Search

This paper proposes a novel texture based method for visualization of electromagnetic fields. According to the characteristics of electric eddy current fields, a scalar potential is constructed to express the field on the surfaces. Then the potential is used in deducing an extension form of the Sinusoidalfunction Modulated Potential (SMP) image of general surface eddy current fields. The SMP Image can display the distribution of both the direction and amplitude of the field. The paper also presents a new algorithm based on hardware accelerated texture mapping. The algorithm could render the SMP image in real time and can be further used in dynamic visualization of time-varying electromagnetic field. The rendered dynamic images can reflect the variation of the field with space and time simultaneously. The only requirements for the method are that the vector field is distributed on surface and that its divergent is free, which makes the method suitable for visualization of general divergent-free vector field distributed on surface. [ABSTRACT FROM AUTHOR]

This paper studies how the built-in natural weakness in the image processing algorithm or system can be searched and found with the evolutionary algorithms. In this paper, we show how the genetic algorithm finds the weakness in the Bayer's dispersed dot dithering method. We also take a closer look at the method and identify why this weakness is relatively easy to find with synthetic images. Moreover, we discuss the importance of comprehensive testing before the results with some image processing methods are reliable. [ABSTRACT FROM AUTHOR]

An algorithm for alternate pattern fill on arbitrary line segments, circular arcs and elliptical arcs is proposed in this paper. The algorithm creates an external loop which is a minimal loop that contains a given point. And then, all loops that delimit the nested regions for the alternate pattern fill inside the external loop are built. The loop detection is based on a new method for searching the leftmost edges. It requires only values of positions, tangent vectors, curvature radii, and the first derivatives of curvature radii at intersection points and areas of circles and ellipses. After all necessary loops are built, the given pattern is filled in the nested loops inside the external loop. Filling the given pattern in this paper is simplified into filling line segment patterns on some parallel lines. [ABSTRACT FROM AUTHOR]

This paper presents a method to reduce features less contributing to the classification of user preferred news groups among several news groups by the use of fuzzy inference and coefficient of determination. To this end, we extract a number of representative keywords from example documents through fuzzy inference. From the observation of training patterns, we found that lots of keywords in training patterns are empty. Thus, a new method to train neural network through reduction of unnecessary dimensions by the statistical coefficient of determination is proposed in this paper. Experimental results show that the proposed method is superior to the method using lots of input attributes in terms of within-cluster variance and its standard deviation. [ABSTRACT FROM AUTHOR]

In this paper, we compare a well-known semantic spacemodel, Latent Semantic Analysis (LSA) with another model, Hyperspace Analogue to Language (HAL) which is widely used in different area, especially in automatic query refinement. We conduct this comparative analysis to prove our hypothesis that with respect to ability of extracting the lexical information from a corpus of text, LSA is quite similar to HAL. We regard HAL and LSA as black boxes. Through a Pearson's correlation analysis to the outputs of these two black boxes, we conclude that LSA highly co-relates with HAL and thus there is a justification that LSA and HAL can potentially play a similar role in the area of facilitating automatic query refinement. This paper evaluates LSA in a new application area and contributes an effective way to compare different semantic space models. Keywords: Correlation Analysis, Hyperspace Analogue to Language, Latent Semantic Indexing, Automatic Query Refinement. [ABSTRACT FROM AUTHOR]

This paper proposes an approach to data integration and migration from a collection of heterogeneous and independent data sources into a data warehouse schema. Current methodology assumes that the data is loaded into data warehouse using queries in order to extract data from multiple data sources. Extracting data from various data sources requires the establishment of generic data integration methodologies. Sometimes, various data anomalies arise when using complex queries across multiple data sources. A data warehouse can be abstractly seen as a set of materialized views. Selecting views for materialization in a data warehouse is one of the most important decisions making tasks in its design. However, there are few facilities in data integration systems that consider these important issues. In this paper, we propose the approach of introducing staging table's schema that can be utilized to load required data from source tables into corresponding staging tables. Staging tables will be assumed to be temporary tables and each staging table will be empty before loading new data and using simple non-join queries to load data. We then focus on an approach to load data into data warehouse materialized views through staging tables including maintenance of materialized views. Keywords: Data Integration, Data Migration, Staging Tables, Global Schema, and Relational Schema. [ABSTRACT FROM AUTHOR]

Computational Grids, coupling geographically distributed resources, have emerged as a next generation computing platform for solving large-scale problems in scientific arena. Monitoring these Grid computing resources, which not only helps end-users monitor the status of Grid resources, but also provides quick and comprehensive information for resource scheduling and management, becomes extremely important for efficiently using Grid Computing resource. This paper presents architecture and implementation of a Grid Monitoring System based on Globus Toolkit. The Grid Monitoring System presented in this paper is flexible, extensible and, most important, secure. The architecture and implementation presented in this paper has important practical value for the monitoring of China Education and Research Grid. [ABSTRACT FROM AUTHOR]

