high cell throughput. Here we present fluorescence imaging flow cytometry with an ultrahigh imaging sensitivity, which is enabled by virtual motion freezing. In this method, we prepare a wide-field imaging system with a CMOS camera and scan images of flowing cells by a scanning device, such as a polygon scanner, equipped in the imaging system so that the motion of the cells is canceled in the imaging plane, thus significantly extending the exposure time of the camera without suffering from motion blur. Additionally, we scan a loosely focused excitation beam during the exposure time of the camera in the direction opposite to the cell flow using a beam scanner such as an acousto-optic deflector, which significantly reduces motion cancellation errors caused by the image distortion of the imaging system and hence allows further extension of the exposure time. Consequently, our method improves imaging sensitivity by a factor of ~1,000 compared with a conventional wide-field excitation method, enabling acquisition of microscopy-grade images of fast flowing cells. As a proof-of-concept, we obtained fluorescence images of nuclei of murine white blood cells stained by SYTO16 at a flow speed of 1 m/s (corresponding to a cell throughput of 10,000 cells/s assuming the 100-μm cell spacing) and determined the population of nuclear lobulation from the high-signal-to-noise-ratio images obtained.

Continuing development of image-based bioassay is mainly hampered by the lack of throughput to systematically screen a large cell/tissue population under extensive experimental conditions; and the overwhelming reliance on biochemical markers, which are not always effective, especially when there is poor prior knowledge of the markers. Here we demonstrate ultralarge-scale, high-resolution “on-the-fly” quantitative phase imaging (QPI) of single-cells and whole-tissue-slide on a spinning-disk assay platform at an imaging rate of at least 100-times faster than current assays – mitigating the imaging throughput limitation hindered by the fundamental space-bandwidth-product limit of classical optical imaging. The concept takes advantage of the high-speed spinning motion, which naturally provides imaging at an ultrafast rate (<10MHz) that can only be made possible with time-stretch imaging. To demonstrate the capability of the system, we imaged both label-free adherent cells and tissue slices, prepared on the functionalized digital versatile discs (DVDs), across a giga-pixel-FOV exceeding 10mm2 at a resolution of ~ 1μm. Both bright-field and QPI images are generated in real-time with this FOV at a spinning speed of <1,000 rpm. In contrast to the vast majority of current QPI modalities, our platform requires no interferometry and no computationally-intensive iterative method for phase retrieval, favouring continuous high-speed QPI operation in real-time. More importantly, this spinning imaging platform allows generation of a catalogue of label-free biophysical phenotypes of cells/tissues, e.g. cell size, dry mass density, morphology as well as light scattering properties, which could enable a new generation of large-scale in-depth label-free image-based bioassays.

Hyperspectral Stimulated Raman scattering (hsSRS) microscopy has recently emerged as a powerful non-destructive technique for label-free chemical imaging of biological samples. In most hsSRS imaging experiments, the SRS spectral range is limited by the total bandwidth of the excitation laser to ~300 cm^-1 and spectral resolution of ~25 cm^-1. Here we present a novel approach for broadband hsSRS microscopy based on parabolic fiber amplification to provide linearly chirped broadened Stokes pulses. This novel hsSRS instrument provides >600 cm^-1 spectral coverage and ~10 cm^-1 spectral resolution. We further demonstrated broadband hsSRS imaging of the entire Raman fingerprint region for resolving distribution of major biomolecules in fixed cells. Moreover, we applied broadband hsSRS in imaging amyloid plaques in human brain tissue with Alzheimer’s disease.

