检索结果-内蒙古大学图书馆

Comparative analysis of image projection-based descriptors in Siamese neural networks

ADVANCES IN ENGINEERING SOFTWARE 2021年 154卷 102963-102963页

作者： Kertesz, Gabor Szenasi, Sandor Vamossy, Zoltan Obuda Univ John von Neumann Fac Informat Budapest Hungary J Selye Univ Fac Econ Komarno Slovakia

Low-level object matching can be done using projection signatures. In case of a large number of projections, the matching algorithm has to deal with less significant slices. A trivial approach would be to do statistical analysis or apply machine learning to determine the significant features. To take adjacent values of the projection matrices into account, a convolutional neural network should be used. To compare two matrices, a Siamese structure of convolutional heads can be applied. In this paper, an experiment is designed and implemented to analyze the object matching performance of Siamese Convolutional Neural Networks based on multi-directional image projection data. A backtracking search-based Neural Architecture Generation method is used to create convolutional architectures, and a Master/Worker structured distributed processing with highly efficient scheduling based on the Longest processing Times-heuristics is used for parallel training and evaluation of the models. Results show that the projection-based methods are Pareto optimal in terms of one-shot classification accuracy and memory consumption.

关键词： Siamese neural networks Multi-directional image projections Neural architecture generation one-shot classification accuracy Vehicle re-identification

来源：评论

学校读者我要写书评

暂无评论

Research and analysis of methods for generating and processing new code structures for the problems of detection, synchronization and noise-resistant coding 26

Research and analysis of methods for generating and processi...

引用

conference on image and Signal processing for Remote Sensing XXVI

作者： Grigoriev, Evgeniy K. Nenashev, Vadim A. Sergeev, Alexander M. Nenashev, Sergey A. St Petersburg State Univ Aerosp Instrumentat 67 Bolshaya Morskaia Str St Petersburg 190000 Russia

ISBN: (纸本)9781510638808

The article provides an analysis of new code structures of maximum length obtained from rows of quasi-orthogonal matrices. Codes of maximum length (m-sequences, codes constructed on the basis of Legendre symbols or quadratic residues, codes generated on the basis of Jacobi symbols and others) have the potential to be widely used both in remote sensing radar systems and in noise-immune and noise-proof systems high-speed communications as a replacement for widely used Barker codes. The questions of searching and researching new noise-resistant codes constructed on the basis of persymmetric circulants are considered. Comparison of performance ratings obtained with the new code sequences is given. The advantages of the codes obtained in the work are discussed in the aspects of increasing the correlation characteristics, their detection, and noise immunity in the radio channels of distributed systems. Since these codes have potential applications in radar, not only the correlation properties (autocorrelation and periodic correlation function) are considered, as for communication systems, but also the ambiguity function, which takes into account not only the correlation of signals in time, but also in frequency. The results show the possibility for their use in distributed location systems, not only in the case of remote sensing of the earth, but also in the field of application of the transmission of generated signals to points of joint information processing.

关键词： quasi-orthogonal matrices remote sensing of the earth distributed systems correlation function uncertainty function new code constructions

来源：评论

学校读者我要写书评

暂无评论

An Efficient parallel CNN Inference Framework for Multi-zone Processor

An Efficient Parallel CNN Inference Framework for Multi-zone...

引用

IEEE International conference on High Performance Computing and Communications (HPCC)

作者： Jie Chen Zhong Liu College of Computer National University of Defense Technology Changsha China

Deep convolutional neural networks have been widely used in various fields, and how to accelerate and optimize the performance of neural network models according to the architectural characteristics of accelerator hardware has become the research hot spot. We propose and design an efficient parallel convolutional neural network inference framework named PCIF for multi-zone processor. Its core is a multi-core parallel and vectorized C++ operator library. PCIF has designed optimized data layout and transmission strategies, layer fusion and multi-image fusion methods, which can automatically parse and efficiently map existing neural network models to multi-core vector accelerators to accelerate inference computation. Users can also build new neural network models through the C++ and Python APIs provided by PCIF to achieve high performance inference computing. The results of the experiment show that the AlexNet, VGG-16, GoogleNet, and ResNet-50 network models achieve efficient computing performance using PCIF.

关键词： Quantization (signal) Multicore processing Computational modeling Neural networks C++ languages parallel processing Libraries

来源：评论

学校读者我要写书评

暂无评论

Weighted Node Mapping and Localisation on a Pixel Processor Array

Weighted Node Mapping and Localisation on a Pixel Processor ...

