We report on new findings to which the observation of vascular dynamics using full-field swept-source OCT with a large field of view has led. In addition to the pulsatile expansions already known from previous measurements, pulsatile contractions of the retina were observed. These may be explained by a longitudinal expansion of retinal arteries. The motion of vessels is actually much more irregular than previously assumed, which renders the determination of pulse wave velocities challenging. Whether these irregularities and their transmission to the tissue can be associated with a clinically relevant biomechanical parameter needs to be clarified by further investigations.

We recently demonstrated high-speed, high-resolution structural imaging of the human eye in vivo by spatiotemporal optical coherence tomography (STOC-T). STOC-T extends the Fourier-Domain Full-Field Optical Coherence Tomography (FD-FF-OCT) by the spatial phase modulation to improve the imaging depth and suppress coherent noises. Here, we show that the dataset produced by STOC-T can be processed differently to reveal blood flow in the superficial and deep retina layers. Our method, denoted as multiwavelength LDH (MLDH) enables noninvasive visualization and quantification of the blood flow deep into the human retina at high speeds and high transverse resolution in vivo.

We present methods based on Optical Coherence Tomography to quantify longitudinal changes in murine cortical vasculature. To demonstrate our methods, we tracked age-related changes in vascular structure and function of 3xTg Alzheimer’s disease (AD) and age-matched wild-type (WT) mice over the course of 7 months. In total, we measured 27 longitudinal parameters related to the morphology, topology, and function of the cortical vasculature across all scales: large pial vessels, penetrating vessels, and capillaries. Ten of these parameters showed different time-courses between AD and WT mice, with significant alterations preceding the onset of cognitive decline.

We demonstrate a novel decorrelation-based localized transverse flow measurement using a line-scan OCT system. We take advantage of the phase stability within each B-scan acquired with a line-scan OCT and digitally generate a low-resolution OCT volume. The ratio of the temporal autocorrelations of the original and low-resolution OCT signals only depends on the system resolution and the flow velocity component along the line-illuminating direction. The ratio is free from the diffusion motion of the flow particles, translation motion orthogonal to the line-illumination direction, and noise-induced distortion. The results from a glass capillary phantom experiment are highly correlated with the digital subaperture Doppler OCT ground truth.

In this study, we present the development of sensorless adaptive optics swept-source optical coherence tomographic angiography (sAO-SS-OCTA) imaging system for mice. GPU-based real-time OCTA image acquisition and processing software was applied to guide wavefront correction using a deformable mirror. High-resolution OCTA images with high capillary resolution and contrast have been successfully acquired. 45-degree field of view high-resolution montaged OCTA image was also acquired.

This Conference Presentation, Comparative study of optical coherence tomography angiography algorithms for rodent retinal imaging, was recorded at SPIE Photonics West held in San Francisco, California, United States.

We present Superfast Polarization-sensitive Off-axis Full-field Optical Coherence Microscopy (SPoOF OCM) as a novel all-optical technique for neurophysiology. Both the optical path length and birefringence induced by the millisecond-scale electrical activity of neurons are captured by SPoOF OCM at 4000 frames per second and with a field-of-view of 200×200 µm sq., 1 µm transverse resolution, 4.5 µm axial resolution, and 300 pm phase sensitivity. With an ability to capture responses spanning three orders of magnitude in both space and time, SPoOF OCM meets the exacting needs of a comprehensive neurophysiology tool and overcomes the existing limitations of traditional electrophysiology and fluorescence microscopy.

In this study, we propose to combine miniaturized optical coherence tomography (OCT) catheter with a residual neural network (ResNet)-based deep learning model for differentiation of normal from cancerous colorectal tissue in fresh ex vivo specimens. The OCT catheter has an outer diameter of 3.8 mm, a lateral resolution of ~10 um, and an axial resolution of 6 um. A customized ResNet-based neural network structure was trained on both benchtop and catheter images. An AUC of 0.97 was achieved to distinguish between normal and cancerous colorectal tissue when testing on the rest of catheter images.

