【摘 要】
：
Optical deep learning based on diffractive optical elements offers unique advantages for parallel process-ing,computational speed,and power efficiency.One landmark method is the diffractive deep neural net-work(D2NN)based on three-dimensional printing tec
【机 构】
：
Center of Ultra-precision Optoelectronic Instrument,Harbin Institute of Technology,Harbin 150001,Chi
论文部分内容阅读
Optical deep learning based on diffractive optical elements offers unique advantages for parallel process-ing,computational speed,and power efficiency.One landmark method is the diffractive deep neural net-work(D2NN)based on three-dimensional printing technology operated in the terahertz spectral range.Since the terahertz bandwidth involves limited interparticle coupling and material losses,this paper extends D2NN to visible wavelengths.A general theory including a revised formula is proposed to solve any contradictions between wavelength,neuron size,and fabrication limitations.A novel visible light D2NN classifier is used to recognize unchanged targets(handwritten digits ranging from 0 to 9)and tar-gets that have been changed(i.e.,targets that have been covered or altered)at a visible wavelength of 632.8 nm.The obtained experimental classification accuracy(84％)and numerical classification accuracy(91.57％)quantify the match between the theoretical design and fabricated system performance.The pre-sented framework can be used to apply a D2NN to various practical applications and design other new applications.
其他文献
为了明确爆破冲击波对高层建筑结构构件及其施工安全性的影响,以广东以色列理工学院二期校区爆破工程为例,基于建筑物地表及典型结构构件监测数据,采用回归分析对高层建筑典型结构构件振动规律进行研究.研究表明:随着距爆源水平距离的增加,垂向、径向和切向振动速度逐渐减小且衰减趋势相似;结构局部振速存在放大效应,在距地高程相同的情况下,随着结构构件与地基的自振频率比值的减小,其振速比逐渐增大;在结构构件与地基的自振频率比值相同的情况下,随距地高程的增加,结构构件与地基的振速比呈线性增长.
In the defense-in-depth concept employed for the safety of nuclear installations,maintaining integrity of containment as the last barrier is of high importance to limit the release of radioactivity to the environment in case of a severe accident.The activ
In a severe accident of a nuclear power reactor,coolant channel blockage by solidified molten core debris may significantly influence the core degradations that follow.The moving particle semi-implicit(MPS)method is one of the Lagrangian-based particle me
The reliable operation of flexible display devices poses a significant engineering challenge regarding the metrology of high barriers against water vapor.No reliable results have been reported in the range of 10-6g·(m2·d)-1,and there is no standard ultra-
To retrieve the fuel debris in Fukushima Daiichi Nuclear Power Plants(1F),it is essential to infer the fuel debris distribution.In particular,the molten metal spreading behavior is one of the vital phenomena in nuclear severe accidents because it determin
为了控制石灰石生产加工过程中岩粉的产生及分析其成因,利用邦德功指数和研磨性分析了石灰岩的主要特性,结合灰岩粉的主要来源,提出了针对性的岩粉控制措施,并通过灰岩矿爆破试验进行了验证.研究结果表明:石灰石相较于其他岩石更易产生岩粉,其主要来自于爆破形成的粉碎区,机械破碎中的矿石整形,钻孔岩屑及软弱夹层;通过优化爆破破碎过程,采用空气间隔不耦合装药、现场混装装药和合理优化爆破参数等方式,可以控制粉碎区大小,使爆破级配更为均匀,从而减少爆破岩粉.
It is difficult to identify suspected cases of atypical patients with coronavirus disease 2019(COVID-19),and data on severe or critical patients are scanty.This retrospective study presents the clinical,labora-tory,and radiological profiles,treatments,and
A comparative computational fluid dynamics(CFD)study was conducted on the three different types of pressurized water reactor(PWR)upper plenum,named TYPE 1(support columns(SCs)and control rod guide tubes(CRGTs)with two large windows),TYPE 2(SCs and CRGTs w
The Sichuan-Tibet Railway is facing extra-ordinary challenges in terms of construction,operation,and maintenance because of its extremely complicated natural environment and geological conditions.Conse-quently,countermeasures are necessary and urgent to e
为了解决爆破振动信号的处理精度极易受到噪声信号的干扰的问题,在局部均值分解法(LMD)和总体平均局部均值分解法(ELMD)的基础上,提出了利用自适应互补集合局部均值分解方法(CELMDAN)分解原始信号,然后采用小波阈值法对获得的乘积函数(PF)分量进行去噪,提高了爆破振动信号去噪的精度.结果表明CELMDAN-WT方法有效地识别破振动信号中的噪声信号,实现了信号与噪声的分离,具有良好的自适应性,在爆破信号去噪中有良好的效果.