grasviq: an image analysis framework for automatically quantifying vein number and morphology in grass leaves.

Beyond facilitating transport and providing mechanical support to the leaf, veins have important roles in the performance and productivity of plants and the ecosystem. In recent decades, computational image analysis has accelerated the extraction and quantification of vein traits, benefiting fields of research from agriculture to climatology. However, most of the existing leaf vein image analysis programs have been developed for the reticulate venation found in dicots. Despite the agroeconomic importance of cereal grass crops, like Oryza sativa (rice) and Zea mays (maize), a dedicated image analysis program for the parallel venation found in monocots has yet to be developed. To address the need for an image-based vein phenotyping tool for model and agronomic grass species, we developed the grass vein image quantification (grasviq) framework. Designed specifically for parallel venation, this framework automatically segments and quantifies vein patterns from images of cleared leaf pieces using classical computer vision techniques. Using image data sets from maize inbred lines and auxin biosynthesis and transport mutants in maize, we demonstrate the utility of grasviq for quantifying important vein traits, including vein density, vein width and interveinal distance. Furthermore, we show that the framework can resolve quantitative differences and identify vein patterning defects, which is advantageous for genetic experiments and mutant screens. We report that grasviq can perform high-throughput vein quantification, with precision on a par with that of manual quantification. Therefore, we envision that grasviq will be adopted for vein phenomics in maize and other grass species.

Time-dependent in situ structural and cellular aberrations in rabbit cornea in vivo after mustard gas exposure.

An array of corneal pathologies collectively called mustard gas keratopathy (MGK) resulting from ocular exposure to sulfur mustard (SM) gas are the most prevalent chemical warfare injury. MGK involves chronic ocular discomfort that results in vision impairment. The etiology of MGK remains unclear and poorly understood primarily due to a lack of scientific data regarding structural and cellular changes in different layers of the cornea altered by mustard vapor exposure in vivo. The goals of this study were to (a) characterize time-dependent changes in different layers of corneal epithelium, stroma, and endothelium in live animals in situ by employing state-of-the-art multimodal clinical ophthalmic imaging techniques and (b) determine if SM-induced acute changes in corneal cells could be rescued by a topical eye drop (TED) treatment using in an established rabbit in vivo model. Forty-five New Zealand White Rabbit eyes were divided into four groups (Naïve, TED, SM, and SM + TED). Only one eye was exposed to SM (200 mg-min/m3 for 8 min), and each group had three time points with six eyes each (Table-1). TED was topically applied twice a day for seven days. Clinical eye examinations and imaging were performed in live rabbits with stereo, Slit-lamp, HRT-RCM3, and Spectralis microscopy system. Fantes grading, fluorescein staining, Schirmer's tests, and applanation tonometry were conducted to measure corneal haze, ocular surface aberrations, tears, and intraocular pressure respectively. H&E and PSR staining were used for histopathological cellular changes in the cornea. In vivo confocal and OCT imaging revealed significant changes in structural and morphological appearance of corneal epithelium, stroma, and endothelium in vivo in SM-exposed rabbit corneas in a time-dependent manner compared to naïve cornea. Also, SM-exposed eyes showed loss of corneal transparency characterized by increased stromal thickness and light-scattering myofibroblasts or activated keratocytes, representing haze formation in the cornea. Neither naive nor TED-alone treated eyes showed any structural, cellular, and functional abnormalities. Topical TED treatment significantly reduced SM-induced abnormalities in primary corneal layers. We conclude that structural and cellular changes in primary corneal layers are early pathological events contributing to MGK in vivo, and efficient targeting of them with suitable agents has the potential to mitigate SM ocular injury.

Impact of Limb Phenotype on Tongue Denervation Atrophy, Dysphagia Penetrance, and Survival Time in a Mouse Model of ALS.

