Novel AI-based HRV analysis (NAIHA) in healthcare automation and related applications.

BACKGROUND: Heart rate variability (HRV) analysis computed on R-R interval series of ECG records with heavy burden of ectopic beats or non-sinus rhythm can significantly distort HRV parameters and hence clinically ineligible for HRV analysis. Yet, existing algorithmic methods of HRV analysis do not check such eligibility and require manual identification of eligible window (portion of ECG record) to ensure reliability. OBJECTIVE: We aimed to propose a robust algorithm with a sliding window feature to automate the identification of an eligible window, if available, which compute HRV parameters within that window obviating manual input. METHODS: The proposed algorithm classifies each window as either eligible or ineligible. With a window classified eligible, we stop sliding through the record, otherwise we move to the next window and repeat the eligibility identification process, until either an eligible window is found, or all windows are exhausted. RESULTS: When evaluated on random subset of 100 records from MIMIC-III waveform database, the proposed algorithm excluded every ineligible record, and missed only 1.25% of eligible ones. The HRV parameters computed using proposed method closely approximated the standard HRV analysis with Pearson correlation coefficients (ideally one) and fractions of variance unexplained (ideally zero) ranging from 96.3% to 99.8% and 0.34% to 7.43%, respectively. CONCLUSIONS: When translated into practice, proposed algorithm will reduce clinicians'' burden without compromising the accuracy of HRV analysis, potentially leading to its wider adoption.

Statistical modeling of cell-to-cell variability in viral infection during passaging in suspension cell culture: Application in Monte-Carlo simulation.

Packaging during the passaging of viruses in cell cultures yields various phenotypes and is regulated by viral protein expression in infected cells. Although such a packaging mechanism has a profound effect in controlling the virus yield, little is known about the underlying statistical models followed by virus packaging and protein expression among cells infected with the virus. A predictive framework combining identification of the probability density function (PDF) based on log-likelihood and using the PDF for Monte-Carlo simulations is developed. The Birnbaum-Saunders distribution was found to be consistent with all three-virus packaging levels, including nucleocapsids/occlusion-derived virus (ODV), ODVs/polyhedra, and polyhedra/cell for both wild-type and genetically modified AcMNPV. Next, it was demonstrated that PDF fitting could be used to compare two viruses having distinctly different genetic configurations. Finally, the identified PDF can be incorporated in RNA synthesis parameters for baculovirus infection to predict the cell-to-cell variability in protein expression using Monte-Carlo simulations. The proposed tool can be used for the estimation of uncertainty in the kinetic parameter and prediction of cell-to-cell variability for other biological systems.

CHOROIDAL VASCULARITY INDEX AND CHOROIDAL THICKNESS IN EYES WITH RETICULAR PSEUDODRUSEN.

PURPOSE: To evaluate choroidal vascularity index (CVI), choroidal thickness, choroidal volume, and choroidal intensity in subjects with nonneovascular age-related macular degeneration (NNVAMD) with and without reticular pseudodrusen (RPD). METHODS: We included 60 eyes of 35 subjects with NNVAMD (including 30 eyes of 18 subjects with RPD) and 30 eyes of 17 age-matched healthy individuals from the ongoing Amish Eye study. The choroid was segmented from dense volume spectral domain optical coherence tomography scans and choroidal thickness (microns), choroidal intensity (log units), and choroidal volume (mm) from the entire macula (6 × 6 mm) were computed. A central horizontal B-scan was binarized and the luminal and stromal portions of the choroid were segmented. Choroidal vascularity index (%) was calculated as the ratio of luminal area to total choroid area. Choroidal parameters were compared between the groups by pairwise comparisons using the Student's t-test. RESULTS: The CVI was significantly lower in healthy eyes compared to those with RPD (53.43 ± 8.51 vs. 54.76 ± 4.83, P < 0.001). The CVI was also significantly lower in NNVAMD eyes without RPD compared to those with RPD (50.09 ± 7.51 vs. 54.76 ± 4.83, P = 0.006). There was no difference in CVI between healthy eyes and NNVAMD eyes without RPD (P = 0.84). Choroidal thickness and choroidal volume were significantly higher in NNVAMD without RPD (P < 0.05); and significantly lower in NNVAMD with RPD (P < 0.05) when compared with normal eyes. Choroidal intensity was significantly higher in NNVAMD with RPD when compared with normal eyes (P = 0.02) and NNVAMD eyes without RPD (P = 0.001). CONCLUSION: Multiple choroidal parameters reflecting the status of the choroidal vasculature and stroma seem to be altered in eyes with RPD compared with both normal eyes and NNVAMD eyes without RPD. These findings may provide insights into the pathophysiology of RPD.

