Search

Due to melanoma is one of the skin cancers with the highest mortality rate and have a large amount of data during the collection and diagnosis, there is an urgent need to improve the diagnostic efficiency and accuracy. However, there remain problems in analyzing medical big data for skin lesion application, such as the intra-class variation and inter-class similarity in skin lesion images and the lacks of ability to focus on the lesion area affecting the classification results of the model. To address these dilemmas, in this paper, we proposed a novel machine learning-based approach that builds on top of DenseNet. It combines the attention mechanism and large margin loss to enhance the classification accuracy in terms of intra-class compactness and inter-class separability. We evaluated our model on ISIC 2017 (International Skin Imaging Collaboration) dataset, which has achieved 92% of Mean AUC. The experimental results show the effectiveness of our solution outperforms the state-of-the-art significantly in classify skin lesion and can accurately classify malignant melanoma on medical images.

Ceramics have been widely used in the field of tribology due to their superior corrosion resistance and low friction coefficient. The tribological properties of ceramics can be improved by machining proper surface texture. This work uses a nanosecond laser to process dimples on different ceramic surfaces. The morphology and size of dimples as a function of the number of laser shots and laser power are investigated. A theoretical model relating laser parameters and dimple size is established. In the initial point contact friction test, based on the experimental parameters mentioned in this paper, the existence of surface texture reduces the wear volume of Si3N4 balls in water by 41.5% and the wear volume of ZrO2 balls in oil by 51%.

Li-rich layered cathodes based on Li2MnO3 have exhibited extraordinary promise to satisfy the rapidly increasing demand for high-energy density Li-ion batteries. However, besides cationic redox, oxygen anionic redox contributing to extra charge transfer often suffers from local structural transformation through O–O bond formation, leading to partial irreversibility during discharge. Four typical transition metal (TM) substitution models Li2Mn0.75TM0.25O3 (TM = Nb, Mo, Ru and Rh) are selected to calculate reversible capacity and perform electronic structure analysis. Our calculations reveal that the inherent t2g orbital of TMs around the Fermi level plays a key role in stabilizing the anionic lattice and operating cationic redox, which further favors anionic redox reversibility. An electron occupancy number Nt2g = 4 of the t2g orbital in substitutional TM cations is identified as the optimal value to achieve a maximum reversible capacity from cationic and anionic redox. Experimentally, Li2Mn0.75Ru0.25O3 exhibits improved reversible capacity by 18.1% at 0.1C compared to Li2MnO3, confirming the suppression effect of Ru on the oxidation of lattice oxygen. The present study provides a new insight into developing high-capacity layered electrodes with good reversibility.

Methane catalytic abilities are considered to be an important and direct form of experimental evidence for evaluating the efficacy of mixed oxide-ion/electron conductors (MIEC) as anode materials for solid oxide fuel cells (SOFCs) with CH4 as the fuel gas. Therefore, a direct link between the quantifiable methane-conversion abilities and CH4-based SOFC electrochemical performance is urgently needed. Here, H2-reduced Sr2ZnMoO6 (R-SZMO) as an anode example exhibits excellent catalytic abilities and tolerance to coking when using CH4 as the fuel gas, which is dominated by the formation of intrinsic anion-Frenkel defect pairs. DFT calculations reveal that the efficient electron transfer process and optimal selectivity towards CH4 conversion are both on the R-SZMO (110) surface. Therefore, applying high CH4 activity R-SZMO as the anode catalyst gives excellent anti-carbon deposition performance, high power output, and long-term stability for SOFCs with CH4 as the fuel gas.

Anionic oxygen redox chemistry in Li-excess transition metal oxides has emerged as a new paradigm to increase the energy density of rechargeable batteries. However, the anionic redox of oxygen mostly forms the O–O bond, leading to partly irreversible discharge. To address this issue, we demonstrate a new strategy to achieve dual functions of activating cation activity and enhancing oxygen anion redox activity in Li2MnO3 through a light metal (Mg, Ca and Al) substitution of Li. Our first-principle thermodynamic calculations show that Mg2+ substitution in Li2MnO3 prefers locating at the Li+ site, which induces the formation of reductive Mn-ions with electrochemical activity. The substituted local structure of (Li–O–Mg)/2(Li–O–Mn) becomes more electrochemically stable than the specific Li–O–Li local structure in Li2MnO3, which effectively retards O–O combination. Therefore, the reversible capacity (245 mA h g−1) of the optimal substituent compound Li1.5Mg0.5MnO3 is significantly improved compared with that (115 mA h g−1) of Li2MnO3. More importantly, this substitution strategy can be effectively extended to other Li-rich electrodes such as Li3VO4 and Li3NbO4 with Mg-substitution activating cationic redox. The study opens a new avenue in improving the reversible capacity of Li-rich cathode materials through low-cost element substitution.

