Showing papers by "General Electric published in 2017"

Time series classification from scratch with deep neural networks: A strong baseline

Abstract: We propose a simple but strong baseline for time series classification from scratch with deep neural networks. Our proposed baseline models are pure end-to-end without any heavy preprocessing on the raw data or feature crafting. The proposed Fully Convolutional Network (FCN) achieves premium performance to other state-of-the-art approaches and our exploration of the very deep neural networks with the ResNet structure is also competitive. The global average pooling in our convolutional model enables the exploitation of the Class Activation Map (CAM) to find out the contributing region in the raw data for the specific labels. Our models provides a simple choice for the real world application and a good starting point for the future research. An overall analysis is provided to discuss the generalization capability of our models, learned features, network structures and the classification semantics.

Review of Silicon Carbide Power Devices and Their Applications

Abstract: Silicon carbide (SiC) power devices have been investigated extensively in the past two decades, and there are many devices commercially available now. Owing to the intrinsic material advantages of SiC over silicon (Si), SiC power devices can operate at higher voltage, higher switching frequency, and higher temperature. This paper reviews the technology progress of SiC power devices and their emerging applications. The design challenges and future trends are summarized at the end of the paper.

The Visual Object Tracking VOT2017 Challenge Results

Abstract: The Visual Object Tracking challenge VOT2017 is the fifth annual tracker benchmarking activity organized by the VOT initiative. Results of 51 trackers are presented; many are state-of-the-art published at major computer vision conferences or journals in recent years. The evaluation included the standard VOT and other popular methodologies and a new "real-time" experiment simulating a situation where a tracker processes images as if provided by a continuously running sensor. Performance of the tested trackers typically by far exceeds standard baselines. The source code for most of the trackers is publicly available from the VOT page. The VOT2017 goes beyond its predecessors by (i) improving the VOT public dataset and introducing a separate VOT2017 sequestered dataset, (ii) introducing a realtime tracking experiment and (iii) releasing a redesigned toolkit that supports complex experiments. The dataset, the evaluation kit and the results are publicly available at the challenge website1.

High power density superconducting rotating machines—development status and technology roadmap

Abstract: Superconducting technology applications in electric machines have long been pursued due to their significant advantages of higher efficiency and power density over conventional technology. However, in spite of many successful technology demonstrations, commercial adoption has been slow, presumably because the threshold for value versus cost and technology risk has not yet been crossed. One likely path for disruptive superconducting technology in commercial products could be in applications where its advantages become key enablers for systems which are not practical with conventional technology. To help systems engineers assess the viability of such future solutions, we present a technology roadmap for superconducting machines. The timeline considered was ten years to attain a Technology Readiness Level of 6+, with systems demonstrated in a relevant environment. Future projections, by definition, are based on the judgment of specialists, and can be subjective. Attempts have been made to obtain input from a broad set of organizations for an inclusive opinion. This document was generated Superconductor Science and Technology Supercond. Sci. Technol. 30 (2017) 123002 (41pp) https://doi.org/10.1088/1361-6668/aa833e Original content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 0953-2048/17/123002+41$33.00 © 2017 IOP Publishing Ltd Printed in the UK 1 through a series of teleconferences and in-person meetings, including meetings at the 2015 IEEE PES General meeting in Denver, CO, the 2015 ECCE in Montreal, Canada, and a final workshop in April 2016 at the University of Illinois, Urbana-Champaign that brought together a broad group of technical experts spanning the industry, government and academia.