We report results from in-vivo measurements of a human retina photoreceptors layer response to a flicker stimulus. We performed our experiments with the Spatio-Temporal Optical Coherence-Tomography (STOC-T) setup. We show that the phase analysis facilitates spatially resolved detection of the retina's response to different stimulus frequencies.

Despite the emerging market in fractional-pixel CO2 lasers to treat Genitourinary Syndrome of Menopause (GSM) in menopausal women, the effect of the laser treatment on vaginal tissue remains poorly understood. We developed an intravaginal OCT endoscope that can obtain structural and vascular information simultaneously during the vaginal laser procedure. Based on the monitoring and statistical analysis of the Vaginal Epithelial Thickness (VET) and Blood Vessel Density (BVD) along with the treatment for months, laser treatment shows a positive impact on vaginal health. This system can serve as a noninvasive biopsy tool in gynecological studies.

We present initial results of OCT images of human fallopian tubes obtained from miniature OCT catheters. Two OCT catheters were fabricated to image from the outside and inside of the fallopian tube. The OCT catheter used to image from outside has an outer diameter of 3.8 mm, a lateral resolution of ~10 um, and an axial resolution of 6 um. Special attention was paid to the fimbriated end. The smaller OCT catheter used to image inner mucosa layer has an outer diameter of 1.5 mm. 3D structures of the normal and malignant human fallopian tubes were revealed.

Endoscopic forward viewing optical coherence tomography is analogous to conventional endoscopic views and can be achieved by resonant scanning of an optical fiber. Multi-beam endoscopic resonant scanning is demonstrated by a set of optical fibers mounted to a piezoelectric bender actuator and depth-multiplexed using a long coherence length swept laser at 200 kHz sweep rate and 8 mm imaging range. A compact translation mechanism adjusting the distance between the imaging fibers and lenses enabled the precise tuning of optical magnification. Scalable fields of view between 1.3 mm and 2.8 mm were demonstrated.

Optical coherence tomography (OCT) has revolutionized diagnostics in ophthalmology. Montaging of multiple OCT volumes allows for wide field retinal volumes. However, OCT requires an operator to align the scanner and requires patient cooperation to fixate on multiple targets, one for each volume in the montage. We have developed a robot-mounted OCT scanner that automatically aligns with the subjects’ eye by compensating motion and gaze error at multiple entry angles, allowing acquisition of volumes from multiple regions of interest without chin or fore rest stabilization or a fixation target. We demonstrate our system by montaging a retinal volume acquired from a free-standing subject.

In this contribution, we report on in vivo retinal and choroid tissue imaging with Spatio-Temporal Optical Coherence Tomography (STOC-T) with a large field of view (9 x 4.6 mm2). We present en-face images of the retina's microstructure and choroid of the human eye with resolution enabling observation of single photoreceptors and choriocapillaris.

We developed a novel combined SD-OCT + TD-FF-OCT device that provides cell-resolution view of TD-FF-OCT without compromising SD-OCT performance. SD-OCT gives global view for eye exploration and FF-OCT shows cell-detail in the central region of the OCT scan. Eye imaging is fast enough to be part of the routine clinical exam (10 min/patient). Four patients with different eye pathologies were imaged. FF-OCT resolved: striae (stromal mechanical folds), guttata, loss of endothelial cells and stromal cuts following the surgery. Additionally, we could access the trabecular meshwork region of the eye and obtain the first images of meshwork fibers at micron resolution.

We present our progress on multi-modal imaging in the mouse retina using OCT and Two-Photon Excited Fluorescence (TPEF). In order to progress this modality towards a clinical setting, the power incident on the retina must be reduced. With a significantly dimmer TPEF signal, motion corrected registration and averaging becomes difficult. We have developed an approach to utilize multi-modal simultaneous acquisition to non-rigidly register both datasets solely using the OCT signal. Image quality is further enhanced by correcting wavefront aberrations introduced from the high NA configuration through a sensorless image-based hill-climbing algorithm.

