最新刊期

    39 4 2024

      Material Physics

    • 纤维素纳米晶体(CNC)在材料科学领域的研究取得了新进展。CNC是由硫酸催化的纤维素水解制备的纳米材料,直径约为5-15纳米,长度约为100-300纳米。研究表明,CNC在水中可以形成具有手性向列型结构的液晶相,并且在干燥成固体薄膜时,CNC能够保持螺旋手性向列顺序并组装成层状结构。这一发现为CNC的自组装过程提供了新的思路。在CNC自组装过程中,可以添加溶胶-凝胶前体,如Si(OMe)4,随着溶剂的蒸发,这些前体会发生水解和缩合,从而形成手性向列相二氧化硅/CNC复合材料。这种复合材料经过煅烧处理,可以破坏纤维素模板,留下高表面积手性向列结构的介孔二氧化硅膜。此外,研究还探索了利用类似方法制备有机二氧化硅或二氧化钛薄膜的可能性。这些薄膜在滤光片、反射器和薄膜等领域具有潜在应用价值。更重要的是,合成的手性介孔二氧化硅被用作手性色谱固定相材料,用于分离手性药物。实验结果显示,手性分离效果显著,这为CNC在手性新材料开发方面提供了新的方法和思路。这项研究的成果不仅丰富了CNC的应用领域,也为材料科学领域的研究开辟了新的方向,为未来手性材料的研发和应用奠定了基础。
      XUE Xudong,HOU Yanli,ZHANG Yan
      Vol. 39, Issue 4, Pages: 409-419(2024) DOI: 10.37188/CJLCD.2024-0012
      摘要:Cellulose nanocrystals (CNC) are prepared by hydrolysis of cellulose catalyzed by sulfuric acid. According to the source and hydrolysis conditions of cellulose, the diameter of the nanocrystals is about 5~15 nm and the length is about 100~300 nm.CNC organizes into chiral nematic liquid crystals in water. When dried into a thin solid film, the CNCs retain the helicoidal chiral nematic order and assemble into a layered structure. In the process of CNC self-assembly, it is possible to add sol-gel precursors, such as Si(OMe)4. With the evaporation of solvent, these precursors undergo hydrolysis and condensation, leading to a chiral nematic silica/CNC composite material.In addition, porous, rainbow colored organic silica or titanium dioxide films are also obtained using this methods. These thin films can be used as filters, reflectors, and membranes. The most important thing is that the synthesized chiral mesoporous silica can be used as a chiral chromatographic stationary phase material to separate chiral drugs. The results shows that the effect of chiral separation is remarkable, providing a new method for the development of chiral new materials based on CNC.  
      关键词:cellulose nanocrystals;chiral nematic structure;mesoporous silica;chiral chromatographic stationary phase;chiral separation   
      142
      |
      102
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 49218255 false
      发布时间:2024-05-15
    • 液晶光子器件的研究取得了重要进展。液晶的方位角和预倾角的调制广泛应用于各类液晶光子器件中,然而传统的液晶预倾角调控方法存在稳定性差、易受外场干扰的问题。为此,研究人员提出了一种基于液晶聚合物预倾角可控制备的新方法。液晶聚合物因其高相变温度、耐热、耐强光、耐机械振动等特性,成为预倾角调控的理想选择。然而,由于液晶单体受液晶与空气界面的锚定能影响,液晶聚合物的预倾角与液晶层厚度相关,难以实现稳定可控。为解决这一问题,研究人员采用两步法曝光偶氮染料亮黄(BY)薄膜诱导液晶单体RM257倾斜,并通过多层旋凃及液晶单体RM257光聚合的方式降低液晶与空气界面锚定能的影响。这一创新方法使得RM257主链型液晶聚合物的预倾角可在0°~16°之间连续可调,并具备光热稳定的优点。这一研究成果为液晶光子器件的制备提供了新的材料设计思路,有望推动液晶光子与显示器件领域的发展。
      LI Chuan,HU Chao,FANG Qipeng,LIU Minxuan,JIAO Binzhang,SHEN Jun,YIN Shaoyun
      Vol. 39, Issue 4, Pages: 420-426(2024) DOI: 10.37188/CJLCD.2024-0044
      摘要:The modulation of azimuth and pretilt angle of liquid crystals (LCs) is widely used in the study of various liquid crystal photonic devices. Comparing to unstability and unreliability of small molecule LC devices fabricated by pretilt angle modulation, LC polymers have some prominent advantages like high phase transition temperature, heating and glaring resistance, stability of mechanical vibration, and so on. Since the LC monomer is affected by the anchoring energy at the interface between LC and air, the pretilt angle of LC polymers is related to the thickness of the LC layer, which makes it difficult to achieve a stable and controllable pretilt angle. In this study, a two-step exposure of azo-dye bright yellow (BY) film was used to guide the tilt of LC monomer RM257, and the effect of the anchoring force at the LC-air interface was reduced by using multilayer spin-coating process and the photopolymerization of LC monomer RM257. Through the controllable preparation of pretilt angle of the LC polymers, it shows that the range of pretilt angle can be continuously adjusted between 0° and 16° with the photo-thermal stability and the potential patterning capability for LC photonic devices. This study is expected to provide a new way for the preparation of LC photonic and display devices.  
