摘要：
Peer-to-peer (P2P) network provides an efficient way for resource sharing. However, due to lacking centralized control and the anonymity of peers, security becomes a serious problem in P2P network, especially in P2P file-sharing network, which heavily restricts the development of P2P file-sharing systems. Trust models have been widely used in estimating the trustworthiness of peers in P2P file-sharing systems in order to minimize threats of such open communities. However, current P2P trust models almost measure whether a peer is good or bad only by a trust value, which results in low efficiency in resisting the attacks of malicious peers. Additionally, there are more and more free-riders in P2P systems. This paper proposes a new trust model based on a 5-tuple to deal with these two problems, which integrate the set of uses, trust, experience, sharing files and heterogeneity levels. The simulation experiments are carried out to evaluate the performance and validity of this trust model in a P2P file-sharing system. The result proves that it is effective to prevent malicious act and encourage peers to share there resources in P2P network.
摘要：
Online learning is an effective incremental learning method. Compared with the conventional off-line learning method, online learning updates the original classifier continuously with new samples and improves its performance. In this paper, we propose a novel online learning framework for head detection in video sequences. At first, an off-line classifier is trained with a few labeled samples. And it was used to object detection in video sequences. Based on online boosting algorithm, the detected objects will be used to train the classifier as new samples. Instead of using another detection algorithm to label the new sample automatically like other online learning framework, we ensure the correct label from tracking. Furthermore, the weights of new samples can be obtained from tracking directly. Thus the training speed of the classifier can be improved. Experimental results on two video datasets are provided to show the efficient and high detection rate of the framework.
摘要：
The most existing systems of CBIR fulfill the tasks of retrieving the similar images through computing the degree of similarity of different images. Furthermore, the quality of the outcomes provided by color histogram-based image retrieval is usually rather limited. In this paper, a new image retrieval method is presented which is integrated the blocking wavelet-histogram with particle swam optimization (PSO). The innovative approach is used as solution to the problem of intelligent retrieval of images in large image databases. The problem is recast to a discrete optimization one, where a suitable speed and position of particle is defined through a customized PSO. Farther on, in virtue of the new computation model, a fitness function which combines blocking wavelet transformation information and the Euclidean distance of color histogram is constructed. The innovative approach based on blocking wavelet-histogram image similarity retrieval and particle swam optimization (PSO) is used for testing, and the experimental results show that our method is feasible and effective to image retrieval.
关键词：
Low-contrast enhancement;illuminace;reflectance;Retinex Theory
摘要：
A simple method for enhancing low-contrast gray image is proposed. The method decomposes an image into the layers of illuminance and reflectance, by using an edge-preserving low-pass filter, and separately enhances each layer. Finally, the enhanced results of these two layers are recomposed as the output. Experiment shows the good results on real images.
摘要：
The proposed vehicle information management system based on Struts architecture is designed for the management of stock, customer, and financial information for used vehicle dealerships. Struts architecture and UML extension mechanism are applied to provide a safe and efficient MVC framework, to reduce the complexity of authorization management, to enhance the system stability, to realize the entire sales data sharing, and to support the security tactics of port in more flexible way. The product not only tracks and reports all information relevant to vehicle acquisition, preparation of vehicle for sale and disposal of vehicle but also provides management information regarding customers, suppliers, salesman, and vehicle financing arrangements.
摘要：
Online boosting algorithm has been used in many vision-related applications, such as object detection. However, in order to obtain good detection result, combining a large number of weak classifiers into a strong classifier is required. And those weak classifiers must be updated and improved online. So the training and detection speed will be reduced inevitably. This paper proposes a novel online boosting based learning method, called self-learning cascade classifier. Cascade decision strategy is integrated with the online boosting procedure. The resulting system contains enough number of weak classifiers while keeping computation cost low. The cascade structure is learned and updated online. And the structure complexity can be increased adaptively when detection task is more difficult. Moreover, most of new samples are labeled by tracking automatically. This can greatly reduce the effort by labeler. We present experimental results that demonstrate the efficient and high detection rate of the method.