The disordered Li-rich oxyfluoride cathodes not only offer potentials for high energy density through cooperative cationic and anionic redox but also for facile lithium diffusion by formatting activated percolating channels. Generally, the d0 transition metals (TM) are considered as an essential facilitator for stabilization of the cation-disordered phase, but no redox activity. Nevertheless, their influence on reversible cationic and anionic redox is not fully understood, and consequently preventing further material design. Here, our studies for typical material Li2Mn1/2Ti1/2O2F unequivocally demonstrate that the local configurations containing redox-inactive d0 TM make more contribution to reversible capacity than those containing electrochemically active Mn2+. The peculiar effect is attributed to forming low-energy charge-ordering state that d0 TM prefers staying with Li–O–Li and pushes down O 2p unbonding state into a low-energy level. The present study reveals that a disordered oxyfluoride cathode design achieving a high reversible capacity may necessarily consider a suitable ratio of active low-charge TM to high-charge d0 TM.

Data for face analysis often exhibit highly-skewed class distribution, i.e., most data belong to a few majority classes, while the minority classes only contain a scarce amount of instances. To mitigate this issue, contemporary deep learning methods typically follow classic strategies such as class re-sampling or cost-sensitive training. In this paper, we conduct extensive and systematic experiments to validate the effectiveness of these classic schemes for representation learning on class-imbalanced data. We further demonstrate that more discriminative deep representation can be learned by enforcing a deep network to maintain inter-cluster margins both within and between classes. This tight constraint effectively reduces the class imbalance inherent in the local data neighborhood, thus carving much more balanced class boundaries locally. We show that it is easy to deploy angular margins between the cluster distributions on a hypersphere manifold. Such learned Cluster-based Large Margin Local Embedding (CLMLE), when combined with a simple k-nearest cluster algorithm, shows significant improvements in accuracy over existing methods on both face recognition and face attribute prediction tasks that exhibit imbalanced class distribution., Comment: 14 pages, 10 figures, 8 tables. Accepted to TPAMI

A colorimetric and fluorometric dual probe based on a water-soluble polythiophene derivative (PMTPBA) was designed and synthesized. It can be applied to determination of picric acid (PA) in 100% aqueous solution. The approach relies on the formation of supramolecular polythiophene assemblies in the presence of PA through electrostatic, charge transfer and π-π stacking interactions. This probe could be utilized for the rapid and visual detection of PA both in aqueous solution and solid support with high specificity and sensitivity. The detection limit of this sensor is as low as 5.0 × 10-8 mol/L.

The incorporation of point defects such as cationic vacancies into electrode materials has been considered as an effective strategy to improve the charge-transfer and ion-diffusion kinetics and allow insertion and migration of multivalent ions. However, they suffer from low specific capacity and electrochemical irreversibility. To elucidate the origin of these issues, we investigated the F-doped and cation-deficient anatase TiO2 through Mg2+ insertion for understanding the redox activity of vacancy structure. Our first-principles calculations showed that charge transfer inductively occurs from anion O2−/F− to cation Ti4+ near vacancies, forming oxidized anions F(1−x)−/O(2−y)− and reduced cations Ti(4−z)+. We further found that cooperative cationic and anionic redox reactions, Ti3.83+ + 0.19e− → Ti3.64+, O1.94− + 0.06e− → O2− and F0.93− + 0.07e− → F−, take place during Mg2+ insertion. The peculiar anionic redox reaction of oxidized F 2p states is attributed to low-coordination fluorine ions, which was demonstrated by the previous NMR characterization. Our calculations showed that the fluorine redox reaction contributes 26.5% of the total redox capacity. The present results provided chemical clues to use the vacancy structure design to develop efficient cationic and anionic redox materials for improving the energy density and cyclic stability of battery materials.

Despite much progress achieved in the past decades in the process of advancing the low-temperature sintering technologies for Solid oxide fuel cells (SOFCs), such as via the structure design of the electrode materials, the practical application of low-temperature sintered SOFCs (with disqualified mechanical strength) remains challenging. In this work, first, we demonstrate that the appropriate amount of CuO as sintering aids can successfully reduce the co-firing temperature of conventional micron size NiO-YSZ (yttrium-stabilized zirconia (Y2O3)0.08–(ZrO2)0.92) anode from about 1400 °C to only 1100 °C. Second, the quantitative structure-activity relationship among the mechanical strength (low-temperature sintering ability) of anode cermets with the inclusion of CuO contents and the densification of YSZ electrolyte was synthetically evaluated, and the optimal Cu–NiO-YSZ anode composition demonstrates almost the equal mechanical strength when compared with the traditional NiO-YSZ anode (sintering at 1400 °C). At last, by comprehensive assessment, 8%Cu–52NiO-40YSZ (8%CuO–NiO-YSZ) shows excellent low-temperature sintering ability, high mechanical strength, optimal power output, and anti-carbon deposition when using as hydrocarbon-based anode for SOFCs.

Garnet-type ceramic electrolyte with exceptionally high Li+ conductivity has attracted wide interests for the development of safe electrochemical devices. However, the factors affecting the ionic migration still deserve more attentions. In this study, five garnet-types ceramic electrolyte were chosen to investigate the factors. The thermal behavior of garnet-type electrolyte during sintering was studied to achieve high conductivity. The electrochemical properties are correlated with crystal structure, migrating route and lattice parameters. Nine influening factors were discussed to analyze the ionic migration in garnet-type ceramic electrolyte.

