Search

Video temporal segmentation plays a vital role in video analysis since many higher-level computer vision tasks rely on it. Some recent efforts have been dedicated to generating temporal action proposals for long and untrimmed videos, which requires methods to generate accurate boundaries for video semantics. In this paper, we propose a novel and efficient Temporal Distribution Network (TDN), to model the long-term distribution of video semantic units (video dictionary). Firstly, we encode the semantics and context relations of video segments with a boundary-specified video embedding method. Then based on temporal convolutional layers, we design a Temporal Distribution Network (TDN) enumerating all the possible temporal locations in one pass and generating proposals that have high action confidence scores by capturing the long-term distributions of video semantics. We validate our method on temporal action proposal generation tasks and action detection tasks. Experimental results on two benchmark datasets, THUMOS14 and ActivityNet-1.3, show that the proposed method can significantly outperform the state-of-the-art approaches. Our model could obtain high-quality action proposals with a much faster speed.

In this study, the variations of the seawater carbonate system parameters and air-sea CO2 flux (FCO2) of Shen’ao Bay, a typical subtropical aquaculture bay located in China, were investigated in spring 2016 (March to May). Parameters related to the seawater carbonate system and FCO2 were measured monthly in 3 different aquaculture areas (fish, oyster and seaweed) and in a non-culture area near the bay mouth. The results showed that the seawater carbonate system was markedly influenced by the biological processes of the culture species. Total alkalinity was significantly lower in the oyster area compared with the fish and seaweed areas, mainly because of the calcification process of oysters. Dissolved inorganic carbon (DIC) and CO2 partial pressure (pCO2) were highest in the fish area, followed by the oyster and non-culture areas, and lowest in the seaweed area. Oysters and fish can have indirect influences on DIC and pCO2by releasing nutrients, which facilitate the growth of seaweed and phytoplankton and therefore promote photosynthetic CO2 fixation. For these reasons, Shen’ao Bay acts as a potential CO2 sink in spring, with an average FCO2 ranging from -1.2 to -4.8 mmol m-2 d-1. CO2 fixation in the seaweed area was the largest contributor to CO2 flux, accounting for ca. 58% of the total CO2 sink capacity of the entire bay. These results suggest that the carbonate system and FCO2 of Shen’ao Bay were significantly affected by large-scale mariculture activities. A higher CO2 sink capacity could be acquired by extending the culture area of seaweed.

This article describes a wideband low-noise frequency synthesizer implemented in 0.13- $\mu \text{m}$ SiGe BiCMOS process for 5G millimeter-wave applications. To extend the frequency range while reducing the phase noise, a fundamental voltage-controlled oscillator (VCO) array including four Colpitts VCO cores with switchable bias circuits is adopted in the proposed frequency synthesizer. A ring-oscillator-based injection locked frequency divider is utilized as the wideband divide-by-2 prescaler, and its bandwidth is optimized based on a new injection-locked behavior model. This fabricated frequency synthesizer can be locked in a range from 23 to 36.8 GHz (46.2%) by a 100-MHz step. It achieves a phase noise of −94.7 dBc/Hz at the 1-MHz offset and an output power of −3.5 dBm measured at 36.8 GHz. The chip consumes 360.6 mW from 3.3 and 1.8 V supplies and has an area of $2.7\times3.1$ mm2.

Bladder cancer is the fourth most common cancer in men, and it is becoming a prevalent malignancy. Most of the regular clinical examinations are prompt evaluations with cystoscopy, renal function testing, which require high-precision instrument, well-trained operators, and high cost. In this study, a microfluidic paper-based analytical device (μPAD) was fabricated to detect nuclear matrix protein 22 (NMP22) and bladder cancer antigen (BTA) from the urine samples. Urine samples were collected from 11 bladder cancer patients and 10 well-beings as experiment and control groups, respectively, to verify the working efficiency of μPAD. A remarkable checkout efficiency of up to 90.91% was found from the results. Meanwhile, this method is feasible for home-based self-detection from urine samples within 10 min for the total process, which provides a new way for quick, economical, and convenient tumor diagnosis, prognosis evaluation, and drug response.

The COVID-19 pandemic has caused a huge impact on global service sector. In the pandemic background, to understand the disparity in service sector outputs at the global level is crucial for assessing the effectiveness of development policies in different countries. This study investigate the global service sector and it aims to investigate the transitional dynamics of the output from the service sector by adopting stochastic kernel analyses. Distribution dynamics analyses are carried out for all the countries in the world. The data are then divided into different regional and income groups to evaluate the impacts of geographical location and income on the development of the service sector. The results show that the Global North will continue to make greater strides, while the output capacity in many Global South countries struggles to reach the global average. Moreover, it is shown that countries with higher per capita income will perform better in the development of their service sector than those with low per capita income, thereby highlighting the persistence of global inequality. Finally, this study shows that the Sub-Saharan Africa region and the South Asia region both are very important in the alleviation of global inequality.

The spread of COVID-19 has significantly dampened global economic activity and has also wreaked havoc on the industrial sector. Industrial development plays an important role in national economic growth. Although industrialisation is not the only feasible way to attain the living standards of developed countries, it has been proven a potent force. In the past few decades, enormous efforts have been made in developing countries to pursue industrial development. Countries such as China, India, and South Korea have achieved success in alleviating income inequality and maintaining a good progress in industrial development. However, in many countries, severe underdevelopment and inequality still exist. Understanding the disparity and convergence of global industrial outputs is important in assessing the effectiveness of concurrent development policies. This study investigates the spatial distribution of global industrial output to unveil the disparity in industrial development and the feasibility of achieving convergence over time. Stochastic kernel analyses are carried out for national regimes to study the overall pattern of industrialisation for all the countries in the world. Countries are then classified into different groups to further analyse the geographical and income effects on industrial development.