AGE-RELATED CHANGES IN CHOROIDAL VASCULAR DENSITY OF HEALTHY SUBJECTS BASED ON IMAGE BINARIZATION OF SWEPT-SOURCE OPTICAL COHERENCE TOMOGRAPHY.

PURPOSE: To analyze the vascular density of the choroid in a healthy population using swept-source optical coherence tomography. METHODS: A cross-sectional, noninterventional study. INCLUSION CRITERIA: best-corrected visual acuity between 20/20 and 20/25, spherical equivalent between ±3 diopters, no systemic or ocular diseases, and ages ranging between 3 and 85 years. One hundred and thirty-six eyes from 136 subjects were analyzed, 86 eyes (63.2%) were from male and 50 eyes (36.8%) from female subjects. The eyes were divided into different age groups to analyze the possible age-related changes. Twelve-millimeter horizontal, fovea-centered B-scans were used. Choroidal stroma and vessel area analysis involved automated segmentation and binarization using validated algorithms. RESULTS: Mean age was 33.1 ± 24.5 years. Mean choroidal area was 0.5554 ± 0.1377 mm. Mean stromal area was 0.2524 ± 0.0762 mm, and mean vascular region area was 0.3029 ± 0.0893 mm. The percentage of choroidal vascularity (vascular area/total area) was 54.40 ± 8.35%. Choroid area, vascular region, and percentage of choroidal vascular density were statistically higher in the <18-year-old group versus the >18-year-old group (P < 0.001). The stromal region was not different (P = 0.46). In the same way, choroid area, vascular region, and percentage of choroidal vascular density between the 5 age groups were statistically different (P < 0.001), showing larger figures in the 0 to 10-year-old group, but not stromal region (P = 0.71). There were no gender-related differences. CONCLUSION: The luminal area and the percentage of vascular/total area decrease with increasing age, while the stromal area remains stable.

Semi-automated quantification of hard exudates in colour fundus photographs diagnosed with diabetic retinopathy.

BACKGROUND: Hard exudates (HEs) are the classical sign of diabetic retinopathy (DR) which is one of the leading causes of blindness, especially in developing countries. Accordingly, disease screening involves examining HEs qualitatively using fundus camera. However, for monitoring the treatment response, quantification of HEs becomes crucial and hence clinicians now seek to measure the area of HEs in the digital colour fundus (CF) photographs. Against this backdrop, we proposed an algorithm to quantify HEs using CF images and compare with previously reported technique using ImageJ. METHODS: CF photographs of 30 eyes (20 patients) with diabetic macular edema were obtained. A robust semi-automated algorithm was developed to quantify area covered by HEs. In particular, the proposed algorithm, a two pronged methodology, involved performing top-hat filtering, second order statistical filtering, and thresholding of the colour fundus images. Subsequently, two masked observers performed HEs measurements using previously reported ImageJ-based protocol and compared with those obtained through proposed method. Intra and inter-observer grading was performed for determining percentage area of HEs identified by the individual algorithm. RESULTS: Of the 30 subjects, 21 were males and 9 were females with a mean age of the 50.25 ± 7.80 years (range 33-66 years). The correlation between the two measurements of semi-automated and ImageJ were 0.99 and 0.99 respectively. Previously reported method detected only 0-30% of the HEs area in 9 images, 30-60% in 12 images and 60-90% in remaining images, and more than 90% in none. In contrast, proposed method, detected 60-90% of the HEs area in 13 images and 90-100% in remaining 17 images. CONCLUSION: Proposed method semi-automated algorithm achieved acceptable accuracy, qualitatively and quantitatively, on a heterogeneous dataset. Further, quantitative analysis performed based on intra- and inter-observer grading showed that proposed methodology detects HEs more accurately than previously reported ImageJ-based technique. In particular, we proposed algorithm detect faint HEs also as opposed to the earlier method.

