基于SSD的眨眼与打哈欠检测

提出了一种新的基于SSD网络的眨眼与打哈欠检测。首先,对原SSD架构进行四个方面的修改。第一,主干网用MobileNetV3替换了VGG,使得架构尺寸缩减三分之一。第二,重新设计了输出检测器,实现了目标检测与分类、目标定位与高宽比回归多任务的融合。第三,网络直接回归高宽比,比传统的从关键点离线计算高宽比的方法鲁棒性要高。第四,省去了非最大抑制层(NMS)中的重叠度(IOU)计算和阈值设定,使得单目标检测速度提高了40%。另外,在数据集YawDD上验证了所提出方法的有效性,眨眼与打哈欠的测全率为95.1%与100.0%,在单卡2080Ti上实现了41fps的高检测速度。     

Antibacterial activity of V- doped rod-like MgO crystals decorated with nanoflake layer

In present study, we report a V doping fabrication method for obtaining rod-like MgO crystals decorated with a nanoflake layer. This novel structure has only been minimally reported in literature. Pure MgO and Mg₂V₂O₇–MgO composite materials were obtained by precipitation and impregnation methods, with vanadium added concentrations of 0–9%. The influence of V doping on crystal structure and particle morphology of MgO was investigated by scanning electron microscopy (SEM). X-ray diffraction (XRD) analysis demonstrated that MgO has a cubic structure, while X-ray photoelectron spectroscopy (XPS) revealed that V⁵⁺ exists on the surface of MgO. The specific surface areas and pore sizes of MgO composites were calculated by BET and BJH analysis. These techniques revealed that specific surface area and pore size of MgO increased due to vanadium doping. The antibacterial effects of Mg₂V₂O₇–MgO composite materials against Escherichia coli (E. coli) and Staphylococcus aureus (S. aureus) were assessed using a bacterial killing/colony-forming unit (CFU) assay and bacteriostatic ring method. Our results demonstrate that V doping dramatically improved antimicrobial properties of MgO, with 7 mol% doping inducing the best antibacterial activity. The antibacterial mechanisms of Mg₂V₂O₇–MgO composite material were also proposed.     

下扬子地区下志留统风暴沉积特征及风暴作用对泥页岩品质的影响

近年来在下扬子地区页岩气地质调查井所获取的地层岩心资料，揭示该地区下志留统风暴沉积较为发育。基于精细的岩心观察描述，详细分析了下扬子地区下志留统高家边组下段中上部风暴沉积的特征，并阐明了风暴作用对泥页岩的影响。研究表明：研究区风暴岩岩性偏细，以粉砂岩为主，石英含量较高，结构成熟度较低，具2种不同的粒度概率曲线；主要发育底部冲刷面、块状层理、平行层理、丘状交错层理、波纹交错层理、准同生变形构造等沉积构造；垂向上主要识别出4种风暴沉积序列，均具有从底到顶粒度变细的沉积特征，不同的沉积序列与风暴流能量的强弱有关。下扬子地区早志留世频繁的风暴活动与研究区的古地理背景和古气候条件具有内在联系。特大型风暴对浅海陆棚深水区的细粒泥质沉积物特征产生重要影响，包括降低有机质含量、减少单层厚度、破坏垂向连续性、改变原始沉积构造和侧向稳定性等，总体上降低了泥页岩的品质。因此，研究认为在页岩气的勘探和评价过程中应重视风暴因素对陆棚深水区原始泥质沉积物的影响。     

Co-precipitation of nano Mg–Y/ZrO2 ternary oxide eutectic system: Effects of calcination temperature

In the family of inorganic nanomaterials, zirconia is a highly promising functional ceramic with a high refractive index, hardness, and dielectric constant, as well as excellent chemical inertness and thermal stability. These properties are enhanced in nano-zirconia ceramics, because nanopowders have a small particle size, good morphology, and uniform and dispersive distribution. In this study, a co-precipitation process was proposed to synthesise highly dispersed MgO–Y₂O₃ co-stabilized ZrO₂ nanopowders. The effects of different calcination temperatures on the crystallisation degree and particle dispersion of zirconia nanopowders were characterised by X-ray diffraction (XRD), thermogravimetry-differential scanning calorimetry (TG-DSC), Raman spectroscopy, Fourier transform infrared spectroscopy (FTIR), nitrogen adsorption using the Brunauer–Emmett–Teller (BET) theory, transmission electron microscopy (TEM), and field emission scanning electron microscopy (FESEM). The optimum synthesis conditions were obtained as follows: 6 h of high-energy planetary grinding and calcination at 800 °C in an electric furnace. Under these optimum conditions, the average particle size of the prepared powder was 28.7 nm. This process enriches the literature on the controllable preparation of Mg–Y/ZrO₂ nanopowders obtained by the co-precipitation method.     