We introduce a new method, temporal phase contrast (TPC) OCT, to measure sub-micron tissue motion in-vivo in the retina over an extended timeframe, i.e. over several seconds. The analysis is based on the calculation of the phase differences between an initial reference B-scan and each of the subsequent B-scans. In this study, retinal nerve fiber (RNF) tissue deformations induced by retinal vessels pulsating with the heartbeat were investigated in healthy volunteers. We show reproducible results of tissue expansion enabling to quantify the extent of the deformation and access the delay between expansion near the arteries and the veins.

Intraoperative OCT (iOCT) provides real-time imaging data that can be used to aid clinical decision-making and verify completion of surgical goals. However, video-rate 4D iOCT imaging of surgical dynamics is limited by the need to manually align the OCT field-of-view (FOV) to the region-of-interest, thus significantly impacting surgical workflow. Here, we demonstrate automated instrument-tracking at over 120 Hz. We present video-rate 4D imaging and tracking of 25G internal limiting membrane forceps at 16 volumes/second. The proposed method and improvements will facilitate the broad adoption of iOCT technology by providing real-time volumetric feedback on surgical dynamics and instrument-tissue interactions.

We present a method for determining the optical and thermal properties of layered materials, applicable to retinal laser therapy, using phase-resolved OCT. Transient heating of a tissue phantom is achieved by focusing a laser pulse onto a buried absorbing layer. Optical path length changes between the top of the phantom and the scattering absorbing layer induced by material expansion are extracted from the sequential B-scans. The absorption coefficient, heat conductivity and thermal expansion coefficient of the polymer are determined by matching the experimental data to a thermomechanical model of the tissue, yielding a temperature precision <2%, well below damage threshold.

We report on a swept-source OCT system based on a photonic-electronic integrated circuit. It enables a parallelization of data acquisition resulting in an effective A-scan rate of 4x100 kHz at a central wavelength of 840 nm. The monolithic co-integration of photonic elements forming the multiplexed interferometers and the system electronics on one chip allows a very compact OCT engine in a photonic package. Integrated in an ophthalmic system, the maximum sensitivity was estimated to be 91 dB with an optical power of 4x520 µW at the model eye. An eye phantom was imaged at 400 kHz showing its layered structure.

Point-scan swept-source optical coherence tomography systems (SS-OCTs) are subject to volumetric framerate limits governed by source sweep rate and scanner dynamics. For scenes with dynamic features on static backgrounds, adaptive scanning escapes these limits by visiting scan positions only as needed. We implemented adaptive scanning using a probabilistic approach that balanced re-imaging of known dynamic positions with exploration for undiscovered ones. We evaluated our approach in model systems that simulated ophthalmic surgery and multi-target tracking using a 200 kHz SS-OCT system. We demonstrated framerate speedups of 6.7x and 8.0x, respectively, performance that would have otherwise required a significantly faster source.

Intraoperative OCT can markedly enhance visualization in both posterior and anterior eye procedures. In these applications, imaging speed is paramount, as slower systems interfere with surgical workflow. We have previously introduced a circular-ranging (CR) OCT architecture optimized for high-speed intraoperative applications. Here, we demonstrate retinal imaging by CR-OCT for the first time. We achieved a 13.5 MHz A-line rate and performed high-quality wide-field and video-rate normal-field imaging in human subjects. The compressive properties of CR allow each of these imaging modes to operate with reduced data capture, easing acquisition and processing requirements that are critical to achieving continuous and low-latency imaging.

The flexible membranes used in MEMS tunable VCSELs are so small and light that thermally-induced vibrations can impact performance. We measure the thermal vibration spectrum of such a membrane showing peaks at the spatial vibration mode resonant frequencies of the membrane/plate. These vibrations result in a theoretical floor to the linewidth of the VCSEL. Frequency domain LiDAR and optical coherence tomography systems can get around this thermal linewidth limit with adequate clock measurement and processing. Essentially an OCT/LiDAR sweep with a concomitantly measured clock is a feed-forward linewidth reduction scheme. LiDAR ranging out to 10 meters has been demonstrated.