Nanosecond pulsed electric field (nsPEF) exposure to cells causes a myriad of bioeffects with great potential to translate into beneficial technology. However, a general lack of fundamental knowledge of how the field is interacting with the cell limits the advancement of predictive models and maximal exploitation. Despite 30 years of research, this same dearth of mechanistic understanding remains for longer pulse exposures. Fundamental to determining what is occurring as these strong electric fields are applied to cells is measuring the induced change in membrane potential on the time scale of the exposure. Such measurements are critical to validating commonly used electric circuit-based continuum models for electroporation, but have remained elusive due to limits in signal-to-noise and fluorescent reporters. In a previous publication, we described a high-speed fluorescent imaging modality that combined a streak camera and a high power laser source termed a high speed streak camera microscope (SCM) to resolve membrane charging during a single nsPEF. In this paper, we use the SCM to quantify changes in membrane potential in CHO-K1 cells exposed to unipolar and bipolar 600ns PEF within the context of the recently discovered “bipolar cancellation” phenomenon. Immediately after a unipolar pulse exposure, we see a prolonged “depolarization” of the cell that is roughly 50-100mV in amplitude. Such a prolonged depolarization is not seen in bipolar exposures nor is it predicted by membrane charging models. We postulate that this lasting membrane depolarization, seen only in unipolar pulse exposure, is either the cause of later uptake of impermeable ions or signifies the acute (during the pulse) breakdown of the plasma membrane (nanoporation). The lack of lasting depolarization in bipolar pulse exposures may be fundamental to “bipolar cancellation” and explain why uptake of ions is substantially reduced as compared to unipolar pulse exposures.

It is not difficult to be able to get a machined immersion grating in the market, and to obtain high spectrum-resolution (λ/Δλ > 200,000) in the middle infrared (3~5 micrometers) any longer. This wavelength region is a part of the molecule fingerprint region and is a very important wavelength region for the analysis of human / living activities. Compared with FT-IR in popular use, the spectrum by such classic device is an important means to a high-speed spectrum. However, although the dispersing potential of an immersion element is excellent, in offer of a device only, it does not mean in practical use for the researcher and the developer who want to use a high-performance spectroscope. So, in this paper, the high-resolution (λ/Δλ ~ 30,000 at 3-5 micrometer) spectrum unit which used Germanium (Ge) immersion grating is proposed. The optical plane size in this spectrum unit is about 200mm x 100mm, it is almost same as the iPad/mini4 (203.2mmx134.8mm) of the Apple Inc. It has fitted in less than 1/10 size compared with the reflected type spectroscope. This means dramatic improvement in the chance of an opportunity using a high-resolution spectroscope including ultrafast application. A large diversion is obtained in smaller beam size according to the refractive index of material in the immersion grating. In this paper, it introduces about the practical design of the high distribution immersion spectroscope. Especially, we introduce in detail about the manufacture of Ge immersion grating as the key device.

Endoscope size is a major design constraint that must be managed with the clinical demand for high-quality illumination and imaging. Existing commercial endoscopes most often use an arc lamp to produce bright, incoherent white light, requiring large-area fiber bundles to deliver sufficient illumination power to the sample. Moreover, the power instability of these light sources creates challenges for computer vision applications. We demonstrate an alternative illumination technique using red-green-blue laser light and a data-driven approach to combat the speckle noise that is a byproduct of coherent illumination. We frame the speckle artifact problem as an image-to-image translation task solved using conditional Generative Adversarial Networks (cGANs). To train the network, we acquire images illuminated with a coherent laser diode, with a laser diode source made partially- coherent using a laser speckle reducer, and with an incoherent LED light source as the target domain. We train networks using laser-illuminated endoscopic images of ex-vivo, porcine gastrointestinal tissues, augmented by images of laser-illuminated household and laboratory objects. The network is then benchmarked against state of-the-art optical and image processing speckle reduction methods, achieving an increased peak signal-to-noise ratio (PSNR) of 4.1 db, compared to 0.7 dB using optical speckle reduction, 0.6 dB using median filtering, and 0.5 dB using non-local means. This approach not only allows for endoscopes with smaller, more efficient light sources with extremely short triggering times, but it also enables imaging modalities that require both coherent and incoherent sources, such as combined widefield and speckle ow contrast imaging in a single image frame.

To overcome the drawbacks of the commonly used lookup table inverse models, we propose a novel custom OpenCL™- accelerated artificial neural network inverse model for spatial frequency domain imaging (https://bitbucket.org/xopto /rftroop). Utilizing a mid-range graphics processing unit, the proposed inverse model can estimate high-definition (1920 × 1080) maps of the absorption and reduced scattering coefficients and two scattering phase function related quantifiers at a rate of more than 50 frames per second. We show that the artificial neural network inverse model can be seamlessly extended to estimate multiple optical properties independently, thus providing a versatile framework that allows introduction of new quantifiers.