In this paper, we propose an approach for the real-time performance analysis of distributed programs, called Eidolon. The approach is based on the capacity of communication links. In Eidolon, some algorithms are designed to find all the transmission paths of each data file needed while the program executes, count the transmission time for each the transmission path, then get the minimum and the maximum transmission time by analyzing all the transmission paths of each data file, and at last, calculate the fastest and the slowest response time of distributed programs. A case study is given at the end of this paper to justify the validity of this approach. [ABSTRACT FROM AUTHOR]

In this paper a networked control system (NCS) design problem is discussed where a single digital controller and multiple plants interconnected by a common computer network with transfer time delays. The network-induced delays are assumed to be unknown constant in this paper, and a measuring method for delays in TCP/IP network is proposed to compensate the performance degradation due to the delay effect. An NCS is designed with LQR algorithm and applied to an inverted pendulum accounting for the multiple time delays as a benchmark plant to check its performance under time delays induced by the network. [ABSTRACT FROM AUTHOR]

We propose a novel multicast communications model using a RingNet hierarchy, called the RingNet model, which is a combination of logical trees and logical rings for multicast communications. The RingNet hierarchy consists of four tiers: Border Router Tier (BRT), Access Gateway Tier (AGT), Access Proxy Tier (APT), and Mobile Host Tier (MHT). Within the hierarchy, the upper two tiers are dynamically organized into logical rings with network entities. In this paper, based on the RingNet model, local group concept is proposed. For simple illustration, we choose each AG in AGT as the controller of each local group. Each local group has its own independent local group key. The member's join or leave in a local group only affects the local group, which makes multicast communications potentially scalable to very large groups. In this paper, we propose a novel key management scheme for secure multicast with the RingNet model. [ABSTRACT FROM AUTHOR]

A dual-cube DC(m) has m + 1 links per node where m is the degree of a cluster (m-cube), and one extra link is used for connection between clusters. The dual-cube mitigates the problem of increasing number of links in the large-scale hypercube network while keeps most of the topological properties of the hypercube network. In this paper, we propose efficient algorithms for finding a nonfaulty routing path between any two nonfaulty nodes in the dual-cube with a large number of faulty nodes. A node v ∈ DC(m) is called k-safe if v has at least k nonfaulty neighbors. The DC(m) is called k-safe if every node in DC(m) is k-safe. The first algorithm presented in this paper is an off-line algorithm that uses global information of faulty status. It finds a nonfaulty path of length at most d(s,t) + O(k2) in O(F + m) time for any two nonfaulty nodes s and t in the k-safe DC(m) with number of faulty nodes F < 2k(m + 1 - k), where 0≤ k ≤ m/2. The second algorithm is an online algorithm that uses local information only. It can find a fault-free path with high probability in an arbitrarily faulty dual-cube with unbounded number of faulty nodes. [ABSTRACT FROM AUTHOR]

Interruptible load management (ILM) is an important part of Demand Side Management (DSM), especially in power market. It has deep influences on reliability and economical operation of power system. How to use a rational model to simulate the ILM becomes a hot topic in China. This paper gives the optimal purchase model of interruptible contract and sets a reasonable selected rule, i.e. the interruptible load should be priorly scheduled in the period with maximal shortage capacity. Based on the rule, the heuristic method is employed to solve the problem. Simulation results of real power system demonstrate the usefulness of the proposed model. [ABSTRACT FROM AUTHOR]

Web user patterns can be used to create a more robust web information service in personalization. But the user interests are changeable, that is, they differ from one user to another, and they are constantly changing for a specific user. This paper presents a dynamic mining approach based on Markov model to solve this problem. Markov model is introduced to keep track of the changes of user interest according to his or her navigational behaviors. Some new concepts in the model are defined. An algorithm based on the model is then designed to learn the user's favorite navigation paths. The approach is implemented in an example website, and the experimental results proved the effective of our approach. [ABSTRACT FROM AUTHOR]

With the rapidly increasing competitiveness in global market, dynamic alliances and virtual enterprises are becoming essential components of the economy in order to meet the market requirements for quality, responsiveness, and customer satisfaction. Partner selection is a key stage in the formation of a successful virtual enterprise. The process can be considered as a multi-class classification problem. In this paper, The Support Vector Machine (SVM) technique is proposed to perform automated ranking of potential partners. Experimental results indicate that desirable outcome can be obtained by using the SVM method in partner selections. In comparison with other methods in the literatures, the SVM-based method is advantageous in terms of generalization performance and the fitness accuracy with a limited number of training datasets. [ABSTRACT FROM AUTHOR]