Currently, most of the physics-based magnetization models for giant magnetostrictive actuator are complicated in their mathematical models. These complicated models pose considerable challenges for real-time control. Focus on the above issue, we proposed a fitting magnetization model using Arctangent function, as the inverse of the Arctangent function can solve it very quickly. This function was suitable to model the magnetic behavior of giant magnetostrictive actuator serving in an electronic controlled injector or other types of on-off valves. Applicable range of the model was determined through computing the maximum magnetic field. Then the deviations of the proposed fitting model and Jiles-Atherton model under different parameters were studied to verify the computing effects of the model. Then the effectiveness of the fitting model was verified by experimental results. With a high computing precision and a concise form, the proposed model shows great potentials for the real-time control of giant magnetostrictive actuators.

A simple and sensitive electrochemical immunoassay method was developed for the quantitative monitoring of lipoprotein-associated phospholipase A2 (Lp-PLA2) on a low-cost microtiter plate by using a portable handheld pH meter. Bovine serum albumin (BSA)-doped CaCO3 nanospheres were utilized to regulate the pH value of an acidic solution on the basis of classical chemical reactions. To construct such an immunosensing system, the synthesized BSA-CaCO3 nanospheres were initially conjugated covalently to an anti-Lp-PLA2 detection antibody, and then a sandwich-type immunoreaction was carried out on the anti-Lp-PLA2 capture antibody-coated microplate in the presence of the target analyte. Accompanying the sandwiched immunocomplexes, the labeled CaCO3 nanospheres were dissolved in the presence of acid (CaCO3 + 2H+ → Ca2+ + CO2 + H2O), thus resulting in the pH change of the detection solution. Introduction of CaCO3 nanospheres including numerous carbonate ions was expected to enhance the sensitivity of the electrochemical immunoassay. Under optimum conditions, BSA-CaCO3-based immunoassay displayed good pH responses for the determination of target Lp-PLA2 within the dynamic linear range from 0.1 ng mL−1 to 300 ng mL−1. The limit of experimental detection achieved was ∼78 pg mL−1 Lp-PLA2 with good coefficients of variation for the inter- and intra-assays. Also, this system gave good anti-interfering capacity toward other biomarkers and enzyme proteins. Importantly, our strategy was applied for the analysis of human serum specimens with satisfactory results in comparison with commercialized Lp-PLA2 ELISA kits.

A simple and portable potentiometric immunoassay was designed for the quantitative detection of matrix metalloproteinase-7 (MMP7) in acute kidney injury and renal cancer on a copper ion-selective electrode (Cu-ISE). To construct such an assay system, a sandwich-type immunoreaction was initially carried out on monoclonal anti-MMP7 capture antibody-coated microplates by using an initiator DNA strand-labeled anti-MMP7 secondary antibody. Thereafter, a hybridization chain reaction (HCR) was used for the signal amplification between copper nanocluster-labeled two hairpin DNA probes. Accompanying the formation of the sandwiched immunocomplex, the labeled DNA initiator strand on the secondary antibody triggered the HCR between two alternating DNA hairpins to form a nicked double-helix with copper nanoclusters. The subsequent potentiometric measurement of copper ions released from the concatenated copper nanoclusters under acidic conditions was conducted on a portable handheld potentiometer. The immuno-HCR assay combined an easy-operation potentiometric assay with signal amplification by using an HCR-induced copper nanocluster concatemer. Under optimum conditions, the developed immuno-HCR assay exhibited a dynamic linear range of 0.01–100 ng mL−1 with a detection limit of 5.3 pg mL−1 and good reproducibility. Also, the immuno-HCR assay gave good anti-interfering capacity toward other biomarkers. The analysis for clinical human serum specimens revealed good accordance with the results obtained by an enzyme-linked immunosorbent assay. Importantly, our strategy is promising for enzyme-free and cost-effective analysis of low-abundance cancer biomarkers.

Nuclei segmentation occupies an important position in medical image analysis, which helps to predict and diagnose diseases. With the further research of deep learning, the task of nuclei segmentation has been automated. However, most existing methods require a great deal of manually marked full masks for training, which is time-consuming and labor-intensive, and can only be done by professional personnel. For the purpose of reducing the cost of labeling, we propose a weakly supervised method using generative adversarial training for segmentation of nucleus. In the case of no boundary, but only the centroid of the nucleus, the proposed method segmented the nucleus region with blurred boundaries. We first use the generative adversarial network(GAN) to generate the likelihood map of the nuclear centroid, then use Guided Backpropagation to visualize the pixels that contributes to the detection of the centroid of each nucleus, and finally obtain the segmentation mask of the nucleus by graph-cut. In addition, for the purpose of training the network better, we performed stain normalization on each pathological image. We have verified the proposed method on a multi-organ nuclei dataset. The final experiment results show that our advanced method achieves better segmentation performance than other weakly supervised methods, and can even reach the level of full supervision.