      关键词:Main-chain liquid crystal polymers;photoalignment;Azo dye materials;pretilt angle;Photothermal stability   
      290
      |
      108
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 49948779 false
      发布时间:2024-05-15

      Device Physics and Device Preparation

    • 液晶偏振体光栅领域取得重要突破。研究团队成功设计并实现了一种三层结构液晶偏振体光栅,并对其衍射特性及在光波导中的耦合情况进行了深入研究。通过理论计算和实验验证,该器件在450 nm、532 nm、633 nm下的衍射效率最高可达80%以上,展现出良好的光波导耦合性能。这一研究成果不仅丰富了液晶偏振体光栅的理论体系,还为光波导耦合器件的应用提供了新的可能,对推动相关领域的技术进步具有重要意义。
      KAN Lina,MU Quanquan,PENG Zenghui,YANG Chengliang,YU Yang,LIU Yonggang,WANG Qidong
      Vol. 39, Issue 4, Pages: 427-435(2024) DOI: 10.37188/CJLCD.2024-0048
      摘要:In this paper, a three-layer liquid crystal polarization grating is designed and implemented, and its diffraction characteristics and coupling in optical waveguide are studied. In theory, the transverse period of 424 nm is selected, and the corresponding three longitudinal periods is calculated according to the specific working wavelength of 450 nm, 532 nm and 633 nm to achieve the superposition of the thickness direction under the condition of Bragg diffraction. Secondly, the finite element method is used to simulate the device. Finally, the relationship between the incident angle and the diffraction angle coupled in the waveguide is simulated according to the red, green and blue light dispersion relationship. The liquid crystal polarization grating device is fabricated and its system performance is characterized. A preliminary optical waveguide coupling experiment is given to demonstrate its coupling effect. The results show that the diffraction efficiency of the prepared three-layer liquid crystal polarization grating device can reach more than 80% at 450 nm, 532 nm and 633 nm, and the angular bandwidth of the corresponding wavelength is 12°, 8° and 6°, respectively. The field of view of the device coupled in the waveguide can reach 27.81° by simulation, which is a good optical waveguide coupler.  
      关键词:optical devices;Angular response;diffraction efficiency;Liquid crystal polarizer grating   
      155
      |
      101
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 50187538 false
      发布时间:2024-05-15
    • 科技新闻播报:柔性显示技术作为当前显示领域的重要研究方向,电子纸作为其理想载体之一备受关注。在众多柔性电子纸显示技术中,电润湿显示以其低能耗、户外使用舒适、快速响应、广色域和易于柔性化等特点脱颖而出,展现出巨大的发展潜力。研究者对柔性电润湿显示原理、制备工艺及研究现状进行了系统性介绍,并对未来发展进行了展望。该研究不仅厘清了柔性电润湿显示的发展现状与未来方向,还指出了电润湿电子纸柔性化所面临的挑战,为柔性电润湿电子纸显示技术的未来发展指明了方向。这一成果在显示技术领域具有重要意义,有望推动柔性显示技术的进一步突破。
      YANG Bin,QIU Ying,LIN Zhe,YANG Tinghong,TANG Biao,YUAN Dong,LUO Dongxiang
      Vol. 39, Issue 4, Pages: 436-446(2024) DOI: 10.37188/CJLCD.2024-0019
      摘要:Flexible display is an important research direction in the field of display, in which electronic paper is one of the ideal carriers for flexible display. Compared with other flexible e-paper display technologies, electrowetting display has the advantages of low energy consumption, comfortable outdoor viewing, fast response speed, wide color gamut and easy to realize flexibility, etc., so it is very promising for development. The display principle, preparation process and research status of flexible electrowetting display are systematically introduced. At the same time, the future development of the flexible electrowetting display is also prospected, and the development status and future direction are clarified. At last, the problems that still need to be solved for the flexibility of electrowetting e-paper are summarized, and the direction for the future development of flexible electrowetting e-paper display technology is pointed out.  
