检索结果-内蒙古大学图书馆

arXiv 2018年

作者： Deǧerli, Ayşen Aslan, Sinem Yamaç, Mehmet Sankur, Bülent Gabbouj, Moncef Tampere University of Technology Laboratory of Signal Processing Tampere Finland Ca' Foscari University of Venice European Centre for Living Technology Venice Italy Ege University International Computer Institute Izmir Turkey Boǧaziçi University Electrical and Electronics Engineering Department Istanbul Turkey

—Compressive Sensing (CS) theory asserts that sparse signal reconstruction is possible from a small number of linear measurements. Although CS enables low-cost linear sampling, it requires non-linear and costly reconstruction. Recent literature works show that compressive image classification is possible in CS domain without reconstruction of the signal. In this work, we introduce a DCT base method that extracts binary discriminative features directly from CS measurements. These CS measurements can be obtained by using (i) a random or a pseudo-random measurement matrix, or (ii) a measurement matrix whose elements are learned from the training data to optimize the given classification task. We further introduce feature fusion by concatenating Bag of Words (BoW) representation of our binary features with one of the two state-of-the-art CNN-based feature vectors. We show that our fused feature outperforms the state-of-the-art in both cases. Copyright © 2018, The Authors. All rights reserved.

关键词： Compressed sensing

来源：评论

学校读者我要写书评

暂无评论

A computationally efficient detector: Using both the test and training data for disturbance correlation estimation

A computationally efficient detector: Using both the test an...

引用

2017 International Conference on Radar Systems, Radar 2017

作者： Liu, Jun Zhao, Hong-Yan Liu, Weijian Zhu, Shengqi Liu, Hongwei Li, Hongbin National Laboratory of Radar Signal Processing Xidian University Xi'an710071 China Wuhan Electronic Information Institute Wuhan430019 China Department of Electrical and Computer Engineering Stevens Institute of Technology HobokenNJ07030 United States

ISBN: (纸本)9781785615153

This paper examines a target detection problem in colored Gaussian disturbance with an unknown covariance matrix. In many classic adaptive detectors, the covariance estimator is formed by using only the training data. This necessitates calculating a new covariance estimator for each cell under test (CUT) during the cell-by-cell target search process. We consider herein an alternative approach that forms the covariance matrix estimate by using both test and training data for detection in homogeneous environments. This approach is computationally much more efficient since the covariance matrix estimator is computed only once and can be applied for target detection at each CUT. Using this estimator, we propose a new detector with two tunable parameters, which includes several existing detectors as special cases. Closed-form expressions for the probabilities of false alarm and detection are derived in the matched and mismatched cases. Simulation results reveal that the rejection capability of mismatched signals of the proposed detector can be flexibly controlled by adjusting its tunable parameters. © 2017 Institution of engineering and Technology. All rights reserved.

关键词： Covariance matrix

来源：评论

学校读者我要写书评

暂无评论

Secure quantum communication in the presence of phase- and polarization-dependent loss

引用

Physical Review A 2018年第4期98卷 042324-042324页

作者： Chenyang Li Marcos Curty Feihu Xu Olinka Bedroya Hoi-Kwong Lo Center for Quantum Information and Quantum Control Department of Electrical & Computer Engineering and Department of Physics University of Toronto Toronto Ontario M5S 3G4 Canada Escuela de Ingeniería de Telecomunicación Department of Signal Theory and Communications University of Vigo Vigo E-36310 Spain Shanghai Branch Hefei National Laboratory for Physical Sciences at the Microscale University of Science and Technology of China Shanghai 201315 China

Silicon photonics holds the promise of the miniaturization of quantum communication devices. Recently, silicon chip optical transmitters for quantum key distribution (QKD) have been built and demonstrated experimentally. Nonetheless, these silicon chips suffer substantial phase- and polarization-dependent loss (PDL) which, if unchecked, could compromise the security of QKD systems due to overestimation of the secret key rate. Here, first, we restore the security by regarding single photons without phase and polarization dependence as untagged and secure qubits. Next, by using a postselection technique, we implement a secure QKD protocol that provides a high key generation rate even in the presence of severe phase- and polarization-dependent loss. Our solution is simple to realize in a practical experiment, as it does not require any hardware modification.