引用

IEEE International conference on Robotics and Automation (ICRA)

作者： Hector Castillo-Elizalde Yanan Liu Laurie Bose Walterio Mayol-Cuevas Bristol Robotics Laboratory University of Bristol UK Visual Information Laboratory University of Bristol UK

This paper implements and demonstrates visual route mapping and localisation upon a Pixel Processor Array (PPA). The PPA sensor comprises of an array of processing Elements (PEs), each of which can capture and process visual information directly. This provides significant parallel processing power allowing novel ways in which information can be processed on-sensor. Our method predicts the correct node within a topological map generated from an image sequence by measuring image similarities, spatial coherence, and exploiting the parallel nature of the PPA. Our implementation runs at +300Hz on large public datasets with +2K locations requiring 2.5W at 500 GOPS/W. We compare vs traditionally implemented methods demonstrating better F-1 performance even on simulation. As far as we are aware, we present the first on-sensor mapping and localisation system running entirely on-sensor.

关键词： Visualization Spatial coherence Prototypes parallel processing Robot sensing systems Hardware Computational efficiency

来源：评论

学校读者我要写书评

暂无评论

Synthesis and Segmentation Method of Cross-Staining Style Nuclei Pathology image Based on Adversarial Learning

Synthesis and Segmentation Method of Cross-Staining Style Nu...

引用

IEEE International conference on Big Data and Cloud Computing (BdCloud)

作者： Ziheng Zhou Hong Guo Yuwen Guo Huanhuan Sheng College of Computer Science and Technology Wuhan University of Science and Technology Wuhan China

In modern clinical diagnosis, nuclei is an important basis for obtaining disease information, and rapid and accurate nuclei segmentation technology can effectively help doctors diagnose diseases. However, the difference in staining equipment or organs will cause staining differences between histopathological images, which will affect the effect of nuclei segmentation. Aiming at the problem of staining differences, we propose a cross-staining style nuclei pathology image synthesis and segmentation method based on adversarial learning. The algorithm is composed of a synthetic network and a segmentation network. The synthetic network adds a content discriminator to the cycle generative adversarial networks. Used to generate an image with the same staining style as the target domain, but the same content. The segmentation network segmented the generated image to verify the necessity of staining normalization and the performance of synthesis. In addition, the synthetic network also uses content loss function and style loss function to optimize network training. Experiments show that we propose a cross-staining style nuclei pathology image synthesis and segmentation method based on adversarial learning can effectively solve the problem of staining differences and optimize the effect of segmentation network.

关键词： Training image segmentation image synthesis Histopathology Generative adversarial networks Adversarial machine learning Clinical diagnosis

来源：评论

学校读者我要写书评

暂无评论

Cooperative Co-Evolutionary Genetic Programming for High Dimensional Problems 16th

Cooperative Co-Evolutionary Genetic Programming for High Dim...

引用

16th International conference on parallel Problem Solving from Nature (PPSN)

作者： Rodriguez-Coayahuitl, Lino Morales-Reyes, Alicia Escalante, Hugo Jair Coello, Carlos A. Coello Inst Nacl Astrofis Opt & Electr Tonantzintla 72840 Mexico CINVESTAV IPN Dept Comp Mexico City DF Mexico

ISBN: (纸本)9783030581145;9783030581152

We propose a framework for Cooperative Co-Evolutionary Genetic Programming (CCGP) that considers co-evolution at three different abstraction levels: genotype, feature and output level. A thorough empirical evaluation is carried out on a real-world high dimensional ML problem (image denoising). Results indicate that GP's performance is enhanced only when cooperation happens at an output level (ensemblea-like). The proposed co-evolutionary ensemble approach is compared against a canonical GP implementation and a GP customized for image processing tasks. Preliminary results show that the proposed framework obtains superior average performance in comparison to the other GP models. Our most relevant finding is the empirical evidence showing that the proposed CCGP model is a promising alternative to specialized GP implementations that require knowledge of the problem's domain.

关键词： Genetic Programming Evolutionary machine learning Co-evolutionary algorithms Ensemble methods image processing.

来源：评论

学校读者我要写书评

暂无评论

Design and analysis of A 2-DOF parallel mechanism for space large deployable antenna 4

Design and analysis of A 2-DOF parallel mechanism for space ...

引用

4th International conference on Vision, image and Signal processing, ICVISP 2020

作者： Huajun, Hu Bo, Pan Haowei, Wang Haiyuan, Wang Fan, Zhao Jing, Sun Beijing Spacecrafts Co. Ltd. Beijing100094 China Beijing Institute of Spacecraft System Engineering Beijing100094 China China Satellite Communications Co. Ltd. Beijing100048 China

ISBN: (纸本)9781450389532

The space thermal environment and sophisticated deployable structure lead to the error of the satellite antenna. In this paper, a parallel-kinematic symmetrical stage to eliminate the error of antenna is proposed to realize the 2-DOF rotation with compact physical size. The flexure Hooke hinge is utilized to replace traditional U-joint with its characteristic of distributed compliance to eliminate the rotation clearance and to improve precision. This parallel mechanism is constituted of two symmetrical chains which are UPS kinematic chains with one spherical joint(S), one ball screw prismatic(P) and one flexure Hooke hinge(U) and one fix chain. Geometric analytical method is implemented to establish the kinematic model and the resolution of the parallel mechanism. A prototype stage is fabricated for experiment study. The open-loop testing is conducted to validate the stage's performance. Result show that the 2-DOF parallel mechanism possesses a compact dimension, which offers a workspace of ±2°×±2° with the repeatability precision of 0.01°. © 2020 ACM.

