Abstract:Data cube is the central mechanism in multi-dimension database and online analytical processing (OLAP) based on multi-dimensional analysis. In this paper, an algebra for OLAP data cube is proposed, which supports multi-dimensional database and analysis. It can be the theoretical foundation of semantic specification of data warehousing and OLAP manipulations. Some applications of the new mathematical tool are presented as well to show the power of the contribution.

Abstract:A method of image zooming-in is presented in this paper. It is based on Bézier interpolation surface. First, it constructs a piecewise bicubic Bézier interpolation C¹ surface for each color component of a digital image, and then, samples the surface with different sampling rate to zoom in. The experimental results show that the method can improve the quality of zoomed image greatly.

Abstract:A segmentation method in which texture and gray variation information is integrated by means of function-linked neural network is proposed in this paper. Based on this method, a region-based region growing algorithm is designed and applied to medical image segmentation. The experimental results show that this method can produce good segmentation result when applied to certain images.

Abstract:In a traditional network, it is difficult to deploy new network protocols and services, and this problem could not be overcome by itself. Active networking, however, is a feasible way. In this paper, the authors introduce a new router operating system——TH-AOSR (Tsinghua active operating system for router), based on the investigation of active network. In addition to be compatible to traditional routers, TH-AOSR has active networking and active computing capacity, which enables any specific user or application to customize their services, while forwarding data packets securely and rapidly.

Abstract:An algorithm for triangle mesh simplification based on super-envelopes is presented in this paper. This algorithm can not only deal with arbitrary meshes, but also control the global error bound of the simplification. It runs fast and the result is perfect. The authors also give a method for constructing continuous level of detail models. Examples illustrate efficiency of the algorithm. Finally, the authors compare the algorithm with those algorithms with global error control.

Abstract:In this paper, the technical characteristics of text/graphics segmentation of engineering drawings, its critical steps and basic processing framework are analyzed. The SegChar, a practical tool for automatic text/graphics segmentation developed by the authors, is presented. The emphasis is put on two technical respects: (1) the automatic text size threshold method, which makes the processing procedure automatic and intelligent; (2) the probing strategy for extraction of text strings of arbitrary direction and length, which can extract Chinese/Western character strings integrally, and enhances its performance in processing speed, space complexity by introducing the concepts of accurate HOUGH space requirements, collinear relaxation and the string based HOUGH field refreshing method. The performance evaluation of SegChar is given finally.

Abstract:Network accounting management is the important management function of commercialized computer network. With the commercialization of Internet and the occurrence of lots of Intranets, it's urgent to develop network accounting management system. A principal requirement of such systems is efficiency and reliability. A fault-tolerant accounting system based on RPC (remote procedure call) mechanism——FTCharge(fault-tolerant charge) is proposed in this paper. After the brief discussion of the accounting principle, the design of the architecture and the fault-tolerance mechanism of FTCharge system are discussed.

Abstract:Pointer alias analysis plays an important role in the parallelizing optimization of C program, however, all the previous analyzing algorithm can only be used to analyze the pointer scalars. In this paper, an extended representation for point-to information is presented, which can represent not only the point-to information of pointer scalars, but also the point-to information of pointer arrays. Furthermore, an algorithm of intraprocedural alias analysis for pointer array is presented. This algorithm comprises both the pointer scalar analysis and pointer array analysis. It can be used to efficiently solve some problems which couldn't be solved by the previous algorithm.

Abstract:Based on the fault-tolerant bankcheck recognition system (FBRS), design and implementation of the preprocessing methods are proposed in this paper, which include software locating algorithm, extracting and erasing background of each sub-image, skew correction of each sub-image, deleting restricting rectangle frame of some sub-images and segmentation of characters' string. Among them, the software locating algorithm and skew correction are discussed in detail.

Abstract:CONET V2.0 is a network system under microkernel environment. Compared with the traditional monolithic kernel, microkernel provides a different approach to construct operating system, which leads to different implementation of those modules running under such environment. In this paper, the architecture and design of CONET V2.0 are outlined, and then the design and the implementation of SOCKET, which is one of the application interfaces for the network system, are presented in detail.

Abstract:In concurrent object-oriented languages, the inheritance anomaly is an important and difficult problem, which makes synchronization codes difficult to reuse. Based on the two-layered-class model, a new method for solving the inheritance anomaly is proposed. It can provide the flexible and sufficient support to the reuse of synchronization codes.

Abstract:Gibbs phenomenon, which occurs in the wavelet-based image compression algorithms under low bit rates, remains an open question for many years. The main cause is that purely-pixel-value-based MSE(mean square error) criteria can not allocate enough bits to the wavelet coefficients corresponding to edges in image. With detail analysis of zero-tree wavelet image compression algorithm originally proposed by Shapiro and then well-modified by Said and Pearlman, the algorithm is improved by suppressing high frequency noises as well as adaptively quantizing coefficients around edges. Experimental results are comparatively given. The main contribution of this paper is the idea of combination of recognition and compression. With the aid of the spatial localization property of wavelet transform, a very flexible bit allocation scheme can be realized, and therefore Gibbs phenomenon is reduced to some extent.