关键词： Quantum cryptography

来源：评论

学校读者我要写书评

暂无评论

Bifocal Technique Applied to Dual Transmitarray Antennas

Bifocal Technique Applied to Dual Transmitarray Antennas

引用

European Conference on Antennas and Propagation

作者： Eduardo Martinez-de-Rioja Jose A. Encinar Antonio Pino Borja Gonzalez-Valdes Sean V. Hum Carolina Tienda Giovanni Toso Information Processing and Telecomm. Center Universidad Politécnica de Madrid Spain Department of Signal Theory and Communications Universidad de Vigo Spain Dept. of Electrical and Computer Engineering University of Toronto Canada Antenna Group Airbus Defence & Space Stevenage United Kingdom Electromagnetics Division European Space Agency Noordwijk The Netherlands

The bifocal technique has been applied to dual transmitarray antennas in order to improve the performance of multi-beam satellite antennas in Ka-band. The use of a transmitarray design allows centered and rotationally symmetrical geometries to be used without blockage. Different dual transmitarray configurations have been studied for operation in transmission in Ka-band (20 GHz), and the bifocal technique has been evaluated for the generation of closely-spaced beams. The results of the simulations show that it is possible to achieve a high degree of beam compression, but at the cost of a reduction in the antenna radiation efficiency.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multilinear class-specific discriminant analysis

arXiv

引用

arXiv 2017年

作者： Tran, Dat Thanh Gabbouj, Moncef Iosifidis, Alexandros Laboratory of Signal Processing Tampere University of Technology Tampere Finland Department of Engineering Electrical and Computer Engineering Aarhus University

There has been a great effort to transfer linear discriminant techniques that operate on vector data to high-order data, generally referred to as Multilinear Discriminant Analysis (MDA) techniques. Many existing works focus on maximizing the inter-class variances to intra-class variances defined on tensor data representations. However, there has not been any attempt to employ class-specific discrimination criteria for the tensor data. In this paper, we propose a multilinear subspace learning technique suitable for applications requiring class-specific tensor models. The method maximizes the discrimination of each individual class in the feature space while retains the spatial structure of the input. We evaluate the efficiency of the proposed method on two problems, i.e. facial image analysis and stock price prediction based on limit order book data. Copyright © 2017, The Authors. All rights reserved.

关键词： Tensors

来源：评论

学校读者我要写书评

暂无评论

Feasibility of Self-backhauling in Full-Duplex Radio Access Systems under QoS Constraints

Feasibility of Self-backhauling in Full-Duplex Radio Access ...

引用

IEEE International Conference on communications Workshops

作者： Dani Korpi Taneli Riihonen Mikko Valkama Laboratory of Electronics and Communications Engineering Tampere University of Technology Finland Department of Signal Processing and Acoustics Aalto University School of Electrical Engineering Finland

ISBN: (纸本)9781509015269

This paper investigates the feasibility of a radio access system with a self-backhauling access node under full-duplex and half-duplex operation modes. In particular, after making certain simplifying assumptions, closed-form solutions for the feasibility conditions of such a radio access system are derived for both of the considered operation modes. Furthermore, the analysis incorporates given quality of service (QoS) constraints for the system, defined in terms of minimum data rates. The numerical results show that the full-duplex scheme outperforms the corresponding half-duplex scheme under most circumstances, both in terms of the highest achievable rates and the highest tolerable path losses, when given the same QoS target. However, this requires a certain amount of self-interference attenuation in the access node. Performing a similar feasibility analysis without any simplifications in the system model is an important future work item.

关键词： Interference Quality of service Receiving antennas High definition video Silicon Antenna arrays signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Efficient and Robust Machine Learning for Real-World Systems

arXiv

引用

arXiv 2018年

作者： Pernkopf, Franz Roth, Wolfgang Zöhrer, Matthias Pfeifenberger, Lukas Schindler, Günther Fröning, Holger Tschiatschek, Sebastian Peharz, Robert Mattina, Matthew Ghahramani, Zoubin Department of Electrical Engineering Laboratory of Signal Processing and Speech Communication Graz University of Technology Austria Institute of Computer Engineering Ruperts Karls University Heidelberg Germany Microsoft Research Cambridge United Kingdom Machine Learning Group Department of Engineering University of Cambridge United Kingdom Machine Learning Group Department of Engineering University of Cambridge United Kingdom Uber AI Labs CA United States Arm Research Arm Ltd. Cambridge United Kingdom

While machine learning is traditionally a resource intensive task, embedded systems, autonomous navigation and the vision of the Internet-of-Things fuel the interest in resource efficient approaches. These approaches require a carefully chosen tradeoff between performance and resource consumption in terms of computation and energy. On top of this, it is crucial to treat uncertainty in a consistent manner in all but the simplest applications of machine learning systems. In particular, a desideratum for any real-world system is to be robust in the presence of outliers and corrupted data, as well as being "aware" of its limits, i.e. the system should maintain and provide an uncertainty estimate over its own predictions. These complex demands are among the major challenges in current machine learning research and key to ensure a smooth transition of machine learning technology into every day’s applications. In this article, we provide an overview of the current state of the art of machine learning techniques facilitating these real-world requirements. First we provide a comprehensive review of resource-efficiency in deep neural networks with focus on techniques for model size reduction, compression and reduced precision. These techniques can be applied during training or as post-processing and are widely used to reduce both computational complexity and memory footprint. As most (practical) neural networks are limited in their ways to treat uncertainty, we contrast them with probabilistic graphical models, which readily serve these desiderata by means of probabilistic inference. In that way, we provide an extensive overview of the current state-of-the-art of robust and efficient machine learning for real-world systems. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Improving efficiency in convolutional neural network with multilinear filters

arXiv

引用

arXiv 2017年

作者： Tran, Dat Thanh Iosifidis, Alexandros Gabbouj, Moncef Laboratory of Signal Processing Tampere University of Technology Tampere Finland Department of Engineering Electrical and Computer Engineering Aarhus University Aarhus Denmark