Three-Dimensional Tracking of Intracellular Calcium and Redox State during Real-Time Control in a Hypoxic Gradient in Microglia Culture: Comparison of the Channel Blocker and Reoxygenation under Ischemic Shock.

Real-time three-dimensional (3-D) imaging is crucial for quantifying correlations among various molecules under acute ischemic stroke. Insights into such correlations may be decisive in selecting molecules capable of providing a protective effect within a shorter period. The major bottleneck is maintaining the cultures under severely hypoxic conditions while simultaneously 3-D imaging intracellular organelles with a microscope. Moreover, comparing the protective effect of drugs and reoxygenation remains challenging. To address this, we propose a novel workflow for the induction of gas-environment-based hypoxia in the HMC-3 cells along with 3-D imaging using laser-scanning-confocal microscopy. The imaging framework is complemented with a pipeline for quantifying time-lapse videos and cell-state classification. First, we show an imaging-based assessment of the in vitro model for hypoxia using a steep gradient in O2 with time. Second, we demonstrate the correlation between mitochondrial superoxide production and cytosolic calcium under acute hypoxia. We then test the efficacy of an L-type calcium channel blocker, compare the results with reoxygenation, and show that the blocker alleviates hypoxic conditions in terms of cytosolic calcium and viability within an acute window of one hour. Furthermore, we show that the drug reduces the expression of oxidative stress markers (HIF1A and OXR1) within the same time window. In the future, this model can also be used to investigate drug toxicity and efficacy under ischemic conditions.

An ML prediction model based on clinical parameters and automated CT scan features for COVID-19 patients.

Outcome prediction for individual patient groups is of paramount importance in terms of selection of appropriate therapeutic options, risk communication to patients and families, and allocating resource through optimum triage. This has become even more necessary in the context of the current COVID-19 pandemic. Widening the spectrum of predictor variables by including radiological parameters alongside the usually utilized demographic, clinical and biochemical ones can facilitate building a comprehensive prediction model. Automation has the potential to build such models with applications to time-critical environments so that a clinician will be able to utilize the model outcomes in real-time decision making at bedside. We show that amalgamation of computed tomogram (CT) data with clinical parameters (CP) in generating a Machine Learning model from 302 COVID-19 patients presenting to an acute care hospital in India could prognosticate the need for invasive mechanical ventilation. Models developed from CP alone, CP and radiologist derived CT severity score and CP with automated lesion-to-lung ratio had AUC of 0.87 (95% CI 0.85-0.88), 0.89 (95% CI 0.87-0.91), and 0.91 (95% CI 0.89-0.93), respectively. We show that an operating point on the ROC can be chosen to aid clinicians in risk characterization according to the resource availability and ethical considerations. This approach can be deployed in more general settings, with appropriate calibrations, to predict outcomes of severe COVID-19 patients effectively.

Automated Quantification of Inflamed Lung Regions in Chest CT by UNet++ and SegCaps: A Comparative Analysis in COVID-19 Cases.

During the current COVID-19 pandemic, a high volume of lung imaging has been generated in the aid of the treating clinician. Importantly, lung inflammation severity, associated with the disease outcome, needs to be precisely quantified. Producing consistent and accurate reporting in high-demand scenarios can be a challenge that can compromise patient care with significant inter- or intra-observer variability in quantifying lung inflammation in a chest CT scan. In this backdrop, automated segmentation has recently been attempted using UNet++, a convolutional neural network (CNN), and results comparable to manual methods have been reported. In this paper, we hypothesize that the desired task can be performed with comparable efficiency using capsule networks with fewer parameters that make use of an advanced vector representation of information and dynamic routing. In this paper, we validate this hypothesis using SegCaps, a capsule network, by direct comparison, individual comparison with CT severity score, and comparing the relative effect on a ML(machine learning)-based prognosis model developed elsewhere. We further provide a scenario, where a combination of UNet++ and SegCaps achieves improved performance compared to individual models.