Abstract:In this paper, the authors present a novel method to incorporate temporal correlation into a speech recognition system based on conventional hidden Markov model (HMM). The temporal correlation is considered to be useful for recognition because of the fact that the speech features of the present frame are highly informative about the feature characteristics of neighboring frames. An obvious way to incorporate temporal correlation is to condition the probability of the current observation on the current state as well as on the previous observation and the previous state. But using this method directly must lead to unreliable parameter estimation for the number of parameters to be estimated may increase too excessively to limited train data. In this paper, the authors approximate the joint conditional PD by non-linear estimation method. As a result, they can still use mixture Gaussian density to represent the joint conditional PD for the principle of any PD can be approximated by mixture Gaussian density. The HMM incorporated temporal correlation by non-linear estimation method, which they called FC (frame correlation) HMM does not need any additional parameters and it only brings a little additional computing quantity. The results of the experiment show that the top 1 recognition rate of FC HMM has been raised by 6 percent compared to the conventional HMM method.

Abstract:An inexpensive satellite receiver can receive high bandwidth traffic from a satellite, while no bandwidth from the receiver to the satellite is provided. Therefore the connection between the satellite and the satellite receiver is unidirectional. The existing routing protocols stand on the fact that any links are bi-directional. They can not handle unidirectional links. Some scholars use a tunneling method to solve this problem, but the tunneling method needs a fixed back channel from satellite receiver to the satellite uplink station. It can not adapt to dynamic topology change. In this paper, the authors propose a dynamic routing algorithm that can be applied to networks with unidirectional links. The routing algorithm can also adapt to the change of network topology.

Abstract:In this paper, based on the LOTOS specification language, the authors present the transformation which starts from the functional specification and the distributed properties of target system to the corresponding modular specification. The authors present the tagged specification to represent the complex distributed properties of the full LOTOS specifications and study the decomposition transformation of full LOTOS specification which using the broadcasting communication gates in the system's internal coordinating communications and directly decomposing the multiple subsets.

Abstract:In this paper, a method of combining particle system with density fields is presented to simulate smog movement and diffusion . By attaching a density function to the particle attributes, smog diffusion is simulated through the variance of particle effect radius and density functions, effect of turbulence is imitated through particle movement and fragmentation. Compared with the traditional particle system, this method can describe smog diffusion and produce continuous density fields with less particles, the computation is thus greatly speeded up. A quick algorithm for smog rendering is also presented for real-time simulation.

Abstract:An automatic segmentation and classification method of magnetic resonance image of human brain based on a computerized brain atlas-TT atlas is introduced in this paper. The algorithm could be divided into two steps. Firstly, the object brain dataset is registered with the TT atlas by a 3D image matching method. With the help of anatomy structure information of the computerized atlas, the brain image is segmented and classified coarsely. Secondly, the segmentation result is taken as the input of a fuzzy cluster segmentation. In order to use the anatomic structure information directly, a new fuzzy cluster method with a shape factor is introduced, the method is mostly and automatically except selection of some intrinsic point pairs for global linear matching method in the first step.

Abstract:Premature convergence and low converging speed are the distinct weaknesses of the genetic algorithms. Using the living things' growth pattern for reference, a new model called ECM(ecological competition model) is proposed, in which the competition is considered to be in important position. In the ECM model, the congenital genetic evolution and the postnatal competition learning on individuals' level are realized in each sub-population, moreover, the competition reinforcement learning on population level is realized. The experimental results show the ECM model's effectiveness.

Abstract:Aiming at the defects of traditional clustering model, a kind of clustering model and algorithm are put forward and researched by use of deterministic annealing. The model takes account of the interactions of clusters, some models which were put forward previously are special cases of this one. Temperature parameter is introduced, and the clustering problem as a physical system is considered. Finding the optimal solution to clustering problem is transformed into simulating the equilibrium state of a physical system. The equilibrium state is simulated by solving a series of problems to minimize the free energy which varies with temperature, and finally, the ground state of the system is attained. That is the optimal solution of clustering problem.

Abstract:The management of modern computer networks is a complex and labour-intensive task that requires the assimilation of vast amounts of information and application of operator's judgment and expertise. The complexity is further increased because of the multi-vendor heterogeneous elements that make up many of these networks, and the high quality of user service requirements. In this paper, the authors give a novel system model for building an intelligent management of network (IMN) in order to solve related problems. In the IMN architecture proposed, expert systems are integrated in the network management system to handle particularly difficult problems, especially an intelligent fault management system is presented. By spreading intelligent agents through network segments or domain, the IMN system can bring the management closer to application and user requirements. The authors also give a manager model in the IMN system, which controls each subsystems and coordinates different management task. The IMN system is well suited to manage the current computer network intelligently, effectively and efficiently.