Low-Frequency Power Decoupling in Single-Phase Applications: A Comprehensive Overview

Abstract: This paper presents a literature overview of all techniques proposed until the submission of this paper in terms of mitigating power oscillation in single-phase applications. This pulsating energy is the major factor for increasing the size of passive components and power losses in the converter and can be responsible for losses or malfunctioning of the dc sources. Reduction of power ripple at twice the fundamental frequency is one of the key elements to increase power converter density without lack of dc stiffness. Pulsation reduction is achieved by incorporating control techniques or auxiliary circuitries with energy storage capability in reactive elements to avoid this oscillating power to propagate through the converter, creating what is called as single-phase power decoupling. The topologies are divided as: rectifiers, inverters, and bidirectional. Among them, it is possible to classify as isolated and nonisolated converters. The energy storage method may be classify as: capacitive and inductive. For the power decoupling technique, it is convenient to divide as control and topology. The power decoupling technique may be implemented as series or parallel with respect to the ac, dc or link side. This paper represents the best reference on this topic.

Unsupervised Trajectory Analysis of Single-Cell RNA-Seq and Imaging Data Reveals Alternative Tuft Cell Origins in the Gut.

Abstract: Summary Modern single-cell technologies allow multiplexed sampling of cellular states within a tissue. However, computational tools that can infer developmental cell-state transitions reproducibly from such single-cell data are lacking. Here, we introduce p-Creode, an unsupervised algorithm that produces multi-branching graphs from single-cell data, compares graphs with differing topologies, and infers a statistically robust hierarchy of cell-state transitions that define developmental trajectories. We have applied p-Creode to mass cytometry, multiplex immunofluorescence, and single-cell RNA-seq data. As a test case, we validate cell-state-transition trajectories predicted by p-Creode for intestinal tuft cells, a rare, chemosensory cell type. We clarify that tuft cells are specified outside of the Atoh1 -dependent secretory lineage in the small intestine. However, p-Creode also predicts, and we confirm, that tuft cells arise from an alternative, Atoh1 -driven developmental program in the colon. These studies introduce p-Creode as a reliable method for analyzing large datasets that depict branching transition trajectories.

Environmental Dimensions of Additive Manufacturing: Mapping Application Domains and Their Environmental Implications

Abstract: Additive manufacturing (AM) proposes a novel paradigm for engineering design and manufacturing, which has profound economic, environmental, and security implications. The design freedom offered by this category of manufacturing processes and its ability to locally print almost each designable object will have important repercussions across society. While AM applications are progressing from rapid prototyping to the production of end-use products, the environmental dimensions and related impacts of these evolving manufacturing processes have yet to be extensively examined. Only limited quantitative data are available on how AM manufactured products compare to conventionally manufactured ones in terms of energy and material consumption, transportation costs, pollution and waste, health and safety issues, as well as other environmental impacts over their full lifetime. Reported research indicates that the specific energy of current AM systems is 1 to 2 orders of magnitude higher compared to that of conventional manufacturing processes. However, only part of the AM process taxonomy is yet documented in terms of its environmental performance, and most life cycle inventory (LCI) efforts mainly focus on energy consumption. From an environmental perspective, AM manufactured parts can be beneficial for very small batches, or in cases where AM-based redesigns offer substantial functional advantages during the product use phase (e.g., lightweight part designs and part remanufacturing). Important pending research questions include the LCI of AM feedstock production, supply-chain consequences, and health and safety issues relating to AM.

Lung nodule detection in CT using 3D convolutional neural networks

Abstract: We propose a new computer-aided detection system that uses 3D convolutional neural networks (CNN) for detecting lung nodules in low dose computed tomography. The system leverages both a priori knowledge about lung nodules and confounding anatomical structures and data-driven machine-learned features and classifier. Specifically, we generate nodule candidates using a local geometric-model-based filter and further reduce the structure variability by estimating the local orientation. The nodule candidates in the form of 3D cubes are fed into a deep 3D convolutional neural network that is trained to differentiate nodule and non-nodule inputs. We use data augmentation techniques to generate a large number of training examples and apply regularization to avoid overfitting. On a set of 99 CT scans, the proposed system achieved state-of-the-art performance and significantly outperformed a similar hybrid system that uses conventional shallow learning. The experimental results showed benefits of using a priori models to reduce the problem space for data-driven machine learning of complex deep neural networks. The results also showed the advantages of 3D CNN over 2D CNN in volumetric medical image analysis.