Mapping of structural arrangement of cells and collective calcium transients: an integrated framework combining live cell imaging using confocal microscopy and UMAP-assisted HDBSCAN-based approach.

Live cell calcium (Ca2+) imaging is one of the important tools to record cellular activity during in vitro and in vivo preclinical studies. Specially, high-resolution microscopy can provide valuable dynamic information at the single cell level. One of the major challenges in the implementation of such imaging schemes is to extract quantitative information in the presence of significant heterogeneity in Ca2+ responses attained due to variation in structural arrangement and drug distribution. To fill this gap, we propose time-lapse imaging using spinning disk confocal microscopy and machine learning-enabled framework for automated grouping of Ca2+ spiking patterns. Time series analysis is performed to correlate the drug induced cellular responses to self-assembly pattern present in multicellular systems. The framework is designed to reduce the large-scale dynamic responses using uniform manifold approximation and projection (UMAP). In particular, we propose the suitability of hierarchical DBSCAN (HDBSCAN) in view of reduced number of hyperparameters. We find UMAP-assisted HDBSCAN outperforms existing approaches in terms of clustering accuracy in segregation of Ca2+ spiking patterns. One of the novelties includes the application of non-linear dimension reduction in segregation of the Ca2+ transients with statistical similarity. The proposed pipeline for automation was also proved to be a reproducible and fast method with minimal user input. The algorithm was used to quantify the effect of cellular arrangement and stimulus level on collective Ca2+ responses induced by GPCR targeting drug. The analysis revealed a significant increase in subpopulation containing sustained oscillation corresponding to higher packing density. In contrast to traditional measurement of rise time and decay ratio from Ca2+ transients, the proposed pipeline was used to classify the complex patterns with longer duration and cluster-wise model fitting. The two-step process has a potential implication in deciphering biophysical mechanisms underlying the Ca2+ oscillations in context of structural arrangement between cells.

En-face choroidal vascularity map of the macula in healthy eyes.

PURPOSE: To report the en-face choroidal vascularity index in healthy eyes. METHODS: Thirty eyes of 30 healthy individuals were studied. Multiple high-density cross-sectional swept source optical coherence tomography scans were obtained to create a volume scan. The choroid was segmented for the whole volume scan and choroidal inner boundaries were flattened. Subsequently, multiple en-face scans separated by 25 µm were obtained and binarized. Choroidal vascularity index was calculated at level of choriocapillaris, medium, and large choroidal vessels. RESULTS: The mean age of the study cohort was 35.6 ± 8.8 years. The overall mean en-face choroidal vascularity index was 54.25 ± 0.55%. There was a statistically significant difference of choroidal vascularity index in choriocapillaris (53.16 ± 0.43%), medium choroidal vessel (51.38 ± 0.27%), and large choroidal vessel (55.69 ± 0.87%) (p < 0.01). Choroidal vascularity index analysis in three subgroups based on subfoveal choroidal thickness (low: <300 µm, medium: 300-400 µm, high: >400 µm) showed a statistically significant difference (p = 0.001). Choroidal vascularity index showed a significant correlation with subfoveal choroidal thickness (r = 0.441; p = 0.015), whereas there was no significant correlation of age (p = 0.21), refraction (p = 0.20), and gender (p = 0.67) with en-face choroidal vascularity index. CONCLUSION: En-face choroidal vascularity index shows a significant variation at the level of choriocapillaris, medium choroidal vessel, and large choroidal vessel in normal eyes. Choroidal vascularity index reaches a nadir at the level of medium choroidal vessel and reaches the maximum value at large choroidal vessel near choroidoscleral interface. En-face choroidal vascularity index shows a significant physiological variation and appears to increase with increase in subfoveal choroidal thickness.

Automated iris volume analysis and trabecular meshwork length using anterior segment optical coherence tomography - Application in pseudoexfoliation and pseudoexfoliation glaucoma.

