深度学习目标检测方法综述

随着深度学习在目标检测领域的大规模应用,目标检测技术的精度和速度得到迅速提高,已被广泛应用于行人检测、人脸检测、文字检测、交通标志及信号灯检测和遥感图像检测等领域.本文在基于调研国内外相关文献的基础上对目标检测方法进行了综述.首先介绍了目标检测领域的研究现状以及对目标检测算法进行检验的数据集和性能指标.对两类不同架构的...     

Dynamic reconfiguration architectures for multi-context FPGAs

Field-programmable gate arrays (FPGAs) are being integrated with processors on the same motherboard or even chip in order to achieve flexible high-performance computing, and this may become main stream in chip multi-core architectures. However, the expensive FPGA area is often used inefficiently, with much of the logic idle at any given time. This work, motivated by the Dynamic-Link Library (DLL) concept in software, explores the possibility of “hardware DLLs” by finding ways for fast dynamic incremental reconfiguration of FPGAs. So doing would, among other things, enable same-function replication at any given time, with functions changing quickly over time, thereby enabling efficient exploitation of data parallelism at no additional hardware cost.We present two new multi-context FPGA architectures based on two different configuration storage architectures: local and centralized. Problems such as configuration storage and reconfiguration (time, power and space) overhead are considered. Well known area and power models are used in evaluating various approaches and in order to provide guidelines for matching architectures to target applications. Lastly, we provide insights into resulting scheduling issues. Our findings provide the foundation and “rules of the game” for subsequent development of reconfiguration schedulers and execution environments.     

Embedded OSs gain the inside track

Almost every automated system has an embedded OS, and so there is an increasing demand for embedded OS. No single vendor dominates the marketplace. A growing number of companies have released open-source embedded OS, primarily those based on BSD Unix and Linux. These OS have the same appeal that open-source software has in other markets: they are relatively inexpensive and users have access to and can customize the source code. It is thought that they will become the accepted standards. Developments from Windows are also outlined. Developments of new architectures are also considered     

基于WebService的WebGIS的设计与应用

基于WebGIS目前存在的一些不足,提出构建基于WebService的WebGIS系统,并说明这种系统的优势和特点。通过对正在开发的基于WebService的WebGIS系统的设计和一些已经完成的工作的介绍,来说明系统的可行性和先进性。     

RAID-RMS: A fault tolerant stripped mirroring RAID architecture for distributed systems

Disk arrays, or RAIDs, have become the solution to increase the capacity, bandwidth and reliability of most storage systems. In spite of its high redundancy level, disk mirroring is a popular RAID paradigm, because replicating data also doubles the bandwidth available for processing read requests, improves the reliability and achieves fault tolerance. In this paper, we present a new RAID architecture called RAID-RMS in which a special hybrid mechanism is used to map the data blocks to the cluster. The main idea behind the proposed algorithm is to combine the data block striping and disk mirroring technique with a data block rotation. The resulting architecture improves the parallelism reliability and efficiency of the RAID array. We show that the proposed architecture is able to serve many more disk requests compared to the other mirroring-based architectures. We also argue that a more balanced disk load is attained by the given architecture, especially when there are some disk failures.     

Second thoughts on parallel processing

A variety of advanced, high-throughput “parallel” and “associative” computer architectures are discussed. In comparing the architectures we argue that it is useful to view them simply as alternative ways of increasing the amount of processing hardware relative to the amount of memory hardware. Another useful criterion is the duty cycle of the processing hardware, i.e. the proportion of the total amount of arithmetic and logical activity that is meaningful.Digital signal processing is considered as an example application. We conclude that cellular-logic-in-memory arrays are inappropriate for signal processing.     

Getting on Board the Enterprise Service Bus

One of the biggest challenges facing businesses today is getting their diverse applications, often built on different platforms, to work together when necessary. To accomplish this, companies must implement effective application integration flexibly, quickly, and easily. Businesses initially turned to manual integration and then enterprise application integration (EAI) but subsequently have focused on service-oriented architectures. To leverage SOAs' benefits effectively, companies are beginning to use an approach known as the enterprise service bus, offered by a growing number of large and small vendors. ESB is the middleware glue that holds an SOA together and enables communication between Web-based enterprise applications. However, ESB also faces challenges, such as implementation costs and complex migration and management     