Public transport infrastructure is essential to city development, improving accessibility and allowing people to engage in activities, allowing travel by environmentally-friendly modes, and also creating economic benefits. This study investigates the impacts of public transport accessibility on development, examining changes in housing prices, and applying hedonic modelling. London’s Docklands Light Railway (DLR) is selected as a case study. The findings show that residential properties in station catchment areas of the south-eastern and northern branches of the DLR have a premium respectively of 0.352% and 0.093% per 100m proximity to a station. By assessing inclusive variables, it is found that housing value is also determined by the property’s own features and the neighbourhood, particularly the property tenure. The detected housing value uplift implies that the provision of good public transport is crucial in policy making, especially in areas with poor public transport accessibility. Moreover, since transport investment is partly capitalised in housing projects, we suggest that future transport infrastructure investment and housing development should be set within a framework of value capture policies.

Representation of videos is essential since it conveys an understanding of video content and enables many higher level tasks to be tackled efficiently. However, it is challenging to propose a rational representation for complex event videos, as most video information is either noisy or redundant. In this paper, we propose a compact event representation method that can concisely describe the inner modes of events. We deem that an optimal event representation scheme should reflect the long-term and high-level visual semantics (visual topics) of events, so different from previous frame-level video semantics representation methods and concept-based video representation methods, we investigate the problem from the perspective of segment-level video representations. We then present three appealing properties of segment-level visual semantics. Based on the observation, we propose different algorithms that rely on a novel deep-visual-word-based video encoding method to discover latent discriminative patterns of events. Finally, our multi-mode event representation is obtained by concatenating the discovered patterns as inner modes. We adopt our event representation for representative event parts mining, which can highlight the visual topics of events and remarkably prune the raw videos. We validate our event representation method based on complex event detection task. Experimental results on two standard benchmarking datasets, MED11 and CCV Dataset, show that the proposed method can significantly outperform the state-of-the-art approaches.

This paper describes a 22-to-36.8 GHz low phase noise Colpitts voltage-controlled oscillator (VCO) array using 0.13-μm SiGe BiCMOS technology. The VCO array consists of four Colpitts VCO cores and three millimeter-wave (MMW) selectors. In the VCO core, in order to achieve the best phase noise and power consumption trade-off, an optimization method based on symbolic expressions is presented. In a further research, a switchable bias current technique is proposed to achieve lower phase noise. The MMW selectors are designed to select the desired VCO core and cover the whole frequency band with sufficient output power. The fabricated VCO array features a wide tuning range of 50.3%, low phase noise between −100.7 dBc/Hz and −95.3 dBc/Hz at 1 MHz offset and high output power of 4.5 dBm, which results in the FOMT between −183.2 dBc/Hz and −177.8 dBc/Hz.

In this paper, a nonparametric kernel prediction algorithm in machine learning is applied to predict CO2 emissions. A literature review has been conducted so that proper independent variables can be identified. Traditional parametric modeling approaches and the Gaussian Process Regression (GPR) algorithms were introduced, and their prediction performance was summarized. The reliability and efficiency of the proposed algorithms were then demonstrated through the comparison of the actual and the predicted results. The results showed that the GPR method can give the most accurate predictions on CO2 emissions.

Deep Reinforcement Learning (DRL) has recently obtained considerable attentions. It empowers Reinforcement Learning (RL) with Deep Learning (DL) techniques to address various difficult tasks. In this paper, a novel approach called the Genetic Algorithm of Neuron Coverage (GANC) is proposed. It is motivated for improving the robustness and performance of a DRL network. The GANC uses Genetic Algorithm (GA) to maximise the Neuron Coverage (NC) of a DRL network by producing augmented inputs. We apply this method in the self-driving car applications, where it is crucial to accurately provide a correct decision for different road tracking views. We evaluate our method on the SYNTHIA-SEQS-05 databases in four different driving environments. Our outcomes are very promising – the best driving accuracy reached 97.75% – and are superior to the state-of-the-art results.

BACKGROUND: This study sought to examine fluid resuscitation in septic shock patients by monitoring their sublabial point of care microcirculation score (POEM) scores (a 3.5 cut-off value was used as the end point of recovery). It also sought to explore the effectiveness and safety of using the POEM score in the fluid resuscitation of septic shock. METHODS: Patients were randomly allocated to the experimental group or the control group. In the experimental group, a POEM score >3 was used as the end point of fluid resuscitation. In the control group, the doctor just monitor, don't know the data. Patients' heart rates, mean arterial pressure (MAP), Acute Physiology and Chronic Health Disease Classification System II (APACHE II) scores, Sequential Organ Failure Assessment (SOFA) scores, and oxygenation index scores were recorded at 2, 24, 48, 72 h, and on the 7th day after admission to the study. Statistically significant differences between the 2 groups were examined. RESULTS: Thirty-one septic shock patients (comprising 14 patients in the experimental group and 17 patients in the control group) participated in our study. Patients' parameters upon admission to the study, including MAP, blood lactate and APACHE score, SOFA score, POEM score, cardiac output (CO), and central venous pressure (CVP), were recorded at 2 h; There was no significant difference in the APACHE II scores, SOFA scores, and oxygenation index scores at 48 h between the 2 groups; however, at 72 h, the scores of the experimental group were significantly better than those of the control group (P

In recent years, online education technology has developed rapidly, and the market scale of online education platform is growing. However, with the widespread application and development of this technology, the requirements of online education users are gradually becoming stricter. Personalized course recommendation system can greatly improve the satisfaction and learning efficiency of student users, making the education platform stand out from many competitors. Based on the analysis of the advantages and disadvantages of the existing learning resource recommendation system on the online education platform, this paper proposes a new learning resource recommendation model based on hybrid recommendation algorithm, which includes course recommendation submodule based on statistics and personalized course recommendation submodule based on professional training requirements. It can not only help users to find high-quality information that they are interested in, save users' time cost, but also effectively solve the problem of cold boot. This paper provides inspiration for the improvement of online education course recommendation system.