Purpose: The aim of this study was to evaluate differences in the iris and angle parameters in psuedoexfoliation syndrome (PXF) and pseudoexfoliation glaucoma (PXG) using anterior segment optical coherence tomography (ASOCT). Methods: Patients with PXF or PXG were compared using ASOCT with primary open-angle glaucoma POAG eyes as controls in this noninterventional comparative study conducted at a tertiary eye care center in East India. All angle parameters, TM length, and iris thickness were analyzed from the enhanced depth imaging (EDI) single scans obtained. Quadrant scans were used for the calculation of iris volume using a custom-built in-house software. In particular, the software performs multiple operations including edge detection, connected components, and thresholding to localize and segment the iris. Differences in the iris volume/thickness and TM length in PXF and PXG with POAG were analyzed. Results: A total of 225 eyes were included, which included 75 PXG and 98 PXF cases and 52 POAG with a mean age of 67 ± 9.7 years at presentation. The algorithm repeatability and reproducibility was also established with correlation coefficients more than 99% which was substantiated with Bland-Altman plots. The iris volume (calculated in 197 images of 225 eyes) did not differ significantly in PXF and PXG eyes, although both had significantly greater volume compared to POAG eyes. The iris volume or other angle parameters including TM length did not correlate with clinical variables such as IOP, age, or visual field indices. Conclusion: Iris parameters or TM length do not explain pathogenesis of glaucoma in pseudoexfoliation.

RESILIENT: A robust statistical method for estimating multiple VOC sources from limited field measurements.

Air pollution due to haphazard industrialization has become a major concern in developing countries. Yet, enforcement of related norms remains problematic because violators cannot easily be pinpointed among closely situated industrial units. Accordingly, it has become imperative to equip regulatory authorities with an economical yet accurate tool that quickly locates emission sources and estimates emission rates. Against this backdrop, we propose RESILIENT, a method for Robust Estimation of Source Information from LImited field measuremENTs, which exhibits significant statistical robustness and accuracy even when the data are collected using a low-cost error-prone sensor. In our field experiment, where ground truth was unavailable, the sources estimated to be inactive based on the complete set of measurements were found inactive (up to three decimal places of accuracy) at least 72% of the time even when estimated using just 54% of random measurements. In that setting, rate estimates of active sources were also found to be statistically robust. For direct validation of RESILIENT, we considered a separate public dataset involving 10 tracer experiments, and obtained a significant correlation coefficient of 0.89 between estimated and recorded emission rates, and that of 0.99 between predicted and measured concentration levels at sensor locations.

Cell-to-Cell Variability in Protein Expression during Viral Infection: Monte-Carlo Simulation and Validation based on Confocal Imaging.

One of the major challenges is to identify the statistical model underlying the heterogeneity in viral protein expression in single cells. In this endeavor, we propose a computational tool to address the cell-to-cell variability in protein expression by random variate generation following probability distributions. Here, we show that statistical modeling using the probability density function of various distribution offers considerable potential for providing stochastic inputs to Monte Carlo simulation. Specifically, we present the ranking between three distribution families including gamma, normal and Weibull distribution using a comparison of cumulative frequency obtained from experiment and simulation. The major contribution of the proposed simulation method is to identify the underlying statistical model in kinetic parameters that capture the variability in protein expression in single cells obtained through imaging using confocal microscopy.

Accurate Cross-Section Estimation of Blood Vessels in Choroidal Haller's Layer: An Iterative Method based on 3D Tensor Voting.

Various eye diseases, including polypoidal choroidal vasculopathy (PCV) and age-related macular degeneration (AMD), affect choroidal vasculature early, but possibly minutely. However, due to the complex networked structure of the vasculature, it becomes hard to visualize, analyze and detect such changes in 2D OCT B-scan images. In contrast, algorithmic evaluation of cross-section facilitates clinicians in tracing minute variations in the vessel network, and quantifying those correlated with pathologies, potentially leading to early diagnosis. In this context, we proposed a novel method of estimating vessel cross-sections in choroidal Haller's layer. Accuracy of our method was evaluated on synthetic as well as clinical data by trained optometrists, and earned a confidence score of 90%, marking about 60% improvement over estimates based on a well-known tree-based method.

Automatic Identification of Mixed Retinal Cells in Time-Lapse Fluorescent Microscopy Images using High-Dimensional DBSCAN.