Background: Bone grafts are in high demand due to the increase in the cases of bone defects mainly caused by trauma, old age, and disease-related bone damages. Tissue-engineered calcium phosphate (CaP) biomaterials match the major inorganic contents of bone, thereby could be the potential bone graft substitute. However, CaP-bone grafts lack the osteoinductivity that is vital for effective bone regeneration. In this study, we aimed to test the bone defect healing potential of biomimetically fabricated low dose BMP2-doped CaP (BMP2.BioCaP) grafts in a large animal model.Methods: Low dose BMP2 was doped internally (BMP2-int.BioCaP) or on the surface of CaP (BMP2-sur.BioCaP) grafts during the fabrication process. Our previous study showed the robust bone regenerative potential of BMP2-int.BioCaP and BMP2-sur.BioCaP grafts in the rat ectopic model. In this study, we investigated the bone defect healing potential of BMP2.BioCaP grafts in sheep humerus/femoral defects, as well as compared with that of autologous bone graft and clinically used deproteinized bovine bone (DBB) xenograft.Results: Different ways of BMP2 doping did not affect the surface morphology and degradation properties of the graft materials. Micro-CT and histology results showed robustly higher bone defect-healing potential of the BMP2.BioCaP grafts compared to clinically used DBB grafts. The bone defect healing potential of BMP2.BioCaP grafts was as effective as that of the autologous bone graft. Although, BMP2-int.BioCaP doped half the amount of BMP2 compared to BMP2-sur.BioCaP, its' bone defect healing potential was even robust. The BMP2.BioCaP grafts showed less immunogenicity compared to BioCaP or DBB grafts. The volume density of blood vessel-like and bone marrow-like structures in both BMP2.BioCaP graft groups were in a similar extent to the autologous group. Meticulous observation of higher magnification histological images showed active bone regeneration and remodeling during bone defect healing in BMP2.BioCaP graft groups.Conclusion: The robust bone regenerative potential of BMP2.BioCaP grafts in the ectopic model and in-situ bone defects in small and large animals warrant the pre-clinical studies on large animal critical-sized segmental bone defects.

In this report, we have established a novel approach for the rapid detection of folic acid (FA) with a water soluble perylene bisimide-based probe in 100% aqueous media. The sensing performance and mechanism of this probe were investigated systematically using spectroscopic techniques. After optimizing conditions, the limit of detection in this assay is as low as 29.2 nmol/L and linear operation range is from 0.1 to 1.2 μmol/L. It relies on the supramolecular recognition between the probe and FA under the synergistic effects of non-covalent interactions such as electrostatic, hydrophobic and π−stacking interactions. Monitoring the FA concentrations in real-world samples is also confirmed.

Image understanding is becoming a vital feature in ever more applications ranging from medical diagnostics to autonomous vehicles. Many applications demand for embedded solutions that integrate into existing systems with power constraints and tight real-time. Convolutional Neural Networks (CNNs) presently achieve record-breaking accuracies in all image understanding benchmarks, but have a very high computational complexity. Modern high-end FPGA generations feature hundreds of thousands of configurable logic blocks, and additionally include an abundance of hardened functional units which enable fast and efficient implementations of common functions. Many researchers have proposed their CNN accelerator prototypes on FPGA. But one problem of the stateof-the-art designs is that they have not solved the data dependence problem well. Data dependency is an important factor affecting accelerator performance. Current designs solve data dependence problem by adding hardware module on FPGA. But this approach has little effect and leads to increased hardware complexity. In this paper, we propose an optimization on the data arrangement in CNN. Which solves the data dependence in CNN by rearranging the data. The rearranged data is stored in a hardware-friendly form. By this way, our accelerator can apply pipeline technology better than current designs. We validate our approach on Xilinx Zynq XC-7Z045 board. The experimental results show that our approach has obvious advantages in terms of hardware resource consumption and bandwidth compare to state-of-the-art designs.

Purpose In real working condition, signal is highly disturbed and even drowned by noise, which extremely interferes in detecting results. Therefore, this paper aims to provide an effective de-noising method for the debris particle in lubricant so that the ultrasonic technique can be applied to the online debris particle detection. Design/methodology/approach For completing the online ultrasonic monitoring of oil wear debris, the research is made on some selected wear debris signals. It applies morphology component analysis (MCA) theory to de-noise signals. To overcome the potential weakness of MCA threshold process, it proposes fuzzy morphology component analysis (FMCA) by fuzzy threshold function. Findings According to simulated and experimental results, it eliminates most of the wear debris signal noises by using FMCA through the signal comparison. According to the comparison of simulation evaluation index, it has highest signal noise ratio, smallest root mean square error and largest similarity factor. Research limitations/implications The rapid movement of the debris particles, as well as the lubricant temperature, may influence the measuring signals. Researchers are encouraged to solve these problems further. Practical implications This paper includes implications for the improvement in the online debris detection and the development of the ultrasonic technique applied in online debris detection. Originality value This paper provides a promising way of applying the MCA theory to de-noise signals. To avoid the potential weakness of the MCA threshold process, it proposes FMCA through fuzzy threshold function. The FMCA method has great obvious advantage in de-noising wear debris signals. It lays the foundation for online ultrasonic monitoring of lubrication wear debris.