Service-Oriented Computing: Bringing Business Systems to the Web

The notion of "service" has spurred major evolutions for both information systems and the Web. A software application is no longer considered a monolithic component; it can be divided into services that are smaller components defined by their function and accessible through well-defined interfaces and protocols. As a result, IT actors are using service-oriented architectures (SOAs) to remodel the information systems of many companies while the Web is increasingly becoming a programmable place. In both domains, developers build composite client applications to consume these services. Even boundaries between enterprise services and Internet services are vanishing. Some companies, such as Strikelron provide enterprise services that were previously always hosted internally - like customer relationship management solutions. As a consequence, companies now have the technologies required to bring their business online. With Web services, private business processes can be exposed to partners through public composite Web applications. When new projects emerge, companies need guidance to properly handle such work. In this context, we aim to provide companies solutions - through a methodology, an architecture, and technical choices - that will help them solve generic problems such as security and application conception     

Parallel computer architectures for image processing

Image processing problems frequently involve large structured arrays of data and a need for very rapid computation. Special parallel processing schemes have evolved over the last 20 years to deal with these problems. In this paper many parallel systems which have been developed for image processing are outlined and the features of their underlying architectures are discussed. Most of these special architectures may be loosely classified as either SIMD or pipeline structures although some MIMD structures have been designed for high level image analysis. In recent years several multiple SIMD (MSIMD) schemes have been proposed as suitable architectures for image processing. The fundamental problems of developing an effective MSIMD system are discussed and a simple SIMD/MIMD computational model for comparison with such systems is proposed.     

Synthesizing MPI Implementations from Functional Data-Parallel Programs

Distributed memory architectures such as Linux clusters have become increasingly common but remain difficult to program. We target this problem and present a novel technique to automatically generate data distribution plans, and subsequently MPI implementations in C++, from programs written in a functional core language. The main novelty of our approach is that we support distributed arrays, maps, and lists in the same framework, rather than just arrays. We formalize distributed data layouts as types, which are then used both to search (via type inference) for optimal data distribution plans and to generate the MPI implementations. We introduce the core language and explain our formalization of distributed data layouts. We describe how we search for data distribution plans using an adaptation of the Damas–Milner type inference algorithm, and how we generate MPI implementations in C++ from such plans.     

Efficient control generation for mapping nested loop programs onto processor arrays

Processor array architectures are optimal platforms for computationally intensive applications. Such architectures are characterized by hierarchies of parallelism and memory structures, i.e. processor arrays apart from different levels of cache have a large number of processing elements (PE) where each PE can further contain sub-word parallelism. In order to handle large scale problems, balance local memory requirements with I/O-bandwidth, and use different hierarchies of parallelism and memory, one needs a sophisticated transformation called hierarchical partitioning. Innately the applications are data flow dominant and have almost no control flow, but the application of hierarchical partitioning techniques has the disadvantage of a more complex control flow. In a previous paper, the authors presented first time a methodology for the automated control path synthesis for the mapping of partitioned algorithms onto processor arrays. However, the control path contained complex multiplication and division operators. In this paper, we propose a significant extension to the methodology which reduces the hardware cost of the global controller and memory address generators by avoiding these costly operations.     

Solving the Java object storage problem

Today's Internet driven view of information systems is helping to popularize Java as an application development language. Developers are beginning to use Java to create multi tier application architectures that often integrate relational data stores with new data types, in order to package information in easier to use, dynamic ways. Java's object oriented nature is ideally suited to this new world. Using objects, Java developers can encapsulate both data and data manipulation methods to give applications a runtime dynamism and self-contained intelligence that is difficult to achieve using other methods. Java application developers need to be able to store these Java objects-technically, to give them persistence-in order to take advantage of these capabilities. We examine the development issues surrounding Java object storage, including a brief overview of the ODMG Java binding, a standard that adds object persistence to Java. We compare this with the much greater level of effort required to implement the same application using the lower level JDBC interface, which supports Java object storage in relational databases. The ODMG binding for Java and JDBC are not competitive specifications: ODMG interfaces can be built on top of JDBC     

云环境中Web应用的微服务架构评估

云计算为我们提供了一种全新、高效的方式来部署可扩展的Web应用,这种方式使企业的应用可以按需对计算资源进行分配.微服务架构用于将庞大复杂的应用系统拆分为一系列可独立开发、测试、部署、运行、升级的服务模块.微服务架构为大批互联网企业实现云环境中的应用扩展、降低应用开发复杂度、实现敏捷开发提供了更加有效地方法.本文分析并测试了微服务架构模式,通过一个具体案例——在云环境中开发和部署的企业级应用系统,对两种架构模式实现（单体架构模式和微服务架构模式）进行性能测试,得出评估结果,这些结果对解决企业级应用微服务化中可能遇到的问题具有一定指导意义.     

