检索结果-内蒙古大学图书馆

IEEE International Symposium on Information Theory (ISIT)

作者： Agrnon, Shlomi Benger, Etam Ordentlich, Or Tishby, Naftali Hebrew Univ Jerusalem Sch Comp Sci & Engn Jerusalem Israel Hebrew Univ Jerusalem Edmond & Lily Safra Ctr Brain Sci Jerusalem Israel

ISBN: (纸本)9781538682098

In rate-distortion (RD) problems one seeks reduced representations of a source that meet a target distortion constraint. Such optimal representations undergo topological transitions at some critical rate values, when their cardinality or dimensionality change. We study the convergence time of the Arimoto-Blahut alternating projection algorithms, used to solve such problems, near those critical points, both for the rate-distortion and information bottleneck settings. We argue that they suffer from critical slowing down - a diverging number of iterations for convergence - near the critical points. This phenomenon can have theoretical and practical implications for both machine learning and data compression problems.

关键词： Rate-distortion Data compression Machine learning Distortion projection algorithms Convergence

来源：评论

学校读者我要写书评

暂无评论

The Development of a Head-up Display (HUD) App on the Android Mobile Operating System 27

The Development of a Head-up Display (HUD) App on the Androi...

引用

IEEE 27th International Symposium for Design and Technology in Electronic Packaging (SIITME)

作者： Szabo, Roland Gontean, Aurel Burta, Andrei Politehn Univ Fac Elect Telecommun & Informat Technol Appl Elect Dept Timisoara Romania

ISBN: (纸本)9781665421102

The head-up display (HUD) is a premium feature found mostly in expensive cars. The method how it works is really simple, some information has to be projected on the windscreen of the car, so the driver won't have to look down on the gauges, and he can keep his eyes always on the road. A head-up display (HUD) is useful due to the fact that it can project the speed of the car on the windscreen and this way the driver can avoid speeding tickets. Also the projection on the windscreen it's small and transparent, so it won't bother the driver, which can watch the road, not the gauges. The projection method is straight forward, due to the fact that almost any object is reflected on a car's windscreen, and a light up display, of a mobile phone, on low lighting conditions, it's reflected even more. The goal of this work is to create a head-up display (HUD) using an :Android mobile phone. This way this expensive feature for cars can be used in arty car, by the use of a mobile phone only. The work presents the development process of an app, on a mobile phone, which can be used to project speed and optionally other informations on the windscreen of a car. The main concern of creating such an app is to figure out the mirroring of the mobile phone's screen, to see an image which is not rotated, on the windscreen of a car.

关键词： advanced driver assistance systems displays mobile applications projection algorithms vehicle driving vehicle safety

来源：评论

学校读者我要写书评

暂无评论

Set-membership affine projection algorithm for echo cancellation

Set-membership affine projection algorithm for echo cancella...

引用

IEEE International Symposium on Circuits and Systems (ISCAS)

作者： P.S.R. Diniz R.P. Braga S. Werner LPS-Signal Processing Laboratory COPPE/Poli-UFRJ Rio de Janeiro Brazil Amazonia Celular Brazil Signal Processing Laboratory Helsinki University of Technology Finland

This paper proposes a new data-selective affine projection algorithm for echo cancellation. The algorithm generalizes the concepts of the conventional set-membership affine-projection by incorporating a lower bound on the output error in order to prevent undesirable attenuation of the far-end signal. It is shown that the echo signal can more reliably be removed from the far-end user signals by employing the new algorithm in double talk situations. In addition, the proposed algorithm retains the fast convergence of the conventional SM-AP algorithm while keeping a reduced number of coefficient updates. Simulation results, using the ITU-T G.168 recommendation setup parameters, are presented in order to confirm the good features of the proposed algorithm

关键词： projection algorithms Echo cancellers Adaptive filters Computational complexity Error correction Estimation error Filtering Least squares approximation Resonance light scattering Adaptive algorithm

来源：评论

学校读者我要写书评

暂无评论

Medical Image Registration Algorithm Based on Compressive Sensing and Scale-invariant Feature Transform 8

Medical Image Registration Algorithm Based on Compressive Se...

