检索结果-内蒙古大学图书馆

Adaptive random down-sampling data augmentation and area attention pooling for low resolution face recognition

EXPERT SYSTEMS WITH APPLICATIONS 2022年 209卷

作者： Gao, Xuliang Sun, Yubin Xiao, Yao Gu, Yun Chai, Shuiqin Chen, Bin Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuit & Intelligent Chongqing 400715 Peoples R China Chongqing Univ Sci & Technol Coll Chem & Chem Engn 20 East Daxuecheng Rd Chongqing 401331 Peoples R China

State-of-the-art face recognition method have achieved satisfactory performance on various face datasets, but limited to high-resolution (HR) faces. In real scenes, with affects from camera quality or distance etc., low quality face images are often obtained in surveillance scenes, which leads to poor performance of general face recognition approaches. In this work, we explored the reasons why low-resolution (LR) images cannot be coped with the conventional models, and proposed the optimization methods. The main contribution of this work includes two aspects: (1) We proposed an adaptive random down-sampling data augmentation method with the corresponding loss function and training strategy. For the network training, we mapped the face images with different resolutions to the same embedding space, and then explicitly separated different classes of LR and gathered the same class of HR and LR. (2) A pooling module based on attention mechanism, area attention pooling, was proposed. After pooling, the fine-grained information of the identity is retained, and the redundant information of the non-identity is removed. We conducted a series of visualization experiments to prove the effectiveness of our method. We evaluated our methods on the real-world LR face dataset Surveillance Cameras Face Database (SCface) and the public face evaluation dataset labeled Faces in the Wild (LFW). Compared with state-of-the-art methods, the proposed method achieves satisfactory performance at LR images on SCface and LFW dataset.

关键词： Low resolution face recognition Convolutional neural networks Feature extraction Pooling Data augmentation Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

Dynamic behaviors and bandwidth-enhanced time-delay signature suppression of mutually coupled spin-polarized vertical-cavity surface-emitting lasers

引用

OPTICS AND LASER TECHNOLOGY 2022年 149卷 107811-107811页

作者： Liu, Bocheng Xie, Yiyuan Song, Tingting Ye, Yichen Jiang, Xiao Chai, Junxiong Tang, Qianfeng Feng, Manying Southwest Univ Sch Elect & Informat Engn Chongqing 400715 Peoples R China Univ Elect Sci & Technol Chengdu Sch Optoelect Informat Chengdu 611731 Sichuan Peoples R China Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Chongqing Normal Univ Sch Comp & Informat Sci Chongqing 401331 Peoples R China

A novel symmetric model based on the mutually coupled spin-polarized vertical-cavity surface-emitting lasers (MCSP-VCSELs) subject to optically pumped is implemented and analyzed in this paper. Numerous nonlinear dynamic behaviors containing the state of steady, period, period-doubling, and chaos have been observed in the bifurcation diagram, revealing the period-doubling route to chaos. The characteristics of these dynamic behaviors are analyzed in detail along with clarifying the physical mechanism of their occurrence. The influences of laser parameters on the dynamic behaviors are investigated through mapping diagrams, and the results show that both the pump intensity and pump ellipticity are crucial to the complicated dynamic behaviors including chaos. Additionally, the increase of the coupling strength, coupling time delay, and frequency detuning have beneficial impacts to chaos while the spin relaxation rate has a harmful effect on the contrary. The comparison results on different parameter planes illustrate that the time-delay signature (TDS) suppression and bandwidth enhancement can be achieved simultaneously when coupling strength, frequency detuning and pump strength are relatively large. Therefore, the MCSP-VCSELs have an application prospect in chaos-based fields.

关键词： nonlinear dynamics Optical chaos Time delay signature suppression Bandwidth enhancement Mutually coupled spin-polarized VCSELs

来源：评论

学校读者我要写书评

暂无评论

EEG classification by filter band component regularized common spatial pattern for motor imagery

引用

BIOMEDICAL SIGNAL PROCESSING AND CONTROL 2020年第0期59卷 101917-101917页

作者： Guo, Yao Zhang, Yuan Chen, Zhiqiang Liu, Yi Chen, Wei Fudan Univ Sch Informat Sci & Technol Shanghai 200433 Peoples R China Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Civil Aviat Gen Hosp Dept Resp Med Beijing 100123 Peoples R China