关键词： Kinematics

来源：评论

学校读者我要写书评

暂无评论

Multiprecision Block-Jacobi for Iterative Triangular Solves 26th

Multiprecision Block-Jacobi for Iterative Triangular Solves

引用

26th International conference on parallel and distributed Computing (Euro-Par)

作者： Goebel, Fritz Anzt, Hartwig Cojean, Terry Flegar, Goran Quintana-Orti, Enrique S. Karlsruhe Inst Technol Steinbuch Ctr Comp Karlsruhe Germany Univ Tennessee Innovat Comp Lab ICL Knoxville TN USA Univ Jaume 1 Dept Ingn & Ciencia Comp Castellon de La Plana Spain Univ Politecn Valencia Dept Informat Sistemas & Comp Valencia Spain

ISBN: (纸本)9783030576752;9783030576745

Recent research efforts have shown that Jacobi and blockJacobi relaxation methods can be used as an effective and highly parallel approach for the solution of sparse triangular linear systems arising in the application of ILU-type preconditioners. Simultaneously, a few independent works have focused on designing efficient high performance adaptive-precision block-Jacobi preconditioning (block-diagonal scaling), in the context of the iterative solution of sparse linear systems, on manycore architectures. In this paper, we bridge the gap between relaxation methods based on regular splittings and preconditioners by demonstrating that iterative refinement can be leveraged to construct a relaxation method from the preconditioner. In addition, we exploit this insight to construct a highly-efficient sparse triangular system solver for graphics processors that combines iterative refinement with the block-Jacobi preconditioner available in the Ginkgo library.

关键词： Sparse linear algebra Incomplete factorization preconditioning Graphics processing units (GPUs) Multiprecision Block-Jacobi

来源：评论

学校读者我要写书评

暂无评论

A Study of Classifying Advanced Persistent Threats With Multi-Layered Deep Learning Approaches

A Study of Classifying Advanced Persistent Threats With Mult...

引用

IEEE International conference on Big Data and Cloud Computing (BdCloud)

作者： Yen-Hung Frank Hu Chung-Chu George Hsieh Norfolk State University Norfolk Virginia USA

Advanced persistent threats have caused serious negative impacts on today’s cyber defense systems because of their stealthy characteristics and sophisticated attacking strategies. Many approaches have been proposed and applied to identify and classify advanced persistent threats. However, there is still a lack of promising solutions for defending against such threats. Research has shown that deep learning can enhance image categorization and classification and provide a potential solution for detecting anomalies. To extend study and investigation of advanced persistent threats, we have proposed multiple-layer deep learning approaches adopting image generation and convolution neural network technology. In this research, more than 3000 portable executable samples belonging to 12 advanced persistent threats were converted into images, then categorized into a training and validation dataset for developing and utilizing the proposed deep learning models. Performance studies and contributions of such models were studied and measured as well.

关键词： Deep learning Training Analytical models Adaptation models Data analysis image synthesis Convolution

来源：评论

学校读者我要写书评

暂无评论

Accelerating Geo-distributed Transaction processing with Fast Logging

Accelerating Geo-Distributed Transaction Processing with Fas...

引用

IEEE International conference on Big Data

作者： Takuto Ogura Yoshiki Akita Yuki Miyazawa Hideyuki Kawashima Mitsubishi UFJ Information Technology Tokyo Faculty of Environment and Information Studies Keio University Kanagawa

ISBN: (纸本)9781665445993

We herein propose three novel optimization methods to accelerate distributed transaction processing into a geographically distributed database. The first optimization involves the parallelization of the write-ahead logging protocol. It allows multiple worker threads to synchronize log entries to the storage device simultaneously without any dependencies. The second optimization involves the grouped transfer of log entries from the leader to followers. This reduces the number of transmissions and effectively uses the network bandwidth. The third optimization involves the separation of the worker thread logic. By breaking the logic into the prepare phase and the commit phase, the worker threads at the leader node run asynchronously in parallel without waiting for responses from the follower nodes. The experimental results demonstrated that the proposed method achieved more than 10 million tps and less than 100 ms with client interactions through the network. The CPU utilization was almost 100%, which implied a dramatic reduction in synchronization in worker threads.

关键词： Protocols Codes Instruction sets conferences Optimization methods distributed databases Bandwidth

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：