Learning and Incorporating Shape Models for Semantic Segmentation

Abstract: Semantic segmentation has been popularly addressed using Fully convolutional networks (FCN) (e.g. U-Net) with impressive results and has been the forerunner in recent segmentation challenges. However, FCN approaches do not necessarily incorporate local geometry such as smoothness and shape, whereas traditional image analysis techniques have benefitted greatly by them in solving segmentation and tracking problems. In this work, we address the problem of incorporating shape priors within the FCN segmentation framework. We demonstrate the utility of such a shape prior in robust handling of scenarios such as loss of contrast and artifacts. Our experiments show \(\approx 5\%\) improvement over U-Net for the challenging problem of ultrasound kidney segmentation.

Contributions to magnetic susceptibility of brain tissue

Abstract: This review discusses the major contributors to the subtle magnetic properties of brain tissue and how they affect MRI contrast. With the increased availability of high-field scanners, the use of magnetic susceptibility contrast for the study of human brain anatomy and function has increased dramatically. This has not only led to novel applications, but has also improved our understanding of the complex relationship between MRI contrast and magnetic susceptibility. Chief contributors to the magnetic susceptibility of brain tissue have been found to include myelin as well as iron. In the brain, iron exists in various forms with diverse biological roles, many of which are now only starting to be uncovered. An interesting aspect of magnetic susceptibility contrast is its sensitivity to the microscopic distribution of iron and myelin, which provides opportunities to extract information at spatial scales well below MRI resolution. For example, in white matter, the myelin sheath that surrounds the axons can provide tissue contrast that is dependent on the axonal orientation and reflects the relative size of intra- and extra-axonal water compartments. The extraction of such ultrastructural information, together with quantitative information about iron and myelin concentrations, is an active area of research geared towards the characterization of brain structure and function, and their alteration in disease. Copyright © 2016 John Wiley & Sons, Ltd.

AC Stability Analysis and dq Frame Impedance Specifications in Power-Electronics-Based Distributed Power Systems

Abstract: Small-signal stability in balanced three-phase systems is typically investigated by means of the generalized Nyquist stability criterion (GNC) that involves operations on the source and load subsystems’ impedance matrices. Eigenvalues of the ratio of these impedance matrices should be calculated for stability judgment. This paper shows that for a power-electronics-based distributed power system, impedance matrices of subsystems can be designed as diagonal dominant. Therefore, stability, in this case, is fully determined by the scalar ratios formed by the impedance elements seen across the ${d}$ -axis and ${q}$ -axis interfaces. A three-phase ac system can then be treated as two decoupled dc systems. As a result, the ac stability analysis can also be conducted based on insightful impedance quantities; impedance specification criteria developed for dc systems can be readily applied to ensure stability in ac systems. The major contribution of this paper is related to the simplification of the GNC analysis with good experimental validations. Limitation of the proposed method is that the systems have to work under high power factor condition. However, such limitation is not a problem for many power-electronics-based distributed power systems.

Calculation of levelized costs of electricity for various electrical energy storage systems

Abstract: Installed capacity of renewable energy resources has increased dramatically in recent years, particularly for wind and photovoltaic solar Concurrently, the costs of utility-scale electrical energy storage options have been decreasing, making inevitable a crossing point at which it will become economically viable to couple renewable energy generation with utility-scale storage systems This paper proposes a methodology for calculating Levelized Cost of Electricity (LCOE) for utility-scale storage systems, with the intent of providing engineers, financiers and policy makers the means by which to evaluate disparate storage systems using a common economic metric We discuss the variables influencing LCOE in detail, particularly those pertinent to electrical energy storage systems We present results of LCOE calculations for various storage systems, specifically pumped hydro, compressed air, and chemical batteries, which we then compare with a more traditional arbitrage option, the simple-cycle combustion turbine Federal and State government electrical energy storage tax incentives are considered as well We also analyze the sensitivities of LCOE to several key variables using Monte Carlo analysis Considering the downward-sloping cost trends of storage systems and the increased penetration levels of stochastic and non-dispatchable renewable resources, large-scale storage is becoming a significant issue for utilities, thus justifying the development of a levelized costing algorithm