Despite providing high spatial resolution, functional imaging remains largely unsuitable for high-throughput experiments because current practices require cells to be manually identified in a time-consuming procedure. Against this backdrop, we seek to integrate such high-resolution technique in high-throughput workflow by automating the process of cell identification. As a step forward, we attempt to identify mixed retinal cells in time-lapse fluorescent microscopy images. Unfortunately, usual 2D image segmentation as well as other existing methods do not adequately distinguish between time courses of different spatial locations. Here, the task gets further complicated due to the inherent heterogeneity of cell morphology. To overcome such challenge, we propose to use a high-dimensional (HiD) version of DBSCAN (density based spatial clustering of applications with noise) algorithm, where difference in such time courses are appropriately accounted. Significantly, outcome of the proposed method matches manually identified cells with over 80% accuracy, marking more than 50% improvement compared to a reference 2D method.

Classification and Quantification of Retinal Cysts in OCT B-Scans: Efficacy of Machine Learning Methods.

The automatic segmentation of fluid spaces in optical coherence tomography (OCT) imaging facilitates clinically relevant quantification and monitoring of eye disorders over time. Eyes with florid disease are particularly challenging to segment, as the anatomy is often highly distorted from normal. In this context, we propose an end-to-end machine learning method consisting of near perfect detection of retinal fluid using random forest classifier and an efficient DeepLab algorithm for quantification and labeling of the target fluid compartments. In particular, we achieve an average Dice score of 86.23% with reference to manual delineations made by a trained expert.

Diagnostic Quality Assessment of Ocular Fundus Photographs: Efficacy of Structure-Preserving ScatNet Features.

Various ophthalmic procedures critically depend on high-quality images. For instance, efficiency of teleophthalmology, a framework to bring advanced eye care to remote regions, is determined by the capability of assessing diagnostic quality of ocular fundus photographs (FPs), and rejecting poor-quality ones at the source. In this context, we study algorithmic methods of classifying high- and low-quality FPs. Crucially, diagnostic quality (DQ) - determined by clinically, but not necessarily perceptually, significant structures - is not synonymous with perceptual appeal. Yet, traditional methods handpick features individually (or in small subsets) to meet certain ad hoc perceptual requirements. In contrast, we investigate the efficacy of a comprehensive set of structure-preserving features, systematically generated by a deep scattering network (ScatNet). Specifically, we consider three advanced machine learning classifiers, train each using ScatNet as well as traditional features separately, and demonstrate that the former ensure significantly superior performance for each classifier under multiple criteria including classification accuracy.

A Robust Model System for Retinal Hypoxia: Live Imaging of Calcium Dynamics and Gene Expression Studies in Primary Human Mixed Retinal Culture.

The detailed mechanisms underlying oxidative stress that leads to neuroinflammation and neurodegeneration in retinal vascular conditions, including diabetic retinopathy, retinopathy of prematurity etc., remain largely unexplored mainly due to a lack of suitable disease models that can simulate the inherent neuron-glia interactions in human retina. Specifically, establishment of a mixed retinal culture (MRC) containing both neuron and glial cell types remains a challenge due to different conditions required for their optimal growth and differentiation. Here, we establish a novel primary MRC model system containing neurons, astrocytes, Müller glia, and microglia from human donor retina that can be used to study the neuromodulatory effects of glial cells under the stress. The cell characterization based on immunostaining with individual cell type-specific markers and their presence in close vicinity to each other further underscores their utility for studying their cross talk. To the best of our knowledge, this is the first instance of an in vitro model obtained from human donor retina containing four major cell types. Next, we induce hypoxic stress to MRC to investigate if hypoxia activated neuroglia modulates altered gene expression for inflammatory, apoptotic, and angiogenic markers and Ca2+ transients by live cell imaging. Further, we performed k-means clustering of the Ca2+ responses to identify the modification of clustering pattern in stressed condition. Finally, we provide the evidence that the altered Ca2+ transient correlates to differential expression of genes shown to be involved in neuroinflammation, angiogenesis, and neurodegeneration under the hypoxic conditions as seen earlier in human cell lines and animal models of diabetic retinopathy. The major features of the hypoxic conditions in the proposed human MRC model included: increase in microglia activity, chemokine and cytokine expression, and percentage of cells having higher amplitude and frequency of Ca2+ transients. Thus, the proposed experimental system can potentially serve as an ideal in vitro model for studying the neuroinflammatory and neurodegenerative changes in the retina and identifying newer drug targets.