Current treatments for dysphagia in ALS do not target the underlying tongue weakness and denervation atrophy that is prevalent in spinal and bulbar ALS cases. To address this clinical gap, we studied the low copy number SOD1-G93A (LCN-SOD1) mouse model of ALS to quantify the impact of limb phenotype on tongue denervation atrophy, dysphagia penetrance, and survival time in preparation for future treatment-based studies. Two male LCN-SOD1 breeders and 125 offspring were followed for limb phenotype inheritance, of which 52 (30 LCN-SOD1 and 22 wild-type/WT, both sexes) underwent characterization of dysphagia penetrance (via videofluoroscopic swallow study; VFSS) and survival time at disease end-stage (15-20% body weight loss). From these, 16 mice (8/genotype) underwent postmortem histological analysis of the genioglossus for evidence of denervation atrophy. Results revealed that both breeders displayed a mixed (hindlimb and forelimb) ALS phenotype and sired equal proportions of hindlimb vs. mixed phenotype offspring. Dysphagia penetrance was complete for mixed (100%) versus incomplete for hindlimb (64%) phenotype mice; yet survival times were similar. Regardless of limb phenotype, LCN-SOD1 mice had significantly smaller genioglossus myofibers and more centralized myonuclei compared to WT mice (p < 0.05). These biomarkers of denervation atrophy were significantly correlated with VFSS metrics (lick and swallow rates, p < 0.05) but not survival time. In conclusion, both LCN-SOD1 phenotypes had significant tongue denervation atrophy, even hindlimb phenotype mice without dysphagia. This finding recapitulates human ALS, providing robust rationale for using this preclinical model to explore targeted treatments for tongue denervation atrophy and ensuing dysphagia.

Advancing Laryngeal Adductor Reflex Testing Beyond Sensory Threshold Detection.

Flexible endoscopic evaluation of swallowing with sensory testing (FEESST) is a promising clinical tool to assess airway integrity via the laryngeal adductor reflex (LAR). The current clinical protocol relies on sensory threshold detection, as relatively little is known about the motor response of this sensorimotor airway protective reflex. Here, we focused on characterizing normative LAR motion dynamics in 20 healthy young participants using our prototype high-pressure syringe-based air pulse device and analytic software (VFtrack™) that tracks vocal fold (VF) motion in endoscopic videos. Following device bench testing for air pulse stimulus characterization, we evoked and objectively quantified LAR motion dynamics in response to two suprathreshold air pulse stimuli (40 versus 60 mm Hg), delivered to the arytenoid mucosa through a bronchoscope working channel. The higher air pressures generated by our device permitted an approximate 1 cm endoscope working distance for continual visualization of the bilateral VFs throughout the LAR. Post hoc video analysis identified two main findings: (1) there are variant and invariant subcomponents of the LAR motor response, and (2) only a fraction of suprathreshold stimuli evoked complete glottic closure during the LAR. While the clinical relevance of these findings remains to be determined, we have nonetheless demonstrated untapped potential in the current FEESST protocol. Our ongoing efforts may reveal LAR biomarkers to quantify the severity of laryngeal pathology and change over time with natural disease progression, spontaneous recovery, or in response to intervention. The ultimate goal is to facilitate predictive modeling of patients at high risk for dysphagia-related aspiration pneumonia.

DRPnet: automated particle picking in cryo-electron micrographs using deep regression.

