Presentation attack detection based on two-stream vision transformers with self-attention fusion

Aiming at the performance degradation of the existing presentation attack detection methods due to the illumination variation, a two-stream vision transformers framework (TSViT) based on transfer learning in two complementary spaces is proposed in this paper. The face images of RGB color space and multi-scale retinex with color restoration (MSRCR) space are fed to TSViT to learn the distinguishing features of presentation attack detection. To effectively fuse features from two sources (RGB color space images and MSRCR images), a feature fusion method based on self-attention is built, which can effectively capture the complementarity of two features. Experiments and analysis on Oulu-NPU, CASIA-MFSD, and Replay-Attack databases show that it outperforms most existing methods in intra-database testing and achieves good generalization performance in cross-database testing.     

Utilization of pyrolytic oil and hydrogen enriched syngas from single feedstock (delonix regia) through pyrolysis process and its influence on performance and emission characteristics in CI engine

The main objective of the present investigation is to conduct the performance, combustion and emission analysis of CI engine operated using hydrogen enriched syngas (pyrolytic gas) and biodiesel (pyrolytic oil) as dual fuel mode condition. Both the pyrolytic oil and syngas is obtained from single feedstock delonix regia fruit pod through pyrolysis process and then pyrolytic oil is converted into biodiesel through esterification. Initially biomass is subjected to thermal degradation at various pyrolysis temperature ranges like 350–600 °C. During the pyrolysis process syngas, pyrolytic oil and char are produced. The syngas is directly used in the CI engine and pyrolytic oil is converted into biodiesel and then used in the CI engine. The pyrolytic oil and syngas is subjected to FTIR and GC/TCD analysis respectively. The syngas analysis confirms the presence of various gases like H₂, CH₄, CO₂, CO and C₂H₄ in different proportions. The various proportions of the syngas is mainly depending upon the reactor temperature and moisture content in the biomass. The syngas composition varies with increase in the temperature and at 400 °C, higher amount of hydrogen is present and its composition are H₂ 28.2%, CO is 21.9%, CH₄ is 39.1% and other gases in smaller amounts. The biodiesel of B20 and syngas of 8lpm produced from the same feedstock are considered as test sample fuels in the CI engine under dual fuel mode operation to study the performance and emission characteristics. The study reveals that BTE has slight increase than diesel of 1.5% at maximum load. On the another hand emission like CO, HC and smoke are reduced by 15%,25% and 32% respectively at full load condition, whereas NO_x emission is increased at all loads in the range of 10–15%. Therefore B20+syngas of 8lpm can be used as an alternative fuel in CI engine without any modification and major products from pyrolysis process with waste biomass is fully used as fuel in the CI engine.     

一种用于电气设备状态监测的新式FBG传感器北大核心CSCD

为了监测绕组变压器的静态应力场和发生短路等故障时的动态应力变化,设计了一种用于电气设备状态监测的新式FBG传感器。该传感器由聚醚醚酮材料封装的FBG构成,通过内部圆锥形空腔结构实现将轴向应力集中于FBG敏感位置。通过仿真对不同压力强度下传感器结构的应力场部分及形变趋势进行了计算与分析,论证了设计的合理性。实验分别对静态载荷和动态冲击进行测试,结果显示,在静态压载测试中,当100 N相似文献   

Korean automatic spacing using pretrained transformer encoder and analysis

Automatic spacing in Korean is used to correct spacing units in a given input sentence. The demand for automatic spacing has been increasing owing to frequent incorrect spacing in recent media, such as the Internet and mobile networks. Therefore, herein, we propose a transformer encoder that reads a sentence bidirectionally and can be pretrained using an out-of-task corpus. Notably, our model exhibited the highest character accuracy (98.42%) among the existing automatic spacing models for Korean. We experimentally validated the effectiveness of bidirectional encoding and pretraining for automatic spacing in Korean. Moreover, we conclude that pretraining is more important than fine-tuning and data size.     

贾燕1,赵永涛2,张红松2

为确定ZrO2/莫来石热障涂层的单层最佳厚度,采用有限元技术研究单层厚度对ZrO2/莫来石热障涂层热冲击应力的影响。结果表明,莫来石层厚度为0.25mm时,涂层表面及ZrO2/莫来石界面处具有最大热应力,而其它莫来石层厚度时涂层内部热应力则受莫来石层厚度影响不明显。当ZrO2层厚度为0.3mm时,ZrO2/莫来石涂层中的热应力在整个横向距离范围内保持稳定,而其它情况下涂层中的热应力则与ZrO2层厚度成正比。涂层表面处的径向热应力大于ZrO2/莫来石界面处的对应值。     

Single nanowire assembled by liquid dielectrophoresis for on-chip technologies

ABSTRACT

Study on dynamic tensile properties and atomic chain fabrication of single nanowire, for understanding its dynamic tensile properties and unique physical properties of atomic chain to fabricate atom scale devices, is one of frontier research issues in nano-scale science. However, how to assemble single nanowire on a tensible micro-structure becomes one of the most difficult problems, which severely restricts the development of this research field. In this paper, after the ultrahigh tensible microelectrode chip is fabricated by micro-electromechanical systems technology, hexamethyldisilazane is utilized to improve hydrophobicity of the chip, and then a micro-droplet dielectrophoresis experimental platform and technology is developed to assemble single nanowire on the sensible microelectrode. Experimental results show that accurate and efficient assembly of single Cu nanowire is realized, which contributes greatly to the further research of dynamic tensile properties and atomic chain fabrication. And for guiding the assembly experiments, finite element technology is also utilized to analyze the local micro electro-field around the microelctrodes during dieletrophoresis experiments.     