      关键词:flexible display;electronic paper display;reflective display;electrowetting display   
      174
      |
      93
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 48716823 false
      发布时间:2024-05-15
    • 科技新闻播报:随着平板显示和柔性显示产业的快速发展,金属氧化物半导体薄膜晶体管(MOS TFTs)因其卓越性能已成为关键技术。经过30多年的研究,非晶镓锌氧化物(a-IGZO)成功替代非晶硅(a-Si)应用于TFTs。然而,显示产业对更高生产效益、更佳显示性能(如高分辨率、高刷新率)和更低功耗的需求,推动了更高迁移率的MOS TFTs技术的发展。本文深入探讨了通过多元MOS材料实现MOS TFTs高迁移率特性的研究进展,并从固体物理学角度分析了迁移率与器件稳定性之间的关系。这一研究不仅为提升显示产业的技术水平提供了有力支持,也为MOS TFTs的未来发展指明了方向。随着研究的深入和技术的不断进步,我们有理由相信,MOS TFTs将在未来显示产业中发挥更加重要的作用,推动整个行业向更高水平迈进。
      LI Qiang,GE Chunqiao,CHEN Lu,ZHONG Weiping,LIANG Qiying,LIU Chunxi,DING Jinduo
      Vol. 39, Issue 4, Pages: 447-465(2024) DOI: 10.37188/CJLCD.2024-0032
      摘要:Thin-film transistor (TFT) based on metal oxide semiconductor (MOS) has become a key technology to boost the development of the flat panel display or flexible display industry due to their high field-effect mobility (μFE), extremely low cut-off leakage current and good large-area electrical uniformity. After more than 30 years of research, amorphous indium gallium zinc oxide (a-IGZO) is the first to be popularized in TFT by replacing the amorphous silicon (a-Si). However, in order to simultaneously meet the multiple upgrade requirements of the display industry for higher productivity, better display performance (such as high resolution, high refresh rate, etc.) and lower power consumption, MOS TFTs technology with higher mobility is required. From the perspective of solid-state physics, this paper reviews the research progress of MOS TFTs to achieve high mobility characteristics through multi-component MOS materials, and discusses the relationship between mobility and device stability. Finally, the status quo and development trend of MOS TFTs are summarized and prospected.  