Giant magnetostrictive material is a kind of excellent engineering material for its large magnetostriction, fast response speed and high energy density. And giant magnetostrictive injector which employs giant magnetostrictive material driving an electronic controlled injector may be one new promising injector. With more adjustable injection rate, giant magnetostrictive injector can promote energy utilization and reduce pollution emission. This paper provides a detailed overview of the research status of the giant magnetostrictive injector. Some important background information as the properties of giant magnetostrictive material, designing methods of giant magnetostrictive actuator and working circumstances of an electronic controlled injector are introduced firstly. Then the structure design, physical model, numerical simulation and power driver design for the giant magnetostrictive injector are reviewed especially. The summaries may support much useful information for the researchers of giant magnetostrictive injector. Furthermore, some general techniques like bias field utilizing skill, fast driving and simple modeling methods etc. are available to more research fields beyond the limitations of giant magnetostrictive injector.

Recent studies suggested that phytochemicals as natural antioxidants in food could alleviate nanoparticle (NP) toxicity. This study investigated the combined toxicity of ZnO NPs and a panel of polyphenols. Surprisingly, polyphenols with both high and almost no radical scavenging activities could elicit cytoprotective effects against NP exposure in Caco-2 cells, which were primarily influenced by the positions of the hydroxyl group. Polyphenols with different chemical structures variously influenced the hydrodynamic size, zeta potential, and solubility of ZnO NPs as well as NP-induced intracellular superoxide and Zn ions, which could all contribute to the combined effects. Responses of human endothelial cells appeared to be different from the responses of Caco-2 cells, which may indicate cell-type dependent responses to combined exposure of NPs and phytochemicals. In conclusion, the data from this study suggested a pivotal role of chemical structures of phytochemicals in determining their capacity to affect ZnO NP toxicity.

The study of cathode-side poisoning in solid-oxide fuel cells due to Cr, H2O, and CO2, etc., has attracted wide interest from the viewpoint of marketability. However, overall perspectives of poisoning/protection of the entire cathode side and multivariate analyses of the cathode side are rare and deserve more attention. The multiple factors, such as contamination, the cathode contact material (CCM) layer, the protective layer, poisoning, and the cathode lifetime that govern the overall performance of the cathode side should be considered comprehensively/integrally rather than solely (as single parameters). In this work, with experimental results as proof, a systematic study of the quantifiable correlation between cathode-side poisoning/protection and contaminations has been performed, and the novelty and significant contributions are listed here: (1) with objective and consistent external factors, a systematic comparison of the most popular CCM layer and protective layer was carried out and a quantifiable correlation versus cathode-side poisoning and protection was demonstrated; (2) a systematic comparative study of the cathode-side multifunctional system by quantification of cathode-side poisoning with ferritic steels as interconnects at different contaminant contents; (3) a clarification of the distribution of various contaminants in the cathode-side 3D configuration and a demonstration of valid solutions for slowing down and even preventing poisoning by chromium, H2O and CO2; (4) a summary of the individual or multidimensional synergistic mechanism of the multi-tiered CCM/coating structure that acts as an accessory device at the cathode side; (5) optimization of the CCM appraisement system, making a reassessment of existing CCM candidates and proving their additional benefit, for instance, as a Cr vapor getter. Finally, the area specific resistance (ASR) of cell specimens with a multipurpose protective/connective structure was measured for 4000 h at different levels of contaminants to determine the long-term stability of the cathode. The degradation rates of Rp and Rohm values of the cathode drastically decreased from 484.3 and 99.2 to 11.2 and 4.3 μΩ cm−2 h−1 (from 484.3 and 99.2 (1000 h) to 44.8 and 17.3 mΩ cm−2 (4000 h)), which represented decline by factors of 43 and 23, respectively.

Surface texturing can change the friction and wear properties of the material surface. It shows outstanding performance in reducing friction coefficients and wear in conformal flat contact. However, during initial point contact, most studies have found that it reduces the friction coefficient and increases wear. In this paper, ball-on-disc sliding tests were conducted to verify whether surface texturing could reduce wear of silicon nitride sliding in water for tests starting with initial point contact. The effects of texture parameters (depth, textured area, and diameter) were studied. Results showed that most textures could reduce Si3N4 wear after running-in process. Multi-scale texture, which was combined with large (11 μm depth, 2% textured area, and 22 μm diameter) and small (5 μm depth, 1% textured area, and 15 μm diameter) textures, could reduce the wear by 93% and increase the bearing capacity by 244%. Comparative experiments of oil and water lubricants were carried out to explore the mechanism of texture reducing wear in the running-in process, and two competitive mechanisms of hydrodynamic effect and stress concentration were found. This work provides a practical foundation for optimizing texture design and reducing wear of ceramic materials under water lubrication.