BACKGROUND: Identification and selection of protein particles in cryo-electron micrographs is an important step in single particle analysis. In this study, we developed a deep learning-based particle picking network to automatically detect particle centers from cryoEM micrographs. This is a challenging task due to the nature of cryoEM data, having low signal-to-noise ratios with variable particle sizes, shapes, distributions, grayscale variations as well as other undesirable artifacts. RESULTS: We propose a double convolutional neural network (CNN) cascade for automated detection of particles in cryo-electron micrographs. This approach, entitled Deep Regression Picker Network or "DRPnet", is simple but very effective in recognizing different particle sizes, shapes, distributions and grayscale patterns corresponding to 2D views of 3D particles. Particles are detected by the first network, a fully convolutional regression network (FCRN), which maps the particle image to a continuous distance map that acts like a probability density function of particle centers. Particles identified by FCRN are further refined to reduce false particle detections by the second classification CNN. DRPnet's first CNN pretrained with only a single cryoEM dataset can be used to detect particles from different datasets without retraining. Compared to RELION template-based autopicking, DRPnet results in better particle picking performance with drastically reduced user interactions and processing time. DRPnet also outperforms the state-of-the-art particle picking networks in terms of the supervised detection evaluation metrics recall, precision, and F-measure. To further highlight quality of the picked particle sets, we compute and present additional performance metrics assessing the resulting 3D reconstructions such as number of 2D class averages, efficiency/angular coverage, Rosenthal-Henderson plots and local/global 3D reconstruction resolution. CONCLUSION: DRPnet shows greatly improved time-savings to generate an initial particle dataset compared to manual picking, followed by template-based autopicking. Compared to other networks, DRPnet has equivalent or better performance. DRPnet excels on cryoEM datasets that have low contrast or clumped particles. Evaluating other performance metrics, DRPnet is useful for higher resolution 3D reconstructions with decreased particle numbers or unknown symmetry, detecting particles with better angular orientation coverage.

Collagen matrix perturbations in corneal stroma of Ossabaw mini pigs with type 2 diabetes.

Purpose: Diabetes mellitus (DM) is a metabolic disorder that affects over 450 million people worldwide. DM is characterized by hyperglycemia, causing severe systemic damage to the heart, kidneys, skin, vasculature, nerves, and eye. Type 2 diabetes (T2DM) constitutes 90% of clinical cases and is the most common cause of blindness in working adults. Also, about 70% of T2DM patients show corneal complications including delayed wound healing, often described as diabetic keratopathy (DK). Despite the increasing severity of DM, the research on DK is bleak. This study investigated cellular morphology and collagen matrix alterations of the diabetic and non-diabetic corneas collected from Ossabaw mini pigs, a T2DM animal model with a "thrifty genotype." Methods: Pig corneas were collected from six-month-old Ossabaw miniature pigs fed on a western diet (WD) for ten weeks. The tissues were processed for immunohistochemistry and analyzed using hematoxylin and eosin staining, Mason Trichrome staining, Picrosirus Red staining, Collage I staining, and TUNEL assay. mRNA was prepared to quantify fibrotic gene expression using quantitative reverse-transcriptase PCR (qRT-PCR). Transmission electron microscopy (TEM) was performed to evaluate stromal fibril arrangements to compare collagen dynamics in WD vs. standard diet (SD) fed Ossabaw pig corneas. Results: Ossabaw mini pigs fed on a WD for 10 weeks exhibit classic symptoms of metabolic syndrome and hyperglycemia seen in T2DM patients. We observed significant disarray in cornea stromal collagen matrix in Ossabaw mini pigs fed on WD compared to the age-matched mini pigs fed on a standard chow diet using Masson Trichome and Picrosirius Red staining. Furthermore, ultrastructure evaluation using TEM showed alterations in stromal collagen fibril size and organization in diabetic corneas compared to healthy age-matched corneas. These changes were accompanied by significantly decreased levels of Collagen IV and increased expression of matrix metallopeptidase 9 in WD-fed pigs. Conclusions: This pilot study indicates that Ossabaw mini pigs fed on WD showed collagen disarray and altered gene expression involved in wound healing, suggesting that corneal stromal collagens are vulnerable to diabetic conditions.

Comparison of two cleaning and sterilization protocols of diamond burr tips used in debridement for canine superficial chronic corneal epithelial defects.