      关键词:Metal oxide semiconductor;thin-film transistor;field-effect mobility;Bias stability   
      106
      |
      105
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 49218222 false
      发布时间:2024-05-15
    • 一项关于超大尺寸IGZO(InGaZnO)产品的研究显示,在高温高湿环境下,该产品容易发生异常显示不良。经研究发现,这是由于集成栅极驱动电路中的关键器件M2的转移特性曲线在评价中发生了严重正移。研究团队通过脉冲实验模拟了M2的实际工作环境,并揭示了造成不良的主要影响因素:在M2器件关闭时,漏极与源极之间的压差VDS过大,导致IGZO膜层内的氧空位VO+迁移,并最终使转移特性曲线发生正移。研究还发现,迁移的氧空位VO+经过加热后可以复原。此外,研究团队在不改变IGZO成膜条件下,通过减小M2器件关闭时的VDS压差,成功将超大尺寸IGZO产品导入,并在高温高湿信赖性评价中未出现AD不良。这一研究为解决超大尺寸IGZO产品在高温高湿环境下的显示不良问题提供了解决方案,为相关领域的研究和应用开辟了新方向。
      QIU Heyuan,XIE Xin,LI Zongxiang,CHEN Zhouyu,WANG Baoqiang,WANG Wenchao,LIU Zheng,LIU Yao,LIU Nani,WANG Yang
      Vol. 39, Issue 4, Pages: 466-471(2024) DOI: 10.37188/CJLCD.2024-0042
      摘要:Ultra-large IGZO (InGaZnO) products were prone to Abnormal Display (AD) during reliability evaluation at high temperature and humidity (50 ℃/80%). The main reason was that the key device’s (M2) transfer characteristic curve (IDS-VGS) of GOA (Gate Driver on Array) had a serious positive shift in the evaluation. In this paper, the real working environment of GOA’s key device(M2) was simulated by pulse experiment, and the bad phenomenon of serious positive shift of transfer characteristic curve was reproduced. By setting different pulse experiments, the main influencing factor was revealed. When the device(M2) was turned off, the VDS (pressure difference between Drain and Source) was too large, so that the oxygen vacancy VO+ in the IGZO layer migrated to the boundary of IGZO and GI (Gate Insulator) and to the Source at the same time under the electric field. Due to the electrons trapping effect of oxygen vacancy VO+, the transfer characteristic curve would eventually shift positively. It was found that the migrated oxygen vacancy VO+ can be restored after heating. In addition, without changing the IGZO film forming conditions, by reducing the VDS pressure when the device(M2) was turned off, the ultra-large IGZO product did not occur AD defects during the reliability evaluation at high temperature and humidity for 2 000 h.  
      关键词:InGaZnO thin film transistor;gate driver on array;abnormal display;threshold voltage shift   
      93
      |
      87
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 50633813 false
      发布时间:2024-05-15
    • 针对数字驱动型硅基OLED微显示器在显示动态图像时出现的视觉感知问题,如动态假轮廓和闪烁现象,研究者提出了一种创新的扫描策略——集中式融合扫描。这种策略运用灰度权值重分配和融合子场的概念,通过重新分配整数子场数目和权值,以及将融合子场固定于调制周期中间位置,显著改善了显示器的图像质量。实验结果显示,与传统的扫描方法相比,集中式融合扫描在峰值信噪比方面平均提升了约13%,均方误差降低了约10%,且结构相似度评分接近1,展现出显著的优势。此外,在JEITA闪烁评估中,集中式融合扫描的表现也优于19子场扫描法,闪烁量化值降低了约22%。这一研究不仅为改善数字驱动型硅基OLED微显示器的图像显示质量提供了有效解决方案,还为未来显示技术的研究和创新指明了新的方向。
      JI Yuan,XU Yiqing,CHEN Baoliang,ZHANG Yin,HUANG Xinjie
      Vol. 39, Issue 4, Pages: 472-481(2024) DOI: 10.37188/CJLCD.2024-0014
      摘要:A novel scanning strategy——centralized fusion (CF) scanning is proposed to address the visual perception issues encountered in OLED-on-silicon microdisplay driven digitally, especially the dynamic false contour(DFC) and flickering phenomena. This strategy incorporates the concepts of grayscale weight redistribution and the integration of subfields. It improves the quality of display images by reallocating the number and weights of integer subfields, and by positioning the integrated subfield at the midpoint of the modulation cycle. Experimental results indicate that the CF scanning approach achieves an approximate 13% increase in peak signal-to-noise ratio and a 10% reduction in mean square error compared to traditional scanning methods. Moreover, the structural similarity index score approaches 1, significantly surpassing existing scanning techniques. CF scanning outperforms the 19-subfield scanning method in JEITA flicker assessment, with a reduction in flicker quantification value by about 22%. The CF strategy offers an effective solution for enhancing the image display quality of digital-drive OLED-on-silicon microdisplay, paving the way for future research and innovation in display technology.  