Objective Sleep breathing disorder may affect skeletal muscle decline in the elderly, but the mechanism is not clear. Therefore, this study explores the mechanism of skeletal muscle aging in chronic intermittent hypoxia (CIH) rats. Methods In vitro and in vivo CIH models were constructed in L6 cells and SD rats by treating chronic intermittent hypoxia. Pathological changes of skeletal muscle in vivo were measured by hematoxylin-eosin (HE) staining. Cell proliferation and apoptosis were detected by CCK-8 and Flow cytometer, respectively. The expression of KLC1/GRX1 and the proteins related to the Wnt/β-catenin pathway were measured by qRT-PCR and western blot. Results CIH model was successfully established induced by chronic intermittent hypoxia with lower skeletal muscle index (SMI), increased inward migration of muscle fiber cell nucleus, and muscle cells’ distance. The results showed that Wnt/β-catenin signalling was activatedin both L6 cells and CIH rats’ model. KLC1 and GRX1 were significantly downregulated in the CIH model. Loss of function showed that downregulation of KLC1 promoted L6 cell and skeletal muscle aging in vitro and in vivo, respectively. Conclusion Our results demonstrated that CIH aggravated skeletal muscle aging by down-regulating KLC1/GRX1 expression via the Wnt/β-catenin pathway.

Organic semiconductor small molecules IHIC and ITIC have been developed as solar cell materials, and because of their strong near-infrared absorption capabilities, they are promising for cancer phototherapy. This article reports the application of semiconductor small molecule IHIC/ITIC liposomes in photothermal therapy and photoacoustic imaging of tumors firstly. Experiments show that the liposome-loaded IHIC/ITIC material has good biocompatibility and can be effectively enriched in tumor sites. After being irradiated with laser, it can emit strong photoacoustic signals, and has achieved good results in the photothermal treatment of breast cancer mice. We believe that organic semiconductor small molecule IHIC/ITIC will become a promising photothermal agent with wonderful development possibilities.

With the development of deep neural networks (DNNs) and the publicly available source code repositories, deep code comment generation models have demonstrated reasonable performance on test datasets. However, it has been confirmed in computer vision (CV) and natural language processing (NLP) that DNNs are vulnerable to adversarial examples. In this paper, we investigate how to maintain the performance of the models against these perturbed samples. We propose a simple, but effective, method to improve the robustness by training the model via data augmentation. We conduct experiments to evaluate our approach on two mainstream sequence-sequence (seq2seq) architectures which are based on the LSTM and the Transformer with a large-scale publicly available dataset. The experimental results demonstrate that our method can efficiently improve the capability of different models to defend the perturbed samples.

As the key technology of the new generation IP network, Segment Routing IPv6 (SRv6) has gradually become the direction of network evolution of the 5G and cloud networks. While it brings powerful network programming capabilities, due to IPv6’s long 128 bits address native SRv6 also introduces some new problems, excessive overhead, low bearing efficiency than IPv4, and hard to migrate without replacing hardware. In order to fix those problems, we proposed a Generalized Segment Identifier (G-SID) header compression solution based on common prefix stitching. In forwarding plane, we created some new ideas, G-SID container, COC flavor, and Segment Identifier (SID) index, and with these new things we further created the solution which support mixed deployment of compressed and uncompressed SRv6 networks elements. And the end, we have also verified function and the interoperability in multi-vendor scenario. All in all, by creating G-SID, we achieve the same overhead as SR-MPLS and retain all the advantages of existing SRv6. By conducting multi-vendor interoperability tests, we lay down solid foundation for the implementation of SRv6 the existing network.

The advent of the era of big data has brought opportunities for the application of artificial intelligence and mental health. From "virtual psychotherapists" to social robots for dementia and autism treatment to robots for treatment of disorders, AI is innovating traditional models of mental illness prevention and treatment with high levels of treatment and interventions. This paper summarizes the research progress of artificial intelligence in mental illness group, including the current situation of the application of mental illness prevention, diagnosis, treatment and nursing, and discusses the advantages, disadvantages and prospects of the application of artificial intelligence in the field of mental illness, hoping to provide reference for the sustainable development of this field.

History and experience of the international community show that long-term aging has a huge impact on both economic and social development. China's ageing population is rising, and the country faces a "getting old before getting rich" and "getting old before getting prepared" situation. If effective measures are not taken in time, the impact of aging on China's economy may be more severe than in other countries. With the increasingly mature application of Internet technology, artificial intelligence and Internet of Things technology are more and more applied in the field of health management. Relying on the Internet and the Internet of Things, artificial intelligence provides real-time, safe and fast intelligent elderly care services for the elderly through intelligent, structured, classified and integrated health data of the elderly.

Background A large number of research studies have explored the health effects of exposure to atmospheric particulate matter. However, limited quantitative evidence has linked specific sources of personal PM2.5 directly to adverse health effects. This study was conducted in order to examine the association between airway inflammation and personal exposure to PM2.5 mass, components, and sources among two healthy cohorts living in both urban and rural areas of Beijing, China. Methods We conducted a follow-up study during the summer of 2016 and the winter of 2016/2017 among 92 students and 43 guards. 24-h personal and ambient exposure to PM2.5 and fractional exhaled nitric oxide (FeNO) were measured at least twice for each participant. Chemical components of 385 personal PM2.5 exposure samples were analyzed, and pollution sources were resolved by a positive matrix factorization (PMF) receptor model. We have constructed linear mixed effect models to evaluate the association between ambient/personal PM2.5 mass, chemical constituents, and source specific PM2.5 with FeNO after controlling for temperature, relative humidity, sites, season, and potential individual confounders. Results Interquartile range (IQR) increase in household heating sources was associated with increased FeNO (2.72%; 95% CI = 1.26–4.17%) across two sites. IQR increase in roadway transport was associated with increased FeNO (9.84%; 95% CI = 2.69–17%) in urban areas; IQR increase in Secondary inorganic sources and Industrial/Combustion sources were associated with increased FeNO (7.96%; 95% CI = 1.47–14.4%% and 7.85%; 95% CI = 0.0676–15.6%, respectively) in rural areas. Personal exposure to EC, OC, and some trace elements (Se, Pb, Bi, Cs) were also estimated to be significantly associated with the increase of FeNO. In addition, there was no significant difference (P > 0.05) between the effects of ambient and personal PM2.5 mass. Conclusions Although personal PM2.5 mass was not significantly associated with the health effects, airway inflammation can be linked to source-resolved exposures.

