Search

Detecting hot regions plays an important role in urban traffic planning and analytics, which is also useful in self-driving car routing and navigation. In this light, we propose and study a novel urban hot-region detection method by using massive geo-tagged image data. Given a set Q of regions (each region q is made up by a set of geo-tagged images), and a matching threshold θ , if a region q is matched with m other regions, its hot degree is defined by m. The hot-region detection (HRD) search finds the regions with the highest hot degrees. We believe that this type of search may benefit many applications in self-driving cars, including route planning and navigation, and traffic management and analytics in general. The HRD search is challenging due to two reasons. First, how to evaluate the similarity when matching different regions. Second, how to compute the HRD search efficiently, since its time complexity is O ( | Q | 2 ) . To overcome the challenges, we define a novel spatial-density correlation measure to evaluate the similarity between two regions, and develop a parallel search framework to process the HRD efficiently. In addition, a series of optimization techniques, e.g., pruning techniques, are defined to further enhance the query efficiency. Finally, we conduct extensive experiments on real data sets to study the performance of the developed methods.

For a car self-driving system, it is vital to accurately recognize the objects on the road. This is often achieved by analyzing the images captured by various cameras. However, in real applications, we often have limited number of labeled images. Besides, it is very expensive to manually annotate objects in images. It is hard to learn reliable classifiers when only few-labeled images are available. To alleviate these problems, multi-view information is used. However, only using visual information is not enough for reliable classification. Besides, we often have limited images with labels. To cope with these problems, in this paper, we propose a novel multi-view visual-semantic representation method for few-labeled visual recognition ( MV 2 S ). We make use of the state-of-the-art deep convolutional neural networks by viewing them as different views to extract semantic representations of images. This is achieved by using the learned deep convolutional neural networks to make predictions of the semantics of images. Both the visual and semantic representations of images are then used to predict the categories of images by combining the predictions of multi-views with visual and semantic consistency constraints. Experiments on four public available datasets prove the effectiveness of the proposed MV 2 S method.

As a promising anode for potassium-ion batteries (PIBs), graphite has drawn great attention due to easy availability, low cost and formation of potassium-based intercalation compound. However, poor cyclic and rate performances due to severe structural degradation limit the practical application of graphite anode. Herein, a low cost mildly-expanded graphite (MEG) with adjustable interlayer lattice distance is synthesized through a wet chemical oxidation route. As anode for PIBs, the MEG exhibited stable intercalation/deintercalation plateaus, outstanding cyclic stability and rate performance. By optimizing the mass ratio between graphite and potassium permanganate oxidant, the MEG anode retaining long-range-ordered domains of graphite delivers a capacity of 88 mAh/g at 1500 mA/g and a capacity retention of 72.3% after 200 cycles at 100 mA/g. The potassium ion full cell assembled with MEGs anode exhibites high energy density and outstanding cyclic stability. Our results demonstrate a promising graphite-based anode for PIBs and provide guidance for the designing of novel carbonic electrodes for PIBs.

Along with the anti-cavitation performance, the high speed and the high power density, are the main trends in the development of centrifugal pumps. At present, the most effective method is to install an inducer in front of the impeller. However, the tip leakage of the inducer results in the vortex cavitation at the blade leading edge of the inducer, and the cavitating flow inside the inducer seriously interferes with the hydraulic behavior of the inducer as well as the impeller with the development of the cavitation, thus to badly affect the operational reliability of the high-speed centrifugal pump. In the present paper, the cavitating flow in a high-speed centrifugal pump with an inducer is investigated by numerical simulations and visual experiments for different cavitation numbers. A typical evolution process of the cavitation is shown, including the inception, the development and the deterioration. A general description of the pump head-drop phenomenon is made through the study of the local and global flow fields, and the relationship between the vapor distribution and the static pressure distribution along the inducer is determined to describe the evolution of the cavitation. This paper intends to provide the foundation for studying the overall cavitation state of a high-speed centrifugal pump, and designing the inducer with a better cavitation resistance.

Although great process has been achieved to accurately classify images, many methods only use labeled images while ignoring the large quantity of unlabeled images. To make use of unlabeled images, in this letter, we propose a novel transductive visual recognition method using the prediction consistency of multiple views (T-PCMV). Both labeled and unlabeled images are used in a unified framework. The predictions of unlabeled images are learned by linearly combining the discriminative information of multiple views. We ensure the smooth constraint that visually similar images should be predicted with similar labels. To learn the classifier, we jointly minimize the classification loss and the discrepancy of predicted labels. To evaluate the usefulness of the proposed method, we conduct transductive visual recognition experiments on four image datasets. Experimental results well demonstrate the effectiveness of the proposed T-PCMV method.

Higher frictional heat will be produced during the service process of artificial hip joint prosthesis in vivo than natural joint, which would cause the rise of the temperature of synovial fluid and the change of the wear resistance of prosthetic materials. In this work, the biotribological properties of ZrO2/Ti6Al7Nb and ZrO2/UHMWPE pairs at different lubricant temperatures were investigated. The results show that there is a strong correlation between the wettability and average friction coefficient of Ti6Al7Nb, while there is also a strong correlation between the wettability and mass loss of UHMWPE. The wear loss of Ti6Al7Nb and UHMWPE decreases gradually and the friction coefficient increases gradually as the lubricant temperature rises from 20 ℃ to 46 ℃, which is mainly affected by the precipitation process of protein. As a protective layer, it can reduce the wear rate, but as a third body, it can participate in the friction process and increase the friction coefficient. The characteristics of lubricant change greatly at 60 ℃, but the increase of temperature is not the only reason for the precipitation of synovial protein. The synergistic effect with the friction process results in a significant change on the wear mechanism of Ti6Al7Nb and UHMWPE under this condition. The wear mechanism of Ti6Al7Nb is mainly abrasive wear and adhesion wear at different lubricant temperature. With the increase of lubricant temperature, the proportion of adhesion wear mechanism becomes larger. Due to the low thermal deformation temperature and thermal conductivity, the wear mechanism of UHMWPE changed obviously at 60 ℃, and the wear morphology is mainly plastic deformation.