Load balancing and Poisson equation in a graph

We present a fully distributed dynamic load balancing algorithm for parallel MIMD architectures. The algorithm can be described as a system of identical parallel processes, each running on a processor of an arbitrary interconnected network of processors. We show that the algorithm can be interpreted as a Poisson (heath) equation in a graph. This equation is analysed using Markov chain techniques and is proved to converge in polynomial time resulting in a global load balance. We also discuss some important parallel architectures and interconnection schemes such as linear processor arrays, tori, hypercubes, etc. Finally we present two applications where the algorithm has been successfully embedded (process mapping and molecular dynamic simulation).     

Enabling technologies for e-business

While distributed object concepts can be difficult to grasp at first, applying them to build reusable components eventually leads to faster time to market. And now a host of enabling technologies is available, making it easier to use distributed objects for Web based e-business projects. These enabling technologies provide the technical underpinnings for Web-enabled object architectures. Web-enabled object architectures integrate the power of objects with the Web's distribution capabilities. But we need special tools to help minimize the complexities of using distributed object technologies. The advent of application servers and the benefits of distributed object technologies remove the bounds that previously restricted the Web. Companies can use these new technologies to embrace an expanded scope of business that will in turn continue to generate further requirements for new technology     

RISC-V架构硬件辅助用户态内存安全防御方案概览

传统的用户态内存安全防御机制基于x86架构和纯软件方式实现,实现内存安全保护的运行时开销很高,难以部署在生产环境中.近年来,随着主流商业处理器开始提供硬件安全扩展,以及RISC-V等开源处理器架构的兴起,内存安全保护方案开始面向x86-64、ARM、RISC-V等多种体系架构和硬件辅助实现方式.我们对RISC-V架构上实现的内存安全防御方案进行了讨论,并对x86-64、ARM、RISC-V等处理器架构在安全方案设计上的特点进行了比较.得益于开放的指令集架构生态, RISC-V架构的内存安全防御方案相较于其他架构有一些优势.一些低成本的安全防御技术有望在RISC-V架构上实现.     

An Integrated Approach to the Design of Processor Arrays with Systolic Organization of Computations

An integrated approach is proposed to the design of economically efficient and high-performance processor arrays with systolic organization of computations. The approach includes the construction of VLSI-oriented versions of locally recursive algorithms and synthesis of new architectures of processor arrays for transforming algorithms that maximally take into account fundamental restrictions of VLSI technology. Within the framework of this approach, strategies are developed for obtaining the above-mentioned algorithms and architectures.     

Analysis on the evolution of shipping logistics service supply chain market structure under the application of blockchain technology

With the digital transformation of the shipping industry, the business potential of shipping companies is gradually expanding. In recent years, shipping companies have begun operating freight forwarding businesses powered by the advantages of cost and network. As a result, their business expansion and gradually occupying the market share of agency services. At the same time, the development requirements of digitization are further highlighted, while the problems of opaque information and inaccurate data in the operation of the shipping logistics service supply chain have become more prominent. Thus, the shipping industry has begun to apply Blockchain Technology (BCT) to the shipping logistics. However, the promotion of BCT, an emerging technology in the shipping logistics service market, is not systematic. The members of the supply chain do not have a deep understanding of it, which have led to changes in market demand. We take the shipping logistics service supply chain composed of shipping companies and freight forwarders as the research object in order to investigate the vertical competition between shipping companies and freight forwarders as well as the impact of BCT application on market structure changes. From the perspectives of freight rate competition and the promotion of BCT, the Stackelberg game theory is applied to create a mathematical model that can be used to discuss the evolution of the shipping logistics service market. The results indicate that, the impact of freight rate competition on market evolution is reduced after the application of BCT. Shipping companies should actively promote BCT to reduce the impact of freight competition on the market. Furthermore, freight forwarders can actively participate in the initial stage of application to increase market demand, and the cost of promoting BCT should be reduced in the later stage.     

CMOS图像传感技术及其在影像产品中的应用

随着超大规模集成技术的发展，CMOS图像传感器显示出强劲的发展趋势。CMOS图像传感器具有在单芯片内集成时序和控制电路、A／D转换、信号处理等功能。本文介绍了CMOS图像传感器的结构和工作原理，并与CCD图像传感器进行比较，综述了CMOS图像传感器的最新进展及其在影像产品中的应用。