The accurate identification of motor imagery (MI), a scheme to transfer intention into control commands based on brain-computer interface, remains to be a vital problem of rehabilitation treatment in smart health. As a most common-used algorithm in feature extraction, common spatial pattern (CSP) has a good effect in MI electroencephalogram (EEG) classification. However, the robustness of CSP is poor with small training samples. Furthermore, obvious performance variability occurs between frequency bands for different individuals. To address these two problems simultaneously, filter band component regularized common spatial patterns (FCCSP) is proposed by reducing estimation variance, adding priors on the spatial filters, and considering more local frequency domain information. Firstly, EEG signals are divided into four sub-bands using wavelet packet. Secondly, the component regularized common spatial patterns (CRCSP) that could extract key features by regularizing the covariance matrix estimation and the objective function is proposed. CRCSP makes the best of neurophysiology priors which is useful in improving the robustness of the covariance matrix. Thereafter, minimal-redundancy-maximal-relevance (mRMR) is applied to discard the irrelevant and redundant features. The final feature set as input is feed into linear discriminant analysis (LDA). Our proposed algorithm has been evaluated using public dataset BCI Competition III dataset IVa. The evaluation of our experiment involving 5 different subjects demonstrates the significant superiority of the proposed algorithm over its classical counterparts: the mean classification accuracy was increased by 15.72%. Experimental results verify particularly excellent performance as compared to the other state-of-the-art methods. (C) 2020 Elsevier Ltd. All rights reserved.

关键词： Brain-computer interfaces (BCI) Common spatial patterns (CSP) Electroencephalography (EEG) Feature selection Regularization

来源：评论

学校读者我要写书评

暂无评论

Novel terahertz optical switch based on PIT phenomenon and Lorentz theory

引用

ISCIENCE 2024年第12期27卷 111301页

作者： Zhu, Jun Chen, Xiner Qin, Liuli Guangxi Normal Univ Sch Elect & Informat Engn Guangxi Key Lab Brain Inspired Comp & Intelligent Guilin 541004 Peoples R China Guangxi Normal Univ Sch Elect & Informat Engn Sch Integrated Circuits Guilin 541004 Peoples R China Suzhou Univ Sci & Technol Sch Chem & Life Sci Key Lab Adv Electrode Mat Novel Solar Cells Petr & Suzhou 215009 Jiangsu Peoples R China Guangxi Normal Univ Sch Math & Stat Guilin 541004 Peoples R China

We propose and demonstrate a structure consisting of graphene rings and square rings that enables broadband and tunable plasmon-induced transparency (PIT) effects. Through coupled Lorentz model analysis, we attribute the transmission window at 2.1 THz to the interference between the equipartitioned exciton resonance of the graphene ring pairs and the inductive-capacitive resonance of the graphene square ring pairs. We also investigate the effect of the variation of the rotation angle of the internal graphene square ring pair on the transmission characteristics. The structure not only achieves a maximum modulation depth (MDA) of 91%, insertion loss (IL) and extinction ratio (ER) of 0.3 dB and 10.94 dB, respectively, but also achieves a maximum detection sensitivity of 0.96 THz/refractive index unit (RIU). In contrast, this study achieves more than 90% modulation amplitude in the range of 0.3 THz with a simple design structure, providing new insights for research and applications in related fields.