Nowadays, alkali metal-oxygen batteries such as Li-, Na-, and K-O2 batteries have been investigated extensively because of their ultrahigh energy density. However, the oxygen crossover of oxygen ba...

In recent years, research on image generation has been developing very fast. The generative adversarial network (GAN) emerges as a promising framework, which uses adversarial training to improve the generative ability of its generator. However, since GAN and most of its variants use randomly sampled noises as the input of their generators, they have to learn a mapping function from a whole random distribution to the image manifold. As the structures of the random distribution and the image manifold are generally different, this results in GAN and its variants difficult to train and converge. In this paper, we propose a novel deep model called generative adversarial networks with decoder-encoder output noises (DE-GANs), which take advantage of both the adversarial training and the variational Bayesian inference to improve GAN and its variants on image generation performances. DE-GANs use a pre-trained decoder-encoder architecture to map the random noise vectors to informative ones and feed them to the generator of the adversarial networks. Since the decoder-encoder architecture is trained with the same data set as the generator, its output vectors, as the inputs of the generator, could carry the intrinsic distribution information of the training images, which greatly improves the learnability of the generator and the quality of the generated images. Extensive experiments demonstrate the effectiveness of the proposed model, DE-GANs.

Early and precise cancer diagnosis substantially improves patient survival. Recent work has revealed that the levels of multiple microRNAs in serum are informative as biomarkers for the diagnosis of cancers. Here, we designed a DNA molecular computation platform for the analysis of miRNA profiles in clinical serum samples. A computational classifier is first trained in silico using miRNA profiles from The Cancer Genome Atlas. This is followed by a computationally powerful but simple molecular implementation scheme using DNA, as well as an effective in situ amplification and transformation method for miRNA enrichment in serum without perturbing the original variety and quantity information. We successfully achieved rapid and accurate cancer diagnosis using clinical serum samples from 22 healthy people (8) and people with lung cancer (14) with an accuracy of 86.4%. We envision that this DNA computational platform will inspire more clinical applications towards inexpensive, non-invasive and rapid disease screening, classification and progress monitoring.

Potassium-ion battery (KIBs) as an economical and high energy candidate for grid-scale applications has drawn significant attention recently. Given the reversible formation of K-graphite intercalation compound, mass production and low price, the graphite is one of the most promising anode materials for KIBs. However, the issue on the fast capacity fading seriously limits the real application of graphite. The formation and stability of solid electrolyte interphase (SEI) at electrode/electrolyte interfaces during the potassiation/depotassiation or plating/striping processes is closely correlated to the above issue. Herein, by employing typical K/graphite batteries the effect of SEI on the capacity fading were systematically investigated in two electrolyte systems (i.e., ester (EC/DEC) and ether (DEGDME) based electrolytes). The decomposition of EC/DEC electrolyte led to the continuous thickening of SEI on K anode, which should be mainly responsible for the fast capacity decay of K/graphite batteries in EC/DEC electrolyte. The gradual distortion of graphite structure is the secondary cause. By comparison, the more stable DEGDME electrolyte and slight graphitic structure degradation lead to better cyclic stability. Additionally, the components of SEI on graphite anodes in full-cells are basically consistent with those in half-cells. In a word, our work provides a guidance for understanding the scientific issues on how the formation and the stability of SEI influences on the capacity decay in KIBs with graphite anodes.

As the smallest unit of life, cells attract interest due to their structural complexity and functional reliability. Protocells assembled by inanimate components are created as an artificial entity to mimic the structure and some essential properties of a natural cell, and artificial reaction networks are used to program the functions of protocells. Although the bottom-up construction of a protocell that can be considered truly ‘alive’ is still an ambitious goal, these man-made constructs with a certain degree of ‘liveness’ can offer effective tools to understand fundamental processes of cellular life, and have paved the new way for bionic applications. In this review, we highlight both the milestones and recent progress of protocells programmed by artificial reaction networks, including genetic circuits, enzyme-assisted non-genetic circuits, prebiotic mimicking reaction networks, and DNA dynamic circuits. Challenges and opportunities have also been discussed., In this review, the milestones and recent progress of protocells programmed by various types of artificial reaction networks are highlighted.

Solar cells are the key part of photovoltaic power system, and the online quality inspection of solar cells ensures conversion efficiency and usable lifetime of photovoltaic modules. However, automatic crack detection in EL images of solar cells has been a challenging task, owing to the heterogeneously textured background, low contrast between crack and surrounding background including randomly distributed crystal grains, the diversity of crack types, and so on. To address these challenges, this paper presents a new accurate and robust crack detection scheme for multicrystalline solar cells. Firstly, a novel steerable evidence filter is developed to generate the crack saliency map, which significantly enhances the contrast between crack and surrounding background. Secondly, a segmentation-based method including local threshold and minimum spanning tree is employed, which ensures the extraction of complete crack. Next, the crack can be accurately located in the inspection image by computing the crack skeleton. Finally, experimental results on defective and defect-free EL images show that the proposed scheme is accurate and robust, and various types of cracks can be correctly detected. The proposed scheme achieved average detection rate of 94.4% on the data set, which performs better than the previous methods.