Understanding the Mechanisms of Deep Transfer Learning for Medical Images

Abstract: The ability to automatically learn task specific feature representations has led to a huge success of deep learning methods. When large training data is scarce, such as in medical imaging problems, transfer learning has been very effective. In this paper, we systematically investigate the process of transferring a Convolutional Neural Network, trained on ImageNet images to perform image classification, to kidney detection problem in ultrasound images. We study how the detection performance depends on the extent of transfer. We show that a transferred and tuned CNN can outperform a state-of-the-art feature engineered pipeline and a hybridization of these two techniques achieves 20\% higher performance. We also investigate how the evolution of intermediate response images from our network. Finally, we compare these responses to state-of-the-art image processing filters in order to gain greater insight into how transfer learning is able to effectively manage widely varying imaging regimes.

Metal artifacts in computed tomography for radiation therapy planning: dosimetric effects and impact of metal artifact reduction

Abstract: A significant and increasing number of patients receiving radiation therapy present with metal objects close to, or even within, the treatment area, resulting in artifacts in computed tomography (CT) imaging, which is the most commonly used imaging method for treatment planning in radiation therapy. In the presence of metal implants, such as dental fillings in treatment of head-and-neck tumors, spinal stabilization implants in spinal or paraspinal treatment or hip replacements in prostate cancer treatments, the extreme photon absorption by the metal object leads to prominent image artifacts. Although current CT scanners include a series of correction steps for beam hardening, scattered radiation and noisy measurements, when metal implants exist within or close to the treatment area, these corrections do not suffice. CT metal artifacts affect negatively the treatment planning of radiation therapy either by causing difficulties to delineate the target volume or by reducing the dose calculation accuracy. Various metal artifact reduction (MAR) methods have been explored in terms of improvement of organ delineation and dose calculation in radiation therapy treatment planning, depending on the type of radiation treatment and location of the metal implant and treatment site. Including a brief description of the available CT MAR methods that have been applied in radiation therapy, this article attempts to provide a comprehensive review on the dosimetric effect of the presence of CT metal artifacts in treatment planning, as reported in the literature, and the potential improvement suggested by different MAR approaches. The impact of artifacts on the treatment planning and delivery accuracy is discussed in the context of different modalities, such as photon external beam, brachytherapy and particle therapy, as well as by type and location of metal implants.

Learning from Synthetic Data: Addressing Domain Shift for Semantic Segmentation

Abstract: Visual Domain Adaptation is a problem of immense importance in computer vision. Previous approaches showcase the inability of even deep neural networks to learn informative representations across domain shift. This problem is more severe for tasks where acquiring hand labeled data is extremely hard and tedious. In this work, we focus on adapting the representations learned by segmentation networks across synthetic and real domains. Contrary to previous approaches that use a simple adversarial objective or superpixel information to aid the process, we propose an approach based on Generative Adversarial Networks (GANs) that brings the embeddings closer in the learned feature space. To showcase the generality and scalability of our approach, we show that we can achieve state of the art results on two challenging scenarios of synthetic to real domain adaptation. Additional exploratory experiments show that our approach: (1) generalizes to unseen domains and (2) results in improved alignment of source and target distributions.