Single transition metal atom catalysts on Ti2CN2 for efficient CO2 reduction reaction

Electrochemical CO₂ reduction reaction (CO₂RR) is an efficient way in the utilization of CO₂. In this work, single transition-metal (TM) atom (TM = Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn) anchored on two-dimensional (2D) Ti₂CN₂ are designed for CO₂RR using density-functional-theory (DFT) calculation. We show that Ti₂CN₂ serves as an excellent substrate to support single atom catalysts (SACs), compared to Ti₂CO₂ and Ti₂CF₂. We find that the Sc, Ti and V supported on Ti₂CN₂ show high catalytic activities to produce CO with a low overpotential of 0.37, 0.27, and 0.23 eV, respectively. Differently, the Mn and Fe on Ti₂CN₂ are catalytically active for the production of HCOOH with a low overpotential of 0.32 and 0.43 eV, respectively. We further show that the negatively charged TM-Ti₂CN₂ can capture and activate CO₂ more effectively, and the catalytic activity and selectivity can be significantly tuned by injecting extra electrons.     

Tropfenbewegung und Stofftransport in technischen Flüssig/flüssig-Systemen. Teil 2: Auswirkung von Grenzflächeneffekten und Verunreinigungen

The design of a liquid/liquid contact apparatus necessitates the knowledge of fluid dynamics and mass transfer in the dispersion. Prediction of process parameters is challenging due to swarm and interfacial effects. Single drop investigations are suitable to enable a sophisticated dimensioning based on few simple experiments combined with published theories and correlations. In the first part of this article, the fundamentals of fluid dynamics and mass transfer in absence of interfacial effects were explained. Here, an overview of relevant interfacial phenomena and the influence of different substances typically occurring in technical applications as impurities or accompanying components is given.     

响应面法优化农产品中抗性淀粉的体外检测

目的 采用响应面法优化农产品中抗性淀粉(Resistant Starch, RS)的体外检测方法。方法 选择谷类、薯类、豆类3类抗性淀粉主要来源的农产品，设计单因素试验和Box-Behnken试验。单因素试验考察了水浴时间、淀粉葡萄糖苷酶（AMYLOGLUCOSIDASE, AMG）作用温度、AMG作用时间、AMG添加量和离心力与离心时间5个条件，并选择AMG作用温度、AMG作用时间、AMG添加量来设计Box-Behnken试验。结果 根据单因素试验，确定了水浴时间为14 h，离心条件选择4500 g离心5 min；根据响应面实验确定对RS含量检测结果影响程度排序为AMG添加量＞AMG作用时间＞AMG作用温度。结合单因素响应面试验，确定最佳检测方法为水浴14 h，4500 g离心5 min，AMG作用温度49℃、AMG添加量0.2 mL、AMG作用时间38 min。结论 优化后的方法重复性良好，适用于玉米、马铃薯和红豆3类样品，且有望适用于其他主要农产品。     

2011～2020年华北平原气溶胶光学厚度时空分布特征及潜在源分析

随着经济高速增长和城市化进程不断加快,华北平原区域性空气污染问题愈演愈烈。针对该区域开展长时序气溶胶光学厚度时空分布特征和潜在源分析研究,对华北平原大气污染治理具有重要意义。基于长时序MODIS/Terra C6.1 MOD04_L2气溶胶光学厚度产品,分析华北平原气溶胶光学厚度的时空分布特征,并利用后向轨迹聚类分析讨论华北平原7个重点城市气团输送的季节变化,并以污染较为严重的河北石家庄为例进行潜在源分析和浓度权重分析,探究影响其大气质量的污染物潜在源区。结果表明:2011～2020年华北平原气溶胶光学厚度月均值呈显著的周期性变化,以年为周期,每个周期内峰值一般出现在6月至8月; 气溶胶光学厚度月际年内呈单峰分布,峰值出现在6月(0.75),最小值出现在12月(0.37); 气溶胶光学厚度季均值从大到小依次为夏季(0.67)、春季(0.59)、冬季(0.49)、秋季(0.46); 10年间气溶胶光学厚度呈下降趋势,整体下降幅度达36.84%,其中2011年最高(0.72),2018年最低(0.45); 华北平原7个重点城市春、夏、秋、冬四季主要受短距离气团输送影响较大,长距离气团输送影响较小; 2014～2020年河北石家庄的空气质量优良天数占比相对较小,空气质量状况差,影响其空气质量的污染物多为本地生成,同时也受周边省市近距离输送的影响。