Recent study suggested that the presence of phytochemicals in food could interact with nanoparticles (NPs) and consequently reduce the toxicity of NPs, which has been attributed to the antioxidant properties of phytochemicals. In this study, we investigated the interactions between ZnO NPs and two flavonoids baicalein (Ba) or baicalin (Bn) as well as the influence of the interactions on the toxicity of ZnO NPs to Caco-2 cells. The antioxidant properties of Ba and Bn were confirmed by 1,1-diphenyl-2-picrylhydrazyl (DPPH) and 2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) assays, with Ba being stronger. However, the presence of Ba or Bn did not significantly affect cytotoxicity, intracellular superoxide or release of inflammatory cytokines of Caco-2 cells after ZnO NP exposure. When Ba was present, the cellular viability of Caco-2 cells after exposure to ZnO NPs was slightly increased, associated with a modest decrease of intracellular Zn ions, but these effects were not statistically different. Ba was more effective than Bn at changing the hydrodynamic sizes, Zeta potential and UV-Vis spectra of ZnO NPs, which indicated that Ba might increase the colloidal stability of NPs. Taken together, the results of the present study indicated that the anti-oxidative phytochemical Ba might only modestly protected Caco-2 cells from the exposure to ZnO NPs associated with an insignificant reduction of the accumulation of intracellular Zn ions. These results also indicated that when assessing the combined effects of NPs and phytochemicals to cells lining gastrointestinal tract, it might be necessary to evaluate the changes of colloidal stability of NPs altered by phytochemicals.

To characterise the resilience of a control system in cyber-physical systems (CPSs) against actuator attacks and sensor attacks, we introduce new notions of controllability and observability to derive conditions under whether the state can be stabled and reconstructed by the existing control signal and measurements over a period of time. With the understanding that some of the available system's inputs and outputs may have been compromised by the adversary, we derive the necessary and sufficient conditions to utilise standard controllability and observability matrix tests to be calculated in checking whether a CPS is `controllable and observable under attacks', and verify the theorems by a simulation of 9-bus power grid system under attacks.

With the rapid development of nanotechnologies, nanoparticles (NPs) are increasingly produced and used in many commercial products, which could lead to the contact of human blood vessels with NPs. Thus, it is necessary to understand the adverse effects of NPs to relevant cells lining human blood vessels, especially endothelial cells (ECs) that cover the lumen of blood vessels. Human umbilical vein endothelial cells (HUVECs) are among one of the most popular models used for ECs in vitro. In the present review, we discussed studies that have used HUVECs as a model to investigate the EC-NP interactions, the toxic effects of NPs on ECs and the mechanisms. The results of these studies indicated that NPs could be internalized into HUVECs by the endocytosis pathway as well as transported across HUVECs by exocytosis and paracellular pathways. Exposure of HUVECs to NPs could induce cytotoxicity, genotoxicity, eNOS uncoupling and endothelial activation, which could be explained by NP-induced oxidative stress, inflammatory response and dysfunction of organelles. In addition, some studies have also evaluated the influences of microenvironment (e.g. the presence of proteins and excessive nutrients), the physiological and/or pathological stimuli related to the diversity of ECs (e.g. shear stress, cyclic stretch and inflammatory stimuli), and the physicochemical properties of NPs on the responses of ECs to NP exposure. In conclusion, it has been suggested that HUVECs could be considered as a relatively reliable and simple in vitro model for ECs to predict and evaluate the toxicity of NPs to endothelium. Copyright © 2017 John Wiley & Sons, Ltd.

Inorganic nanoparticles (NPs) are among the most produced NPs that could be used in consumer products and as healthcare materials, however, the intrinsic toxicity particularly through the mechanism associated oxidative stress raises the health concern about inorganic NP exposure. Phytochemicals are bioactive metabolites derived from plants as well as non-pathogenic microorganisms living within plants and have been shown to be beneficial to human health with their anti-aging, anti-cancer, anti-inflammation and anti-oxidant properties. In the present review, the influence of on the biocompatibility of inorganic NPs was discussed. It has been shown that phytochemicals could be used as bio-friendly capping agents for green synthesis of inorganic NPs, and phytochemical coated inorganic NPs were remarkable stable and biocompatible with high therapeutic efficiency. Meanwhile, the presence of phytochemicals was also able to reduce the side effects and enhance the therapeutic abilities of inorganic NPs, which is likely attributed to the anti-oxidative properties of phytochemicals. Thus, using phytochemicals could be a promising and plausible way to reduce side effects and increase the biocompatibility of inorganic NPs for biomedical applications.

Fast-convolution multicarrier (FCMC), the asynchronous waveform with ultra-low sidelobe, has appeared to be a promising waveform technique for future wireless communications. In this paper, we investigate the filter bank optimization as well as receiver design including low-complexity channel estimator and equalizer for FCMC based frequency division multiple access (FDMA). Starting from the conventional multi-carrier signals, we first derive a vectorized signal model as a framework to systematically design the FCMC transceiver. For nearly perfect reconstruction (NPR) filter banks design, an optimization criterion which consists of minimizing received signal segments mean square error (MSE) is proposed. From the fact that Toeplitz matrices can be asymptotically diagonalized by discrete Fourier transform (DFT) matrix, the channel equalizer can be simplified to one-tap frequency domain equalizer when DFT size is large enough. The minimum mean square error (MMSE) criterion is then applied to calculate the coefficients of one-tap frequency domain equalizer. In practice, due to the channel fading, the channel estimation has to be performed to obtain the channel state information (CSI) which is required by the channel equalization. To this end, we propose a combined cyclic prefix (CP) and cyclic suffix (CS) pilot structure which facilitates to estimate the frequency domain CSI directly in the receiver end. The proposed FCMC based FDMA features low-complexity receiver, adjustable users bandwidth and low peak-to-average power ratio. Simulation results confirm the performance of the proposed scheme.