Fine-grained image classification is a hot topic that has been widely studied recently. Many fine-grained image classification methods ignore misclassification information, which is important to improve classification accuracy. To make use of misclassification information, in this paper, we propose a novel fine-grained image classification method by exploring the misclassification information (FGMI) of prelearned models. For each class, we harvest the confusion information from several prelearned fine-grained image classification models. For one particular class, we select a number of classes which are likely to be misclassified with this class. The images of selected classes are then used to train classifiers. In this way, we can reduce the influence of irrelevant images to some extent. We use the misclassification information for all the classes by training a number of confusion classifiers. The outputs of these trained classifiers are combined to represent images and produce classifications. To evaluate the effectiveness of the proposed FGMI method, we conduct fine-grained classification experiments on several public image datasets. Experimental results prove the usefulness of the proposed method.

License plate recognition (LPR) remains a challenging task in face of some difficulties such as image deformation and multi-line character distribution. Text rectification that is crucial to eliminate the effects of image deformation has attracted increasing attentions in scene text recognition. However, current text rectification methods are not designed specifically for LPR, which did not take the features of plate deformation into account. Considering the fact that a license plate (LP) can only generate perspective distortion in the image due to its rigid feature, in this paper we propose a novel perspective rectification network (PRN) to automatically estimate the perspective transformation and rectify the distorted LP accordingly. For recognition, we propose a location-aware 2D attention based recognition network that is capable of recognizing both single-line and double-line plates with perspective deformation. The rectification network and recognition network are connected for end-to-end training. Experiments on common datasets show that the proposed method achieves the state-of-the-art performance, demonstrating the effectiveness of the proposed approach.

Whole blood analysis reveals crucial information about various physiological and pathological conditions, including cancer metastasis, infection, and immune status, among others. Despite this rich information, the complex composition of whole blood usually required multiple sample preparation steps to purify targeted analytes. Traditionally, whole blood preparation processes, including centrifugation, lysis, dilution, or staining, are usually manually operated by well-trained technicians using bench-top instruments. This preparation can require a large blood volume and cannot be directly integrated with detection systems. Recently, various studies have integrated microfluidics with electrical sensors for whole blood analysis, with a focus on cell-based analysis, such as cell type, number, morphology, phenotype, and secreted molecules. These miniaturized systems require less sample and shorter reaction times. Besides, the sample processing and analysis can be fully integrated and automated with minimal operations. We believe these systems can transfer the current whole blood analysis from hospitals or laboratories into clinics or home settings to enable real-time and continuous health condition monitoring in point-of-care settings.

Biological membrane channels, considered as molecular gatekeepers, control the transportation of molecules and ions across live cell membranes. Developing synthetic passable channels with predictable structures, high transport efficiency, and low cytotoxicity on live cells is of great interest for replicating the functions of endogenous protein channels, but remains challenging. The development of DNA nanotechnology provides possible solutions for making synthetic channels with precise structures and controllable functionalization. Therefore, in this work, we constructed a phosphorothioate-modified DNA nanopore able to structurally mimic biological channels for molecular transport across live cell membranes. With its stable structure with small hollow size (

This paper presents the competition on Offline Recognition and Spotting of Handwritten Mathematical Expressions (OffRaSHME) held at the 17th International Conference on Frontiers in Handwriting Recognition (ICFHR 2020). Handwritten Mathematical Expression Recognition (HMER) has wide potential applications and is capturing increasing attention in recent years. Previous HMER competitions mainly focused on online datasets or offline datasets that are converted from online data. In this competition, we have collected a dataset of offline handwritten mathematical expressions by scanning papers that contain expressions. Moreover, we labeled the offline dataset at symbol level, i.e., the bounding boxes of each symbol are also provided, to facilitate the research of HMER. At last, 19,749 offline handwritten mathematical expressions are collected for training, and 2,000 ones are provided for evaluating the participating systems. In the competition, 7 teams submitted 8 systems for the task of offline HMER, among which 5 systems only use the provided datasets without any extra data while 3 systems use extra data. The winner team achieved a recognition accuracy of 79.85% (without extra data) and 81.85% (with extra data) on the offline formula recognition task.

Nucleic acids are natural biopolymers of nucleotides that store, encode, transmit and express genetic information, which play central roles in diverse cellular events and diseases in living things. The analysis of nucleic acids and nucleic acids-based analysis have been widely applied in biological studies, clinical diagnosis, environmental analysis, food safety and forensic analysis. During the past decades, the field of nucleic acids analysis has been rapidly advancing with many technological breakthroughs. In this review, we focus on the methods developed for analyzing nucleic acids, nucleic acids-based analysis, device for nucleic acids analysis, and applications of nucleic acids analysis. The representative strategies for the development of new nucleic acids analysis in this field are summarized, and key advantages and possible limitations are discussed. Finally, a brief perspective on existing challenges and further research development is provided.