Adaptive RNN Tree for Large-Scale Human Action Recognition

Abstract: In this work, we present the RNN Tree (RNN-T), an adaptive learning framework for skeleton based human action recognition. Our method categorizes action classes and uses multiple Recurrent Neural Networks (RNNs) in a treelike hierarchy. The RNNs in RNN-T are co-trained with the action category hierarchy, which determines the structure of RNN-T. Actions in skeletal representations are recognized via a hierarchical inference process, during which individual RNNs differentiate finer-grained action classes with increasing confidence. Inference in RNN-T ends when any RNN in the tree recognizes the action with high confidence, or a leaf node is reached. RNN-T effectively addresses two main challenges of large-scale action recognition: (i) able to distinguish fine-grained action classes that are intractable using a single network, and (ii) adaptive to new action classes by augmenting an existing model. We demonstrate the effectiveness of RNN-T/ACH method and compare it with the state-of-the-art methods on a large-scale dataset and several existing benchmarks.

Optimized multiplex immunofluorescence single-cell analysis reveals tuft cell heterogeneity

Abstract: Intestinal tuft cells are a rare, poorly understood cell type recently shown to be a critical mediator of type 2 immune response to helminth infection Here, we present advances in segmentation algorithms and analytical tools for multiplex immunofluorescence (MxIF), a platform that enables iterative staining of over 60 antibodies on a single tissue section These refinements have enabled a comprehensive analysis of tuft cell number, distribution, and protein expression profiles as a function of anatomical location and physiological perturbations Based solely on DCLK1 immunoreactivity, tuft cell numbers were similar throughout the mouse small intestine and colon However, multiple subsets of tuft cells were uncovered when protein coexpression signatures were examined, including two new intestinal tuft cell markers, Hopx and EGFR phosphotyrosine 1068 Furthermore, we identified dynamic changes in tuft cell number, composition, and protein expression associated with fasting and refeeding and after introduction of microbiota to germ-free mice These studies provide a foundational framework for future studies of intestinal tuft cell regulation and demonstrate the utility of our improved MxIF computational methods and workflow for understanding cellular heterogeneity in complex tissues in normal and disease states

Conductors for commercial MRI magnets beyond NbTi: requirements and challenges

Abstract: Magnetic Resonance Imaging (MRI), a powerful medical diagnostic tool, is the largest commercial application of superconductivity. The superconducting magnet is the largest and most expensive component of an MRI system. The magnet configuration is determined by competing requirements including optimized functional performance, patient comfort, ease of siting in a hospital environment, minimum acquisition and lifecycle cost including service. In this paper, we analyze conductor requirements for commercial MRI magnets beyond traditional NbTi conductors, while avoiding links to a particular magnet configuration or design decisions. Potential conductor candidates include MgB2, ReBCO and BSCCO options. The analysis shows that no MRI-ready non-NbTi conductor is commercially available at the moment. For some conductors, MRI specifications will be difficult to achieve in principle. For others, cost is a key barrier. In some cases, the prospects for developing an MRI-ready conductor are more favorable, but significant developments are still needed. The key needs include the development of, or significant improvements in: (a) conductors specifically designed for MRI applications, with form-fit-and-function readily integratable into the present MRI magnet technology with minimum modifications. Preferably, similar conductors should be available from multiple vendors; (b) conductors with improved quench characteristics, i.e. the ability to carry significant current without damage while in the resistive state; (c) insulation which is compatible with manufacturing and refrigeration technologies; (d) dramatic increases in production and long-length quality control, including large-volume conductor manufacturing technology. In-situ MgB2 is, perhaps, the closest to meeting commercial and technical requirements to become suitable for commercial MRI. Conductor technology is an important, but not the only, issue in introduction of HTS / MgB2 conductor into commercial MRI magnets. These new conductors, even when they meet the above requirements, will likely require numerous modifications and developments in the associated magnet technology.