OBJECTIVES: To serially evaluate morphologic and elemental composition changes to diamond burr tips (DBTs) comparing two sterilization protocols. ANIMALS STUDIED: A total of 300 fresh cadaver porcine globes. PROCEDURES: Six DBTs were randomly, equally assigned into Group 1 or 2, and then analyzed using Scanning Electron Microscopy (SEM) and Energy Dispersive Spectroscopy (EDS) at 0, 25, 50, and 100 cycles. Diamond burr debridement (DBD) was performed for 120 seconds on corneal stroma using the Algerbrush®. DBTs were cleaned, and then: Group 1 was sterilized by Germinator 500™; and Group 2 underwent ultrasonic cleaning and pre-vacuum autoclave. A cycle is defined as one DBD, cleaning and sterilization protocol. Data were quantified using custom MatLab program. RESULTS: Energy Dispersive Spectroscopy revealed minor buildup of sulfur on both groups. Group 1 displayed major buildup of carbon and calcium. All DBTs were stippled with inorganic particulate at baseline. Particulates were no longer present on Group 2 by 25 cycles, but remained on Group 1 at all time points. There was significantly more buildup on Group 1 at all time points (P = 0.0000, 0.0009, and 0.0003 for 25, 50, and 100 cycles, respectively). More damage to Group 2 at all time points (P = 0.003, 0.002, and 0.003 for 25, 50, and 100 cycles, respectively) was observed. CONCLUSIONS: No significant damage to Group 1 DBTs was noted after 100 cycles, however, particulate matter is not adequately removed using this sterilization technique. Ultrasonic cleaning is warranted between DBDs to achieve adequate particulate removal prior to sterilization; greater damage occurs with this technique which supports replacing DBTs regularly.

Flexible Hinges in Bacterial Chemoreceptors.

Transmembrane bacterial chemoreceptors are extended, rod-shaped homodimers with ligand-binding sites at one end and interaction sites for signaling complex formation and histidine kinase control at the other. There are atomic-resolution structures of chemoreceptor fragments but not of intact, membrane-inserted receptors. Electron tomography of in vivo signaling complex arrays lack distinct densities for chemoreceptor rods away from the well-ordered base plate region, implying structural heterogeneity. We used negative staining, transmission electron microscopy, and image analysis to characterize the molecular shapes of intact homodimers of the Escherichia coli aspartate receptor Tar rendered functional by insertion into nanodisc-provided E. coli lipid bilayers. Single-particle analysis plus tomography of particles in a three-dimensional matrix revealed two bend loci in the chemoreceptor cytoplasmic domain, (i) a short, two-strand gap between the membrane-proximal, four-helix-bundle HAMP (histidine kinases, adenylyl cyclases, methyl-accepting chemoreceptors, and phosphatases) domain and the membrane-distal, four-helix coiled coil and (ii) aligned glycines in the extended, four-helix coiled coil, the position of a bend noted in the previous X-ray structure of a receptor fragment. Our images showed HAMP bends from 0° to ∼13° and glycine bends from 0° to ∼20°, suggesting that the loci are flexible hinges. Variable hinge bending explains indistinct densities for receptor rods outside the base plate region in subvolume averages of chemotaxis arrays. Bending at flexible hinges was not correlated with the chemoreceptor signaling state. However, our analyses showed that chemoreceptor bending avoided what would otherwise be steric clashes between neighboring receptors that would block the formation of core signaling complexes and chemoreceptor arrays.IMPORTANCE This work provides new information about the shape of transmembrane bacterial chemoreceptors, crucial components in the molecular machinery of bacterial chemotaxis. We found that intact, lipid-bilayer-inserted, and thus functional homodimers of the Escherichia coli chemoreceptor Tar exhibited bends at two flexible hinges along their ∼200-Å, rod-like, cytoplasmic domains. One hinge was at the short, two-strand gap between the membrane-proximal, four-helix-bundle HAMP (histidine kinases, adenylyl cyclases, methyl-accepting chemoreceptors, and phosphatases) domain and the membrane-distal, four-helix coiled coil. The other hinge was at aligned glycines in the extended, four-helix coiled coil, where a bend had been identified in the X-ray structure of a chemoreceptor fragment. Our analyses showed that flexible hinge bending avoided structural clashes in chemotaxis core complexes and their arrays.

Vasoactive agonists exert dynamic and coordinated effects on vascular smooth muscle cell elasticity, cytoskeletal remodelling and adhesion.

