检索结果-内蒙古大学图书馆

Progressive Masking Oriented Self-Taught Learning for Occluded Facial Expression Recognition

IEEE Transactions on Affective Computing 2025年

作者： Kang, Bin Wang, Shuangshuang Wang, Zongyu Li, Xin Dou, Haie Wang, Lei Xia, Zhijie College of Internet of Things Nanjing University of Posts and Telecommunications Nanjing China Key Laboratory of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications Ministry of Education Nanjing China School of Mechanical Engineering Southeast University Nanjing China

Self-taught learning (STL) is a promising solution that reduces the performance gap between weakly supervised and fully supervised learning for easily accessible, lab.l-free images. The success of traditional STL solutions relies on the assumption that the target appearance is completely visible and well-defined. In real-world facial expression recognition scenarios, however, saliency regions are often partially occluded, which significantly hampers the generalization capability of STL methods. Nevertheless, few studies have investigated the impact of occlusion on STL. In this paper, we propose an interweaved autoencoder network for weakly supervised facial expression recognition in occlusion scenarios. The key innovation of our network lies in the Residual Connection Union (RCU) blocks that can integrate the Convolutional Neural network (CNN) and Transformer layers into a multi-scale structure. The RCU enables a progressive masking strategy to accurately identify and focus on contributive yet often overlooked image patches by analyzing the relationships among region-level target representations. In addition, we introduce a self-knowledge distillation module for the effective training of the proposed autoencoder network. Extensive experiments are conducted on four public datasets to demonstrate the superiority of our method over related works. © 2010-2012 IEEE All rights reserved.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Improved finite-size analysis for measurement-device-independent quantum digital signature

arXiv

引用

arXiv 2025年

作者： Zhu, Jia-Li Zhang, Chun-Hui Wang, Yue-Ying Wang, Qin Institute of quantum information and technology Nanjing University of Posts and Telecommunications Nanjing210003 China "Broadband Wireless Communication and Sensor Network Technology" Key Lab of Ministry of Education NUPT Nanjing210003 China "Telecommunication and Networks" National Engineering Research Center NUPT Nanjing210003 China

Quantum digital signatures (QDS), based on the principles of quantum mechanics, provide information-theoretic security, ensuring the integrity, authenticity, and non-repudiation of data transmission. With present QDS protocols, measurement-device-independent QDS (MDI-QDS) can resist all attacks on detections, yet it suffers from finite-size effect. In this work, we present and compare three parameter estimation models for finite-size analysis of two-decoy MDI-QDS. The first model is a commonly used model in previous schemes, and we propose two new models to improve the performance. Subsequently, we perform numerical simulations to evaluate the performance of the three models. The results demonstrate that the proposed methods are less affected by finite-size effect, thereby effectively enhancing the signature rate. This work contributes to the practical development of QDS. Copyright © 2025, The Authors. All rights reserved.

关键词： Authentication

来源：评论

学校读者我要写书评

暂无评论

Hybrid Implementation for Untrusted-node-based Quantum key Distribution network

arXiv

引用

arXiv 2025年

作者： Liu, Jingyang Zhou, Xingyu Ding, Huajian Xu, Jiaxin Zhang, Chunhui Li, Jian Wang, Qin Institute of quantum information and technology Nanjing University of Posts and Telecommunications Nanjing210003 China "Broadband Wireless Communication and Sensor Network Technology"Key Lab of Ministry of Education Ministry of Education Nanjing210003 China "Telecommunication and Networks"National Engineering Research Center NUPT Nanjing210003 China

Quantum key distribution (QKD) serves as a cornerstone of secure quantum communication, providing unconditional security grounded in quantum mechanics. While trusted-node networks have facilitated early QKD deployment, their vulnerability to node compromise underscores the need for untrusted-node architectures. Measurement-device-independent QKD (MDI-QKD) and twin-field QKD (TF-QKD) have emerged as leading candidates, addressing security vulnerabilities and extending transmission distances. Despite the wide adoptions in various fiber scaling, no integrated implementation of these two protocols has been demonstrated to date. Here, we present a hybrid system that seamlessly integrates TF-QKD and MDI-QKD into one untrusted-node-based architecture. Utilizing an efficient phase estimation method based on asymmetric interferometers, we convert twin-field global phase tracking to relative phase calibration, allowing near continuous running of both protocols. Experiments demonstrate secure finite-size key rates for sending-or-not-sending QKD and MDI-QKD over fiber distances of 150 to 431 km. The results align with theoretical simulations and show the ability to surpass the absolute repeaterless key capacity. Our work offers an unified framework for deploying multi-protocol QKD networks, laying the foundation for adaptable and scalab.e quantum infrastructures that can meet a wide range of security and performance needs. Copyright © 2025, The Authors. All rights reserved.