Solid electrolytes are considered as strong alternatives for conventional liquid electrolytes to overcome the safety issues of next-generation high-energy-density lithium metal batteries (LMBs). Although Li1.5Al0.5Ge1.5(PO4)3 (LAGP) has satisfied ionic conductivity at room temperature (∼10-4 S cm-1), high stability in air, and can be easily sintered, it still suffers from instability of the lithium metal. Moreover, the large interfacial resistance between solid electrolytes and solid electrodes and the stress generated by the volumetric change of lithium metal anodes during cycling would deteriorate the performance of LMBs. Here, we report an effective solution to overcome the abovementioned problems by introducing a three-dimensional gel polymer electrolyte at the interface between LAGP pellets and lithium metal anodes, achieving stable cycling of LiFePO4//Li cells at room temperature for 300 cycles. Besides, the degeneration mechanisms of the interfaces of LAGP pellets under different conditions are compared, and peculiar properties different from their counterparts were found.

Since adoption of the Paris Climate Agreement, efforts to reduce greenhouse gases have been conducted worldwide. Particularly in the transportation sector, researches on electric vehicles are actively underway as an effort to reduce air pollutants, and electric vehicles are expected to become a major transportation vehicle by 2030. If electric vehicles replace existing internal combustion engines, much more electricity will be required to charge the electric vehicles, which will affect the power system. To deal with this growing power demand, it is necessary to forecast and analyze the charging pattern of electric vehicles. This paper uses Bayesian inference methods with convolution to forecast the 24-hour charging pattern for electric vehicles. The results were compared with that by linear regression analysis.

This study reports a microfluidic device for whole blood processing. The device uses the bifurcation law, cross-flow method, and hydrodynamic flow for simultaneous extraction of plasma, red blood cells, and on-chip white blood cell trapping. The results demonstrate successful plasma and red blood cell collection with a minimum dilution factor (0.76x) and low haemolysis effect. The extracted red blood cells can also be applied for blood type tests. Moreover, the device can trap up to ~1,800 white blood cells in 20 minutes. The three components can be collected simultaneously using only 6 μL of whole blood without any sample preparation processes. Based on these features, the microfluidic device enables low-cost, rapid, and efficient whole blood processing functionality that could potentially be applied for blood analysis in resource-limited environments or point-of-care settings.

A novel redox precipitation route was developed for the preparation of Ce Mn composite oxides, where KMnO4 and Ce(NO3)3 were adopted as precursors. The samples were tested for selective catalytic reduction (SCR) of NO by NH3 at low temperature, and characterized using X-ray diffraction (XRD), Brunauer-Emmett-Teller (BET), field emission scanning electron microscope (FESEM), transmission electron microscope (TEM), thermogravimetric analysis (TG), hydrogen temperature-programmed reduction (H2-TPR), X-ray photoelectron spectroscopy (XPS) and in situ diffuse reflectance infrared transform spectroscopy (in situ DRIFT). This novel method produced samples with large specific surface area, low crystallinity, high oxidation degree and superior performances in low-temperature NH3-SCR. Ce 3Mn containing almost 100% Mn4+ ion on its surface exhibited superior low-temperature reducibility, reaching up to above 95% NO conversion almost throughout the entire testing range of 353–453 K. The surface concentration of Mn4+, the strong synergistic effects between MnO2 and CeO2-x and the large surface area played a key role for its excellent catalytic activity for the complete reduction of nitrogen oxides. A primary Eley-Rideal (E-R) reaction mechanism for Ce 3Mn catalyst was proposed with Lewis acid sites involved in the SCR reaction.

Well-crystalline monoclinic Mn5O8, tetragonal Mn3O4 and cubic Mn2O3 were synthesized via a solvothermal route followed by distinct heat treatment. The formation of different MnOx species observed by X-ray diffraction (XRD) measurements showed temperature dependent phase transformation occurring during the heat treatment process. XRD patterns showed that Mn3O4 and Mn5O8 were formed at calcination temperature range of 573–673 K and a new phase Mn2O3 appeared at 773 K, respectively. And a tendency to the formation of Mn5O8 was found as the solvothermal temperature turns higher. The catalytic activities of various MnOx for selective catalytic reduction (SCR) of nitrogen oxides were evaluated between 353 and 573 K using ammonia as reducer. Mn5O8 appeared to exhibit the highest activity, followed by Mn3O4, Mn2O3.

To develop the symmetrical electrode materials for solid oxide fuel cells (SOFCs) and to explore the facile cell fabrication technique are both meaningful and of great significance. Here a bi-functional hybrid material LaNi0.82Fe0.18O3 (LNF)/NiO was synthesized by a one-pot citrate method and further used as the quasi-symmetrical electrode catalysts for solid oxide fuel cells (SOFCs). LNF and Ni (reduced NiO) functioned as the cathode/anode catalysts. The La0.9Sr0.1Ga0.8Mg0.2O3 (LSGM) based asymmetrical tri-layered substrates were fabricated by a screen-printing assisted co-firing technique. The polarization resistances (Rp) of the infiltrated anode at 700, 650, 600 and 550 °C were only 0.08, 0.12, 0.18 and 0.3 Ω cm2, respectively. Comparably, the Rp of the infiltrated cathode were much larger, e.g., 0.18, 0.35, 0.875 and 2.55 Ω cm2 at 700, 650, 600 and 550 °C, respectively. Encouragingly, these cathode Rp values were largely reduced when discharge due to an activation process. The LNF and NiO reversibly formed and decomposed during the oxidation and reduction processes, suggesting that the LNF/NiO hybrid is a potential quasi-symmetrical SOFC electrode material. When using H2 as fuel and air as oxidant, the maximum power densities of the single cell at 650, 600 and 550 °C were as high as 928, 580 and 329 mW cm−2, respectively.

