In this paper,we present a simulation method to reconstruct the object's imaging in a ghost-imaging system.We use Laguerre-Gaussian(LG) beam as the light source which will produce the Orbital Angular Momentum(OAM)...In this paper,we present a simulation method to reconstruct the object's imaging in a ghost-imaging system.We use Laguerre-Gaussian(LG) beam as the light source which will produce the Orbital Angular Momentum(OAM) states.In the signal arm,the object spatial information is encoded as phase,and added to the phase of the signal light.In the idler arm,the phase of the idler light is changed according to the maximum coincidence rate in the receiver.We get each pixel matched phase of the object and reconstruct the object's imaging in the idle arm.We demostrate some examples of the reconstruction and compare with the other reported method.The results show that our method can improve the resoltuion of the image effectively.展开更多
针对轻量级模型在光刻热点检测中特征提取能力不足的问题,提出以改进重排网络第2版(shuffle net version 2,ShuffleNetV2)为主干网络,引入多尺度双重注意力(multi-scale dual attention,MSDA)模块,同时融合梯度协调机制(gradient harmon...针对轻量级模型在光刻热点检测中特征提取能力不足的问题,提出以改进重排网络第2版(shuffle net version 2,ShuffleNetV2)为主干网络,引入多尺度双重注意力(multi-scale dual attention,MSDA)模块,同时融合梯度协调机制(gradient harmonizing mechanism,GHM)和基于度量学习思想的加性角度边距(additive angular margin,AAM)的光刻热点检测模型——ShuffleNetV2-MSDA-GHM-AAM(SMGA)。该模型提升了对不同尺度上下文信息的建模与感知能力,优化了特征嵌入空间的类间判别性,缓解了数据集的不平衡。在2012年国际计算机辅助设计会议(2012 international conference on computer-aided design,ICCAD 2012)数据集上进行实验,结果显示,SMGA模型在保持98.22%的较高检测召回率的同时,平均误报数量降低到484个。该模型为实现集成电路设计阶段的高效、低成本光刻热点检测提供了可行方案,具有重要的工程应用价值和推广前景。展开更多
文摘In this paper,we present a simulation method to reconstruct the object's imaging in a ghost-imaging system.We use Laguerre-Gaussian(LG) beam as the light source which will produce the Orbital Angular Momentum(OAM) states.In the signal arm,the object spatial information is encoded as phase,and added to the phase of the signal light.In the idler arm,the phase of the idler light is changed according to the maximum coincidence rate in the receiver.We get each pixel matched phase of the object and reconstruct the object's imaging in the idle arm.We demostrate some examples of the reconstruction and compare with the other reported method.The results show that our method can improve the resoltuion of the image effectively.
文摘针对轻量级模型在光刻热点检测中特征提取能力不足的问题,提出以改进重排网络第2版(shuffle net version 2,ShuffleNetV2)为主干网络,引入多尺度双重注意力(multi-scale dual attention,MSDA)模块,同时融合梯度协调机制(gradient harmonizing mechanism,GHM)和基于度量学习思想的加性角度边距(additive angular margin,AAM)的光刻热点检测模型——ShuffleNetV2-MSDA-GHM-AAM(SMGA)。该模型提升了对不同尺度上下文信息的建模与感知能力,优化了特征嵌入空间的类间判别性,缓解了数据集的不平衡。在2012年国际计算机辅助设计会议(2012 international conference on computer-aided design,ICCAD 2012)数据集上进行实验,结果显示,SMGA模型在保持98.22%的较高检测召回率的同时,平均误报数量降低到484个。该模型为实现集成电路设计阶段的高效、低成本光刻热点检测提供了可行方案,具有重要的工程应用价值和推广前景。
