High-resolution Details Research Articles

Insights to Human γD-Crystallin Unfolding by NMR Spectroscopy and Molecular Dynamics Simulations.

Human γD-crystallin (HGDC) is an abundant lens protein residing in the nucleus of the human lens. Aggregation of this and other structural proteins within the lens leads to the development of cataract. Much has been explored on the stability and aggregation of HGDC and where detailed investigation at the atomic resolution was needed, the X-ray structure was used as an initial starting conformer for molecular modeling. In this study, we implemented NMR-solution HGDC structures as starting conformers for molecular dynamics simulations to provide the missing pieces of the puzzle on the very early stages of HGDC unfolding leading up to the domain swap theories proposed by past studies. The high-resolution details of the conformational dynamics also revealed additional insights to possible early intervention for cataractogenesis.

Structural Identification of Individual Helical Amyloid Filaments by Integration of Cryo-Electron Microscopy-Derived Maps in Comparative Morphometric Atomic Force Microscopy Image Analysis

The presence of amyloid fibrils is a hallmark of more than 50 human disorders, including neurodegenerative diseases and systemic amyloidoses. A key unresolved challenge in understanding the involvement of amyloid in disease is to explain the relationship between individual structural polymorphs of amyloid fibrils, in potentially mixed populations, and the specific pathologies with which they are associated. Although cryo-electron microscopy (cryo-EM) and solid-state nuclear magnetic resonance (ssNMR) spectroscopy methods have been successfully employed in recent years to determine the structures of amyloid fibrils with high resolution detail, they rely on ensemble averaging of fibril structures in the entire sample or significant subpopulations. Here, we report a method for structural identification of individual fibril structures imaged by atomic force microscopy (AFM) by integration of high-resolution maps of amyloid fibrils determined by cryo-EM in comparative AFM image analysis. This approach was demonstrated using the hitherto structurally unresolved amyloid fibrils formed in vitro from a fragment of tau (297–391), termed ‘dGAE’. Our approach established unequivocally that dGAE amyloid fibrils bear no structural relationship to heparin-induced tau fibrils formed in vitro. Furthermore, our comparative analysis resulted in the prediction that dGAE fibrils are structurally closely related to the paired helical filaments (PHFs) isolated from Alzheimer’s disease (AD) brain tissue characterised by cryo-EM. These results show the utility of individual particle structural analysis using AFM, provide a workflow of how cryo-EM data can be incorporated into AFM image analysis and facilitate an integrated structural analysis of amyloid polymorphism.

Late Holocene vegetation history of the Western Caucasus inferred from high-resolution pollen record from Lake Karakel

Abstract. The paper presents a new paleoecological evidence for the last 2200 years based on high resolution pollen record and detail radiocarbon dating, obtained from Lake Karakel (Teberda River valley, Western Caucasus). The obtained results showed that vegetation changes from 2200 to stage 1200 cal, yr BP occurred under a climate warming and increase of humidity. The Medieval Climatic Anomaly-MCA (ca. 1200-880 cal. yr BP) was characterized in this region by the predominance of broadleaved forests. The MCA was followed by a cold interval of the Little Ice Age (ca. 880-120 cal yr BP) that marked by expansion of pine, spruce and fir forests to the lower altitude the present time.

Learning Residual Color for Novel View Synthesis.

Scene Representation Networks (SRN) have been proven as a powerful tool for novel view synthesis in recent works. They learn a mapping function from the world coordinates of spatial points to radiance color and the scene's density using a fully connected network. However, scene texture contains complex high-frequency details in practice that is hard to be memorized by a network with limited parameters, leading to disturbing blurry effects when rendering novel views. In this paper, we propose to learn 'residual color' instead of 'radiance color' for novel view synthesis, i.e., the residuals between surface color and reference color. Here the reference color is calculated based on spatial color priors, which are extracted from input view observations. The beauty of such a strategy lies in that the residuals between radiance color and reference are close to zero for most spatial points thus are easier to learn. A novel view synthesis system that learns the residual color using SRN is presented in this paper. Experiments on public datasets demonstrate that the proposed method achieves competitive performance in preserving high-resolution details, leading to visually more pleasant results than the state of the arts.