引用

8th International Conference on Intelligent Computation Technology and Automation (ICICTA)

作者： Sa, Yang Guangdong Univ Educ Dept Phys Guangzhou 510303 Guangdong Peoples R China

ISBN: (纸本)9781467376440

A registration algorithm based on compressive sensing theory and SIFT(Scale-Invariant Feature Transform) is proposed. By the sparse feature representation methods, the feature vector of SIFT is extracted and the high-dimensional gradient derivative is decreased to low-dimensional sparse feature vector. Accordingly, Euclidean distance is introduced to compute the similarity and dissimilarity between feature vectors used for image registration and BBF(Best-Bin-First) data structure is used to avoid exhaustion. The experimental results show that the proposed algorithm has better performance than the traditional SIFT algorithm. Comparing with the current modified SIFT algorithms, the real-time performance of the proposed algorithm is improved obviously.

关键词： Biomedical image processing feature extraction image registration projection algorithms

来源：评论

学校读者我要写书评

暂无评论

A sparse forcing conformal array synthesis method

A sparse forcing conformal array synthesis method

引用

International Symposium of IEEE-Antennas-and-Propagation-Society / USNC/URSI National Radio Science Meeting

作者： Pinchera, Daniele Migliore, Marco Donald Schettino, Fulvio Lucido, Mario Panariello, Gaetano Univ Cassino & Southern Lazio DIEI Via G Di Biasio 43 I-03043 Cassino FR Italy

We propose a sparse forcing synthesis approach for the design of conformal arrays. The method is based on alternate projections and l(1) minimization to achieve a sparse solution. The preliminary results shown are ver... 详细信息

ISBN: (纸本)9781538632840

关键词： Antenna measurements Sensors Compressed sensing projection algorithms Phased arrays

来源：评论

学校读者我要写书评

暂无评论

Low-Complexity Ramp Metering for Freeway Congestion Control via Network Utility Maximization 56

Low-Complexity Ramp Metering for Freeway Congestion Control ...

引用

56th Annual IEEE Conference on Decision and Control (CDC)

作者： Mehr, Negar Horowitz, Roberto Pedarsani, Ramtin Univ Calif Berkeley Dept Mech Engn 410 McLaughlin Hall Berkeley CA 94720 USA Univ Calif Berkeley Dept Mech Engn 6143 Etcheverry Hall Berkeley CA 94720 USA Univ Calif Santa Barbara Dept Elect & Comp Engn 3159 Harold Frank Hall Santa Barbara CA 93106 USA

ISBN: (纸本)9781509028733

In this paper, we present a novel framework for freeway ramp metering that is based on maximizing the aggregate utility of onramp flows. We show how solving the dual problem of maximizing the network utility via a gradient projection algorithm synthesizes a low-complexity control law that is simple enough to be implemented on real platforms, while being robust to measurement noises. Our control algorithm can be partially distributed at each time step, every onramp selects a traffic flow to maximize its own benefit, and the network adjusts unit traffic flow prices for different onramps. We provide theoretical guarantees on the convergence of our algorithm under mild technical assumptions. We further demonstrate the practicality of our method in an example where the state of the art controls fail (due to infeasibility) and introduce multiple interesting future directions.

关键词： Traffic control Mathematical model Optimization Vehicle dynamics Steady-state projection algorithms Predictive control

来源：评论

学校读者我要写书评

暂无评论

Tailored QP Algorithm for Predictive Control with Dynamics Penalty 22

Tailored QP Algorithm for Predictive Control with Dynamics P...

引用

22nd Mediterranean Conference on Control and Automation (MED)

作者： Otta, Pavel Santin, Ondrej Havlena, Vladimir Czech Tech Univ Fac Elect Engn Dept Control Engn Tech 2 Prague 16627 6 Czech Republic

ISBN: (纸本)9781479959013

In order to reduce the computational complexity of solving Quadratic Programming (QP), related to linear Model Predictive Control (MPC), a new approximated formulation of the QP with simple bounds is introduced in this paper. This formulation is based on the idea not to consider model dynamics as a hard constraint but rather modify the objective function of MPC by penalty to capture the violation of model dynamics. The system dynamics is usually uncertain and then it does not make sense to design the control law based on the exact model. Furthermore, the specific sparse structure of the approximated simple bounded QP formulation of the MPC problem is exploited in the new type of combined gradient/Newton step projection algorithm with linear complexity of each iteration with respect to prediction horizon. It is shown by examples that the proposed method is faster on tested problem than other state-of-the-art solvers while retaining a high performance level.