α-, β-, δ-, and γ-MnO2 nanocrystals are successfully prepared. We then evaluated the NH3 selective catalytic reduction (SCR) performance of the MnO2 catalysts with different phases. The NOx conversion efficiency decreased in the order: γ-MnO2 > α-MnO2 > δ-MnO2 > β-MnO2. The NOx conversion with the use of γ-MnO2 and α-MnO2 catalysts reached 90% in the temperature range of 140–200 °C, while that based on β-MnO2 reached only 40% at 200 °C. The γ-MnO2 and α-MnO2 nanowire crystal morphologies enabled good dispersion of the catalysts and resulted in a relatively high specific surface area. We found that γ-MnO2 and α-MnO2 possessed stronger reducing abilities and more and stronger acidic sites than the other catalysts. In addition, more chemisorbed oxygen existed on the surface of the γ-MnO2 and α-MnO2 catalysts. The γ-MnO2 and α-MnO2 catalysts showed excellent performance in the low-temperature SCR of NO to N2 with NH3.

Herein, the Sr2Fe1.5Mo0.5O6 (SFM) precursor solution is infiltrated into a tri-layered “porous La0.9Sr0.1Ga0.8Mg0.2O3 (LSGM)/dense LSGM/porous LSGM” skeleton to form both SFM/LSGM symmetrical fuel cells and functional fuel cells by adopting an ultra-fast and time-saving procedure. The heating/cooling rate when fabricating is fixed at 200 °C/min. Thanks to the unique cell structure with high thermal shock resistance and matched thermal expansion coefficients (TEC) between SFM and LSGM, no SFM/LSGM interfacial detachment is detected. The polarization resistances (Rp) of SFM/LSGM composite cathode and anode at 650 °C are 0.27 Ω·cm2 and 0.235 Ω·cm2, respectively. These values are even smaller than those of the cells fabricated with traditional method. From scanning electron microscope (SEM), a more homogenous distribution of SFM is identified in the ultra-fast fabricated SFM/LSGM composite, therefore leading to the enhanced performance. This study also strengthens the evidence that SFM can be used as high performance symmetrical electrode material both running in H2 and CH4. When using H2 as fuel, the maximum power density of “SFM-LSGM/LSGM/LSGM-SFM” functional fuel cell at 700 °C is 880 mW cm−2. By using CH4 as fuel, the maximum power densities at 850 and 900 °C are 146 and 306 mW cm−2, respectively. Keywords: Sr2Fe1.5Mo0.5O6, LSGM skeleton, Ultra-fast, Symmetrical electrode

This paper introduces a Chinese Ancient Handwritten Characters Database (CASIA-AHCDB) for character recognition research. The database was built by annotating 11,937 pages of Chinese ancient handwritten documents. It consists of more than 2.2 million annotated handwritten character samples of 10,350 categories. According to the source of these documents, the database is divided into two datasets of different styles: Complete Library in Four Sections (AHCDB-style1) and Ancient Buddhist Scriptures (AHCDB-style2). Each dataset can be divided into three parts based on its applications. The first part, called basic category set, contains samples of common categories in two datasets, and is suitable for basic character recognition task. The second part, called enhanced category set, is mainly used for open-set character recognition task based on the basic character recognition. The third part, called the reserved category set, can be used in many pattern recognition tasks in the future. Based on the large category set, the various writing styles and the imbalanced sample number per category, CASIA-AHCDB can also be used for various classification and learning tasks such as transfer learning, few-shot learning. We performed experiments of basic character recognition on the basic category set, and report the results for benchmark. More techniques can be evaluated on this challenging database in the future.

Performances of fine-grained recognition have been greatly improved thanks to the fast developments of deep convolutional neural networks (DCNN). DCNN methods often treat each image region equally. Besides, researchers often rely on visual information for classification. To solve these problems, we propose a novel discriminative semantic region selection method for fine-grained recognition (DSRS). We first select a few image regions and then use the pre-trained DCNN models to predict their semantic correlations with corresponding classes. We use both visual and semantic representations to represent image regions. The visual and semantic representations are then linearly combined for joint representation. The combination parameters are determined by considering both semantic distinctiveness and spatial-semantic correlations. We use the joint representations for classifier training. A testing image can be classified by obtaining the visual and semantic representations and encoded for joint representation and classification. Experiments on several publicly available datasets demonstrate the proposed method's superiority.

Recently, the lithium–oxygen (Li–O2) battery has attracted much interest due to its ultrahigh theoretical energy density. However, its potential application is limited by an unstable electrolyte system, low round-trip efficiency, and poor cyclic performance. In this study, we present a new electrolyte based on N,N′-dimethylpropyleneurea (DMPU) applied for the Li–O2 battery. This electrolyte possesses high ionic conductivity and achieves a low discharge/charge voltage gap of 0.6 V, which is mainly due to the possible one-electron charge transfer mechanism. The introduction of the antioxidant butylatedhydroxytoluene (BHT) as an additive stabilizes the superoxide radical by chemical adsorption and improves the cyclic performance remarkably. Thus, this new electrolyte system may be one of the candidates for Li–O2 batteries.