Advanced active-gas 3D printing of 436 stainless steel for future rocket engine structure manufacture

During this study a novel means to fabricate a rocket engine combustion chamber was developed. Here the shape of this rocket engine component has been fabricated using prototype 3D metal active gas deposition technology. The manufacturing process is based upon an argon gas fused deposition process for producing complex component geometries. This process uses multi-pass deposition employing stainless steel 436 alloy as the feed material. The structures created during this process were metallographically characterised to understand the microstructural properties produced during fabrication. The engine combustion chamber casing was fabricated and later machined to produce a smooth precision finish. The chamber casing features high resolution details with fine control of the process resulting in micron-scale precision. This allowed for the creation of wall structures and hollow internal sections. It was by allowing the structure to cool precisely between each layer that resulted in no loss in precision due to imposed cooling stresses on this large-scale structure. It is this factor which is of significant importance towards the formation of complex components. The process developed was a paramount means towards the fabrication of robust structures.

Predicting high-resolution turbulence details in space and time

Predicting the fine and intricate details of a turbulent flow field in both space and time from a coarse input remains a major challenge despite the availability of modern machine learning tools. In this paper, we present a simple and effective dictionary-based approach to spatio-temporal upsampling of fluid simulation. We demonstrate that our neural network approach can reproduce the visual complexity of turbulent flows from spatially and temporally coarse velocity fields even when using a generic training set. Moreover, since our method generates finer spatial and/or temporal details through embarrassingly-parallel upsampling of small local patches, it can efficiently predict high-resolution turbulence details across a variety of grid resolutions. As a consequence, our method offers a whole range of applications varying from fluid flow upsampling to fluid data compression. We demonstrate the efficiency and generalizability of our method for synthesizing turbulent flows on a series of complex examples, highlighting dramatically better results in spatio-temporal upsampling and flow data compression than existing methods as assessed by both qualitative and quantitative comparisons.

CleftNet: Augmented Deep Learning for Synaptic Cleft Detection From Brain Electron Microscopy.

Detecting synaptic clefts is a crucial step to investigate the biological function of synapses. The volume electron microscopy (EM) allows the identification of synaptic clefts by photoing EM images with high resolution and fine details. Machine learning approaches have been employed to automatically predict synaptic clefts from EM images. In this work, we propose a novel and augmented deep learning model, known as CleftNet, for improving synaptic cleft detection from brain EM images. We first propose two novel network components, known as the feature augmentor and the label augmentor, for augmenting features and labels to improve cleft representations. The feature augmentor can fuse global information from inputs and learn common morphological patterns in clefts, leading to augmented cleft features. In addition, it can generate outputs with varying dimensions, making it flexible to be integrated in any deep network. The proposed label augmentor augments the label of each voxel from a value to a vector, which contains both the segmentation label and boundary label. This allows the network to learn important shape information and to produce more informative cleft representations. Based on the proposed feature augmentor and label augmentor, We build the CleftNet as a U-Net like network. The effectiveness of our methods is evaluated on both external and internal tasks. Our CleftNet currently ranks #1 on the external task of the CREMI open challenge. In addition, both quantitative and qualitative results in the internal tasks show that our method outperforms the baseline approaches significantly.

Image quality enhancement using hybrid attention networks

Image quality enhancement aims to recover rich details from degraded images, which is applied into many fields, such as medical imaging, filming production and autonomous driving. Deep convolutional neural networks (CNNs) have enabled rapid development of image quality enhancement. However, most existing CNN-based methods lack versatility when targeting different subtasks in terms of the design of networks. Besides, they often fail to balance precise spatial representations and necessary contextual information. To deal with these problems, this paper proposes a novel unified framework for low-light image enhancement, image denoising and image super-resolution. The core of this architecture is a residual hybrid attention block (RHAB), which consists of several dynamic down-sampling modules (DDM) and hybrid attention up-sampling modules (HAUM). Specifically, multi-scale feature maps are fully interacted with each other with the help of nested subnetworks so that both high-resolution spatial details and high-level contextual information can be combined to improve the representation ability of the network. Further, a hybrid attention network (HAN) is proposed and evaluations on three separate subtasks demonstrate its good performance. Extensive experiments on the authors' synthetic dataset, a more complex version, show that the authors' method achieve better quantitative and visual results compared to other state-of-the-art methods.

Exposure of Von Willebrand Factor Cleavage Site in A1A2A3-Fragment under Extreme Hydrodynamic Shear.