In this study, we examined the ability of vasoactive agonists to induce dynamic changes in vascular smooth muscle cell (VSMC) elasticity and adhesion, and tested the hypothesis that these events are coordinated with rapid remodelling of the cortical cytoskeleton. Real-time measurement of cell elasticity was performed with atomic force microscopy (AFM) and adhesion was assessed with AFM probes coated with fibronectin (FN). Temporal data were analysed using an Eigen-decomposition method. Elasticity in VSMCs displayed temporal oscillations with three components at approximately 0.001, 0.004 and 0.07 Hz, respectively. Similarly, adhesion displayed a similar oscillatory pattern. Angiotensin II (ANG II, 10(-6) M) increased (+100%) the amplitude of the oscillations, whereas the vasodilator adenosine (ADO, 10(-4) M) reduced oscillation amplitude (-30%). To test whether the oscillatory changes were related to the architectural alterations in cortical cytoskeleton, the topography of the submembranous actin cytoskeleton (100-300 nm depth) was acquired with AFM. These data were analysed to compare cortical actin fibre distribution and orientation before and after treatment with vasoactive agonists. The results showed that ANG II increased the density of stress fibres by 23%, while ADO decreased the density of the stress fibres by 45%. AFM data were supported by Western blot and confocal microscopy. Collectively, these observations indicate that VSMC cytoskeletal structure and adhesion to the extracellular matrix are dynamically altered in response to agonist stimulation. Thus, vasoactive agonists probably invoke unique mechanisms that dynamically alter the behaviour and structure of both the VSMC cytoskeleton and focal adhesions to efficiently support the normal contractile behaviour of VSMCs.

Superresolution microscope image reconstruction by spatiotemporal object decomposition and association: application in resolving t-tubule structure in skeletal muscle.

One key factor that limits resolution of single-molecule superresolution microscopy relates to the localization accuracy of the activated emitters, which is usually deteriorated by two factors. One originates from the background noise due to out-of-focus signals, sample auto-fluorescence, and camera acquisition noise; and the other is due to the low photon count of emitters at a single frame. With fast acquisition rate, the activated emitters can last multiple frames before they transiently switch off or permanently bleach. Effectively incorporating the temporal information of these emitters is critical to improve the spatial resolution. However, majority of the existing reconstruction algorithms locate the emitters frame by frame, discarding or underusing the temporal information. Here we present a new image reconstruction algorithm based on tracklets, short trajectories of the same objects. We improve the localization accuracy by associating the same emitters from multiple frames to form tracklets and by aggregating signals to enhance the signal to noise ratio. We also introduce a weighted mean-shift algorithm (WMS) to automatically detect the number of modes (emitters) in overlapping regions of tracklets so that not only well-separated single emitters but also individual emitters within multi-emitter groups can be identified and tracked. In combination with a maximum likelihood estimator method (MLE), we are able to resolve low to medium density of overlapping emitters with improved localization accuracy. We evaluate the performance of our method with both synthetic and experimental data, and show that the tracklet-based reconstruction is superior in localization accuracy, particularly for weak signals embedded in a strong background. Using this method, for the first time, we resolve the transverse tubule structure of the mammalian skeletal muscle.

Corneal injury is associated with stromal and vascular alterations within cranial dura mater.

The cornea and cranial dura mater share sensory innervation. This link raises the possibility that pathological impulses mediated by corneal injury may be transmitted to the cranial dura, trigger dural perivascular/connective tissue nociceptor responses, and induce vascular and stromal alterations affecting dura mater blood and lymphatic vessel functionality. In this study, using a mouse model, we demonstrate for the first time that two weeks after the initial insult, alkaline injury to the cornea leads to remote pathological changes within the coronal suture area of the dura mater. Specifically, we detected significant pro-fibrotic changes in the dural stroma, as well as vascular remodeling characterized by alterations in vascular smooth muscle cell (VSMC) morphology, reduced blood vessel VSMC coverage, endothelial cell expression of the fibroblast specific protein 1, and significant increase in the number of podoplanin-positive lymphatic sprouts. Intriguingly, the deficiency of a major extracellular matrix component, small leucine-rich proteoglycan decorin, modifies both the direction and the extent of these changes. As the dura mater is the most important route for the brain metabolic clearance, these results are of clinical relevance and provide a much-needed link explaining the association between ophthalmic conditions and the development of neurodegenerative diseases.