期刊：
DCABES 2009: THE 8TH INTERNATIONAL SYMPOSIUM ON DISTRIBUTED COMPUTING AND APPLICATIONS TO BUSINESS, ENGINEERING AND SCIENCE, PROCEEDINGS,2009年:242-244
会议名称：
第八届分布式计算及其应用国际学术研讨会(The 8th International Symposium on Distributed Computing and Applications to Business,Engineering and Science)
会议时间：
2009-10-16
会议地点：
武汉
会议论文集名称：
第八届分布式计算及其应用国际学术研讨会(The 8th International Symposium on Distributed Computing and Applications to Business,Engineering and Science)论文集
关键词：
Intrusion Detection;Data Mining;Apriori
摘要：
This paper presents a method based on data mining technology to set up an intrusion detection system, which has characteristics of scalability, adaptability and accuracy. On the basis of analyzing the association mining algorithm combined with a number of attacks on networks, the eligible item set is identified and rule-base is set up, which has improved the ability of intrusion detection.
作者机构：
[Kang, Zhou] Wuhan Polytech Univ, Dept Math & Phys, Wuhan 430023, Peoples R China.;[Yufang, Huang; Zhen, Cheng; Jin, Xu] Huazhong Univ Sci & Technol, Dept Control Sci & Engn, Wuhan 430074, Peoples R China.
会议名称：
4th International Conference on Bio-Inspired Computing
会议时间：
OCT 16-19, 2009
会议地点：
Peking Univ, Beijing, PEOPLES R CHINA
摘要：
Composition and biochemistry experiments of sticker model and the realization process of fundamental biochemical experiments are discussed. Based on separation technology and electrophoresis experiment, a new detection experiment is put forward, which can be used to detect experimental results containing various kinds of memory complex. DNA algorithm of making all spanning tree problem and its biochemical realization process are put forward, first the initial solution space of spanning subgraph is created, then all spanning trees are selected from the initial solution space. During the creating solution space of spanning subgraph, complete hybridization based on incomplete separation is First put forward, which can separate equably one tube into several tubes according to the requirement of algorithm design. The correctness and complexity of the DNA algorithm are discussed and proved. Finally, the validity and feasibility of the DNA algorithm are explained by a simulate experiment. Two kinds of biotechnology are first put forward, which are detection experiment based on separation technology and complete hybridization based on incomplete separation, and DNA algorithm of making all spanning tree problem is also first put forward.
关键词：
high performace cloud;storage cloud;compute cloud;distributed data application
摘要：
This paper presents a new high performance cloud framework that support simplified data parallel applications. and descripes the design and implementation of it. this cloud implements stream processing paradigm over clusters of computers, either within single data center or across wide arer. Unlike MapReduce, this Cloud's stream processing paradigm is simpler, more straightforward, and reaches better performance in certain cases. The experimental results show this cloud is more straighforword, simpler to use, and it is about twice as fast as Hadoop.
摘要：
In order to maximize the network lifetime for target coverage, we propose a distributed optimum coverage algorithm for point target in wireless sensor networks. In this paper, we first present a 1-hop local target coverage problem, next analyze the critical restraint of this problem and introduced the definition of key target, then designed an energy utility function, last established an adaptive adjustment mechanism of the waiting time. Measurement results show that the new algorithm extends 20% longer network lifetime, has good scalability and stability, and a lower computational and communication complexity.
摘要：
This paper studies the utility-maximization resource allocation problem for soft QoS traffic in wireless networks, which demands certain amount of bandwidth for normal operation but allows some flexibility. With deviations to the known approaches and solutions, we propose a novel utility optimization model, which bounds the bandwidth allocation of each flow around its desired value and allows certain flexibility. We theoretically analyse this model and give a complete procedure to obtain the solution.
摘要：
This paper analyses structure and algorithm of Floating-Point ALU, and implements multiplication and division operation in the homo-hardware circuit. The Floating-Point multiplication and division ALU supports Floating-Point number according with IEEE-754 standard. This ALU adopts 4-Level pipelining structure: '0' operation number check, exponent addition and subtraction operation, fraction multiplication and division operation, result normalization and round. Each step can act as a single module. Among these modules, there are some registers which can prepare necessary data for next operation.
会议名称：
First IITA International Joint Conference on Artificial Intelligence, Hainan Island, China, 25-26 April 2009
关键词：
MIPS;Data Flow;Data Path;Pipeline
摘要：
In this paper, we analyze MIPS instruction format-, instruction data path-, decoder module function and design theory basend on RISC CPUT instruction set. Furthermore, we design instruction fetch (ZF) module of 32-bit CPU based on RISC CPU instruction set. Function of IF module mainly includes fetch instruction and latch module -, address arithmetic module-, check validity of instruction module-, synchronous control module. Function of IF modules are implemented by pipeline and simulated successfully on Quartus ll.