A high speed motion detection technique using SS-OCT system is demonstrated. Acquired OCT signal from high speed reflector result in producing artifacts like axial position shifts and broadening of the OCT signal in final processed images. A methodology using forward and backward wavelength sweeps of swept source laser to correct these artifacts is proposed. Analysis of phase changes of interferograms recorded with bi-directional laser sweeps at high sweep rates can be used to determine the true trajectory of the fast moving object. This technique also helps in monitoring velocity of the object exceeding the velocity range set by the acquisition speed of the OCT system.

Fetal membranes have important mechanical and antimicrobial roles in maintaining pregnancy. However, compared to other pregnancy tissues (e.g., uterus, cervix, placenta), they are understudied. Their low thickness (<800 µm) places them outside the resolution limits of most ultrasound and magnetic resonance scanners. As such, optical imaging methods like OCT have the potential to fill this technical gap. Here, an application of OCT imaging and machine learning for studying (ex vivo) the mechanical properties of the multilayered fetal membranes and correlating them with gestation and birth condition (i.e., labored vs. unlabored), and anatomy (i.e., near vs. far from cervix) is presented.

This Conference Presentation, Real-time histology evaluation by optical coherence tomography (OCT) holds promise to improve the diagnostic anatomic pathology gross evaluation process, was recorded at SPIE Photonics West held in San Francisco, California, United States.

Whereas Full Field OCT (FFOCT) relies on backscattering of light, Full Field Optical Transmission Tomography (FFOTT) relies on forward scattering using the Gouy’s phase shift modulation that is achieved close to the focus of a microscope objective. This new type of endogenous cell imaging technique that offers structural and metabolic contrasts is particularly well suited for imaging cell culture on glass slide or Petri dishes avoiding fringes that mask cells in FFOCT as well as biological structures such as biofilms. The sectioning ability is close to confocal microscopy but no contrast agent is required.

We present 3D intracellular motility imaging in MCD-diet induced non-alcoholic fatty liver disease (NAFLD) model by OCT-based dynamics imaging method, logarithmic intensity variance (LIV). LIV imaging visualizes the label-free intracellular activity. A 1-week and 2-week NAFLD model were investigated. In 1-week NAFLD, formation of large number of highly dynamic small particles at the beneath of the tissue surface were observed in LIV volume rendering image. In 2-week NAFLD model, a thin high LIV layer signal appeared in cross-sectional LIV image just beneath the tissue surface. The LIV projection and volume rendering images also reveal several discontinuous vessel-like structures.

Optogenetics is a powerful tool that allows tissue specific control with light activation. Here, we use Drosophila melanogaster as a tool to optimize and improve optogenetic cardiac pacing. We have obtained several D. melanogaster strains to test the performance of different opsins,light sensitive proteins, to determine which provide high-fidelity control of the heart with minimal power to activate. Using a lab-built optical coherence tomography (OCT) system integrated with an optogenetic set-up, flies were efficiently tested for performance and fidelity. Using a custom convolutional neural network, 2D+Time pacing images were segmented to quantify functional parameters such heart beat rate, change in lumen area, and heart wall velocity.

Organoids play an increasingly important role in in vitro models for studying organ development and disease mechanisms, and drug discovery. Recently, two groups independently developed human heart organoids from human pluripotent stem cells (hPSCs). In this study, we utilized a customized spectral-domain OCT (SD-OCT) to study heart organoids and demonstrated its capability to produce 3D images. Heart organoids formed cavities of various sizes, and complex interconnections were observed as early as on day 6. Heart organoids and the OCT system showed promising insights as an in vitro platform to investigate heart development and diseases mechanisms.

We present a novel approach to actively match the sampling of two linear array spectrometers to a precision of 0.001% of their total spectral range. We show that with a simple scaled subtraction of data from matched spectrometers, we are able to achieve more than two orders-of-magnitude excess noise suppression in balanced spectral domain visible light Optical Coherence Tomography. We demonstrate retinal imaging with visible light OCT at high-speed (70,000 axial scans per second) and low power (125 microwatts) at near the shot noise limit, even using a relatively inexpensive supercontinuum light source with a 30 MHz repetition rate.

