Face anti-spoofing based on multi-modal and multi-scale features fusion

Face anti-spoofing is used to assist face recognition system to judge whether the detected face is real face or fake face. In the traditional face anti-spoofing methods, features extracted by hand are used to describe the difference between living face and fraudulent face. But these handmade features do not apply to different variations in an unconstrained environment. The convolutional neural network (CNN) for face deceptions achieves considerable results. However, most existing neural network-based methods simply use neural networks to extract single-scale features from single-modal data, while ignoring multi-scale and multi-modal information. To address this problem, a novel face anti-spoofing method based on multi-modal and multi-scale features fusion ( MMFF) is proposed. Specifically, first residual network ( Resnet )-34 is adopted to extract features of different scales from each modality, then these features of different scales are fused by feature pyramid network (FPN), finally squeeze-and-excitation fusion ( SEF) module and self-attention network ( SAN) are combined to fuse features from different modalities for classification. Experiments on the CASIA-SURF dataset show that the new method based on MMFF achieves better performance compared with most existing methods.     

基于智能手机平台的计算光学显微成像技术研究综述

计算光学显微成像技术将光学编码和计算解码相结合,通过光学操作和图像算法重建来恢复微观物体的多维信息,为显微成像技术突破传统成像能力提供了强大的助力。这项技术的发展得益于现代光学系统、图像传感器以及高性能数据处理设备的优化,同时也被先进的通信技术和设备的发展所赋能。智能手机平台作为高度集成化的电子设备,具有先进的图像传感器和高性能的处理器,可以采集光学系统的图像并运行图像处理算法,为计算光学显微成像技术的实现创造了全新的方式。进一步地,作为可移动通信终端,智能手机平台开放的操作系统和多样的无线网络接入方法,赋予了显微镜灵活智能化操控能力与丰富的显示和处理分析功能,可用于实现各种复杂环境下多样化的生物学检测应用。文中从四个方面综述了基于智能手机平台的计算光学显微成像技术,首先综述了智能手机平台作为光学成像器件的新型显微成像光路设计,接下来介绍了基于智能手机平台先进传感器的计算光学高通量显微成像技术,然后介绍了智能手机平台的数据处理能力和互联能力在计算显微成像中的应用,最后讨论了这项技术现存在的一些问题及解决方向。     

双脉冲激光诱导铝等离子体的双波长干涉诊断

双脉冲激光诱导等离子体在激光加工、元素检测、材料去除等领域有广阔的应用前景和发展空间,对其进行诊断具有重要意义。针对延迟双脉冲激光诱导铝等离子体的作用效果和影响机理,采用双波长干涉法对其时间演化规律展开研究。基于马赫-曾德尔干涉仪搭建了双波长干涉诊断系统,得到了双脉冲激光诱导等离子体干涉图。通过对干涉图的处理和分析,得到了等离子体电子密度随双脉冲激光延迟时间的变化规律。结果表明,随着双脉冲激光延迟时间的增加,第二束脉冲激光对等离子体电子密度的增强效果先加强后减弱。其中,双脉冲激光延迟时间为10 ns时,对等离子体电子密度的增强效果最强,在30 ns时刻,其中心区域平均电子密度可达6.49×1019 cm?3,相较于同等能量单脉冲激光诱导等离子体提升了26%。同时研究了延迟时间对第二束脉冲激光作用机制的影响。研究结果为双脉冲激光诱导等离子体的优化方向提供了参考。     

基于深度森林的多级特征融合SAR目标识别

大多数传统的合成孔径雷达(SAR)目标识别方法仅仅使用了单一的幅度特征,但是由于斑点噪声的存在,仅仅使用幅度特征会限制识别的性能.为了进一步提高SAR目标识别的性能,该文提出了一个基于深度森林的多级特征融合SAR目标识别方法.首先,在特征提取阶段,提取了多级幅度特征和多级密集尺度不变特征变换(Dense-SIFT)特征.幅度特征反映了目标反射强度,Dense-SIFT特征描述了目标的结构特征.而多级特征可以从局部到全局表征目标.随后,为了更完整、充分地反映SAR目标信息,借鉴深度森林的思想对多级幅度特征和多级Dense-SIFT特征进行联合利用.一方面通过堆叠的方式不断将多级幅度特征和多级Dense-SIFT特征进行融合,另一方面通过逐层的特征变换挖掘深层信息.最后利用得到的深层融合特征对目标进行识别任务.该文在MSTAR数据集上进行对比实验,实验结果表明所提算法在性能方面取得了提升,且其性能对超参数设置具有一定的鲁棒性.     

基于自组织增量-图卷积神经网络的金相图半监督学习

采用深度学习对钢铁材料显微组织图像分类,需要大量带标注信息的训练集.针对训练集人工标注效率低下问题,该文提出一种新的融合自组织增量神经网络和图卷积神经网络的半监督学习方法.首先,采用迁移学习获取图像数据样本的特征向量集合;其次,通过引入连接权重策略的自组织增量神经网络(WSOINN)对特征数据进行学习,获得其拓扑图结构,并引入胜利次数进行少量人工节点标注;然后,搭建图卷积网络(GCN)挖掘图中节点的潜在联系,利用Dropout手段提高网络的泛化能力,对剩余节点进行自动标注进而获得所有金相图的分类结果.针对从某国家重点实验室收集到的金相图数据,比较了在不同人工标注比例下的自动分类精度,结果表明:在图片标注量仅为传统模型12％时,新模型的分类准确度可达到91％.     

Evaluation Method of Classroom Teaching Effect Under Intelligent Teaching Mode

Mobile Networks and Applications - In order to improve the accuracy and performance of classroom teaching effect evaluation, an intelligent teaching mode classroom teaching effect evaluation method...     

A game model for information dissemination in live streaming e-commerce environment

As a new form of social e-commerce, live streaming e-commerce is becoming increasingly popular among Chinese consumers. Live streaming brings huge opportunities, and we can take measures to further expand the dissemination of live streaming information. A live streaming information dissemination game model based on social evolutionary game theory is presented to simulate multiple complex live streaming e-commerce networks. The introduction of incentive mechanism in the model further quantifies the internal relation between rewards and live streaming information dissemination to a certain extent and predicts the ratings of live streaming in networks with different update frequency of social relationships. The simulation results indicate that the reputation environment in social networks, adjusting frequency of relationship, pre-broadcast rewards to post-broadcast rewards ratio all have certain ranges of impact on information dissemination in live streaming e-commerce environment and then affect the live streaming ratings. These findings offer insights into the dissemination of live streaming information in social networks.     

关于模拟法描会聚点电场的讨论

本文论证了“ＱＥ－２”型静电场描迹中模拟聚焦电场的电极在设计上有误的观点。