The excellent performance of deep neural networks has enabled us to solve several automatization problems, opening an era of autonomous devices. However, current deep net architectures are heavy with millions of parameters and require billions of floating point operations. Several works have been developed to compress a pre-trained deep network to reduce memory footprint and, possibly, computation. Instead of compressing a pre-trained network, in this work, we propose a generic neural network layer structure employing multilinear projection as the primary feature extractor. The proposed architecture requires several times less memory as compared to the traditional Convolutional Neural Networks (CNN), while inherits the similar design principles of a CNN. In addition, the proposed architecture is equipped with two computation schemes that enable computation reduction or scalability. Experimental results show the effectiveness of our compact projection that outperforms traditional CNN, while requiring far fewer parameters. Copyright © 2017, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Subpixel Target Detection Approach to Hyperspectral Image Classification

引用

IEEE Transactions on Geoscience and Remote Sensing 2017年第9期55卷 5093-5114页

作者： Xue, Bai Yu, Chunyan Wang, Yulei Song, Meiping Li, Sen Wang, Lin Chen, Hsian-Min Chang, Chein-I Center for Hyperspectral Imaging in Remote Sensing Information and Technology College Dalian Maritime University Dalian China Remote Sensing Signal and Image Processing Laboratory Department of Computer Science and Electrical Engineering University of Maryland Baltimore County BaltimoreMD21250 United States Key Laboratory of Spectral Imaging Technology Chinese Academy of Sciences Xi'an China State Key Laboratory of Integrated Services Networks Xi'an China School of Physics and Optoelectronic Engineering Xidian University Xi'an China Department of Medical Research Taichung Veterans General Hospital Taichung Taiwan Department of Computer Science and Information Management Providence University Taichung02912 Taiwan

Hyperspectral image classification faces various levels of difficulty due to the use of different types of hyperspectral image data. Recently, spectral-spatial approaches have been developed by jointly taking care of spectral and spatial information. This paper presents a completely different approach from a subpixel target detection view point. It implements four stage processes, a preprocessing stage, which uses band selection (BS) and nonlinear band expansion, referred to as BS-then-nonlinear expansion (BSNE), a detection stage, which implements constrained energy minimization (CEM) to produce subpixel target maps, and an iterative stage, which develops an iterative CEM (ICEM) by applying Gaussian filters to capture spatial information, and then feeding the Gaussian-filtered CEM-detection maps back to BSNE band images to reprocess CEM in an iterative manner. Finally, in the last stage Otsu's method is applied to converting ICEM-detected real-valued maps to discrete values for classification. The entire process is called BSNE-ICEM. Experimental results demonstrate BSNE-ICEM, which has advantages over support vector machine-based approaches in many aspects, such as easy implementation, fewer parameters to be used, and better false classification and precision rates. © 1980-2012 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

A Novel Clustering Method Based on Density Peaks and Its Validation by Channel Measurement

A Novel Clustering Method Based on Density Peaks and Its Val...

引用

European Conference on Antennas and Propagation

作者： Hanqing Xue Lei Tian Yuxiang Zhang Chao Wang Jianhua Zhang Wei Li Hewen Wei Key Lab of Universal Wireless Communications Ministry of Education State Key Lab of Networking and Switching Technology Beijing University of Posts and Telecommunications P. O. Box 92 Beijing 100876 China Department of Electrical Engineering Northern Illinois University DeKalb IL 60115 National key laboratory of science and technology on blind signal processing Chengdu Sichuan 610041 China

ISBN: (纸本)9781785618154;9781785618161

In this paper, cluster centers are defined as local maxima in the density which is called density peaks. We use an algorithm based on density peaks to cluster multipath components (MPCs) whose parameters are estimated by the space-alternating generalized expectation-maximization (SAGE) algorithm. To validate its performance, we apply it to the data from a channel measurement in an indoor scenario at 28 GHz and compare it to the widely used KPowerMeans algorithm with the same distance metric of multiple component distance (MCD). Through the results, we find it can cluster MPCs more accurately validated by Calinski-Harabasz (CH) and Davies-Bouldin (DB) indices. More importantly, it overcomes the shortcoming of setting the number of clusters manually because we can determine the cluster centers automatically. Besides, the efficiency of DP clustering is far superior to KPowerMeans because what it needs is only a distance matrix and we only need to compute it once. At the end, statistical analysis of clusters are presented to provide insights in channel modeling of millimeter wave.

关键词：

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：