摘要：
This paper presents a new approach based on maximum entropy and maximum entropy Markov model for web information extraction. This approach is not only able to overcome the shortcoming of the less precision and recall of the hidden Markov model. In addition, this approach can make the most of various kinds of contextual information from web. The experiments are found that the hybrid approach has an average precision rate of 87.516% while the hidden Markov model trained by the Baum-Welch algorithm has an average precision rate of 68.630%. This implies that the hybrid approach is more optimized than the hidden Markov model trained by the Baum-Welch algorithm.
会议名称：
Multimedia and Ubiquitous Engineering, 2008. MUE 2008. International Conference on
摘要：
Designing storage systems to provide high availability in the face of failures needs the use of various data protection techniques, such as dual-controller RAID. The failure of RAID controller may cause RAID storage system to fail to respond to ongoing requests and to no longer be available to new requests. Heartbeat is used to detect controllers whether survival. So, the heartbeat cycle's impact on the high availability of a dual-controller hot-standby system has become the key of current research. To address the problem of fixed setting heartbeat in building high availability system currently, a self-adaptive heartbeat model of dual-controller, which can adjust heartbeat cycle based on the frequency of data read-write request, is designed to improve the high availability of dual-controller RAID storage system. Based on this model, the high availability stochastic Petri net model of fault detection was established and used to evaluate the effect of the availability. In addition, we define a SHA (self-adaptive heart ability) parameter to scale the ability of system heartbeat cycle to adapt to the environment when high availability system is at a changing environment of read and write requests. The results show that, relatively speaking with fixed configuration, the design can enhance dual controller RAID system high availability.
摘要：
As so many novel network-based attacking behaviors have emerged in the Internet, it is increasingly difficult to protect the security of the network with traditional defense facilities, such as the Firewall and IDS. To solve this problem, a new multi-layers architecture of Artificial Immunity-Based Proactive Defense System (AIPDS) is put forward in this paper, which consists of detection layer, defense layer and user layer. According to the grown-up course of the B-Cells and T-cells in human lymphatic tissues, a B-Detector and T-Detector is designed to identify the unknown intrusions and the training algorithm of the Detector has also been given. Finally, a testing is done to prove that the AIPDS has an advantage over the traditional IDS.
摘要：
This paper demonstrates a new training method based on GA and Baum-Welch algorithms to obtain an HMM model with optimized number of states in the HMM models and its model parameters for web information extraction. This method is not only able to overcome the shortcomings of the slow convergence speed of the HMM approach. In addition, this method also finds better number of states in the HMM topology as well as its model parameters. From the experiments with the 2100 webs extracted from our corpus, this method is able to find the optimal topology in all cases. The experiments are found that the GA-HMM approach has an average precision rate of 84.483% while the HMM trained by the Baum-Welch method has an average precision rate of 71.049%. This implies that the GA-HMM method is more optimized than the HMM trained by the Baum-Welch method.
摘要：
Contour information is regarded as important characteristic in computer vision. It is difficult to extract Contour information from the non-structural object due to its complicated structure. This paper present a novel concept of Multi-Scale Entropy (MSE) based on traditional Entropy that can be used to perform reliable extracting contours from non-structured objects such as smoking and rocks. The variety of image information amount was presented dynamically by this means. The Multi-Scale Entropy Difference (MSED) can present the break part of the image gray information and recognize the boundary of object and background effectively. Finally the non-structural object contours was extracted by Maximal Multi-Scale Entropy Difference (MMSED). Experiments have shown that the operator can extract stable contours from non-structural objects and eliminate the interior complex texture structure effectively.
会议论文集名称：
International Symposium on Intelligent Signal Processing and Communication Systems ISPACS
摘要：
In order to reduce the bit rate and to improve the compression performance in video coding in H.264, this paper presents more scanning patterns for entropy coding Firstly, we investigate the two designs of entropy coding, termed CAVLC and CABAC And then according to the DCT transform, we deduce that it is possible to adopt more scanning patterns in entropy coding in H.264 to reduce the coding gain at the cost of bit rate. So this paper deduces some different features of an original or residual block, and proposes eight scanning patterns for entropy coding and the corresponding pattern decision method In the pattern decision method, we make a full search according to the bit rate cost to select the best pattern out of all scanning patterns. Experimental results based on the H.264 JVT reference software of JM10.1 show that employing the proposed scanning patterns bit rate reduction by up to 4 similar to 7% will be obtained with neglectable computation complexity increased and a better performance is also obtained.
摘要：
Motion estimation and motion compensation are included as major technologies into the existing video coding standards, which generate the motion vectors that determine how each motion compensated prediction frame is created from the reference frame. As traditional motion vectors are 2-dimensional and translational, they could not represent the actual moving directions of objects or blocks in frames. This paper proposes the special idea of GTMV (global translational motion vector), and then presents us the nearly Full Search Motion Estimation and Motion Compensation algorithms based on GTMV. Experimental results based on the H.264/AVC reference software of JM10.1 show that the proposed algorithms could get a better video compression performance and could improve the coding efficiency despite increasing some coding complexity.
摘要：
This paper provides an architecture of the distributed sampling measurement system based on mask matching,and the system is divided into three layers which includes message collecting,sampling and sav
摘要：
we propose a distributed wavelet-based algorithm which can transform irregularly sample data using haar wavelet-based compression solutions in this paper. We consider the characteristics and location information of nodes in sensor networks, a new distributed data aggregation mode based on "area" is used firstly. On the basis of these new models, a novel wavelet-based irregularly sample data compression and data transform model DDCM is proposed for sensor networks. Theoretical analyses and simulation results show that, the above new methods have the good ability of approximation, and can compress the data efficiently and can reduce the amount of data greatly, So, it can prolong the lifetime of the whole network to a greater degree.
摘要：
In this paper we propose video coding grid environment (VCGE), which is a project to provide a high performance video-coding platform in a grid-computing environment. As far, computational grids have become a vital emerging platform for high performance computing and video coding applications with high computing complexity have been used widely in many scientific research areas. We adopt the H.264 video compression technologies that are due to get high coding efficiency and high encoding performance in the new JVT while have very high computing complexity. VCGE is a combination of grid and H.264 video compression techniques on which complex video coding and other related applications can be carried out. In this paper, we discuss the infrastructure to provide the integration platform with both techniques, and give the components of the environment, which provides flexible and useful mechanism to achieve high video coding performance for video coding applications. The proposed design is very efficient for more complex video coding applications and presents us a novel video coding technology, which is leading a direction of grid application development