LARNet-STC: Spatio-temporal orthogonal region selection network for laryngeal closure detection in endoscopy videos.

The vocal folds (VFs) are a pair of muscles in the larynx that play a critical role in breathing, swallowing, and speaking. VF function can be adversely affected by various medical conditions including head or neck injuries, stroke, tumor, and neurological disorders. In this paper, we propose a deep learning system for automated detection of laryngeal adductor reflex (LAR) events in laryngeal endoscopy videos to enable objective, quantitative analysis of VF function. The proposed deep learning system incorporates our novel orthogonal region selection network and temporal context. This network learns to directly map its input to a VF open/close state without first segmenting or tracking the VF region. This one-step approach drastically reduces manual annotation needs from labor-intensive segmentation masks or VF motion tracks to frame-level class labels. The proposed spatio-temporal network with an orthogonal region selection subnetwork allows integration of local image features, global image features, and VF state information in time for robust LAR event detection. The proposed network is evaluated against several network variations that incorporate temporal context and is shown to lead to better performance. The experimental results show promising performance for automated, objective, and quantitative analysis of LAR events from laryngeal endoscopy videos with over 90% and 99% F1 scores for LAR and non-LAR frames respectively.

Axon and Myelin Sheath Segmentation in Electron Microscopy Images using Meta Learning.

Various neurological diseases affect the morphology of myelinated axons. Quantitative analysis of these structures and changes occurring due to neurodegeneration or neuroregeneration is of great importance for characterization of disease state and treatment response. This paper proposes a robust, meta-learning based pipeline for segmentation of axons and surrounding myelin sheaths in electron microscopy images. This is the first step towards computation of electron microscopy related bio-markers of hypoglossal nerve degeneration/regeneration. This segmentation task is challenging due to large variations in morphology and texture of myelinated axons at different levels of degeneration and very limited availability of annotated data. To overcome these difficulties, the proposed pipeline uses a meta learning-based training strategy and a U-net like encoder decoder deep neural network. Experiments on unseen test data collected at different magnification levels (i.e, trained on 500X and 1200X images, and tested on 250X and 2500X images) showed improved segmentation performance by 5% to 7% compared to a regularly trained, comparable deep learning network.

A Strength Endurance Exercise Paradigm Mitigates Deficits in Hypoglossal-Tongue Axis Function, Strength, and Structure in a Rodent Model of Hypoglossal Motor Neuron Degeneration.

The tongue plays a crucial role in the swallowing process, and impairment can lead to dysphagia, particularly in motor neuron diseases (MNDs) resulting in hypoglossal-tongue axis degeneration (e.g., amyotrophic lateral sclerosis and progressive bulbar palsy). This study utilized our previously established inducible rodent model of dysphagia due to targeted degeneration of the hypoglossal-tongue axis. This model was created by injecting cholera toxin B conjugated to saporin (CTB-SAP) into the genioglossus muscle of the tongue base for retrograde transport to the hypoglossal (XII) nucleus via the hypoglossal nerve, which provides the sole motor control of the tongue. Our goal was to investigate the effect of high-repetition/low-resistance tongue exercise on tongue function, strength, and structure in four groups of male rats: (1) control + sham exercise (n = 13); (2) control + exercise (n = 10); (3) CTB-SAP + sham exercise (n = 13); and (4) CTB-SAP + exercise (n = 12). For each group, a custom spout with adjustable lick force requirement for fluid access was placed in the home cage overnight on days 4 and 6 post-tongue injection. For the two sham exercise groups, the lick force requirement was negligible. For the two exercise groups, the lick force requirement was set to ∼40% greater than the maximum voluntary lick force for individual rats. Following exercise exposure, we evaluated the effect on hypoglossal-tongue axis function (via videofluoroscopy), strength (via force-lickometer), and structure [via Magnetic Resonance Imaging (MRI) of the brainstem and tongue in a subset of rats]. Results showed that sham-exercised CTB-SAP rats had significant deficits in lick rate, swallow timing, and lick force. In exercised CTB-SAP rats, lick rate and lick force were preserved; however, swallow timing deficits persisted. MRI revealed corresponding degenerative changes in the hypoglossal-tongue axis that were mitigated by tongue exercise. These collective findings suggest that high-repetition/low-resistance tongue exercise in our model is a safe and effective treatment to prevent/diminish signs of hypoglossal-tongue axis degeneration. The next step is to leverage our rat model to optimize exercise dosing parameters and investigate corresponding treatment mechanisms of action for future translation to MND clinical trials.