Finger Texture (FT) images acquired from different spectral lighting sensors reveal various features. This inspires the idea of establishing a recognition model between FT features collected using two different spectral lighting forms to provide high recognition performance. This can be implemented by establishing an efficient feature extraction and effective classifier, which can be applied to different FT patterns. So, an effective feature extraction method called the Surrounded Patterns Code (SPC) is adopted. This method can collect the surrounded patterns around the main FT features. It is believed that these patterns are robust and valuable. The SPC approach proposes using a single texture descriptor for FT images captured under multispectral illuminations, where this reduces the cost of employing different feature extraction methods for different spectral FT images. Furthermore, a novel classifier termed the Re-enforced Probabilistic Neural Network (RPNN) is proposed. It enhances the capability of the standard Probabilistic Neural Network (PNN) and provides better recognition performance. Two types of FT images from the Multi-Spectral CASIA (MSCASIA) database were employed as two types of spectral sensors were used in the acquiring device: the White (WHT) light and spectral 460 nm of Blue (BLU) light. Supporting comparisons were performed, analysed and discussed. The best results were recorded for the SPC by enhancing the Equal Error Rates (EERs) at 4% for spectral BLU and 2% for spectral WHT. These percentages have been reduced to 0% after utilizing the RPNN.

Event detection, which targets the detection of complex events among numerous videos, has attracted growing interest recently. Previous approaches suffered from huge computation costs in multiple feature extraction and classification process. Lately, a discriminative CNN video representation method for event detection is proposed to obtain promising performances. However, this method samples the video frames uniformly for global video representations, without considering that some video parts might be redundant or noisy for the task. Though a multirate sampling solution is proposed later in consideration of the video content motion speed variation, it remains uncertain and unclear that which video part is more important to define the event. In this paper, we propose to mine effective parts for event detection and understanding. After video segmentation, we try to mine the definite parts (Event Patches) that mostly contribute to defining an event. We evaluate our event patches on the TRECVID MED 2011 dataset. Compared with CNN video representation method, which has been recognized as the best video representation for event detection, our method improves the Mean Average Precision (mAP) from 74.7% to 76.2%.

A versatile optofluidic dye laser based on the sub-micron thickness resonator embedded in the Kagome photonic crystal fiber is demonstrated. The hexagonal hollow core supported by the cladding scaffolds serves as the microcavity and microfluidic channel for gain medium in the meantime, offering good compactness and robustness. Lasing emission could be achieved by employing dyes with refractive index either lower or higher than silica background. The lasing thresholds for dye LDS751 dissolved in methanol or DMSO at 0.2 mM are 0.43 and $0.51~\mu \text{J}$ , respectively. In addition, the lasing peaks exhibit obvious redshift as the refractive index of the gain medium increases, which is in agreement with our theoretical expectation.

We present here a capacitive model for the coulometric signal transduction readout of solid-contact ion-selective membrane electrodes (SC-ISE) with a conducting polymer (CP) as an intermediate layer for the detection of anions. The capacitive model correlates well with experimental data obtained for chloride-selective SC-ISEs utilizing poly(3,4-ethylenedioxythiophene) (PEDOT) doped with chloride as the ion-to-electron transducer. Additionally, Prussian blue is used as a simple sodium capacitor to further demonstrate the role of the transduction layer. The influence of different thicknesses of PEDOT as a conducting polymer transducer, different thicknesses of the overlaying ion-selective membranes deposited by drop casting and spin coating, and different compositions of the chloride-selective membrane are explored. The responses are evaluated in terms of current-time, charge-time, and charge-chloride activity relationships. The utility of the sensor with coulometric readout is illustrated by the monitoring of very small concentration changes in solution.

Background and objective Current retinal prostheses can only generate low-resolution visual percepts constituted of limited phosphenes which are elicited by an electrode array and with uncontrollable color and restricted grayscale. Under this visual perception, prosthetic recipients can just complete some simple visual tasks, but more complex tasks like face identification/object recognition are extremely difficult. Therefore, it is necessary to investigate and apply image processing strategies for optimizing the visual perception of the recipients. This study focuses on recognition of the object of interest employing simulated prosthetic vision. Method We used a saliency segmentation method based on a biologically plausible graph-based visual saliency model and a grabCut-based self-adaptive-iterative optimization framework to automatically extract foreground objects. Based on this, two image processing strategies, Addition of Separate Pixelization and Background Pixel Shrink, were further utilized to enhance the extracted foreground objects. Results i) The results showed by verification of psychophysical experiments that under simulated prosthetic vision, both strategies had marked advantages over Direct Pixelization in terms of recognition accuracy and efficiency. ii) We also found that recognition performance under two strategies was tied to the segmentation results and was affected positively by the paired-interrelated objects in the scene. Conclusion The use of the saliency segmentation method and image processing strategies can automatically extract and enhance foreground objects, and significantly improve object recognition performance towards recipients implanted a high-density implant.

We proposed and demonstrated a high sensitivity temperature sensor in an avoided-crossing-based selective-filling photonic crystal fiber (SFPCF) which was designed by selectively filling two symmetrical air holes around the fiber core of an index-guiding PCF with a kind of high index liquid. The avoided-crossing effect caused by the resonant couplings rather than the crossings between the two-high-index-rod modes and the fundamental mode results in unique birefringence characteristics. The temperature response of the modal phase and group birefringence were both calculated in detail. And the phase and group birefringence are highly sensitive to temperature. Then, the transmission characteristics of a Sagnac interferometer (SI) based on the SFPCF were discussed both in theory and in experiments. The experimental results present that the weak mode couplings have no influence on the transmission spectrum of the SFPCF based SI. The interference dips at different wavelengths present distinctive transmission and temperature characteristics. In experiment, the highest temperature sensitivity of -6.9 nm/°C at 36.5°C was achieved.