关键词： Quantum communication

来源：评论

学校读者我要写书评

暂无评论

Simplification of I-Vector Extraction for Speaker Identification

引用

Chinese Journal of Electronics 2016年第6期25卷 1121-1126页

作者： XU Longting YANG Zhen SUN Linhui Broadband Wireless Communication and Sensor Network Technology Key Lab Nanjing University of Posts and Telecommunications

The identity vector(i-vector) approach has been the state-of-the-art for text-independent speaker recognition, both identification and verification in recent years. An i-vector is a low-dimensional vector in the socalled total variability space represented with a thin and tall rectangular matrix. This paper introduces a novel algorithm to improve the computational and memory requirements for the application. In our method, the series of symmetric matrices can be represented by diagonal expression,sharing the same dictionary, which to some extent is analogous to eigen decomposition, and we name this algorithm Eigen decomposition like factorization(EDLF). Similar algorithms are listed for comparison, in the same condition,our method shows no disadvantages in identification accuracy.

关键词： Speaker identification Closed-set I-vector Symmetric matrix factorization

来源：评论

学校读者我要写书评

暂无评论

Cross-layer resource allocation based on equivalent bandwidth in OFDMA systems

引用

Journal of Systems Engineering and Electronics 2016年第4期27卷 754-762页

作者： Su Pan Cheng Li Sheng Zhang Danwei Chen Key Lab of Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts & Telecommunications

A quality of service（QoS） guaranteed cross-layer resource allocation algorithm with physical layer, medium access control（MAC） layer and call admission control（CAC） considered simultaneously is proposed for the full IP orthogonal frequency division multiple access（OFDMA） communication system, which can ensure the quality of multimedia services in full IP *** algorithm converts the physical layer resources such as subcarriers, transmission power, and the QoS metrics into equivalent bandwidth which can be distributed by the base station in all three layers. By this means, the QoS requirements in terms of bit error rate（BER）, transmission delay and dropping probability can be guaranteed by the cross-layer optimal equivalent bandwidth allocation. The numerical results show that the proposed algorithm has higher spectrum efficiency compared to the existing systems.

关键词： resource allocation equivalent bandwidth crosslayer optimization orthogonal frequency division multiple access(OFDMA)

来源：评论

学校读者我要写书评

暂无评论

Speech Signal Processing on Graphs: Graph Topology, Graph Frequency Analysis and Denoising

引用

Chinese Journal of Electronics 2020年第5期29卷 926-936页

作者： WANG Tingting GUO Haiyan LYU Bin YANG Zhen Broadband Wireless Communication and Sensor Network Technology Key Lab Nanjing University of Posts and Telecommunications

The paper investigates the hidden relationships among speech samples by applying graph tools. Specifically, we first estimate an applicable graph topology for unstructured speech signals, which can map speech signals into the vertex domain successfully and construct as Speech graph signals(SGSs). On the basis, we define a new graph Fourier transform for SGSs, which can investigate its related graph Fourier analysis. Moreover,we propose a new Graph structure spectral subtraction(GSSS) method for speech enhancement under different noisy environments. Simulation results show that the performance of the GSSS method can be significantly improved than the classical Basic spectral subtraction(BSS) method in terms of the average Segmental signal-tonoise ratio(SSNR), Perceptual evaluation of speech quality(PESQ) and the computational complexity.

关键词： Graph Fourier transform Graph topology graph frequency analysis speech processing graph topology unstructured speech signals GSSS method Speech graph Fourier analysis graph theory Fourier transforms speech quality classical Basic spectral subtraction method Denoising speech signal processing perceptual evaluation of speech quality computational complexity speech samples Graph structure spectral subtraction method Graph signal processing segmental signal-to-noise ratio speech enhancement graph tools Fourier analysis SGSs

来源：评论

学校读者我要写书评

暂无评论

On Least Squares Weighted Recursive Estimation of Clock Skew and Offset in wireless sensor networks

引用

Chinese Journal of Electronics 2017年第5期26卷 1041-1047页

作者： HU Bing SUN Zhixin Key Laboratory of Broadband Wireless Communication and Sensor Network Technology Ministry of EducationNanjing University of Posts and Telecommunications

Three clock synchronization algorithms for wireless sensor networks(WSNs) in Pairwise broadcast synchronization(PBS) mechanism are *** include the joint Least squares estimator(LS),joint Least squares weighted estimator(LSW) and joint Least squares weighted Recursive estimator(R-LSW).For these estimators,the corresponding algorithms are derived and described by assuming a Gaussian random delay *** PBS,these estimators can achieve the Cramer-Rao lower bound(CRLB) for both listening node and active node without knowledge of the deterministic *** purpose of considering R-LSW is to reduce the use of storage space with the method of estimating while *** and analytical results verify that the estimators are efficient.

