摘要：人脸活体检测是确保人脸识别系统安全不可或缺的一环。当前大多数人脸活体检测算法采用的是深度卷积神经网络,尽管其检测性能卓越,但往往涉及庞大的参数量和计算复杂度,限制了其应用。为克服这些挑战,提出了一种基于蓝图差分卷积和双重辅助监督的轻量级人脸活体检测算法(BDC-DAL)。通过将轻量化的卷积结构和蓝图差分卷积设计相结合,有效降低了模型的参数和计算量,同时保留了捕捉复杂细粒度特征的能力。此外,引入的双重辅助监督机制,利用监督信号重建过程学习细微的欺诈线索,显著提高了模型对未知攻击场景的鲁棒性。为验证算法的有效性,在CelebA-Spoof和OULU-NPU人脸活体检测数据集上进行了广泛测试,并与当前主流的深度学习人脸活体检测算法和其他轻量化网络模型进行了比较,实验结果显示,BDC-DAL算法大幅降低了模型复杂度,参数量减少了近90%,模型的浮点计算量也下降了约10%。BDC-DAL算法在实现高效性和轻量化的同时,成功平衡了性能与资源消耗,为在资源受限环境下的人脸识别提供了可行的解决方案。

摘要：随着对X射线成像,特别是对弯曲和不规则物体的三维成像的需求不断增加,高性能柔性闪烁膜的开发引起了人们的极大关注.镧系(Ⅲ)-有机闪烁体由于其在结构设计和光学可调谐方面的巨大灵活性,最近成为一类新的有前途的X射线闪烁体.然而,设计具有高光产额和良好溶液分散性的镧系元素(Ⅲ)-有机闪烁体仍然具有挑战性.本文报道了一种具有高效X射线激发发光(XEL)和良好溶液分散性的铕-钛氧簇合物(Eu-TiOCs)闪烁体.笼状镧-钛氧原子簇核作为X射线吸收剂,产生高能电子并引发二次电子热弛豫.有机配体作为能量转移桥和天线,通过级联能量转移来敏化Ln^(3+);离子的发光.Eu-TiOCs闪烁体表现出高的稳态X射线光产额(高达26,816±2000 photons MeV^(-1)).此外,EuTiOCs闪烁体和聚合物的羟基之间的弱氢键相互作用增强了Eu-TiOCs的溶液分散性,有利于制备低光散射和均匀闪烁膜.铕-钛氧簇与聚合物之间的通过无机簇笼和有机配体之间的协同作用,所得到的柔性闪烁膜对平面和非平面物体都表现出的高空间分辨率和低剂量X射线成像.本工作为进一步设计和制备高性能X射线闪烁体和高分辨率柔性闪烁膜提供了新思路.

摘要：The possible changes of tropical cyclone（TC） tracks and their influence on the future basin-wide intensity of TCs over the western North Pacific（WNP） are examined based on the projected large-scale environments derived from a selection of CMIP5（Coupled Model Intercomparison Project Phase 5） models. Specific attention is paid to the performance of the CMIP5 climate models in simulating the large-scale environment for TC development over the WNP. A downscaling system including individual models for simulating the TC track and intensity is used to select the CMIP5 models and to simulate the TC activity in the *** assessment of the future track and intensity changes of TCs is based on the projected large-scale environment in the21 st century from a selection of nine CMIP5 climate models under the Representative Concentration Pathway 4.5（RCP4.5）scenario. Due to changes in mean steering flows, the influence of TCs over the South China Sea area is projected to decrease,with an increasing number of TCs taking a northwestward track. Changes in prevailing tracks and their contribution to basin-wide intensity change show considerable inter-model variability. The influences of changes in prevailing track make a marked contribution to TC intensity change in some models, tending to counteract the effect of SST warming. This study suggests that attention should be paid to the simulated large-scale environment when assessing the future changes in regional TC activity based on climate models. In addition, the change in prevailing tracks should be considered when assessing future TC intensity change.

摘要：A natural attapulgite （ATP）‐based catalyst, sulfated In2O3‐ATP （SO42-/In2O3‐ATP）, was obtained by an impregnation‐calcination method and was used to efficiently and selectively produce the useful platform chemical 5‐hydroxymethylfurfural （HMF） from hexoses. Some important reaction param‐eters were studied, revealing that Lewis and Br-nsted acid sites on SO42-/In2O3‐ATP catalyze glu‐cose isomerization and fructose dehydration. The yields of HMF from glucose and fructose were 40.2%and 46.2%, respectively, using the optimal conditions of 180℃ for 60 min with 10 wt%of solid acid catalyst in a mixture of γ‐valerolactone‐water （9：1）.