A selective-filling high birefringent photonic crystal fiber (SF-HBPCF) based Sagnac interferometer (SI) was demonstrated. The SF-HBPCF was achieved by infiltrating a high index liquid into two symmetrical air holes of the innermost layer of an index-guiding un-birefringent PCF. The birefringence characteristics of the SF-HBPCF and the strain sensing characteristics of the SF-HBPCF based SI were theoretically analyzed in detail. The group birefringence presented unique characteristics, and particularly possessed a zero value at a certain wavelength. This directly resulted in the strain sensitivities having ultrahigh even infinite value at the certain wavelength. Besides, with the change of the loaded strain, the sensitivities of the interference dips presented disparate variation trend. In experiments, the results were well matched with theoretical simulation. And the strain sensitivities from 25 pm/με to 12 pm/με were achieved from 61 με to 789 με in the type of the SF-HBPCF.

We examined the separate effect of Portuguese oyster Crassostrea angulata and the interactive effects of oyster and red seaweed Gracilaria lemaneiformis on seawater dissolved inorganic carbon (DIC) systems and the air-sea CO2 flux (FCO2) in Daya Bay, southern China. Respiration and calcification rates of oysters were measured and the effects of oyster aquaculture on marine DIC systems were evaluated. The interactive effects on seawater DIC and air-sea FCO2 were examined using mesocosms containing oyster and seaweed assemblages. Results showed populations of C. angulata cultured in Daya Bay sequestered ca. 258 g C m-2 yr-1 for shell formation, whereas the CO2 released due to respiration and calcification was 349 and 153 g C m-2 yr-1, respectively. This indicates that oyster cultivation in Daya Bay is a CO2 generator, favoring the escape of CO2 into the atmosphere. DIC, HCO3- and CO2 concentrations and the partial pressure of CO2 in oyster-seaweed co-cultured mesocosms were significantly lower than the oyster monoculture mesocosm. These results indicated that G. lemaneiformis effectively absorbs the CO2 released by oysters. The negative values of air-sea FCO2 in the co-culture groups represent a CO2 sink from the atmosphere to the sea. These results demonstrated that there could be an interspecies mutual benefit for both C. angulata and G. lemaneiformis in the integrated culture system. Considering that photosynthesis of seaweed is carbon limited, we suggest that the 2 species are co-cultured at a ratio of ca. 4:1 (based on fresh weight) for efficient utilization of DIC in seawater by G. lemaneiformis, and further to increase the ocean CO2 sink.

6PolSK-QPSK is a promising modulation format in optical fiber communication. Because of the damage suffered during the transmission and reception, a series of algorithms are needed to be adopted to recover the original data. We proposed a novel quadrature imbalance compensation algorithm based on the data statistical properties. Simulation results show that the quadrature imbalance can be well compensated with the proposed algorithm.

Photoelectrochemical water splitting is a promising method to produce H 2 by making use of solar energy. In this paper we report on a photocathode made by p-type crystalline Si covered with an n-type 3C-SiC polycrystalline film, acting as protective layer and transparent emitter. The photoelectrodes exhibit a saturated photocurrent above 30 mA cm −2 . No decay is observed after 9 h under constant current stress at 1 kW m −2 with AM1.5G spectrum illumination. Improvement of the photocurrent value is achieved by covering the 3C-SiC emitter with Au or Pt nanoparticles. Under suitable metal nanoparticles deposition conditions, compared to the samples without nanoparticles, two major effects are observed: first the onset potential is considerably reduced, and second, higher saturated photocurrent is found, up to 38 mA cm −2 , i.e. with a 27% increase. Optical and micro-structural studies on the nanoparticles provide insights on the origin of the observed effects.

Lipoprotein-associated phospholipase A2 (Lp-PLA2) is an independent risk factor for cardiovascular disease. Accordingly, studies from many countries around the world have shown an association between Lp-PLA2 and cardiovascular events in patients with acute coronary syndrome (ACS) undergoing percutaneous coronary intervention (PCI), but this association has not been documented among the Chinese. The aim of this study was to assess the use of Lp-PLA2 as a useful marker for predicting the long-term prognosis of Chinese patients with ACS undergoing PCI. A total of 651 consecutive patients undergoing PCI between September 2013 and January 2015 were divided into 2 groups: the high Lp-PLA2 group (n = 262, Lp-PLA2 > 138 nmol/L) and the low Lp-PLA2 group (n = 389, Lp-PLA2 ≤ 138 nmol/L). The end point was all-cause mortality and rehospitalization. The median follow-up was 24 months. Multivariate analysis showed that high Lp-PLA2 was an independent predictor of all-cause mortality and rehospitalization (hazard ratio: 1.429, 95% confidence interval [CI]: 1.411-1.448; P < .05). The Lp-PLA2 had good accuracy for predicting all-cause mortality and rehospitalization among patients with ACS undergoing PCI (area under the receiver–operating characteristic curve: 0.858, 95% CI: 0.819-0.898; P < .05), and a good correlation with the Global Registry of Acute Coronary Event score ( r = 0.525, P < .05). This study provided evidence that Lp-PLA2 could predict all-cause mortality and rehospitalization risk among patients with ACS undergoing PCI.

Current retinal prostheses can only generate low-resolution visual percepts constituted of inadequate phosphenes which are elicited by a limited number of stimulating electrodes and with unruly color and restricted grayscale. Fortunately, for most retinal prostheses, an external camera and a video processing unit are employed to be essential components, and allow image processing to improve visual perception for recipients. At present, there have been some studies that use a variety of sophisticated image processing algorithms to improve prosthetic vision perception. However, most of them cannot achieve real-time processing due to the complexity of the algorithms and the limitation of platform processing power. This greatly curbs the practical application of these algorithms on the retinal prostheses. In this study, we propose a real-time image processing strategy based on a novel bottom-up saliency detection algorithm, aiming to detect and enhance foreground objects in a scene. Results demonstrate by verification of conducting two eye-hand-coordination visual tasks that under simulated prosthetic vision, our proposed strategy has noticeable advantages in terms of accuracy, efficiency, and head motion range. The study aims to help develop image processing modules in retinal prostheses, and is hoped to provide more benefit towards prosthesis recipients.