An emergency control strategy of electric vehicle charging station (EVCS) is proposed to avoid transmission line overload, where the second order power sensitivity is considered. In this paper, EVCSs are considered to be flexible load, which can be controlled for enhancing the system security. Compared to the traditional sensitivity method, the second order sensitivity can capture the nonlinearity better, especially when the system is with heavy load. Case studies show that the proposed method can provide more effective results, when the system nonlinearity is strong.

We present a novel approach for the task of human pose transfer, which aims at synthesizing a new image of a person from an input image of that person and a target pose. We address the issues of limited correspondences identified between keypoints only and invisible pixels due to self-occlusion. Unlike existing methods, we propose to estimate dense and intrinsic 3D appearance flow to better guide the transfer of pixels between poses. In particular, we wish to generate the 3D flow from just the reference and target poses. Training a network for this purpose is non-trivial, especially when the annotations for 3D appearance flow are scarce by nature. We address this problem through a flow synthesis stage. This is achieved by fitting a 3D model to the given pose pair and project them back to the 2D plane to compute the dense appearance flow for training. The synthesized ground-truths are then used to train a feedforward network for efficient mapping from the input and target skeleton poses to the 3D appearance flow. With the appearance flow, we perform feature warping on the input image and generate a photorealistic image of the target pose. Extensive results on DeepFashion and Market-1501 datasets demonstrate the effectiveness of our approach over existing methods. Our code is available at http://mmlab.ie.cuhk.edu.hk/projects/pose-transfer, Comment: CVPR 2019

Numerous theoretical and experimental works about Ruddlesden-Popper (RP) based oxygen catalysts were based on the critical assumption: oxygen vacancies do exist in the perovskite-like layer, and which with the interstitial O2− (in rock-salt (AO) layers) were the two key influencing factors to determine RP’ oxygen exchange rate and oxygen catalytic ability. But overall, the presence or absence of critical oxygen vacancies in the perovskite-like layer of RP oxides is still unclear. Here, from our experimental results, no oxygen vacancy was detected in the perovskite-like layer of RP series materials. Herein, for the Ruddlesden-Popper based oxygen catalysts, we proposed that more interstitial O2− in the AO layers and high-active lattice oxygen ions in the perovskite-like layers are the determinant factors to activate the ORR and OER. Accordingly, we reveal that the high-valence ion doping strategy for RP electrode can enhances the redox reactions while improving the lattice-oxygen activity and interstitial O2− concentration, and then benefiting the ORR and OER activity.

Porous micro-/nanostructures are of great interest in many current and emerging areas of technology. In this paper, porous SnO 2 nanocubes have been successfully fabricated via a selective leaching strategy using CoSn(OH) 6 as precursor. The structure and morphology of as-prepared samples were investigated by several techniques, such as X-ray diffraction (XRD), scanning electron microscopy (SEM), thermogravimetric and differential scanning calorimeter analysis (TG⿿DSC), transmission electron microscopy (TEM) and N 2 adsorption⿿desorption analyses. On the basis of those characterizations, the mechanism for the formation of porous SnO 2 nanocubes has been proposed. Owing to the well-defined and uniform porous structures, porous SnO 2 nanocubes possessing more adsorbent amount of analytic gas and accelerate the transmission speed so as to enhance the gas-sensing properties. Gas sensing investigation showed that the sensor based on porous SnO 2 nanocubes exhibited high response, short response⿿recovery times and good selectivity to ethanol gas.

High-pressure Raman scattering and angle dispersive synchrotron x-ray diffraction studies have been carried out on magnesium tantalate (MgTa2O6) up to 41 GPa at room temperature. A pressure-induced phase transition at pressures above 10 GPa accompanied by softening of the internal ν11 (Eg) modes has been observed. Another phase transition took place at 20.2 GPa, which indicates that the edge shared TaO6 octahedra are distorted. The equation of state for MgTa2O6 in low pressure phase I was obtained by the second-order Birch–Murnaghan method, yielding a zero-pressure bulk modulus B0 of 205.8 ± 8.4 GPa. The pressure dependencies of the lattice vibration mode were extracted, and it is suggested that the instability of the trirutile structure at high pressure is related to the strong deformation of the TaO6 octahedra. Decompression measurements suggest that the pressure-induced transformation is irreversible.