Symmetrical solid oxide fuel cells (SOFCs) that feature nanostructured SrFe0.75Mo0.25O3−δ (SFMO) electrodes which are impregnated into porous La0.9Sr0.1Ga0.8Mg0.2O3−δ (LSGM) skeletons by solution infiltration technique are prepared. The symmetrical SOFCs are investigated in detail by operating with 5 wt% Fe-loaded activated carbon as fuel and ambient air as oxidant. A typical single cell can give a maximum power density of 405 mW cm−2 at 850 °C, which is comparable to the results with Ni-based anode-supported SOFCs. Also, the direct carbon SOFC (DC-SOFC) with reloaded fuel can recover to its initial performance. The microstructures and XRD of the electrodes after test are also examined and analysed. The results show that the SFMO-LSGM composite anode remains porous and maintains its structure stability, implying that the symmetrical SFMO electrodes can be applied into DC-SOFCs.

Efficient hypothesis generation plays an important role in robust model fitting. In this study, based on the combination of residual sorting and local constraints, we propose an efficient guided hypothesis generation method, called Rapid Hypothesis Generation (RHG). By exploiting the local constraints to guide the hypothesis generation process, RHG raises the probability of generating promising hypotheses and reduces the computational cost during hypotheses generation. Experimental results on homography and fundamental matrix estimation show that RHG can effectively guide hypothesis generation process and rapidly generate promising hypotheses for heavily contaminated multi-structure data.

A delivery service using unmanned aerial vehicles (UAVs) has potential as a future business opportunity, due to its speed, safety and low-environmental impact. To operate a UAV delivery network, a management system is required to optimize UAV delivery routes. Therefore, we create a routing algorithm to find optimal round-trip routes for UAVs, which deliver goods from depots to customers. Optimal routes per UAV are determined by minimizing delivery distances considering the maximum range and loading capacity of the UAV. In order to accomplish this, we propose an algorithm with four steps. First, we build a virtual network to describe the realistic environment that UAVs would encounter during operation. Second, we determine the optimal number of in-service UAVs per depot. Third, we eliminate subtours, which are infeasible routes, using flow variables part of the constraints. Fourth, we allocate UAVs to customers minimizing delivery distances from depots to customers. In this process, we allow multiple UAVs to deliver goods to one customer at the same time. Finally, we verify that our algorithm can determine the number of UAVs in service per depot, round-trip routes for UAVs, and allocate UAVs to customers to deliver at the minimum cost.

Folates are typically present in polyglutamyl form in organisms. In traditional extraction methods, polyglutamyl folates are hydrolyzed to monoglutamates, sacrificing valuable information. To advance folate metabolism research, we developed an accurate, sensitive, and reproducible extraction method for polyglutamyl folate species in maize, the main crop in most parts of the world. Twelve folates, including six polyglutamyl folates, were simultaneously determined in maize for the first time using high-performance liquid chromatography-tandem mass spectrometry. The glutamation states of the folates were protected by boiling, which inactivated the native conjugases. α-Amylase and protease were added to obtain better recoveries and decrease difficulties in centrifugation and filtration. The recoveries (n = 5) of six polyglutamyl folates were between 80.5 and 101%. All calibration curves showed good linear regression (r2 ≥ 0.994) within the working range. The instrumental limits of detection and quantitation ranged from 0.070 to 2.4 ng/mL and 0.22 to 8.0 ng/mL, respectively. Intra- and inter-day precision was below 7.81% and 11.9%, respectively (n = 5). Using this method, changes in poly- and monoglutamyl folates during maize germination were determined for the first time. The results suggest that folates were largely synthesized as germination initiated, and 5-methyltetrahydrofolate was the most abundant species. Tetraglutamyl 5-methyltetrahydrofolate contributed more than 50% of the 5-methyltetrahydrofolate species. Inverse changes in contents of 5,10-methenyltetrahydrofolate, and 10-formyl folic acid, monoglutamate, and diglutamate of 5-formyltetrahydrofolate were also observed, indicating potential regulation. Additionally, polyglutamyl folates in sweet potatoes were determined using this method, indicating its applications in starchy crops.

A DNA reaction network is like a biological algorithm that can respond to "molecular input signals", such as biological molecules, while the artificial cell is like a microrobot whose function is powered by the encapsulated DNA reaction network. In this work, we describe the feasibility of using a DNA reaction network as the computational core of a protocell, which will perform an artificial immune response in a concise way to eliminate a mimicked pathogenic challenge. Such a DNA reaction network (RN)-powered protocell can realize the connection of logical computation and biological recognition due to the natural programmability and biological properties of DNA. Thus, the biological input molecules can be easily involved in the molecular computation and the computation process can be spatially isolated and protected by artificial bilayer membrane. We believe the strategy proposed in the current paper, i.e., using DNA RN to power artificial cells, will lay the groundwork for understanding the basic design principles of DNA algorithm-based nanodevices which will, in turn, inspire the construction of artificial cells, or protocells, that will find a place in future biomedical research.