We demonstrate a novel multi-color switchable dye laser based on cavity-assisted energy transfer induced by microfluidic droplets flowing in side-hole microstructured optical fibers that provide dual channels for, respectively, loading donor and acceptor and offer the high-Q microring resonator for effective optical feedback. The droplet generator produces periodical acceptor droplets flowing by the focus point to activate color switching thanks to the cavity-assisted energy transfer effect. The lasing thresholds of rhodamine B at 1 mM and LDS730 at 1.5 mM are, respectively, 233 and 369 nJ when the flow rates of the LDS730 solution and the carrier fluid are kept at 5 μL/min. The influence of flow rate on switching time of dual-color lasing emission is also investigated. And moreover, experimental results indicate that by varying the acceptor concentration the proposed dye laser could operate in uni-color or dual-color laser switching mode. The compact multi-color switchable dye laser provides a promising candidate to serve as miniaturized laser for biochemical and color display applications.

To explore the prognostic value of the enhanced International Prognostic Index (NCCN-IPI) for Asian patients with diffuse large B-cell lymphoma (DLBCL) treated in the rituximab era.We performed a retrospective analysis of 176 patients with newly diagnosed DLBCL. The estimated overall survival (OS) and progression-free survival (PFS) of the different risk groups were discriminated by the International Prognostic Index (IPI), the revised International Prognostic Index (R-IPI) and the NCCN-IPI.With a median follow-up of 18 months, at 3 years, the OS was 73% and the PFS was 65%. The 3-year OS for the 4 NCCN-IPI risk groups were 91% versus 80% versus 57% versus 45% (P0.001); the 3-year PFS were 77% versus 72% versus 56% versus 26% (P0.001). The 3-year OS of the 4 risk groups discriminated by the IPI ranged from 85-55% (P0.001); the 3-year PFS ranged from 81-41% (P0.001). The 3-year OS of the 3 distinct prognostic groups by the R-IPI ranged from 86-51% (P0.001); the 3-year PFS ranged from 86-47% (P0.001). The 3-year OS and PFS of the high-risk group according to the NCCN-IPI were lower than the IPI and R-IPI. Using the NCCN-IPI, the outcomes among the risk groups spanned a large range, and the survival of the high-risk group was significantly different from the high-intermediate risk group.The NCCN-IPI is a clinically useful prognostic index for patients with DLBCL treated in the rituximab era, especially for high-risk patients.

Benefiting from the high spectrum efficiency and power efficiency, 6PolSK-QPSK (6-ary polarization-shift keying quadrature phase-shift keying) is a promising modulation format in coherent optical communication. We proposed a novel structure to generate the 6PolSK-QPSK with two dual-drive MZMs. Simulation results show that the proposed structure can generate 6PolSK-QPSK effectively and the performance is nearly the same to the traditional one.

Memristors have shown an extraordinary potential to emulate the plastic and dynamic electrical behaviors of biological synapses and have been already used to construct neuromorphic systems with in-memory computing and unsupervised learning capabilities; moreover, the small size and simple fabrication process of memristors make them ideal candidates for ultradense configurations. So far, the properties of memristive electronic synapses (i.e., potentiation/depression, relaxation, linearity) have been extensively analyzed by several groups. However, the dynamics of electroforming in memristive devices, which defines the position, size, shape, and chemical composition of the conductive nanofilaments across the device, has not been analyzed in depth. By applying ramped voltage stress (RVS), constant voltage stress (CVS), and pulsed voltage stress (PVS), we found that electroforming is highly affected by the biasing methods applied. We also found that the technique used to deposit the oxide, the chemical composition of the adjacent metal electrodes, and the polarity of the electrical stimuli applied have important effects on the dynamics of the electroforming process and in subsequent post-electroforming bipolar resistive switching. This work should be of interest to designers of memristive neuromorphic systems and could open the door for the implementation of new bioinspired functionalities into memristive neuromorphic systems.

Building a food traceability system provides an effective way solving the problems arousing in the food safety domain. Blockchain is preferred to construct a commodity traceability system due to its innate immutability and consistency of stored data, maintained through cryptographic means and consensus mechanism. While current blockchain based traceability system just collects and stores verified commodity information in the chain, which is composed of a decentralized ledger architecture. But the information is less accurate or intact which cannot trace the real food source along the food supply-chain. We propose to establish a Hyperledger-based Food Trading and Traceability system called Hyper-FTT, by aggregating all the providers including food warehousing enterprises, food processing enterprises and food retails to reach agreements and conclude business transactions on the chain, then an unbroken food supply-chain can be formed to provide trusted food tracing. Implementation and experiments are conducted to evaluate the performance of the proposed demonstration system.

In this paper, the detailed strain characteristics of a selective-filling birefringent photonic crystal fiber (SF-PCF) were analyzed theoretically. The SF-PCF was achieved by selectively filling two symmetrical air holes around the fiber core of a PCF with a high index liquid. The group birefringence and phase birefringence presented unique characteristics different from the traditional index-guiding birefringence fiber. Hence, the transmission and sensing characteristics of the Sagnac interferometer based on the SF-PCF presented diversity and high sensitivity. Finally, the highest strain sensitivity of −9.56 pm/μe was theoretically achieved.

Deep reinforcement learning has received wide attentions recently. It combines deep learning with reinforcement learning and shows to be able to solve unprecedented challenging tasks. This paper proposes an efficient approach based on deep reinforcement learning to tackle the road tracking problem arisen from self-driving car applications. We propose a new neural network which collects input states from forward car facing views and produces suitable road tracking actions. The actions are derived from encoding the tracking directions and movements. We perform extensive experiments and demonstrate the efficacy of our approach. In particular, our approach has achieved 93.94% driving accuracy, outperforming the state-of-the-art approaches in literature.