Von Willebrand Factor (vWf) is a giant multimeric extracellular blood plasma involved in hemostasis. In this work we present multi-scale simulations of its three-domains fragment A1A2A3. These three domains are essential for the functional regulation of vWf. Namely the A2 domain hosts the site where the protease ADAMTS13 cleavages the multimeric vWf allowing for its length control that prevents thrombotic conditions. The exposure of the cleavage site follows the elongation/unfolding of the domain that is caused by an increased shear stress in blood. By deploying Lattice Boltzmann molecular dynamics simulations based on the OPEP coarse-grained model for proteins, we investigated at molecular level the unfolding of the A2 domain under the action of a perturbing shear flow. We described the structural steps of this unfolding that mainly concerns the -strand structures of the domain, and we compared the process occurring under shear with that produced by the action of a directional pulling force, a typical condition of single molecule experiments. We observe, that under the action of shear flow, the competition among the elongational and rotational components of the fluid field leads to a complex behaviour of the domain, where elongated structures can be followed by partially collapsed melted globule structures with a very different degree of exposure of the cleavage site. Our simulations pose the base for the development of a multi-scale in-silico description of vWf dynamics and functionality in physiological conditions, including high resolution details for molecular relevant events, e.g., the binding to platelets and collagen during coagulation or thrombosis.

All-d-Enantiomeric Peptide D3 Designed for Alzheimer's Disease Treatment Dynamically Interacts with Membrane-Bound Amyloid-β Precursors.

Alzheimer's disease (AD) is a severe neurodegenerative pathology with no effective treatment known. Toxic amyloid-β peptide (Aβ) oligomers play a crucial role in AD pathogenesis. All-d-Enantiomeric peptide D3 and its derivatives were developed to disassemble and destroy cytotoxic Aβ aggregates. One of the D3-like compounds is approaching phase II clinical trials; however, high-resolution details of its disease-preventing or pharmacological actions are not completely clear. We demonstrate that peptide D3 stabilizing Aβ monomer dynamically interacts with the extracellular juxtamembrane region of a membrane-bound fragment of an amyloid precursor protein containing the Aβ sequence. MD simulations based on NMR measurement results suggest that D3 targets the amyloidogenic region, not compromising its α-helicity and preventing intermolecular hydrogen bonding, thus creating prerequisites for inhibition of early steps of Aβ conversion into β-conformation and its toxic oligomerization. An enhanced understanding of the D3 action molecular mechanism facilitates development of effective AD treatment and prevention strategies.

Wide-Field Pixel Super-Resolution Colour Lensfree Microscope for Digital Pathology.

Whole slide imaging enables scanning entire stained-glass slides with high resolution into digital images for the tissue morphology/molecular pathology assessment and analysis, which has increased in adoption for both clinical and research applications. As an alternative to conventional optical microscopy, lensfree holography imaging, which offers high resolution and a wide field of view (FOV) with digital focus, has been widely used in various types of biomedical imaging. However, accurate colour holographic imaging with pixel super-resolution reconstruction has remained a great challenge due to its coherent characteristic. In this work, we propose a wide-field pixel super-resolution colour lensfree microscopy by performing wavelength scanning pixel super-resolution and phase retrieval simultaneously on the three channels of red, green and blue (RGB), respectively. High-resolution RGB three-channel composite colour image is converted to the YUV space for separating the colour component and the brightness component, keeping the brightness component unchanged as well as enhancing the colour component through average filter, which not only eliminates the common rainbow artifacts of holographic colour reconstruction but also maintains the high-resolution details collected under different colour illuminations. We conducted experiments on the reconstruction of a USAF1951, stained lotus root and red bone marrow smear for performance evaluation of the spatial resolution and colour reconstruction with an imaging FOV >40 mm2.

Sensor location design for interdicting mobile travelers with probabilistic space-time trajectories

Most existing sensor location design studies focus on stationary or deterministic objects (e.g., fixed OD demand) yet have not well investigated uncertain mobility of inspected objects. To bridge this gap, this study aims to integrate human mobility characteristics specified by the time geography theory into sensor location design. We first investigate trajectories of moving objects to see how they can reveal detailed information about human travel characteristics and presence probability with high-resolution detail. Then, a space–time network-based modeling framework is proposed to integrate human mobility into network location design problems. We construct a probabilistic network structure to quantify human’s presence probability at different locations and time. Then, a Mixed Integer Nonlinear Programming (MINLP) model is proposed to maximize the spatial and temporal coverage of individual targets. To obtain near optimal solutions for large-scale problems, greedy heuristic, simulated annealing and Lagrangian relaxation algorithms are tested. Theoretical analysis is conducted to show the optimality gap of the greedy algorithm. The proposed algorithms are implemented on hypothetical and real-world numerical examples (some on a high-performance computing platform) to demonstrate the performance and effectiveness of the proposed model on different network sizes and promising results have been obtained.