Well-known limitations of optical coherence tomography (OCT) include deleterious speckle noise and relatively poor lateral resolution (typically >10 μm) due the tradeoff between lateral resolution and depth of focus. To address these limitations, we present 3D optical coherence refraction tomography (OCRT), which computationally combines 3D volumes from two rotational axes to form a 3D reconstruction with substantially reduced speckle noise and enhanced lateral resolution. Our approach features a parabolic mirror as the objective, which enables multi-view OCT volume acquisition over up to ±75° without moving the sample. We demonstrate 3D OCRT on a phantom sample and several biological samples, revealing new structures that are missed in conventional OCT.

Focal size and depth-of-focus (DOF) are dependent by the numerical aperture (N.A.) of the lens. Consequently, a high-resolution image inherently results in a short DOF. In order to extend the DOF of a high N.A. lens, a novel diffractive optical element is developed to generate needle-shaped beams. The DOF can be enhanced from 12μm (two Rayleigh lengths) to 120μm with a constant diameter of 1.5μm (the same as the focal size). When applied to a virtual biopsy of human skin, the needle-shaped beam can reveal the individual cells in the epidermal layer.

Access to clinical OCT systems is currently limited to well-resourced medical centers due to their mechanical footprint, complexity and cost. Smartphone computational power and optical system quality has increased exponentially in recent years, leading to its implementation in various imaging and sensing applications. Here, we demonstrate a line-field visible-light OCT system that utilizes the native camera of a commercial smartphone and a custom phone application to collect, process and visualize 2D OCT cross-sectional data in real-time. We believe smartOCT can lead to significant impact in low-resource areas by making OCT devices accessible to a broader population.

This study aims to explore Denoising Predictive Coding (DN-PC), a new compressed sensing (CS) algorithm for reconstruction of optical coherence tomography (OCT) image volumes. In preliminary work, the algorithm has yielded reconstructions of OCT images of various biological samples, with accuracy and computation time superior to other CS methods. Here we have assembled an image bank of representative OCT images of normal breast such as adipose and stroma, and pathology such as breast cancer. We apply DN-PC for compression at decreasing a-line sampling rates to evaluate a relation between reconstruction behavior and breast tissue structure.

Multimodal optical coherence tomography (OCT) can be implemented using double-clad fiber (DCF). A consequence of using DCF is the introduction of multipath artifacts which deteriorate the quality of OCT imaging. We demonstrate that a w-type DCF, characterized by a depressed cladding layer between the core and the multimode cladding, can eliminate OCT multipath artifacts. The modal contents of the fiber are determined from simulation and verified experimentally. A w-type fiber-based endoscope is used to generate co-registered OCT and autofluorescence imaging (AFI) with reduced artifacts. Results are compared with a DCF-based catheter.

One of the solutions for spectral-domain OCT (SD-OCT) system to improve its functionalities is to have multiple spectrometers to achieve high speed or sensitivity, but the multiple spectrometers require the performance of the spectrometers to be carefully matched. In this study, we introduce and demonstrate a numerical method to calibrate the performance of multiple spectrometers. The calibration was done by remapping the spectrum of the spectrometers and assessed by a merit function based on phase subtraction. This numerical method allowed us to match the performances of the two spectrometers by over 99%.

Conventional methods of spectroscopic Optical Coherence Tomography (OCT) determine depth-resolved spectra. Here, we present a spectroscopic method of assessing hemoglobin in OCT which, rather than determine a depth-resolved spectrum, determines a depth-resolved autocorrelation function. This complex-valued autocorrelation function is then fit with a model that incorporates the spectral absorption characteristics of different chromophores present in tissue. The proposed method does not use windowed Fourier transforms of the OCT data, and is well-suited for assessing chromophores in dynamic scattering environments such as blood vessels. The new autocorrelation spectroscopy method is compared against the conventional windowed Fourier transform method in the retina.

Degradation or migration of melanin within the retina is an important biomarker for ophthalmic diagnostics. Here we investigate spectral analysis of optical coherence tomography (OCT) data for measuring melanin in the retina. We demonstrated that a near infrared clinical OCT system was able to produce spectral signals for melanin in the retinal pigment epithelium and choroid. Our results showed lower melanin concentrations in the choroid for a human compared to a mouse. Because spectral analysis can be performed on standard OCT data, these methods may be readily integrated into clinical systems around the world without the need for new hardware.