Introduction The main objective of this study was to explore the mechanism of nicotine improving cognitive impairments in ischemic rats. Methods Twenty adult male Sprague–Dawley (SD) rats underwent ischemic model surgery by injecting endothelin-1 into the left thalamus, which were classified into four different groups with different intervention: nicotine (1.5 mg/kg/d), dihydro-β-erythroidine (DHβE; 3 mg/kg/d), nicotine (1.5 mg/kg/d) + DHβE (3 mg/kg/d), or saline, after ischemic model surgery. Another five male SD rats also underwent same surgery, while not injecting endothelin-1 but saline, as the control group. Morris water maze (MWM) test was adopted to assess the cognition. All the rats underwent the MWM test, micro positron emission tomography imaging with 2-[18F]-A-85380, and messenger RNA (mRNA) test of α 4 nicotinic acetylcholine receptor (nAChR), β 2 nAChR, tumor necrosis factor-alpha (TNF-α), IL-1β, and IL-6. Results The MWM test showed the rats given nicotine showing better memory than ischemic rats (p < .05), whereas the rats given DHβE or both nicotine and DHβE did not show any statistical difference from the ischemic rats (p > .05). Micro positron emission tomography imaging showed higher uptake of tracer in the left thalamus and whole brain in rats given nicotine than in ischemic rats, but the rats given DHβE or both nicotine and DHβE did not. By real-time PCR test, the mRNA of α 4 nAChR and β 2 nAChR in rats given nicotine increased significantly compared with ischemic rats and decreased TNF-α, IL-1β, and IL-6 mRNA (all ps < .05). Conclusions By activating α 4β 2 nAChRs, nicotine plays a role in inhibiting the inflammatory factors, which contributes to improving cognitive impairment in ischemic rats. Implications It is well acknowledged that vascular cognitive impairment (VCI) is the second most common cause of dementia after Alzheimer’s disease. Cholinergic agents have potential for the symptomatic treatment of the cognitive symptoms of dementia, but the exact mechanism still remains unclear. There are potential complex associations and interactions between VCI and inflammation. This study showed that nicotine had anti-inflammatory potency, which is most likely because of the activation of the nAChRs. By activating α4β2 nAChRs, nicotine played a role in inhibiting the inflammatory factors, which contribute to improving cognitive impairment in ischemic rats.

This work investigates the different sensitivities of an ion-sensitive field-effect transistor (ISFET) based on fully depleted silicon-on-insulator (FDSOI). Using computer-aided design (TCAD) tools, the sensitivity of a single-gate FDSOI based ISFET (FDSOI-ISFET) at different temperatures and the effects of the planar dual-gate structure on the sensitivity are determined. It is found that the sensitivity increases linearly with increasing temperature, reaching 890 mV/pH at 75 °C. By using a dual-gate structure and adjusting the control gate voltage, the sensitivity can be reduced from 750 mV/pH at 0 V control gate voltage to 540 mV/pH at 1 V control gate voltage. The above sensitivity changes are produced because the Nernst limit changes with temperature or the electric field generated by different control gate voltages causes changes in the carrier movement. It is proved that a single FDSOI-ISFET can have adjustable sensitivity by adjusting the operating temperature or the control gate voltage of the dual-gate device.

Background. Pulmonary tuberculoma can mimic lung malignancy and thereby pose a diagnostic dilemma to clinicians. The purpose of this study was to establish an accurate, convenient, and clinically practical model for distinguishing small-sized, noncalcified, solitary pulmonary tuberculoma from solid lung adenocarcinoma. Methods. Thirty-one patients with noncalcified, solitary tuberculoma and 30 patients with solid adenocarcinoma were enrolled. Clinical characteristics and CT morphological features of lesions were compared between the two groups. Multivariate logistic regression analyses were applied to identify independent predictors of pulmonary tuberculoma and lung adenocarcinoma. Receiver operating characteristic (ROC) analysis was performed to investigate the discriminating efficacy. Results. The mean age of patients with tuberculoma and adenocarcinoma was 46.8 ± 12.3 years (range, 28–64) and 61.1 ± 9.9 years (range, 41–77), respectively. No significant differences were observed concerning smoking history and smoking index, underlying disease, or tumor markers between the two groups. Univariate and multivariate analyses showed age and lobulation combined with pleural indentation demonstrated excellent discrimination. The sensitivity, specificity, accuracy, and the area under the ROC curve were 87.1%, 93.3%, 90.2%, and 0.956 (95% confidence interval (CI), 0.901–1.000), respectively. Conclusion. The combination of clinical characteristics and CT morphological features can be used to distinguish noncalcified, solitary tuberculoma from solid adenocarcinoma with high diagnostic performance and has a clinical application value.

// Ting-Ting Han 1,* , Shuang Zheng 1,* , Zeng-Ai Chen 2,* , Wei Liu 1 and Yao-Min Hu 1 1 Department of Endocrinology, Renji Hospital, School of Medicine, Shanghai Jiaotong University, Shanghai, China 2 Department of Radiology, Renji Hospital, School of Medicine, Shanghai Jiaotong University, Shanghai, China * Co-first author Correspondence to: Wei Liu, email: // Yao-Min Hu, email: // Keywords : POEMS syndrome; thrombosis; monoclonal protein; young patient; atypical feature Received : August 15, 2016 Accepted : October 04, 2016 Published : October 11, 2016 Abstract POEMS syndrome is a rare multi-systemic disease characterized by polyneuropathy, organomegaly, endocrinopathy, monoclonal protein and skin changes. Arterial thrombosis is a distinctively unusual feature in patients with POEMS syndrome. We report a 33-year-old man with intermittent amaurosis of left eye and skin changes as the onset manifestations, who was finally confirmed as having POEMS syndrome. Most notably, this was a young man without high risk factors of arterial thrombosis and no monoclonal protein was detected until the repeated measurement later. This case evokes the need to consider the diagnosis of POEMS syndrome for young patients with symptoms of arterial thrombosis but no high risk factors of thrombosis.