Hexagonal close packed Cr2O3, fabricated by an electrospinning technique combined with a heating method, is adopted for the first time as a catalyst for non-aqueous lithium–oxygen (Li–O2) batteries. The synthesized highly mesoporous Cr2O3 nanotubes (Cr2O3-MNT) with a large surface area of 53.4 m2 g−1 are confirmed by field emission scanning electron microscopy (FESEM), field emission high-resolution transmission electron microscopy (TEM) and nitrogen adsorption/desorption isotherms (BET). By using the prepared Super P (SP) (60 wt%)/Cr2O3 (30 wt%)/polyvinylidenefluoride (PVDF) (10 wt%) composite as an oxygen electrode, the Li–O2 battery shows an astonishingly enhanced capacity of 8280 mA h g−1 at a current density of 50 mA g−1. More encouragingly, when the current densities are fixed at 25, 50, 100 and 200 mA g−1 with a limited capacity of 500 mA h g−1, the charging potentials are 3.47, 3.51, 3.78 and 4.01 V, respectively, which are among the lowest charge potentials reported to date. By using a capacity-controlled method (1000 mA h g−1) at a current density of 100 mA g−1, the cell shows excellent cyclic stability up to 50 cycles. The reversible formation and dissociation of Li2O2 are verified by X-ray diffusion (XRD) and SEM, indicating that the as-prepared Cr2O3 nanotubes are a promising catalyst for Li–O2 batteries.

Robust crack defect detection in solar cells has been challenging because of the inhomogeneously textured surface, low contrast between crack defect and background, the diversity of crack types, and so on. To overcome these challenges, this paper presents a new robust crack defect detection scheme for multicrystalline solar cells. Firstly, a steerable evidence filter is designed to process EL image to obtain the response map, which enhances the contrast between crack and background and provides evidence for the presence of crack defect. Secondly, complete crack extraction from the response map is employed. Finally, the complete crack can be located in the inspection image by the crack skeleton extraction. Experimental results on defective and defect-free EL images show that the proposed scheme is robust, and various cracks can be effectively detected, which outperforms the previous methods.

Cells interact with the extracellular environment through molecules expressed on the membrane. Disruption of these membrane-bound interactions (or encounters) can result in disease progression. Advances in super-resolution microscopy have allowed membrane encounters to be examined, however, these methods cannot image entire membranes and cannot provide information on the dynamic interactions between membrane-bound molecules. Here, we show a novel DNA probe that can transduce transient membrane encounter events into readable cumulative fluorescence signals. The probe, which translocates from one anchor site to another, mimicking motor proteins, is realized through a toehold-mediated DNA strand displacement reaction. Using this probe, we successfully monitored rapid encounter events of membrane lipid domains using flow cytometry and fluorescence microscopy. Our results show a preference for encounters within the same lipid domains.

Automatic crack defect detection for multicrystalline solar cells is a challenging task, owing to inhomogeneously textured background, disturbance of crystal grains pseudo defects, and low contrast between crack defect and background. In this paper, a novel structure-aware-based crack defect detection scheme (SACDDS) is proposed. Firstly, the structure features of crack defect and randomly distributed crystal grains are analyzed, and corresponding mathematical models are used to represent two structure features. Secondly, according to Hessian eigenvalues of the above mathematical models, the identification functions of linear-structure and blob-structure are obtained. Then, a novel structure similarity measure (SSM) function is designed by using the identification functions of two structures, which can highlight crack defect, suppress crystal grains simultaneously. It significantly weakens the interference of inhomogeneous texture and obtains uniform background. Further, in order to overcome non-uniform response of crack region and extract crack defect, a tensor voting-based non-maximum suppression (TV-NMS) method is developed. It improves the uniformity of crack defect response and extracts candidate crack defect pixels. Finally, an effective morphological operation is applied to remove non-crack pixels and complete crack defect can be located in the EL images. Experimental results show that the proposed method can completely extract crack defect in the inhomogeneously textured background, which is well effective and outperforms the previous methods.

Two-dimensional nanocarbons are intriguing functional materials for energy storage. However, the serious aggregation problems hinder their wider applications. To address this issue, we developed a unique two-dimensional hierarchical carbon architecture (2D-HCA) with ultrasmall graphene-like carbon nanosheets uniformly grown on hexagonal carbon nanoplates. The obtained 2D-HCA shows an interconnected porous structure and abundant heteroelement doping. When employed as anode for lithium ion batteries, it exhibits a high discharge capacity of 748 m Ah g–1 even after 400 cycles at 2 A g–1.

There are a wide range of applications for scene text detection and recognition due to the increasing popularity of portable digital devices. However, large-scale evaluation benchmark with multilingual and multi-oriented texts is still slow to occur to facilitate the research on scene text detection and recognition. In this paper, a large-scale and well-annotated scene text dataset, namely STV2k, is presented, which can be used for scene text detection as well as scene text recognition. Since all the images are collected from streets by smart phone, the textual scenes are rich of variations in layouts, color, fonts and backgrounds. Two state-of-the-art algorithms for scene text recognition are tested on this newly built dataset. The preliminary experiments demonstrate how challenging the scene text recognition is in real scenario.