Neural UpFlow

We present a novel up-resing technique for generating high-resolution liquids based on scene flow estimation using deep neural networks. Our approach infers and synthesizes small- and large-scale details solely from a low-resolution particle-based liquid simulation. The proposed network leverages neighborhood contributions to encode inherent liquid properties throughout convolutions. We also propose a particle-based approach to interpolate between liquids generated from varying simulation discretizations using a state-of-the-art bidirectional optical flow solver method for fluids in addition with a novel key-event topological alignment constraint. In conjunction with the neighborhood contributions, our loss formulation allows the inference model throughout epochs to reward important differences in regard to significant gaps in simulation discretizations. Even when applied in an untested simulation setup, our approach is able to generate plausible high-resolution details. Using this interpolation approach and the predicted displacements, our approach combines the input liquid properties with the predicted motion to infer semi-Lagrangian advection. We furthermore showcase how the proposed interpolation approach can facilitate generating large simulation datasets with a subset of initial condition parameters.

Noise Removal Algorithm for Out-of-focus Blurred Images Based on Bilateral Filtering

The feature resolution of traditional methods for fuzzy image denoising is low, for the sake of improve the strepitus removal and investigation ability of defocused blurred night images, a strepitus removal algorithm based on bilateral filtering is suggested. The method include the following steps of: Building an out-of-focus blurred night scene image acquisition model with grid block feature matching of the out-of-focus blurred night scene image; Carrying out information enhancement processing of the out-of-focus blurred night scene image by adopting a high-resolution image detail feature enhancement technology; Collecting edge contour feature quantity of the out-of-focus blurred night scene image; Carrying out grid block feature matching design of the out-of-focus blurred night scene image by adopting a bilateral filtering information reconstruction technology; And building the gray-level histogram information location model of the out-of-focus blurred night scene image. Fuzzy pixel information fusion investigation method is used to collect gray features of defocused blurred night images. According to the feature collection results, bilateral filtering algorithm is used to automatically optimize the strepitus removal of defocused blurred night images. The simulation results show that the out-of-focus blurred night scene image using this method for machine learning has better strepitus removal performance, shorter time cost and higher export peak signal-to-strepitus proportion.

Single-particle cryo-EM: alternative schemes to improve dose efficiency.

Imaging of biomolecules by ionizing radiation, such as electrons, causes radiation damage which introduces structural and compositional changes of the specimen. The total number of high-energy electrons per surface area that can be used for imaging in cryogenic electron microscopy (cryo-EM) is severely restricted due to radiation damage, resulting in low signal-to-noise ratios (SNR). High resolution details are dampened by the transfer function of the microscope and detector, and are the first to be lost as radiation damage alters the individual molecules which are presumed to be identical during averaging. As a consequence, radiation damage puts a limit on the particle size and sample heterogeneity with which electron microscopy (EM) can deal. Since a transmission EM (TEM) image is formed from the scattering process of the electron by the specimen interaction potential, radiation damage is inevitable. However, we can aim to maximize the information transfer for a given dose and increase the SNR by finding alternatives to the conventional phase-contrast cryo-EM techniques. Here some alternative transmission electron microscopy techniques are reviewed, including phase plate, multi-pass transmission electron microscopy, off-axis holography, ptychography and a quantum sorter. Their prospects for providing more or complementary structural information within the limited lifetime of the sample are discussed.

A Universal Approach to Analyzing Transmission Electron Microscopy with ImageJ.

Transmission electron microscopy (TEM) is widely used as an imaging modality to provide high-resolution details of subcellular components within cells and tissues. Mitochondria and endoplasmic reticulum (ER) are organelles of particular interest to those investigating metabolic disorders. A straightforward method for quantifying and characterizing particular aspects of these organelles would be a useful tool. In this protocol, we outline how to accurately assess the morphology of these important subcellular structures using open source software ImageJ, originally developed by the National Institutes of Health (NIH). Specifically, we detail how to obtain mitochondrial length, width, area, and circularity, in addition to assessing cristae morphology and measuring mito/endoplasmic reticulum (ER) interactions. These procedures provide useful tools for quantifying and characterizing key features of sub-cellular morphology, leading to accurate and reproducible measurements and visualizations of mitochondria and ER.