Histopathology tissue segmentation by combining fuzzy clustering with multiphase vector level sets.

High resolution, multispectral, and multimodal imagery of tissue biopsies is an indispensable source of information for diagnosis and prognosis of diseases. Automatic extraction of relevant features from these imagery is a valuable assistance for medical experts. A primary step in computational histology is accurate image segmentation to detect the number and spatial distribution of cell nuclei in the tissue, along with segmenting other structures such as lumen and epithelial regions which together make up a gland structure. This chapter presents an automatic segmentation system for histopathology imaging. Spatial constraint fuzzy C-means provides an unsupervised initialization. An active contour algorithm that combines multispectral edge and region informations through a vector multiphase level set framework and Beltrami color metric tensors refines the segmentation. An improved iterative kernel filtering approach detects individual nuclei centers and decomposes densely clustered nuclei structures. The obtained results show high performances for nuclei detection compared to the human annotation.

Ensemble of Deep Learning Cascades for Segmentation of Blood Vessels in Confocal Microscopy Images.

Detection, segmentation, and quantification of microvascular structures are the main steps towards studying microvascular remodeling. Combined with appropriate staining, confocal microscopy imaging enables exploration of the full 3D anatomical characteristics of microvascular systems. Segmentation of confocal microscopy images is a challenging task due to complexity of anatomical structures, staining and imaging issues, and lack of annotated training data. In this paper, we propose a deep learning system for robust segmentation of cranial vasculature of mice in confocal microscopy images. The proposed system is an ensemble of two deep-learning cascades consisting of two coarse-to-fine subnetworks with skip connections in between. One cascade aims to improve sensitivity, while the other aims to improve precision of the segmentation results. Our experiments on mice cranial vasculature showed promising results achieving segmentation accuracy of 92.02% and dice score of 81.45% despite being trained on very limited confocal microscopy data.

DMNet: Dual-Stream Marker Guided Deep Network for Dense Cell Segmentation and Lineage Tracking.

Accurate segmentation and tracking of cells in microscopy image sequences is extremely beneficial in clinical diagnostic applications and biomedical research. A continuing challenge is the segmentation of dense touching cells and deforming cells with indistinct boundaries, in low signal-to-noise-ratio images. In this paper, we present a dual-stream marker-guided network (DMNet) for segmentation of touching cells in microscopy videos of many cell types. DMNet uses an explicit cell marker-detection stream, with a separate mask-prediction stream using a distance map penalty function, which enables supervised training to focus attention on touching and nearby cells. For multi-object cell tracking we use M2Track tracking-by-detection approach with multi-step data association. Our M2Track with mask overlap includes short term track-to-cell association followed by track-to-track association to re-link tracklets with missing segmentation masks over a short sequence of frames. Our combined detection, segmentation and tracking algorithm has proven its potential on the IEEE ISBI 2021 6th Cell Tracking Challenge (CTC-6) where we achieved multiple top three rankings for diverse cell types. Our team name is MU-Ba-US, and the implementation of DMNet is available at, http://celltrackingchallenge.net/participants/MU-Ba-US/.