      关键词:oled-on-silicon;microdisplay;centralized fusion scanning;digital driving method;pulse width modulation   
      69
      |
      77
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 49245337 false
      发布时间:2024-05-15

      Image Processing

    • 实时图像增强处理一直是图像处理领域的热门话题。最近,有专家在灰度等间距密度均衡算法的基础上进行了改进,解决了连续图像序列实时增强处理的难题。该改进算法能够有效避免图像局部变化对整体增强效果的影响,同时保证了处理算法的实时性和图像序列间增强效果的连贯性。实验表明,对于1920×1080分辨率的图像,该算法的单帧处理时间仅需约32毫秒。这一研究成果为实时图像增强处理提供了新的解决方案,具有重要的实际应用价值。
      ZHANG Yanchao,GAO Ce,SONG Congcong,XU Jiaxing,WU Jie
      Vol. 39, Issue 4, Pages: 482-489(2024) DOI: 10.37188/CJLCD.2023-0203
      摘要:In the enhancement processing of real-time image, in addition to ensuring the high contrast of single frame image, the real-time performance of the processing algorithm and the coherence of the enhancement effect between image sequences should be considered.Gray-scale equispacing density equalization algorithm can effectively enhance image contrast by removing redundant gray level and rearranging effective gray level equispacing. In this paper, based on the gray-scale equispacing density equalization enhancement algorithm, the effective gray level effective threshold selection, excessive enhancement and other problems are improved accordingly, in order to meet the real-time enhancement processing requirements of continuous image sequences. Experimental results show that the improved algorithm can effectively avoid the influence of local changes on the overall enhancement effect. For 1 920×1 080 resolution images, the processing time of single frame is about 32 ms. The improved algorithm has good enhancement effect on continuous image sequence, and can better meet the processing requirements of real-time image enhancement for coherence, real-time performance and robustness.  
      关键词:real-time image enhancement;gray-scale equispacing density equalization;histogram   
      141
      |
      84
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 40850342 false
      发布时间:2024-05-15
    • 衍射神经网络的研究取得了重要进展。相比传统数字神经网络,衍射神经网络具有高速、低功耗和并行运算的优势,因此备受关注。为了探索利用衍射神经网络执行多任务图像分类识别的可行性,研究者们设计并搭建了一种衍射神经网络系统。该系统采用空间光调制器作为衍射神经网络的相位及振幅权重的调制和网络层的光学全连接,利用CMOS相机实现衍射神经网络中各衍射层输出的光电非线性激活与输出图像识别结果的判别。实验结果显示,设计的系统模型在MNIST和Fashion-MNIST图像分类识别中的正确率分别达到了94.1%和92.1%。通过搭建光路系统进行的光学实验也取得了令人满意的成果,正确率分别为91%和81.7%。这一研究验证了设计的衍射神经网络系统可以满足多种图像分类识别应用需求,为衍射网络的设计与构建提供了新的思路。这一成果对于推动衍射神经网络在实际应用中的发展具有重要意义。
      WANG Zirong,ZHANG Xingxiang,LONG Yongji,FU Tianjiao,ZHANG Mo
      Vol. 39, Issue 4, Pages: 490-505(2024) DOI: 10.37188/CJLCD.2023-0144
      摘要:To investigate the feasibility of diffraction neural network to perform multi-task image classification recognition, a diffraction neural network system is designed and built. The system uses a spatial light modulator (SLM) to modulate the phase and amplitude weights of the diffraction neural network and the optical full connection of the network layers. A CMOS camera is adopted to realize the optical nonlinear activation of the output of each diffraction layer in the diffraction neural network and discriminate the output image recognition results. The designed system model achieves 94.1% and 92.1% accuracy in MNIST and Fashion-MNIST image classification recognition. Finally, by building optical path system, optical experiments have 91% and 81.7% accuracy respectively, which verifies that the designed diffraction neural network system can meet the requirements of various image classification and recognition applications, and provides a new idea for the design and construction of diffraction networks.  