Fully Convolutional DenseNet with Attention Mechanism for Liver Lesion Segmentation in MRI Images

The goal of this research is to achieve accurate segmentation of liver tumors in noncontrast T2-weighted magnetic resonance imaging. As liver tumors and adjacent organs are represented by pixels of very similar gray intensity, segmentation is challenging, and the presence of different sizes of liver tumor makes segmentation more difficult. Differing from previous work to capture contextual information using multiscale feature fusion with concatenation, attention mechanism is added to our segmentation model to extract precise global contextual information for pixel labeling without requiring complex dilated convolution. This study describe a liver lesion segmentation model derived from FC-DenseNet with attention mechanism. Specifically, a global attention module (GAM) is added to up-sampling path, and high-level features are processed by the GAM to generating weighting information for guiding high resolution detail features recovery. High-level features are very effective for accurate category classification, but relatively weak at pixel classification and predicting restoration of the original resolution, so the fusion of high-level semantic features and low-level detail features can improve segmentation accuracy. A weighted focal loss function is used to solve the problem of lesion area occupying a relatively low proportion of the whole image, and to deal with the disequilibrium of foreground and background in the training liver lesion images. Experimental results show our segmentation model can automatically segment liver tumors from complete MRI images, and the addition of the GAM model can effectively improve liver tumor segmentation. Our algorithms have obvious advantages over other CNN algorithms and traditional manual methods of feature extraction.

Super-resolution reconstruction of infrared images based on a convolutional neural network with skip connections

Image super-resolution technology successfully overcomes the limitation of excessively large pixel size in infrared detectors and meets the increasing demand for high-resolution infrared image information. In this paper, the super-resolution reconstruction of infrared images based on a convolutional neural network with skip connections is reported. The introduction of global residual learning and local residual learning reduces computational complexity and accelerates network convergence. Multiple convolution layers and deconvolution layers respectively implement the extraction and restoration of the features in infrared images. Skip connections and channel fusion are introduced to the network to increase the number of feature maps and promote the deconvolution layers to restore image details. Compared with the other previously proposed methods for infrared information restoration, our proposed method shows obvious advantages in the ability of high-resolution details acquisition.

HIGH-RESOLUTION URBAN MAPPING BY FUSION OF SAR AND OPTICAL DATA

Abstract. Mapping the exact extent of urban areas is a critical prerequisite in many remote sensing applications, such as hazard evaluation and change detection. The usage of Synthetical Aperture Radar (SAR) data has gained popularity due to the unique characteristics of the backscattered radio signal from human-made targets. The Sentinel-1 (S1) constellation, with a global revisit time of 6–12 days in Interferometric Wide Swath (IW) mode and free and open access to the data, allows the development of new applications to monitor urban sites. However, S1 is rarely considered when fine resolution is required due to the large pixel size and the need for spatial averaging to obtain robust estimators. We propose a method to improve Sentinel-1 urban classification performance by exploiting one Multi-Spectral (MS) image acquired by Sentinel-2 (S2). MS data is used for tracing the precise natural boundaries in a scene through superpixels segmentation. A machine learning approach is then applied to interpret the thematic context of each segment from short temporal stacks of coregistered SAR data. We use a short sensing period (around two months), so rapid changes can be traces. The proposed fusion of S1 and S2 data was tested in the area of Milan (Italy), with a total accuracy of about 90%. The ability to follow high-resolution details in a mixed environment is demonstrated, opening the possibility of efficiently tracing the human footprint.

Two-stream network for infrared and visible images fusion

Long-wave infrared(thermal) images distinguish the target and background according to different thermal radiation. They are insensitive to light conditions, and cannot present details obtained from reflected light. By contrast, the visible images have high spatial resolution and texture details, but they are easily affected by the occlusion and light conditions. Combining the advantages of the two sources may generate a new image with clear targets and high resolution, which satisfy requirements in all-weather and all-day/night conditions. Most of the existing methods cannot fully capture the underlying characteristics in the infrared and visible images, and ignore complementary information between the sources. In this paper, we propose an end-to-end model (TSFNet) for infrared and visible image fusion, which is able to handle the sources simultaneously. In addition, it adopts an adaptive weight allocation strategy to capture the informative global features. Experiments on public datasets demonstrate the proposed fusion method achieves state-of-the-art performance, in both global visual quality and quantitative comparison.