The hemoglobin-A1c test, measuring the ratio of glycated hemoglobin (HbA1c) to hemoglobin (Hb) levels, has been a standard assay in diabetes diagnosis that removes the day-to-day glucose level variation. Currently, the HbA1c test is restricted to hospitals and central laboratories due to the laborious, time-consuming whole blood processing and bulky instruments. In this paper, we have developed a microfluidic device integrating dual CMOS polysilicon nanowire sensors (MINS) for on-chip whole blood processing and simultaneous detection of multiple analytes. The micromachined polymethylmethacrylate (PMMA) microfluidic device consisted of a serpentine microchannel with multiple dam structures designed for non-lysed cells or debris trapping, uniform plasma/buffer mixing and dilution. The CMOS-fabricated polysilicon nanowire sensors integrated with the microfluidic device were designed for the simultaneous, label-free electrical detection of multiple analytes. Our study first measured the Hb and HbA1c levels in 11 clinical samples via these nanowire sensors. The results were compared with those of standard Hb and HbA1c measurement methods (Hb: the sodium lauryl sulfate hemoglobin detection method; HbA1c: cation-exchange high-performance liquid chromatography) and showed comparable outcomes. Finally, we successfully demonstrated the efficacy of the MINS device's on-chip whole blood processing followed by simultaneous Hb and HbA1c measurement in a clinical sample. Compared to current Hb and HbA1c sensing instruments, the MINS platform is compact and can simultaneously detect two analytes with only 5 μL of whole blood, which corresponds to a 300-fold blood volume reduction. The total assay time, including the in situ sample processing and analyte detection, was just 30 minutes. Based on its on-chip whole blood processing and simultaneous multiple analyte detection functionalities with a lower sample volume requirement and shorter process time, the MINS device can be effectively applied to real-time diabetes diagnostics and monitoring in point-of-care settings.

In order to culture and analyze individual living cells, microfluidic cultivation and manipulation of cells become an increasingly important topic. Such microfluidic systems allow for exploring the phenotypic differences between thousands of genetically identical cells or pharmacological tests in parallel, which is impossible to achieve by traditional macroscopic cell culture methods. Therefore, plenty of microfluidic systems and devices have been developed for cell biological studies like cell culture, cell sorting, and cell lysis in the past. However, these microfluidic systems are still limited by the external pressure sources which most of the time are large in size and have to be connected by fluidic tubing leading to complex and delicate systems. In order to provide a miniaturized, more robust actuation system a novel, compact and low power consumption digital hydraulic drive (DHD) has been developed that is intended for use in portable and automated microfluidic systems for various applications. The DHD considered in this work consists of a shape memory alloy (SMA) actuator and a pneumatic cylinder. The switching time of the digital modes (pressure ON versus OFF) can be adjusted from 1 s to min. Thus, the DHDs might have many applications for driving microfluidic devices. In this work, different implementations of DHDs are presented and their performance is characterized by experiments. In particular, it will be shown that DHDs can be used for microfluidic large-scale integration (mLSI) valve control (256 valves in parallel) as well as potentially for droplet-based microfluidic systems. As further application example, high-throughput mixing of cell cultures (96 wells in parallel) is demonstrated employing the DHD to drive a so-called 'functional lid' (FL), to enable a miniaturized micro bioreactor in a regular 96-well micro well plate.

This paper investigates resonance vibration suppression under persistent excitation near the first structural resonant frequency of a clamped-clamped (doubly-clamped) piezoelectric flexible beam. In this study, an acceleration sensor is used to measure the resonant vibration. Firstly, the finite element method (FEM) is utilized to derive the dynamics model of the system, and modal analysis is carried out. Secondly, an acceleration feedback-based proportional-integral control algorithm and variable structure control algorithms are designed, and a numerical simulation is performed. Finally, a doubly-clamped piezoelectric flexible beam experimental setup is constructed. Experiments are conducted on resonant vibration suppression using the designed control algorithms. The numerical simulation and experimental results demonstrate that the resonant vibration can be suppressed by using the designed control methods, and the improved variable structure control method shows better control performance in suppressing the resonant vibration.

Multidrug resistance (MDR) resulting from overexpression of P-glycoprotein (Pgp) transporters increases the drug efflux and thereby limits the chemotherapeutic efficacy. It is desirable to administer both an MDR1 gene silencer and a chemotherapeutic agent in a sequential way to generate a synergistic therapeutic effect in multidrug-resistant cancer cells. Herein, we rationally designed an anti-MDR1 molecular beacon (MB)-based micelle (a-MBM) nanosystem, which is composed of a diacyllipid core densely packed with an MB corona. One of Pgp-transportable agents, doxorubicin (DOX), was encapsulated in the hydrophobic core of the micelle and in the stem sequence of MB. The a-MBM-DOX nanosystem showed an efficient self-delivery, enhanced enzymatic stability, excellent target selectivity, and high drug-loading capacity. With its relatively high enzymatic stability, a-MBM-DOX initially facilitated intracellular MDR1 mRNA imaging to distinguish multidrug-resistant and non-multidrug-resistant cells and subsequently downregulated the MDR1 gene expression owing to an antisense effect. After that, the MB corona was degraded, destroying the micellar nanostructure and releasing DOX, which resulted in a high accumulation of DOX in OVCAR8/ADR cells and a high chemotherapeutic efficacy owing to successful restoration of drug sensitivity. This micelle approach has the potential for both visualizing MDR1 mRNA and overcoming MDR in a sequential and synergistic way.

An azobenzene-containing surfactant was synthesized for the phase transfer of a-cyclodextrin (alpha-CD)-capped gold nanoparticles between water and toluene phases by host guest chemistry. With the use of the photoisomerization of azobenzene, the reversible phase transfer of gold nanoparticles was realized by irradiation with UV and visible light. Furthermore, the phase transfer scheme was applied for the quenching of a reaction catalyzed by gold nanoparticles, as well as the recovery and recycling of the gold nanoparticles from aqueous solutions. This work will have significant impact on materials transfer and recovery in catalysis and biotechnological applications.