Materials and Corrosion Trends in Offshore and Subsea Oil and Gas Production

Abstract: The ever-growing energy demand requires the exploration and the safe, profitable exploitation of unconventional reserves. The extreme environments of some of these unique prospects challenge the boundaries of traditional engineering alloys, as well as our understanding of the underlying degradation mechanisms that could lead to a failure. Despite their complexity, high-pressure and high-temperature, deep and ultra-deep, pre-salt, and Arctic reservoirs represent the most important source of innovation regarding materials technology, design methodologies, and corrosion control strategies. This paper provides an overview of trends in materials and corrosion research and development, with focus on subsea production but applicable to the entire industry. Emphasis is given to environmentally assisted cracking of high strength alloys and advanced characterization techniques based on in situ electrochemical nanoindentation and cantilever bending testing for the study of microstructure-environment interactions.

Clearance of Gadolinium from the Brain with No Pathologic Effect after Repeated Administration of Gadodiamide in Healthy Rats: An Analytical and Histologic Study

Abstract: Purpose To measure the levels of gadolinium present in the rat brain 1 and 20 weeks after dosing with contrast agent and to determine if there are any histopathologic sequelae. Materials and Methods The study was approved by the GE Global Research Center Institutional Animal Care and Use Committee. Absolute gadolinium levels were quantified in the blood and brains of rats 1 week after dosing and 20 weeks after dosing with up to 20 repeat doses of gadodiamide (cumulative dose, 12 mmol per kilogram of body weight) by using inductively coupled plasma-mass spectrometry. Treatment groups (n = 6 rats per group) included low-dosage and high-dosage gadodiamide and osmolality-matched saline controls. Brain sections were submitted (blinded) for standard toxicology assessment per Registry of Industrial Toxicology Animal data guidelines. Analysis of variance and Mann-Whitney U tests with post hoc correction were used to assess differences in absolute gadolinium levels and percentage of injected dose, respectively. Results Dose-dependent low levels of gadolinium were detected in the brain, a mean ± standard deviation of 2.49 nmol per gram of brain tissue ± 0.30 or 0.00019% of the injected dose 1 week after dosing. This diminished by approximately 50% (to 1.38 nmol per gram of brain tissue ± 0.10 or 0.00011% of the injected dose) 20 weeks after dosing. As a percentage of injected dose, the levels of gadolinium measured were comparable between different doses, indicating that mechanisms of uptake and elimination were not saturated at the tested doses. There were no histopathologic findings associated with the levels of gadolinium measured. Conclusion Low levels of gadolinium are present in the brain after repeat dosing with gadodiamide, which is partially cleared over 20 weeks with no detectable neurotoxicity.

A SiC Power MOSFET Loss Model Suitable for High-Frequency Applications

Abstract: The reduced chip size and unipolar current conduction mechanism make silicon carbide (SiC) metal–oxide–semiconductor field-effect transistors (MOSFETs) suitable for high-frequency power electronics applications. Modeling the switching process of the SiC power MOSFET with parasitic components is important for achieving higher efficiency and power density system design. Therefore, this paper proposes a new concise yet accurate switching loss model for SiC power MOSFETs. Addressing the limitations in experimental measurements, numerical simulations are conducted to validate the proposed model taking the output capacitance C oss discharge and charge into consideration. The role of the parasitic components in the second-order model is discussed in depth for switching losses. Furthermore, this paper also provides guidelines in designing the gate driver for ultrafast SiC power MOSFETs.

UA-DETRAC 2017: Report of AVSS2017 & IWT4S Challenge on Advanced Traffic Monitoring

Abstract: The rapid advances of transportation infrastructure have led to a dramatic increase in the demand for smart systems capable of monitoring traffic and street safety. Fundamental to these applications are a community-based evaluation platform and benchmark for object detection and multi-object tracking. To this end, we organize the AVSS2017 Challenge on Advanced Traffic Monitoring, in conjunction with the International Workshop on Traffic and Street Surveillance for Safety and Security (IWT4S), to evaluate the state-of-the-art object detection and multi-object tracking algorithms in the relevance of traffic surveillance. Submitted algorithms are evaluated using the large-scale UA-DETRAC benchmark and evaluation protocol. The benchmark, the evaluation toolkit and the algorithm performance are publicly available from the website http://detrac-db.rit.albany.edu.