关键词： Physics Optics Electromagnetics

来源：评论

学校读者我要写书评

暂无评论

Roadmap on nonlinear optics-focus on Chinese research

引用

JOURNAL OF PHYSICS-PHOTONICS 2023年第3期5卷 032501页

作者： Ren, Mengxin Xu, Jingjun Lan, Pengfei Lu, Peixiang Li, Zhi-Yuan Hong, Li-Hong Wang, Yulei Bai, Zhenxu Lv, Zhiwei Zhou, Zhi-Yuan Shi, Bao-Sen Zhang, Yong Zhu, Shining Xiao, Min Aya, Satoshi Lu, Yan-qing Fan, Huixin Luo, Min Ye, Ning Sun, Zeyuan Liu, Wei-Tao Wu, Shiwei Li, Qingyun Hu, Hui Zheng, Yuanlin Chen, Xianfeng Hu, Xiaoyong Tian, Chuanshan Hu, Zixian Li, Guixin Hu, Yi Huang, Kun Zeng, Heping Li, Zhen-Ze Sun, Hong-Bo Dong, Lei Li, Runfeng Yang, Wenkai Shi, Kebin Nankai Univ TEDA Appl Phys Inst Sch Phys Key Lab Weak Light Nonlinear Photon Tianjin 300071 Peoples R China Huazhong Univ Sci & Technol Sch Phys Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Wuhan Natl Lab Opt Elect Wuhan 430074 Peoples R China South China Univ Technol Sch Phys & Optoelect Guangzhou 510640 Peoples R China Hebei Univ Technol Ctr Adv Laser Technol Tianjin 300401 Peoples R China Hebei Key Lab Adv Laser Technol & Equipment Tianjin Peoples R China Univ Sci & Technol China CAS Key Lab Quantum Informat Hefei 230026 Peoples R China Nanjing Univ Coll Engn & Appl Sci Natl Lab Solid State Microstruct Nanjing 210093 Peoples R China Univ Arkansas Dept Phys Fayetteville AR 72701 USA South China Univ Technol South China Adv Inst Soft Matter Sci & Technol AIS Sch Emergent Soft Matter Guangzhou 510640 Peoples R China South China Univ Technol Guangdong Prov Key Lab Funct & Intelligent Hybrid Guangzhou 510640 Peoples R China Nanjing Univ Key Lab Intelligent Opt Sensing & Manipulat Nanjing 210093 Peoples R China Chinese Acad Sci Fujian Inst Res Struct Matter Fuzhou 350002 Peoples R China Tianjin Univ Technol Inst Funct Crystals Tianjin 300384 Peoples R China Fudan Univ State Key Lab Surface Phys Key Lab Micro & Nano Photon Struct MOE Shanghai 200433 Peoples R China Fudan Univ Dept Phys Shanghai 200433 Peoples R China Shandong Univ Sch Phys Jinan 250100 Peoples R China Shanghai Jiao Tong Univ Sch Phys & Astron State Key Lab Adv Opt Commun Syst & Networks Shanghai 200240 Peoples R China Peking Univ State Key Lab Mesoscop Phys Beijing 100871 Peoples R China Peking Univ Frontiers Sci Ctr Nanooptoelectron Sch Phys Beijing 100871 Peoples R China Southern Univ Sci & Technol Dept Mat Sci & Engn Shenzhen 518055 Peoples R China East China Normal Univ State Key Lab Precis Spect Shanghai Peoples R China Chongqing Inst East China Normal Univ Chongqing Key Lab Precis Opt Chongqing 401121 Peoples R China Jilin Univ Coll Elect Sci & Engn State K

In nonlinear optical systems, the optical superposition principle breaks down. The system's response (including electric polarization, current density, etc) is not proportional to the stimulus it receives. Over the past half century, nonlinear optics has grown from an individual frequency doubling experiment into a broad academic field. The nonlinear optics has not only brought new physics and phenomena, but also has become an enabling technology for numerous areas that are vital to our lives, such as communications, health, advanced manufacturing, et al. This Roadmap surveys some of the recent emerging fields of the nonlinear optics, with a special attention to studies in China. Each section provides an overview of the current and future challenges within a part of the field, highlighting the most exciting opportunities for future research and developments.

关键词： nonlinear optics lasers nonlinear materials nonlinear effects applications

来源：评论

学校读者我要写书评

暂无评论

Fabrication of integrated silicon PIN detector based on Al-Sn-Al bonding for ΔE-E telescope application

引用

MICROELECTRONIC ENGINEERING 2021年 247卷 111599-111599页

作者： Zhu, Zhiyuan Yu, Min Jin, Yufeng Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing Peoples R China Peking Univ Inst Microelect Natl Key Lab Sci & Technol Micro Nano Fabricat Beijing Peoples R China

The conventional monolithic integration of silicon PIN detector as Delta E-E telescope suffer from incompatibility with IC process, signal crosstalk, high cost, etc.. In this paper, integrated silicon PIN detector based on Al-Sn-Al bonding is proposed. The intermediate conductive layer between thin and thick PIN structure comprises of metallic bonding layer, which can reduce the signal crosstalk. Moreover, the novel integration process enables known-good-die (KGD) bonding of thin and thick PIN structure, which increases flexibility and reliability of the fabrication process. Besides, the fabrication reduces the cost and increases reliability by utilizing silicon integrated process.

关键词： Silicon PIN detector Al-Sn-Al bonding Integration

来源：评论

学校读者我要写书评

暂无评论

Distributed differentially private average consensus for multi-agent networks by additive functional Laplace noise

引用

JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS 2020年第6期357卷 3565-3584页

作者： Dong, Tao Bu, Xiangyu Hu, Wenjie Southwest Univ Coll Elect & Informat Engn Chongqing 400715 Peoples R China Southwest Univ Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Chongqing Technol & Business Univ Sch Management Chongqing 400067 Peoples R China