摘要：在极端条件下,材料性能退化是引起航天器失效的主要因素,而摩擦运动部件的润滑失效是导致材料性能退化的关键所在.聚酰亚胺以其独特的耐高低温性能、优异的机械性能以及独特的自润滑性能而被广泛应用于航空航天等高新科技领域.因此研制适用于不同工况需求的高性能聚酰亚胺自润滑复合材料具有重大的意义.基于国内外研究进展,本文介绍了中国科学院兰州化学物理研究所近年来在极端条件下的聚酰亚胺自润滑材料研究成果,并对未来聚酰亚胺润滑材料的设计思路与发展趋势进行了展望.

摘要：随着国内通信运营商5G无线网络规模和用户业务快速发展,如何快速响应并解决用户日常各类投诉问题,做好日益复杂的网络和业务场景下用户投诉服务工作,已经成为各个运营商面对的十分重要的课题。要解决要这类问题,必须建设与之相适应的高性能、去中心化、面相服务的信息系统。文章针对这些问题,研究并提出了基于Springboot微服务分布式的移动网络用户投诉处理系统架构设计方案,最后给出了具体技术实现。

摘要：目的 面向儿童历史学习设计数字智能化纪念馆文创产品,解决趣味性不足与教育价值转化率低的问题。方法 基于儿童认知发展规律与智能化学习理论,提出数字情境体验、互动参与、故事化表达的设计方法,构建系统化产品开发体系。以“智能小小讲解员”主题套装为例,整合智能学习工具、数字教育资源与评测系统,开展为期3个月的实践评估。结果 参与儿童的历史知识记忆准确率由65%提升至88%,解说技能水平提升35%,学习兴趣显著增长。85%的儿童掌握智能讲解技巧,形成数字化史料分析方法,建立持续性学习习惯。结论 丰富了智慧纪念馆设计理论,为实现数字化儿童历史学习提供创新方案,对推动文创产品的智能创新发展与文化传承教育具有重要实践意义。

摘要：基于三周期极小曲面(TPMS)设计了孔隙率为50%、55%和60%的均质和梯度Gyroid(G)型、I-Graph-Wrapped Package(IWP)型TPMS多孔结构,并采用熔融沉积成型(FDM)技术制备试样。通过Micro-CT扫描、压缩试验和声阻抗管测试,研究了多孔结构的孔隙特征、压缩和吸声性能。结果表明,熔融沉积技术成型的试样具有良好的连通性,内部局部的缺陷导致孔隙率存在较小偏差;当孔隙率从50%增加到60%,均质多孔结构的弹性模量和屈服强度会逐渐降低;在平均孔隙率均为55%的情况下,G-50-60结构的屈服强度和弹性模量要大于IWP-50-60结构的,而单元尺寸梯度较孔隙率梯度多孔结构具有较好的压缩性能和吸声特性。在揭示多孔结构压缩性能的基础上研究其吸声性能,扩展了TPMS轻质结构的多功能性。

摘要：开放世界目标检测(open world object detection,OWOD)是一个计算机视觉挑战,聚焦于现实世界环境,其不仅要检测出标记出的已知物体,还需要能处理训练过程中被忽视的未知物体。针对已知和未知物体的检测混淆、密集未知目标和小目标遗漏等问题,提出了一种新的基于偏移过滤和未知特征强化的开放世界目标检测器(offset filter and unknown-feature reinforcement for open world object detection,OFUR-OWOD)。首先设计一个未知类特征强化(unknown class feature reinforcement,UCFR)模块,通过自适应未知对象得分的方法来强化未知类目标特征,进而提高模型对未知类对象的训练准确度。然后,将重叠框偏移过滤器(overlapping box offset filter,OBOF)应用于目标预测框,根据目标位置和大小,获得不同偏移得分,以过滤冗余未知框。通过丰富实验证明,该方法在COCO-OOD和COCO-Mix上优于现有一些最先进的方法。

摘要：In detailed aerodynamic design optimization,a large number of design variables in geometry parameterization are required to provide sufficient flexibility and obtain the potential optimum ***,with the increasing number of design variables,it becomes difficult to maintain the smoothness on the surface which consequently makes the optimization process progressively *** this paper,smoothing methods based on B-spline functions are studied to improve the smoothness and design *** wavelet smoothing method and the least square smoothing method are developed through coordinate transformation in a linear space constructed by B-spline basis *** these two methods,smoothing is achieved by a mapping from the linear space to itself such that the design space remains unchanged.A design example is presented where aerodynamic optimization of a supercritical airfoil is conducted with smoothing methods included in the optimization *** results from the design example confirm that these two smoothing methods can greatly improve quality and efficiency compared with the existing conventional non-smoothing method.