Dictionary-based monitoring of premature ventricular contractions: An ultra-low-cost point-of-care service.

While cardiovascular diseases (CVDs) are prevalent across economic strata, the economically disadvantaged population is disproportionately affected due to the high cost of traditional CVD management, involving consultations, testing and monitoring at medical facilities. Accordingly, developing an ultra-low-cost alternative, affordable even to groups at the bottom of the economic pyramid, has emerged as a societal imperative. Against this backdrop, we propose an inexpensive yet accurate home-based electrocardiogram (ECG) monitoring service. Specifically, we seek to provide point-of-care monitoring of premature ventricular contractions (PVCs), high frequency of which could indicate the onset of potentially fatal arrhythmia. Note that the first-generation telecardiology system acquires the ECG, transmits it to a professional diagnostic center without processing, and nearly achieves the diagnostic accuracy of a bedside setup. In the process, such a system incurs high bandwidth cost and requires the physicians to process the entire record for diagnosis. To reduce cost, current telecardiology systems compress data before transmitting. However, the burden on physicians remains undiminished. In this context, we develop a dictionary-based algorithm that reduces not only the overall bandwidth requirement, but also the physicians workload by localizing anomalous beats. Specifically, we detect anomalous beats with high sensitivity and only those beats are then transmitted. In fact, we further compress those beats using class-specific dictionaries subject to suitable reconstruction/diagnostic fidelity. Finally, using Monte Carlo cross validation on MIT/BIH arrhythmia database, we evaluate the performance of the proposed system. In particular, with a sensitivity target of at most one undetected PVC in one hundred beats, and a percentage root mean squared difference less than 9% (a clinically acceptable level of fidelity), we achieved about 99.15% reduction in bandwidth cost, equivalent to 118-fold savings over first-generation telecardiology. In the process, the professional workload is reduced by at least 85.9% for noncritical cases. Our algorithm also outperforms known algorithms under certain measures in the telecardiological context.

Automated quantification of Haller's layer in choroid using swept-source optical coherence tomography.

OBJECTIVE: To develop an algorithm for automated quantification of Haller's layer in choroid using swept-source optical coherence tomography (OCT). BACKGROUND: So far, to understand the association of various diseases with structural changes of choroid, only gross indicators such as thickness, volume and vascularity index have been examined. However, certain diseases affect specific sublayers of the choroid. Accordingly, a need for targeted quantitation arises. In particular, there is significant interest in understanding Haller's layer, a choroidal sublayer comprising relatively large blood vessels. Unfortunately, its intricate vasculature makes, manual quantitation difficult, tedious, and error-prone. To surmount this difficulty, it is imperative to develop an algorithmic method. METHODOLOGY: The primary contribution of this work consists in developing an approach for detecting the boundary between Haller's and Sattler's layers, the latter comprising medium-sized vessels. The proposed algorithm estimates vessel cross-sections using exponentiation-based binarization, and labels a vessel large if its cross-section exceeds certain statistically determined threshold. Finally, the desired boundary is obtained as a smooth curve joining the innermost points of such large vessels. On 50 OCT B-scans (of 50 healthy eyes), our algorithm was validated both qualitatively and quantitatively, by comparing with intra-observer variability. Extensive statistical analysis was performed using metrics including Dice coefficient (DC), correlation coefficient (CC) and absolute difference (AD). RESULTS: The proposed algorithm achieved a mean DC of 89.48% (SD:5.03%) in close agreement with the intra-observer repeatability of 89.12% (SD:5.68%). Corresponding mean AD and mean CC were of 17.54 µm (SD:16.45µm) and 98.10% (SD:1.60%) which too approximate the respective intra-observer repeatability values 19.19 µm (SD:17.69 µm) and 98.58% (SD:1.12%). CONCLUSION: High correlation between algorithmic and manual delineations indicates suitability of our algorithm for clinically analyzing choroid in greater finer details, especially, in diseased eyes.