The ability of natural killer (NK) cells to mediate potent antitumor immunity in clinical adoptive transfer settings relies, in large part, on their ability to retain cytotoxic function following cryopreservation. To avoid potential systemic toxicities associated with infusions of NK cells into patients in the presence of dimethylsulfoxide (DMSO), interest in alternative cryoprotective agents (CPAs) with improved safety profiles has grown. Despite the development of various sugars, amino acids, polyols, and polyampholytes as cryoprotectants, their ability to promote protection from intracellular cryodamage is limited because they mostly act outside of the cell. Though ways to shuttle cryoprotectants intracellularly exist, NK cells' high aversity to manipulation and freezing has meant they are highly understudied as targets for the development of new cryopreservation approaches. Here, the first example of a safe and efficient platform for the intracellular delivery of non‐DMSO CPAs to NK cells is presented. Biocompatible chitosan‐based nanoparticles are engineered to mediate the efficient DMSO‐free cryopreservation of NK cells. NK cells cryopreserved in this way retain potent cytotoxic, degranulation, and cytokine production functions against tumor targets. This not only represents the first example of delivering nanoparticles to NK cells, but illustrates the clinical potential in manufacturing safer allogeneic adoptive immunotherapies “off the shelf.”, The clinical use of natural killer (NK) cells in immunotherapy requires cryopreservation prior to administration to patients. The process of cryopreservation, however, damages NK cells and, when DMSO is used, risks severe toxicities. Here, the development of a safe new nanoparticle‐based approach is described for the intracellular cryoprotection of NK cells, which yields functional effectors post cryopreservation devoid of DMSO.

The ability to ask questions is a powerful tool to gather information in order to learn about the world and resolve ambiguities. In this paper, we explore a novel problem of generating discriminative questions to help disambiguate visual instances. Our work can be seen as a complement and new extension to the rich research studies on image captioning and question answering. We introduce the first large-scale dataset with over 10,000 carefully annotated images-question tuples to facilitate benchmarking. In particular, each tuple consists of a pair of images and 4.6 discriminative questions (as positive samples) and 5.9 non-discriminative questions (as negative samples) on average. In addition, we present an effective method for visual discriminative question generation. The method can be trained in a weakly supervised manner without discriminative images-question tuples but just existing visual question answering datasets. Promising results are shown against representative baselines through quantitative evaluations and user studies., 14 pages, 12 figures, ICCV2017

Data in vision domain often exhibit highly-skewed class distribution, i.e., most data belong to a few majority classes, while the minority classes only contain a scarce amount of instances. To mitigate this issue, contemporary classification methods based on deep convolutional neural network (CNN) typically follow classic strategies such as class re-sampling or cost-sensitive training. In this paper, we conduct extensive and systematic experiments to validate the effectiveness of these classic schemes for representation learning on class-imbalanced data. We further demonstrate that more discriminative deep representation can be learned by enforcing a deep network to maintain both intercluster and inter-class margins. This tighter constraint effectively reduces the class imbalance inherent in the local data neighborhood. We show that the margins can be easily deployed in standard deep learning framework through quintuplet instance sampling and the associated triple-header hinge loss. The representation learned by our approach, when combined with a simple k-nearest neighbor (kNN) algorithm, shows significant improvements over existing methods on both high-and low-level vision classification tasks that exhibit imbalanced class distribution.

We present an approach for recognizing human attributes in unconstrained settings. We train a Convolutional Neural Network (CNN) to select the most attribute-descriptive human parts from all poselet detections, and combine them with the whole body as a pose-normalized deep representation. We further improve by using deep hierarchical contexts ranging from human-centric level to scene level. Human-centric context captures human relations, which we compute from the nearest neighbor parts of other people on a pyramid of CNN feature maps. The matched parts are then average pooled and they act as a similarity regularization. To utilize the scene context, we re-score human-centric predictions by the global scene classification score jointly learned in our CNN, yielding final scene-aware predictions. To facilitate our study, a large-scale WIDER Attribute dataset(Dataset URL: http://mmlab.ie.cuhk.edu.hk/projects/WIDERAttribute) is introduced with human attribute and image event annotations, and our method surpasses competitive baselines on this dataset and other popular ones.

As a commercial interface, three-phase voltage-source inverters (VSI) are commonly equipped for energy conversion to export DC power from most distributed generation (DG) to the AC utility. Not only do voltage-source converters take charge of converting the power to the loads but support the grid voltage at the point of common connection (PCC) as well, which is dependent on the condition of the grid-connected loads. This paper explores the border collision and its interacting mechanism among the VSI, resistive interacting loads and grids, which manifests as the alternating emergence of the inverting and rectifying operations, where the normal operation is terminated and a new one is assumed. Their mutual effect on the power quality under investigation will cause the circuital stability issue and further deteriorate the voltage regulation capability of VSI by dramatically raising the grid voltage harmonics. It is found in a design-oriented view that the border collision operation will be induced within the unsuitable parameter space with respect to transmission lines of AC grid, resistive loads and internal resistance of VSI. The physical phenomenon is also identified by the theoretical analysis. With numerical simulations for various circuit conditions, the corresponding bifurcation boundaries are collected, where the stability of the system is lost via border collision.

For the physical network, such as power physical network, built cyber network to optimise the regulation and control to constitute cyber-physical system, in order to achieve optimal control of large scale distributed systems, and bring along with virus, hacking, denial of service from the threat of cyber network that may lead to malicious damage in physical system. In this paper, a distributed dynamic model for a class of power systems with data attack is built with the assumption of attack detectability. Dynamic state estimators are presented to detect the attacked data and estimate its original value. Finally, a simulation of 9-bus power system shows the effectiveness of our proposed method.