Microscopic optical coherence tomography (OCT) provides three-dimensional, high-resolution imaging but lacks (sub-) cellular contrast. Dynamic-microscopic OCT (dmOCT) is an approach exploiting dynamic changes of the scattering behavior in metabolically active cells. However, the underlying cellular processes responsible for those intensity fluctuations and hence the dynamic signals are not finally identified yet. Here, we present the effects of different temperatures and metabolic reagents on dmOCT images of an in-vitro human skin model. Our data indicates a dependency of the dmOCT signals on metabolic activity rather than Brownian motion and suggests dependency on the metabolic state.

Measuring light scattering properties using OCT stands to enhance its ability to capture clinically-relevant microstructural details, but challenges remain in accurately relating these properties to underlying tissue architecture. In this work, we demonstrate the collection of depth-multiplexed data at diverse scattering angles with a glass annulus to facilitate correction of imaging system-based signal biases and identify multiply scattering areas in tissue. Based on our promising preliminary results from phantoms and healthy excised tissue samples, we hope our approach will enhance the accuracy of quantitative scattering parameter measurements, and help to realize their potential in offering detailed microstructural characterization of biological tissues.

We will present a deep convolutional neural network (DCNN) based estimators for optical coherence tomography (OCT). The DCNNs analyze local OCT speckle patterns and estimate the sample’s scatterer density and OCT resolutions. This estimator is intensity invariant, i.e., it does not use the net signal strength of OCT even to estimate the scatterer density. The DCNN is trained by a huge training dataset that was generated by a simple simulator of OCT imaging. This method is validated either by scattering phantom and in vitro tumor spheroid, and good accuracies of the estimation were shown.

The attenuation coefficient can be calculated from OCT data, but accurate determination requires compensating for the confocal function. We present detailed measurement series for extraction of the focal plane and the apparent Rayleigh length from the ratios of OCT images acquired with different focus depths and compare these results with alternative approaches. The optimal focus depth difference is determined for intralipid and titanium oxide phantoms with different scatterer concentrations and the attenuation coefficients corrected for the confocal function are calculated. We further demonstrate good reproducibility of the determined attenuation coefficient of layers with identical scatter concentrations in a multi-layered titanium oxide phantom.

Polarization Sensitive Optical Coherence Tomography (PS-OCT) measures the intensity and polarization state of backscattered light to provide information about tissue structure, retardation and depolarization. Developing molecular contrast agents for PS-OCT could also provide physiological, cellular, and molecular information. In this study, we utilize the depolarization and spectral signature of anisotropic gold nanobipyramids (GNBPs) and demonstrate how the optical properties of these nanostructures can be used as contrast agents for PS-OCT in living tissue.

A novel polarization state tracing algorithm has been proposed to visualize depth-resolved birefringent information by using the polarization sensitive optical coherence tomography (PSOCT) system. This algorithm is compatible to the widely adopted single input PSOCT system which uses only one circularly polarized incident light. We demonstrate the ability of this method to visualize depth-resolved myocardial architecture in both healthy and infarcted rodent hearts (ex vivo) and collagen structures responsible for skin tension lines at various anatomical locations on the face of a healthy human volunteer (in vivo).

Polarization sensitivity is a powerful extension of conventional OCT, providing additional contrast for birefringent structures. However, its widespread clinical adoption is complicated by increased system complexity and cost. Single-input polarization-sensitive OCT provides an alternative which bypasses some of the hardware requirements but is prone to artifacts. Here we present a method utilizing existing polarization mode dispersion within the catheter in combination with optimization algorithms to accurately estimate the sample retardance. This strategy has been shown to suppress artifacts and improve agreement with conventional two-state PS-OCT, facilitating the clinical adoption using optimization methods to overcome hardware limitations.