Effect of thermal treatment on the antioxidant activity of egg white hydrolysate and the preparation of novel antioxidant peptides

Heat treatment will affect the nutritional properties and potential bioactivity of food materials. The aim of this work was to evaluate the effect of different thermal treatment (4, 56, 65 and 100 ℃) and in vitro gastrointestinal digestion on the antioxidant activity of egg white hydrolysate. The results demonstrated that egg white hydrolysate treated at 65 ℃ exhibited the highest antioxidant. Remarkably, the simulated digestion significantly increased antioxidant activity of egg white hydrolysate. Furthermore, we identified twenty-four potential antioxidant peptides by performing mass spectrometry and bioinformatic analysis. Six peptides were selected based on the activity prediction score of the online tool. The results showed that P6 (ACPECPK) possessed the most outstanding antioxidant properties and had low cytotoxicity and allergenicity. Bioinformatics technology combined with biochemical assays may offer a way for discovering novel antioxidant peptides from different kinds of food under various heat treatment conditions.     

基于多元神经网络融合的分布式资源空间文本分类研究

针对实体产业对科技资源的服务需求,以服务效应作为资源文本分类标准,提出一种基于多元神经网络融合的分布式资源空间文本分类模型。设计了包含词嵌入层、卷积层、双向门控循环单元层、注意力机制层和softmax层的多元神经网络通路;在此基础上采用基于需求—效应—资源分类策略,完成了从定性科技资源需求到定量资源服务效应求解,再到定性科技资源输出的映射变换,重点解决了分布式科技资源局部和全局语义特征形式多样、文本长距离依赖特征显著、重要资源信息难以准确识别的问题,进而从分布式科技资源空间中快速准确地获取效应知识,提升实体产业产品研发效率和创新能力;通过万方专利科技资源数据集验证了所提方法的可行性和有效性,为更加全面地挖掘资源文本特征和按需服务实体产业提供了一种新的思路和手段。     

区分站点状态的两级轮询控制系统

针对提高轮询控制模型工作效率和区分网络优先级的问题，提出了区分站点忙闲状态的完全-门限两级轮询控制模型（ETTPSS）。模型以两级优先级为基础，依据站点的忙闲状态采用并行处理方式只对忙站点进行信息分组发送服务。模型既能区分传输服务优先级又能避开对无信息分组的空闲站点的查询，从而提高了模型资源利用率和工作效率。运用概率母函数与马尔可夫链相结合的方法对该模型进行理论分析研究，精确解析了模型各个重要性能参数。仿真实验结果表明，仿真值与理论值近似相等，说明理论分析正确合理。与普通轮询模型相比，该模型性能大幅度提高。     

Sn-induced CuO–CeO2 catalysts with improved performance for CO preferential oxidation in H2-rich streams

Sn modified CuO–CeO₂ catalysts with different Sn loadings were prepared by a facile, green and solvent-free method. The effect of Sn/Ce ratio over Sn–Cu–Ce-x (x = 0, 1, 2.5, 5, 7.5) samples on CO activity and O₂ selectivity was investigated. The samples were characterized by various techniques using N₂-adsorption/desorption, XRD, H₂-TPR, XPS, Raman and in-situ DRIFTS. It was revealed that stronger interaction between acitve sites and support, higher amounts of Sn²⁺ and Ce³⁺, associated with increased amount of oxygen vacancies, were observed on the catalyst of Sn–Cu–Ce-5. As a result, the optimized catalyst displayed an excellent catalytic performance even in the presence of CO₂ and H₂O. In this sense, probing the Sn modified CuO–CeO₂ catalyst can elucidate some useful keys for the development of high CO₂ and H₂O-resistance catalyst during CO-preferential oxidation in H₂-rich streams.     

Parameter identification of proton exchange membrane fuel cell via Levenberg-Marquardt backpropagation algorithm

It is essential to develop an accurate model of proton exchange membrane fuel cell (PEMFC) for a reliable operation and analysis, in which unknown parameters usually need to be determined. The inherent nonlinear, strong coupling, and diversification of PEMFC model seriously hinder traditional methods to identify the parameters. For the sake of overcoming these thorny obstacles, Levenberg-Marquardt backpropagation (LMBP) algorithm based on artificial neural networks (ANNs) is proposed for PEMFC parameter identification. Furthermore, the performance of LMBP is thoroughly evaluated and compared with four typical meta-heuristic algorithms under three cases. Simulation results indicate that LMBP performs a higher accuracy and faster speed for parameter identification. In particular, accuracy and convergence speed can achieve as much as 99.8% and 95.9% growth via LMBP, respectively.     

Reversible formation-melting of nano-crystals in supercooled oxyfluoride germanate liquids

The nanocrystals play a critical role in generating and affecting functionalities of glass materials. Therefore, scientists have made considerable efforts in clarifying microscopic mechanisms of nanocrystal formation in glass to obtain the desired type of nanocrystals. However, the phase transitions of nanocrystals during heating have not been well understood. Here we report on a discovery of the reversible melting-formation of nanocrystals in an oxyfluoride germanate glass during heating-cooling circles. Using a differential scanning calorimetry (DSC), we detected a striking endothermic event at 925 K during heating, after the glass underwent a DSC upscan to a temperature between 925–986 K and subsequent cooling. Based on Raman spectroscopy, X-ray diffraction and transmission electron microscopy, the endotherm is attributed to the melting of nano-crystal BaGeF₆ (˜20 nm). An exothermal response was observed at 890 K during the DSC downscan, implying the re-formation of BaGeF₆ nano-crystals. This suggests that the melting-formation of BaGeF₆ nano-crystals is a typical first-order transition.