      关键词:Diffraction neural network;Optical Neural Networks;System Design;Image classification recognition   
      236
      |
      87
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 37841971 false
      发布时间:2024-05-15
    • 细粒度图像识别领域迎来新突破。针对微小潜在性特征易忽视、外观差异细微等问题,研究团队提出了一种基于双注意力随机选择全局上下文细粒度识别网络。该网络利用ConvNeXt作为主干,引入双注意力随机选择模块,有效关注到其他潜在微小判别性特征。同时,结合全局上下文注意力模块,将深层语义信息融入中间层,增强了微小特征定位能力。此外,创新性地提出多分支损失,结合不同分支特征,引导网络获取多样性判别特征。在Stanford-cars、CUB-200-2011、FGVC-Aircraft等公开数据集及真实场景车型数据集VMRURS上,该网络分别实现了95.2%、92.1%、94.0%和97.0%的高识别准确率,性能远超其他对比方法,为细粒度图像识别领域的发展奠定了坚实基础。
      XU Shengjun,JING Yang,DUAN Zhongxing,LI Minghai,LI Haitao,LIU Fuyou
      Vol. 39, Issue 4, Pages: 506-521(2024) DOI: 10.37188/CJLCD.2023-0114
      摘要:To address the difficulties of capturing the potential distinguishable features and subtle appearance differences in fine-grained image recognition tasks, dual-attention random selection global context fine-grained recognition network is proposed. Firstly, the ConvNeXt is taken as the backbone network, a dual-attention random selection module is proposed to perform channel random selection and spatial random selection on the features extracted at different stages, so that the network could focus on other potential subtle distinguishable features. Then, by using the global context attention module, the semantic information of top-level is applied to the middle-level to enhance the ability of the middle-level to locate potential subtle distinguishable features. Finally, the multi-branch loss is proposed, and classification loss is imposed on middle-level, top-level and concat-level characteristics. Combining the features extracted from different branches, the network is induced to obtain diverse distinguishable features. The network achieves the accuracies of 95.2%, 92.1%, 94.0% and 97.0% respectively on the three open datasets,Stanford-cars,CUB-200-2011,FGVC-Aircraft and dataset VMRURS in real scenes. The presented method in this paper greatly upgrades the comparison performance.  
      关键词:Fine-grained recognition;ConvNeXt;Dual-attention random selection;Global context attention;Multi-branch loss   
      232
      |
      54
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 37326133 false
      发布时间:2024-05-15
    • 人脸识别领域的研究取得了新突破。针对传统人脸面部表情识别方法存在的参数量大、计算资源消耗大和识别精度低的问题,研究者提出了一种基于条件协调注意力机制的轻量化人脸面部表情识别方法。通过缩减MobileNet V3网络层数、增大通道数和使用Mish激活函数,实现了特征提取后的非线性化。同时,引入改进的协调注意力机制,可以捕捉面部表情在空间和通道位置上的详细信息。在公开数据集FERPlus和RAF-DB上的实验结果显示,新方法参数量降低15.91%,准确率分别达到了88.84%和85.90%,相比改进前模型准确率分别提升了0.83%和1.39%。这一研究成果为人脸表情识别领域提供了新的解决方案,并有望在实际应用中发挥重要作用。
      ZUO Yihai,BAI Wushang,HE Qiusheng
      Vol. 39, Issue 4, Pages: 522-531(2024) DOI: 10.37188/CJLCD.2023-0153
      摘要:At present, facial expression recognition methods have the problems of large number of parameters, large consumption of computing resources and low recognition accuracy. Aiming at the above problems, a lightweight human facial expression recognition method based on conditional coordinated attention mechanism is studied. First, the number of layers of MobileNet V3 network is reduced, while the numbers of intermediate channels and output channels of the inverse residual structure are increased to 1.5~3.2 times of the original number. Mish is used instead of Hardswish activation function to realize the nonlinearization after feature extraction. Secondly, an improved coordinated attention mechanism is introduced to encode the tensor information embedding along horizontal and vertical directions sequentially by maximum pooling and average pooling. And tensor information integration is used to generate features with global sensory field and precise location information to extract detailed information of facial expressions in space and channel location. Finally, experiments are conducted on the publicly available datasets FERPlus and RAF-DB, and the results show that the proposed method reduces the number of parameters by 15.91%, and the accuracy rates are 88.84% and 85.90%, respectively, which are 0.83% and 1.39% higher than the accuracy rates of the model before improvement. The method has good recognition performance and validate the effectiveness of the proposed method.  