Image delivery through multimode fibers (MMFs) suffers from modal scrambling which results in a speckle pattern at the fiber output. In this work, we use Deep Neural Networks (DNNs) for recovery and/or classification of the input image from the intensity-only images of the speckle patterns at the distal end of the fiber. We train the DNNs using 16,000 images of handwritten digits of the MNIST database and we test the accuracy of classification and reconstruction on another 2,000 new digits. Very positive results and robustness were observed for up to 1 km long MMF showing 90% reconstruction fidelity. The classification accuracy of the system for different inputs (phase-only, amplitude-only, hologram intensity etc.) to the DNN classifier was also tested.

Since confocal laser microscopy (CLM) can acquire a high-contrast three-dimensional image, it is widely used in the field of bio-imaging. However, CLM is based on point measurement, it is necessary to mechanically scan the focal spot while keeping the conjugate relation between confocal pinhole and focal spot on the sample. Since such a mechanical scanning is vulnerable to environmental disturbance such as vibration, it needs a stable measurement environment like the active anti-vibration table. Also, for imaging of living samples, CLM cannot visualize clear image because of motion blur due to the difference in scanning time, thus it is limited to get only fixed sample images. We here propose a dual optical comb microscope combining dual optical comb spectroscopy (DCS) and two-dimensional spectral encoding (2D-SE). This combination enables one-to-one correspondence between optical frequency comb (OFC) modes and image pixels. Image information is superimposed on the mode-resolved OFC spectrum waveform by 2D-SE. Simultaneously, the confocality of all the pixels is given in parallel by a single confocal pinhole. The frame rate is limited by the data acquisition time of the interferogram, imaging rate over 1000 frames/s is possible. In this paper, we demonstrate the fast confocal phase imaging of a living paramecium.

Photonics based imaging is a widely utilised technique for the study of biological functions within pre-clinical studies. It is a sensitive and non-invasive technique that is able to detect distributed (biologically informative) visible and nearinfrared light sources providing information about biological function. Compressive Sensing (CS) is a method of signal processing that works on the basis that a signal or image can be compressed without important information being lost. This work describes the development of a CS based hyperspectral Bioluminescence imaging system that can be used to collect compressed fluence data from the external surface of an animal model, due to an internal source, providing lower acquisition times, higher spectral content and potentially better tomographic source localisation.

Hyperspectral imaging is a useful tool for characterization of human tissue. However, the vast amount of data created makes it challenging and tedious to manually select spatial regions of interest for further processing. In this study, a random forest-based method was evaluated on basis of its ability to segment human skin regions from the background. The method was compared to the performance of two alternative methods, spectral angle mapper (SAM) and a K-means clustering-based method. The methods were tested on hyperspectral images of ex vivo and in vivo human skin in the wavelength range 400-1000 nm. The random forest approach was found to be robust and perform well regardless of image type. The method is simple to train, and requires minimal parameter tuning for good skin segmentation results.

Spectroscopy is an important tool having already been applied in various research fields, but still limited in observation of dynamic scenes. In this paper, we propose a video rate spectroscopy via Fourier-spectral-multiplexing (FSM-VRS) which exploits both spectral and spatial sparsity. Under the computational imaging scheme, the hyperspectral datacube is first modulated by several broadband bases, and then mapped into different regions in the Fourier domain. The encoded image compressed both in spectral and spatial are finally collected by a monochrome detector. Correspondingly, the reconstruction is essentially a Fourier domain extraction and spectral dimensional back projection with low computational load. The encoding and decoding process of the FSM-VRS is simple thus can be implemented in a low cost manner. The temporary resolution of the FSM-VRS is only limited by the camera frame rate. We demonstrate the high performance of our method by quantitative evaluation on simulation data and build a prototype system experimentally for further validation.