关键词： Clock synchronization wireless sensor networks(WSNs) Pairwise broadcast synchronization(PBS) Least squares estimator(LS)

来源：评论

学校读者我要写书评

暂无评论

LANDMARC Localization Algorithm Based on Weight Optimization

引用

Chinese Journal of Electronics 2018年第6期27卷 1291-1296页

作者： HU Bing PENG Huijun SUN Zhixin Key Laboratory of Broadband Wireless Communication and Sensor Network Technology Ministry of EducationNanjing University of Posts and Telecommunications

In order to solve the problem that the difference of Received signal strength(RSS) between tags will become large when tags are close to the reader, which exists in LANDMARC system, a LANDMARC localization algorithm based on weight optimization is proposed in the paper. We optimize the weight by redefining the formulas of in-weight and ex-weight. In-weight formula is discussed under the free space propagation model. The definition of ex-weight is obtained by analyzing the relationship between RSS and distance under the log-distance path loss model. We evaluate the performance of the proposed algorithm and compare it with the algorithm based on normalized weight. The simulation results show the superior performance of the proposed algorithm in terms of location accuracy.

关键词： Location identification based on dynamic active RFID calibration(LANDMARC) system Radio frequency identification(RFID) Indoor localization Weight optimization

来源：评论

学校读者我要写书评

暂无评论

Proportional Fairness-Based Energy-Efficient Power Allocation in Downlink MIMO-NOMA Systems with Statistical CSI

引用

China communications 2019年第12期16卷 47-55页

作者： Yonghai Lin Zhen Yang Haiyan Guo Key Laboratory of Ministry of Education in Broadband Wireless Communication and Sensor Network Technology Nanjing University of Posts and Telecommunications

In this paper, proportional fairness(PF)-based energy-efficient power allocation is studied for multiple-input multiple-output(MIMO) non-orthogonal multiple access(NOMA) systems. In our schemes, statistical channel state information(CSI) is utilized for perfect CSI is impossible to achieve in practice. PF is used to balance the transmission efficiency and user fairness. Energy efficiency(EE) is formulated under basic data rate requirements and maximum transmitting power constraints. Due to the non-convex nature of EE, a two-step algorithm is proposed to obtain sub-optimal solution with a low complexity. Firstly, power allocation is determined by golden section search for fixed power. Secondly total transmitting power is determined by fractional programming method in the feasible regions. Compared to the performance of MIMO-NOMA without PF constraint, fairness is obtained at expense of decreasing of EE.

关键词： MIMO-NOMA energy efficiency proportional fairness fractional programming method golden section search

来源：评论

学校读者我要写书评

暂无评论

Load Balanced Coding Aware Multipath Routing for wireless Mesh networks

引用

Chinese Journal of Electronics 2015年第1期24卷 8-12页

作者： SHAO Xing WANG Ruchuan HUANG Haiping SUN Lijuan College of Computer Nanjing University of Posts and Telecommunications College of Information Engneering Yancheng Institutue of Technology Jiangsu High Technology Research Key Laboratory for Wireless Sensor Networks Key Lab of Broadband Wireless Communication and Sensor Network Technology

The growth of network coding opportunities is considered the unique optimization goal by most current network coding based routing algorithms for wireless mesh networks. This usually results in flows aggregation problem in areas with coding opportunities, and degrades the network performance. This paper proposes a Load balanced coding aware multipath routing(LCMR)for wireless mesh networks. To facilitate the evaluation of discovered multiple paths and the tradeoffs between coding opportunity and load balancing, a novel routing metric, Load balanced coding aware routing metric(LCRM) is presented, which considers the load degree of nodes when detects coding opportunities. LCMR could spread traffic over multipath to further balance load. Simulation results demonstrate that LCMR could evenly spread the traffic over the network with increasing network throughput in a heavy load at the expense of some coding opportunities.

关键词： Load balanced Coding aware Multipath Routing wireless mesh networks.

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：