Image restoration is a highly ill-posed problem and requires to be regularized. Many common image priors aim to make full use of natural image prior information. Total variation (TV) regularize prior has good performance of preserving edges but also has drawbacks in arising In this paper, we propose a total variation based image restoration method using hyper-Laplacian prior for image gradient and the overlapping group sparsity prior for sparser image representation constraint. We adopt the alternating direction method of multipliers (ADMM) method to optimize the object function of the proposed model and discuss the parameter selection criterion in the complex formulation. Finally, we carry out experiments on various degrade images and compare our method with several classical state-of-the-art methods. Experimental results show that our method has good performance in convergence and suppressing staircase artifacts, which makes a good balance between alleviating staircase effects and preserving image details. A total variation based image restoration method is proposed that using hyper-Laplacian prior for image gradient and the overlapping group sparsity prior for sparser image representation constraint.The alternating direction method of multipliers (ADMM) method is adopted to optimize the complex object function of the proposed model. Parameter selection criterions are discussed for the complex formulation.The method has good performance in convergence and suppressing staircase artifacts, which makes a good balance between alleviating staircase effects and preserving image details.An image restoration method using hyper-Laplacian prior and overlapping group sparsity.

Some scholars' studies have demonstrated that Pro-kin balance system training is able to promote the recovery of the balance function in stroke patients. The present study has expanded on those studies, and was not merely limited to studying balance, but also encompassed walking and self-care abilities of the patients; furthermore, the association among balance and walking and self-care abilities was also explored.A total of 40 stroke patients were randomly and equally divided into 2 groups: the control group (n = 20) and the treatment group (n = 20). Both groups underwent conventional balance training, although the treatment group also underwent visual feedback balance training with the Pro-kin system. The balance function was assessed using the Berg Balance Scale (BBS), the Timed "Up & Go" (TUG) test, and Pro-kin system parameters. The Pro-kin system parameters included the perimeter and ellipse area, which were both tested once with eyes open (EO) and eyes closed (EC). Walking ability was assessed using the Holden Walking Ability Scale, according to the Functional Ambulation Classification (FAC). The self-care abilities were assessed with the Barthel Index (BI). The tests were conducted prior to training, and 3 weeks after the end of the training programme.No significant differences were noted among the groups before the training. After 3 weeks of training, for both the groups, significant improvements in balance and the walking and self-care abilities were noted: The BBS value was significantly increased (P 

Machine vision is an indispensable part of today's artificial intelligence. The artificial visual systems used in industrial production and domestic daily life rely significantly on cameras and image‐processing components for live monitoring and target identifying. They, however, often suffer from bulky volume, high energy consumption, and more critically, lack of adaptive responsiveness under extreme lighting conditions and thus possible mortal visual disability of flash blinding or nyctalopia for applications such as auto‐piloting. Herein, it is demonstrated that perovskite switchable photovoltaic devices are used to effectively construct all‐in‐one sensory neural network. Arising from the spontaneous and electric field‐induced ion‐migration effect, the photoresponsivity of the perovskite device can be reconfigured over the wide range of 540–1270%, which not only allows high‐fidelity adaptive image sensing of the visual information but also acts as updatable synaptic weight to enable the sensor array for performing machine‐learning tasks. With the bioinspired electronic pupil regulation function achieved through adjustable photoresponsivity of the perovskite sensor array, a proof‐of‐concept adaptive machine vision system with a maximum 263% enhancement of the object recognition accuracy for compact, mobile yet delay‐sensitive applications is demonstrated.

With the breakthroughs in general action understanding, it has become an inevitable trend to analyze the actions in finer granularity. However, related researches have been largely hindered by the lack of fine-grained datasets and the difficulty of capturing subtle differences between fine-grained actions that are highly similar overall. In this paper, we address the above challenges by constructing a fine-grained action dataset, i.e., Figure Skating, which can be used for end-to-end network training and presenting a framework for the joint optimization of classification and similarity constraints. We propose to incorporate the triplet loss into the training of Convolutional Neural Network, which learns a mapping from fine-grained actions to a compact Euclidean space where distances directly correspond to a measure of action similarity. Triplet loss compels actions of distinct classes to have larger distances than actions of the same class. Besides, to boost the discrimination of the fine-grained actions, we further propose a temporal variance embedding network (TVENet) embedding temporal context variances into the feature embeddings during the joint network training. The experimental results on Figure Skating dataset, HMDB51 dataset as well as UCF101 dataset demonstrate the effectiveness of TVENet representation for fine-grained action search.

Finger Texture (FT) is one of the most recent attractive biometric characteristic. It refers to a finger skin area which is restricted between the fingerprint and the palm print (just after including the lower knuckle). Different specifications for the FT can be obtained by employing multiple images spectrum of lights. Individual verification systems are established in this paper by using multiple spectrum FT specifications. The key idea here is that by combining two various spectrum lightings of FTs, high personal recognitions can be attained. Four types of fusion will be listed and explained here: Sensor Level Fusion (SLF), Feature Level Fusion (FLF), Score Level Fusion (ScLF) and Decision Level Fusion (DLF). Each fusion method is employed, examined for different rules and analysed. Then, the best performance procedure is benchmarked to be considered. From the database of Multiple Spectrum CASIA (MSCASIA), FT images have been collected. Two types of spectrum lights have been exploited (the wavelength of 460 nm, which represents a Blue (BLU) light, and the White (WHT) light). Supporting comparisons were performed, including the state-of-the-art. Best recognition performance was recorded for the FLF based concatenation rule by improving the Equal Error Rate (EER) percentages from 5% for the BLU and 7% for the WHT to 2%.

Conductance quantization (QC) phenomena occurring in metal oxide based memristors demonstrate great potential for high-density data storage through multilevel switching, and analog synaptic weight update for effective training of the artificial neural networks. Continuous, linear and symmetrical modulation of the device conductance is a critical issue in QC behavior of memristors. In this contribution, we employ the scanning probe microscope (SPM) assisted electrode engineering strategy to control the ion migration process to construct single conductive filaments in Pt/HfOx/Pt devices. Upon deliberate tuning and evolution of the filament, 32 half integer quantized conductance states in the 16 G0 to 0.5 G0 range with enhanced distribution uniformity was achieved. Simulation results revealed that the numbers of the available QC states and fluctuation of the conductance at each state play an important role in determining the overall performance of the neural networks. The 32-state QC behavior of the hafnium oxide device shows improved recognition accuracy approaching 90% for handwritten digits, based on analog type operation of the multilayer perception (MLP) neural network.