Average consensus is a very useful consensus algorithm for distributed cooperative control and computing, where all the agents in the network communicate with its neighbor and reach the average of the initial states of all agents. The main secure defect of average consensus is that the initial state of agent can be inferred by using the state information sequence of the agent, which leads to the information disclosure. To handle this problem, in this paper, based on functional Laplace noise and differential privacy scheme, a novel differentially private average consensus algorithm is proposed to preserve the privacy of the state of each agent in the whole process of consensus computation. We develop detailed consensus analysis of our algorithm, including convergence rate and the consensus condition for network agents. Moreover, the privacy-preserving analysis is also given, which indicates that privacy of the states of all agents is guaranteed to preserve. Finally, a numerical experiment is used to demonstrate the effectiveness of our algorithm. (C) 2020 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Effects of initial conditions on the synchronization of the coupled memristor neural circuits

引用

nonlinear DYNAMICS 2019年第2期95卷 1269-1282页

作者： Zhang, Jihong Liao, Xiaofeng Southwest Univ Chongqing Key Lab Nonlinear Circuits & Intelligen Coll Elect & Informat Engn Chongqing 400715 Peoples R China

The fourth basic two-terminal circuit element has been called the memristor because its resistance (conductance) depends on the complete past history of the memristor current (voltage), i.e., the initial charge (flux) condition at a given instant. This paper aims to provide some insight into the effects of the initial flux condition of the memristor synapse in the synchronization of two coupled memristor-based neural circuits. First, we build the coupled memristor-based FitzHugh-Nagumo circuits with the memristor synapse, and obtain the initial conditions by means of the flux-charge analysis method in the differential equations. Then, as a result of varying the initial conditions of the coupling memristor in the neural network, the details of synchronization with the parallel shift are derived theoretically by solving the nonhomogeneous error equations. These results of theoretical analyses have been confirmed by numerical simulations. Finally, we focus on the influence of the initial condition of the memristor on chaos generation for individual FitzHugh-Nagumo neuron and how to change chaotic state into stable periodic oscillation for a FitzHugh-Nagumo neuron in the synchronous neural network.

关键词： Initial flux condition Memristor synapse FitzHugh-Nagumo circuit Synchronization Coupled neural network

来源：评论

学校读者我要写书评

暂无评论

Neural networks for power management optimal strategy in hybrid microgrid

引用

NEURAL COMPUTING & APPLICATIONS 2019年第7期31卷 2635-2647页

作者： Wang, Tiancai He, Xing Deng, Ting Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

This paper proposes a more reasonable objective function for combined economic emission dispatch problem. To solve it, Lagrange programming neural network (LPNN) is utilized to obtain optimal scheduling of a hybrid microgrid, which includes power generation resources, variable demands and energy storage system for energy storing and supplying. Combining variable neurons with Lagrange neurons, the LPNN aims to minimize the cost function and maximize the power generated by the renewable sources. The asymptotic stability condition of the neurodynamic model is analyzed, and simulation results show that optimal power of each component with certain time interval can be obtained. In addition, a new method by radial basis function neural network is proposed to predict the power values of renewable energy and load demand, which are used as the input values in the optimal process.

关键词： Hybrid microgrid RBF neural network prediction Quadratic optimization Lagrange programming neural network Renewable energy sources

来源：评论

学校读者我要写书评

暂无评论

Distributed algorithm based on consensus control strategy for dynamic economic dispatch problem

引用

INTERNATIONAL JOURNAL OF ELECTRICAL POWER & ENERGY SYSTEMS 2021年 129卷 106833-106833页

作者： Duan, Yuzhu He, Xing Zhao, You Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Chongqing Univ Coll Comp Chongqing 400044 Peoples R China

In this paper, the dynamic economic dispatch problem (DEDP) of a hybrid power network is studied. The microgrid model is constructed with three types of non-renewable power sources, renewable energy and energy storage batteries. Meanwhile, the pollutant emission costs and benefit function are considered here to maximize the total welfare. Firstly, we transform the benefit-maximum problem into its equivalent minimization problem which the generators connected each other in a directed graph. Then, we introduced a distributed algorithm based on alternating direction method of multipliers (ADMM) to find the optimal solution of different generators within 24 h. To solve the sub-problem, the bisection method and the finite-time distributed algorithm are adopted here. Finally, the convergence of the distributed algorithm is analyzed. Moreover, our experimental results verify that the optimal solution satisfies the constraints of supply?demand balance equality constraint and capacity inequalities well in each time slot, and meanwhile satisfies the constraints of output threshold between different time slots, furthermore, we utilized a practical example to make a comparison with the existed literature, and the superiority of the algorithm is verified to a certain extent.

关键词： ADMM Distributed algorithm Dynamic economic dispatch Micro-grid

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：