关键词： computational complexity predictive control quadratic programming MPC problem dynamics penalty gradient-Newton step projection algorithm linear complexity tailored QP algorithm Approximation methods Complexity theory Convergence Linear programming Optimization Prediction algorithms projection algorithms

来源：评论

学校读者我要写书评

暂无评论

AN EFFICIENT METHOD FOR FINDING INTERSECTIONS OF MANY MANIFOLDS WITH APPLICATION TO PATCH BASED IMAGE PROCESSING

AN EFFICIENT METHOD FOR FINDING INTERSECTIONS OF MANY MANIFO...

引用

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Matviychuk, Yevgen Hughes, Shannon M. Univ Colorado Dept Elect Comp & Energy Engn Boulder CO 80309 USA

ISBN: (纸本)9781479903566

Solving inverse problems in signal processing often involves making prior assumptions about the signal being reconstructed. Here the appropriateness of the chosen model greatly determines the quality of the final result. Recently it has been proposed to model images by representing them as sets of smaller patches arising from an underlying manifold. This model has been shown to be surprisingly effective in tasks such as denoising, inpainting, and superresolution. However, such a representation is fraught with the difficulty of finding the intersection of many presumably non-linear manifolds in high-dimensional space, which precludes much of its potential use. This paper proposes an efficient method of solving this problem using a kernel methods variant of the projection onto convex sets algorithm to quickly find the intersection of many manifolds while learning their non-linear structure. Indeed the final solution can even be expressed in closed form. We foresee our method allowing a patch-based regularization to be applied across a wide variety of inverse problems, including compressive sensing, inpainting, deconvolution, etc. Indeed, as a proof of concept of our approach, we show how it can be employed in the regularization of an image denoising problem. Here, it even outperforms a state-of-the-art denoising technique, non-local means.

关键词： Patch-based image processing inverse problems kernel methods manifold models projection algorithms

来源：评论

学校读者我要写书评

暂无评论

t-SNE-CUDA: GPU-Accelerated t-SNE and its Applications to Modern Data 30

t-SNE-CUDA: GPU-Accelerated t-SNE and its Applications to Mo...

引用

30th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

作者： Chan, David M. Rao, Roshan Huang, Forrest Canny, John F. Univ Calif Berkeley EECS Dept Berkeley CA 94720 USA

ISBN: (纸本)9781538677698

Modern datasets and models are notoriously difficult to explore and analyze due to their inherent high dimensionality and massive numbers of samples. Existing visualization methods which employ dimensionality reduction to two or three dimensions are often inefficient and/or ineffective for these datasets. This paper introduces t-SNE-CUDA, a GPU-accelerated implementation of t-distributed Symmetric Neighbour Embedding (t-SNE) for visualizing datasets and models. t-SNE-CUDA significantly outperforms current implementations with 50-700x speedups on the CIFAR-10 and MNIST datasets. These speedups enable, for the first time, visualization of the neural network activations on the entire ImageNet dataset - a feat that was previously computationally intractable. We also demonstrate visualization performance in the NLP domain by visualizing the GloVe embedding vectors. From these visualizations, we can draw interesting conclusions about using the L2 metric in these embedding spaces. t-SNE-CUDA is publicly available at https://***/CannyLab/tsne-cuda.

关键词： Artificial intelligence Machine learning projection algorithms Dimensionality Reduction t-SNE CUDA

来源：评论

学校读者我要写书评

暂无评论

GENERALIZED ARRAY PATTERN SYNTHESIS BY THE METHOD OF ALTERNATING ORTHOGONAL projectionS

引用

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION 1980年第3期28卷 328-332页

作者： PRASAD, S Department of Electrical Engineering Indian Institute of Technology New Delhi India

A geometric formulation of the "null-steering" array antenna pattern synthesis problems for arbitrary array geometries is proposed. The formulation allows the use of some simple intuitive techniques for the construction of the array weights. In particular it is shown that the optimum weight vector can be obtained in an appropriately selected "weight vector subspace" via a recursive projection algorithm starting from its known projection on the vector lying in the constrained "look direction." The algorithm is computationally very simple and can be used for on-line solution of synthesis problems for complex array geometries.

关键词： Vectors Antenna arrays Subspace constraints Network synthesis Phased arrays Space technology projection algorithms Computational geometry Computer 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：