Common-path probes provide considerable advantages for fiber-based OCT due to intrinsic length and phase matching. However, the polarization state of the reference light is usually arbitrary and variable due to stress-induced birefringence in single-mode fibers, which complicates implementing polarization-sensitive OCT. Here, we present depth-resolved retardation measurements with a single-mode fiber-based common-path probe by utilizing the constrained polarization evolution and the mirror state phenomenon for reconstruction of the round-trip measurements in the case of arbitrary reference states. Thus, a compact and flexible polarization-sensitive OCT implementation is demonstrated.

A robust multi-volume three-dimensional (3D) registration algorithm is introduced to improve the contrast of optical coherence tomography (OCT) volumes. Our method involves registering multiple volumes to a selected reference volume to correct for the translational and rotational differences between each target and the reference volume and averaging the registered volumes. We tested our registration algorithm on the volumes obtained from three OCT systems with different field-of-views and resolutions. To demonstrate its accuracy, our developed method is evaluated using two different metrics, and its advantages over the other registration algorithms and its limitations are discussed.

Ophthalmic OCT image-quality is highly variable and directly impacts clinical diagnosis of disease. Computational methods such as frame-averaging, filtering, deep-learning approaches are generally constrained by either extended imaging times when acquiring repeated-frames, over-smoothing and loss of features, or the need for extensive training sets. Self-fusion is a robust OCT image-enhancement method that overcomes these aforementioned limitations by averaging serial OCT frames weighted by their respective similarity. Here, we demonstrated video-rate self-fusion using a convolutional neural network. Our experimental results show a near doubling of OCT contrast-to-noise ratio at a frame-rate of ~22 fps when integrated with custom OCT acquisition software.

Idiopathic pulmonary fibrosis (IPF) is a fatal form of fibrotic interstitial lung disease (ILD). Early diagnosis of IPF is essential, however, resolution limitations of HRCT prohibit identification and monitoring of early microanatomic alterations. Developing precise imaging biomarkers using quantitative imaging features and artificial intelligence has significant potential for early diagnosis of IPF and non IPF ILDs, as well as for monitoring disease progression and therapeutic response. We demonstrate the feasibility of a deep learning-based algorithm for accurate segmentation and classification of salient microscopic ILD imaging features on endobronchial optical coherence tomography (EB-OCT) imaging.

We present a novel approach of leveraging deep learning to reconstruct high-resolution OCT B-scans from reduced axial resolution data. In this work, the original OCT signal is used as the ground truth, and lower resolution was simulated by windowing the interference fringes. A super-resolution pixel-to-pixel generative adversarial network (GAN) was investigated for reconstructing high-resolution OCT data in the spatial domain and is compared against reconstructing in the spectral domain.

This study is to compare the optical coherence tomography (OCT) amplitude intrinsic optical signal (IOS) and phase IOS change of the human retina after light stimulation. A custom constructed OCT was employed for functional optoretinography imaging. A white LED was used as the retinal stimulator. OCT amplitude and phase IOS were computed by comparing the amplitude and phase before and after light stimulation. Both amplitude IOS and phase IOS were observed right after the stimulus onset, predominantly in the outer retina. The phase IOS is more sensitive to the layer boundaries.

We proposed a dual-GAN-based deep learning to enhance resolution and reduce noise of optical coherence tomography (OCT). The dual GAN was designed with a model that enhances axial resolution and a model that enhances lateral resolution and reduces noise. We demonstrated improvements on the swine coronary artery data used for training, and further validated the performance on other sample data acquired in other systems. Through this, not only the performance but also the feasibility of independent application to a specific system or sample was verified. The current approach will be highly helpful in overcoming existing limitations of OCT.

In its standard implementation, OCT suffers from the well-known “mirror artifact” around the zero-delay plane. Full-range OCT enables imaging on both sides of zero-delay by measuring the interference signal’s in-phase and quadrature components. These signals can be accessed passively using polarization demultiplexing or optical hybrids. However, such optical systems present imperfections, including chromatic and RF variations in phase and amplitude, which result in residual mirror artifacts. In this work, we propose a calibration methodology that relies solely on simple mirror measurements to correct the imperfections of the demultiplexing system and achieve high extinction of the mirror artifacts.