Multi-focus Image Fusion for Confocal Microscopy Using U-Net Regression Map.

Characterizing the spatial relationship between blood vessel and lymphatic vascular structures, in the mice dura mater tissue, is useful for modeling fluid flows and changes in dynamics in various disease processes. We propose a new deep learning-based approach to fuse a set of multi-channel single-focus microscopy images within each volumetric z-stack into a single fused image that accurately captures as much of the vascular structures as possible. The red spectral channel captures small blood vessels and the green fluorescence channel images lymphatics structures in the intact dura mater attached to bone. The deep architecture Multi-Channel Fusion U-Net (MCFU-Net) combines multi-slice regression likelihood maps of thin linear structures using max pooling for each channel independently to estimate a slice-based focus selection map. We compare MCFU-Net with a widely used derivative-based multi-scale Hessian fusion method [8]. The multi-scale Hessian-based fusion produces dark-halos, non-homogeneous backgrounds and less detailed anatomical structures. Perception based no-reference image quality assessment metrics PIQUE, NIQE, and BRISQUE confirm the effectiveness of the proposed method.

Orthogonal Region Selection Network for Laryngeal Closure Detection in Laryngoscopy Videos.

Vocal folds (VFs) play a critical role in breathing, swallowing, and speech production. VF dysfunctions caused by various medical conditions can significantly reduce patients' quality of life and lead to life-threatening conditions such as aspiration pneumonia, caused by food and/or liquid "invasion" into the windpipe. Laryngeal endoscopy is routinely used in clinical practice to inspect the larynx and to assess the VF function. Unfortunately, the resulting videos are only visually inspected, leading to loss of valuable information that can be used for early diagnosis and disease or treatment monitoring. In this paper, we propose a deep learning-based image analysis solution for automated detection of laryngeal adductor reflex (LAR) events in laryngeal endoscopy videos. Laryngeal endoscopy image analysis is a challenging task because of anatomical variations and various imaging problems. Analysis of LAR events is further challenging because of data imbalance since these are rare events. In order to tackle this problem, we propose a deep learning system that consists of a two-stream network with a novel orthogonal region selection subnetwork. To our best knowledge, this is the first deep learning network that learns to directly map its input to a VF open/close state without first segmenting or tracking the VF region, which drastically reduces labor-intensive manual annotation needed for mask or track generation. The proposed two-stream network and the orthogonal region selection subnetwork allow integration of local and global information for improved performance. The experimental results show promising performance for the automated, objective, and quantitative analysis of LAR events from laryngeal endoscopy videos.Clinical relevance- This paper presents an objective, quantitative, and automatic deep learning based system for detection of laryngeal adductor reflex (LAR) events in laryngoscopy videos.

Robust mosaicking of maize fields from aerial imagery.

PREMISE: Aerial imagery from small unmanned aerial vehicle systems is a promising approach for high-throughput phenotyping and precision agriculture. A key requirement for both applications is to create a field-scale mosaic of the aerial imagery sequence so that the same features are in registration, a very challenging problem for crop imagery. METHODS: We have developed an improved mosaicking pipeline, Video Mosaicking and summariZation (VMZ), which uses a novel two-dimensional mosaicking algorithm that minimizes errors in estimating the transformations between successive frames during registration. The VMZ pipeline uses only the imagery, rather than relying on vehicle telemetry, ground control points, or global positioning system data, to estimate the frame-to-frame homographies. It exploits the spatiotemporal ordering of the image frames to reduce the computational complexity of finding corresponding features between frames using feature descriptors. We compared the performance of VMZ to a standard two-dimensional mosaicking algorithm (AutoStitch) by mosaicking imagery of two maize (Zea mays) research nurseries freely flown with a variety of trajectories. RESULTS: The VMZ pipeline produces superior mosaics faster. Using the speeded up robust features (SURF) descriptor, VMZ produces the highest-quality mosaics. DISCUSSION: Our results demonstrate the value of VMZ for the future automated extraction of plant phenotypes and dynamic scouting for crop management.