      关键词:facial expression recognition;lightweight;attention mechanism;feature extraction   
      237
      |
      46
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 37860080 false
      发布时间:2024-05-15
    • 针对无人机低空飞行和视频图像目标像素占比较小的问题,研究者提出了一种名为SiamAU的孪生神经网络目标跟踪算法。该算法在SiamRPN++的基础上,引入改进的主干网络和特征重排机制,通过加入ECA-Net注意力机制网络和改进激活函数,提升复杂背景下的特征表征能力。同时,算法还进行了浅层降维和深层特征融合,得到更适合无人机等小目标跟踪的改进深度融合特征。在DUT Anti-UAV数据集上的测试表明,SiamAU算法的成功率和精确率分别达到了60.5%和88.1%,相比基准算法提升了5.6%和8.1%。这一研究成果为反无人机任务中的目标跟踪提供了新的解决方案,有望为相关领域的研究开辟新方向。
      ZHENG Binxi,YANG Zhigang,DING Yufeng
      Vol. 39, Issue 4, Pages: 532-542(2024) DOI: 10.37188/CJLCD.2023-0150
      摘要:Object tracking for the unmanned aerial vehicle (UAV) in videos is an important part of the Anti-UAV task. The complex background during low-altitude flight and the small imaging size are two difficulties for UAV object tracking. A Siamese neural network object tracking algorithm (SiamAU) is proposed, which is based on SiamRPN++ in combination with an improved backbone and a feature rearrangement technique. Firstly, ECA-Net attention module is integrated into the backbone network, while the activation function is improved to enhance the representation ability of convolution features in complex background. Then, channel number of the last three convolution features is rearranged in order to make full use of low-level features that are conducive for small object tracking. The rearranged feathers are further fused to obtain the improved feature map. Finally, On the DUT Anti-UAV dataset, SiamAU algorithm achieves success and precession scores of 60.5% and 88.1%, an improvement of 5.6% and 8.1% in comparison with the baseline algorithm. Extensive experimental results on two public datasets validate that the proposed SiamAU achieves better UAV tracking performance and outperforms previous methods, especially in small object and complex background scenarios.  
      关键词:Anti-UAV;object tracking;siamese network;attention mechanism;feature rearrangement   
      235
      |
      53
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 37326083 false
      发布时间:2024-05-15
    • 在车道线检测领域取得重要突破。针对现有算法难以平衡检测精度和速度的问题,研究者提出一种全新的基于ARM嵌入式平台的车道线检测算法。该算法通过设计轻量化语义分割网络,结合卡尔曼滤波车道线跟踪模型,重构编码器以及利用TensorRT加速推理,实现了在保证检测精度的同时提升检测速度。实验结果表明,该算法能够应对各种复杂交通场景,检测精度高达98.03%,检测速度达到50 FSP,满足实时性检测要求。这一研究成果为车道线检测领域提供了新的解决方案,具有重要的理论价值和实际应用价值。
      GUAN Tiantian,YANG Fan
      Vol. 39, Issue 4, Pages: 543-552(2024) DOI: 10.37188/CJLCD.2023-0141
      摘要:Aiming at the problem that the existing lane detection algorithms are difficult to balance the detection accuracy and speed in practical application, a new lane detection algorithm based on ARM embedded platform is proposed. Firstly, a lightweight semantic segmentation network is designed. When SegNet structure is optimized, skip connections are added to the first layer of the network, and channel attention mechanism modules are added after every two convolutional layers to ensure detection accuracy and improve detection speed. Secondly, Kalman filter lane tracking model is constructed to improve the robustness of detection in video streams. Then, the encoder is reconstructed and the model is lightweight. The deep separable convolution is used instead of the traditional convolution to reduce the calculation cost and improve the detection speed. Finally, the Trt model is generated by TensorRT accelerated reasoning to facilitate its deployment in ARM embedded platform for real-time lane detection. Experimental results on the self-produced Tusimeple extended data set show that the proposed algorithm can cope with various complex traffic scenarios, and its detection accuracy is 98.03%, which is superior to other algorithms. And its detection speed reaches 50 FSP, which meets the real-time detection requirements. This algorithm has high robustness and good real-time performance in complex traffic scenarios, and has certain theoretical and practical value.  
      关键词:lane detection;semantic segmentation;depth-separable convolution;TensorRT acceleration;ARM embedded platform   
      214
      |
      53
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 37842510 false
      发布时间:2024-05-15
    0