摘要：
The DoS attack is the most popular attack in the network security with the development of network and internet. In this paper, the DoS attack principle is discussed and some DoS attack methods are deeply analyzed. The DoS attack detection technologies which include network traffic detection and packet content detection are presented. The DDoS based on DoS is introduced and some DDoS tools are described and the important TCP flood DoS attack theory is discussed. The DoS attack program and a DoS attack detection program based on Winpcap for experiment are designed and the network packet generation and capture are implemented. The experiment expressed the key progress of DoS attack and detection in detail.
摘要：
Currently existing data access object (DAO) patterns have several limitations. First, the interface of the patterns and business objects is tightly-coupled, which affects seriously the dynamic extensibility of software systems. Second, the patterns have duplicated implementation codes, which add to difficulties of system maintenance. To solve these problems, a new DAO pattern with stronger independency and dynamic extensibility is proposed in this paper. Finally, an example is given to illustrate the using process of the new DAO pattern. The greatest advantages of the new DAO pattern are as follows. If any business object is needed to add to the system, we don't have to modify any codes of the class DAO Factory. All we need to do is to modify the mapping file. Furthermore, because we have only one DAO implementation class to accomplish all the data access to business objects, if some SQL statements are needed to be modified, all we need to do is to modify the DAO implementation class but not need to modify any business objects.
关键词：
access control model;attribute;policy;decision;multi-policy supporting
摘要：
Because of the limitations of condition configuration and execution mechanism, the exiting access control models are not flexible and can not support multi-policy. Focusing on this, the paper presents a policy-based access control model (PBAC). Different from the traditional models which restrict session only with subject authorization, PBAC realizes policy-based access control by defining attribute to describe session property, performing a new policy management method that is free from application logic, and advocating an independent access control decision mechanism. As a consequence, PBAC is more flexible on restricting session, and makes great progress on multi-policy supporting. The comparison analysis indicates that PBAC is superior to the current access control models.
关键词：
network threat;risk assessment;Dempster-Shafer theory
摘要：
Current practice for real-time security risk assessment typically takes Intrusion Detection Systems alerts as the only source of risk factor. Their assessment results are more likely to suffer from the impact of false positive alerts in the increasingly complex and severe network security environment. This paper proposes a novel online fusion model for dynamical network risk assessment by using multiple risk factors. The model is composed by three fusion levels. First, an online alert fusion algorithm is proposed and the redundancy of the raw alerts is dramatically reduced. Then, the model employs Dempster-Shafer theory to handle uncertainties and ignorance existed in the multiple risk factors. Threats in different kinds of severity levels are identified. Finally, the whole network risk distribution is dynamically calculated and reported by using HMM approach. Experiments show the effectiveness and validity of our method.
摘要：
In recent years, with the conversion of the dominant teaching theory from behaviorism to constructivism, the body part of the learning process, the learner, is widely concerned. Many scholars at home and abroad and people in the educational circles treat the ability of the self-study and personalized learning as the ultimate goal of education. The rapid development of computer network technology provides the reliable material condition for the implementation of constructivist theory in teaching. The "online learning" technology established on the basis of computer networks is increasingly becoming one of the development directions of education which makes the network teaching, distance teaching possible. In this paper, the author explores the characteristics, the purpose and significance of the research, the theoretical basis and the research design and procedures of the self-study and personalized learning with the knowledge and experience accumulated in the process of research and development of Academic Support System.
摘要：
This paper presents a method of group decision-making based on CBR. This method involves two steps. Firstly, decision-making problems are expressed by cases and the decision tree is created via ID3 algorithm. Then each decision-maker's preference is incorporated via utility theory to attain group consistent preference; and on the basis of the group preference, the similar case is retrieved in the decision tree to get the solution. At the end of the paper, an example is given, which shows that the method is feasible.
关键词：
three dimensional map;interactive;spatial environment;shortest path;animation
摘要：
This paper is to practice the art of three-dimensional map based on a university campus which can be used for tourism. The spatial elements and animated character are initially built in Maya, and exported into Virtools. This paper explores ways to make the system interactive and view in the web using Virtools. This paper also presents a general system architecture which is able to be reused in other three-dimensional maps with a little modification.