Multimaterial Decomposition Algorithm for the Quantification of Liver Fat Content by Using Fast-Kilovolt-Peak Switching Dual-Energy CT: Clinical Evaluation

Abstract: Purpose To assess the ability of fast-kilovolt-peak switching dual-energy computed tomography (CT) by using the multimaterial decomposition (MMD) algorithm to quantify liver fat. Materials and Methods Fifteen syringes that contained various proportions of swine liver obtained from an abattoir, lard in food products, and iron (saccharated ferric oxide) were prepared. Approval of this study by the animal care and use committee was not required. Solid cylindrical phantoms that consisted of a polyurethane epoxy resin 20 and 30 cm in diameter that held the syringes were scanned with dual- and single-energy 64-section multidetector CT. CT attenuation on single-energy CT images (in Hounsfield units) and MMD-derived fat volume fraction (FVF; dual-energy CT FVF) were obtained for each syringe, as were magnetic resonance (MR) spectroscopy measurements by using a 1.5-T imager (fat fraction [FF] of MR spectroscopy). Reference values of FVF (FVFref) were determined by using the Soxhlet method. Iron concentrations were determined by inductively coupled plasma optical emission spectroscopy and divided into three ranges (0 mg per 100 g, 48.1-55.9 mg per 100 g, and 92.6-103.0 mg per 100 g). Statistical analysis included Spearman rank correlation and analysis of covariance. Results Both dual-energy CT FVF (ρ = 0.97; P < .001) and CT attenuation on single-energy CT images (ρ = -0.97; P < .001) correlated significantly with FVFref for phantoms without iron. Phantom size had a significant effect on dual-energy CT FVF after controlling for FVFref (P < .001). The regression slopes for CT attenuation on single-energy CT images in 20- and 30-cm-diameter phantoms differed significantly (P = .015). In sections with higher iron concentrations, the linear coefficients of dual-energy CT FVF decreased and those of MR spectroscopy FF increased (P < .001). Conclusion Dual-energy CT FVF allows for direct quantification of fat content in units of volume percent. Dual-energy CT FVF was larger in 30-cm than in 20-cm phantoms, though the effect of object size on fat estimation was less than that of CT attenuation on single-energy CT images. In the presence of iron, dual-energy CT FVF led to underestimateion of FVFref to a lesser degree than FF of MR spectroscopy led to overestimation of FVFref. © RSNA, 2016 Online supplemental material is available for this article.

Toward high value sensing: monolayer-protected metal nanoparticles in multivariable gas and vapor sensors.

Abstract: For detection of gases and vapors in complex backgrounds, “classic” analytical instruments are an unavoidable alternative to existing sensors. Recently a new generation of sensors, known as multivariable sensors, emerged with a fundamentally different perspective for sensing to eliminate limitations of existing sensors. In multivariable sensors, a sensing material is designed to have diverse responses to different gases and vapors and is coupled to a multivariable transducer that provides independent outputs to recognize these diverse responses. Data analytics tools provide rejection of interferences and multi-analyte quantitation. This review critically analyses advances of multivariable sensors based on ligand-functionalized metal nanoparticles also known as monolayer-protected nanoparticles (MPNs). These MPN sensing materials distinctively stand out from other sensing materials for multivariable sensors due to their diversity of gas- and vapor-response mechanisms as provided by organic and biological ligands, applicability of these sensing materials for broad classes of gas-phase compounds such as condensable vapors and non-condensable gases, and for several principles of signal transduction in multivariable sensors that result in non-resonant and resonant electrical sensors as well as material- and structure-based photonic sensors. Such features should allow MPN multivariable sensors to be an attractive high value addition to existing analytical instrumentation.