Motion estimation is an important part of video encoding systems, because it can significantly affect the output quality and the compression ratio. Motion estimation which consists of integer pixel motion estimation and half pixel motion estimation is very computationally intensive part. To reduce the computational complexity, many methods have been proposed in both integer pixel motion estimation and half pixel motion estimation. For integer pixel motion estimation, some fast methods could reduce their computational complexity significantly. There remains, however, room for improvement in the performance of current methods for half pixel motion estimation. In this paper, an efficient half pixel motion estimation algorithm based on spatial correlations is proposed to reduce the computational complexity. According to spatially correlated information, the proposed method decides whether half pixel motion estimation is performed or not for the current block. Experimental results show that the proposed method outperforms most of current methods in computation complexity by reducing the number of search points with little degradation in image quality. When compared to full half pixel search method, the proposed algorithm achieves the search point reduction up to 96% with only 0.01 ~ 0.1 (dB) degradation of image quality. [ABSTRACT FROM AUTHOR]

A Chaikin curve is a subdivision curve. Subdivision begins from an initial control polygonal curve. For each subdivision step, all corners of the polygonal curve are cut off, and a new polygonal curve is thus produced as the input of the next subdivision step. In the limit of subdivision, a Chaikin curve is created. In this paper, a multi-step subdivision algorithm for generating Chaikin curves is proposed. For an arbitrary positive integer k, the algorithm builds the resultant polygonal curve of the kth subdivision step directly from the initial polygonal curve. Examples show that the new algorithm speeds up curve generation in several times. Keywords: Subdivision scheme; Chaikin curve; Corner cutting. [ABSTRACT FROM AUTHOR]

To estimate an accuracy motion vector (MV), a two step search is generally used. In the first step, integer pixel points within a search area are examined to find the integer pixel MV. Then, in the second step, 8 half pixel points around the selected integer pixel MV are examined and the best matching point is chosen as the final MV. Many fast integer pixel motion estimation (ME) algorithms can be found by examining less than about 10 search points. However, the half pixel ME requires huge computational complexity. In this paper, We propose a new fast algorithm for half pixel ME that reduces the computational overhead by limiting the number of interpolations of the candidate half pixel points. The proposed method based on the property of MVs and the correlations between integer pixel MVs and half pixel MVs. Experimental results show that the speedup improvement of the proposed algorithm over a full half pixel search (FHPS), horizontal and vertical directions as references (HVDR), chen's half pixel search algorithm (CHPS-1) and a parabolic prediction-based fast half-pixel search (PPHPS) can be up to 1.4 ~ 3.9 times on average. Also the image quality improvement can be better up to 0.05(dB) ~ 0.1(dB) compare with CHPS and PPHPS. [ABSTRACT FROM AUTHOR]

In order to extract the region of an object from an image, the difference image method is attractive since it is computationally inexpensive. However, the difference image is not frequently used due to the noise in the difference image. In this paper, we analyze the noise in an uncalibrated difference image and propose a statistical noise calibration method. In the experiment, the proposed method is verified using a real image. [ABSTRACT FROM AUTHOR]

This paper proposes a method for tracking and identifying persons from video image frames taken by a fixed camera. The majority of conventional video tracking surveillance systems assumes a likeness to a person's appearance for some time, and existing human tracking systems usually consider short-term situations. To address this situation, we use an adaptive background and human body model updated statistically frame-by-frame to correctly construct a person with body parts. The formed person is labeled and recorded in a person's list, which stores the individual's human body model details. Such recorded information can be used to identify tracked persons. The results of this experiment are demonstrated in several indoor situations. [ABSTRACT FROM AUTHOR]

Up to face extrusion (UTFE) is an effective operation to extrude the profile from a sketch plane to a selected face. An up to face extrusion algorithm is presented in this paper. The algorithm first generates the trimmed extrusion through a simple body-surface Boolean operation method, and then generates the resultant solid through a regularized Boolean operation between the original solid and the trimmed extrusion. The algorithm has been implemented in a commercial geometric modeling software system TiGems. Examples are given to illustrate the algorithm. [ABSTRACT FROM AUTHOR]

In this paper, we propose a fairing method of 3D rough meshes based on illuminations and geometric conditions for smooth rendering. In applications of interactive graphics such as virtual reality or augmented reality, rough meshes are widely used for fast optical interactions owing to their simple representation. However, in vertex-based shading, rough meshes may produce non-smooth rendering results; Distinct normal vectors and comparatively longer distance between consecutive vertices increase the difference of radiances. In order to improve the smoothness of the rendering results, the difference of radiances among vertices should be minimized by considering lighting conditions as fairing parameters. We calculated illuminations using diffuse lighting models at each vertex. Then normalized illumination is linearly integrated with the curvedness to prevent the shape distortion. By adapting integrated values to Laplacian weight factors, the difference of radiances is minimized and rendering result is improved, while maintaining the important curved shapes of the rough meshes. The proposed method also improves the compactness of triangles. The comparative study of our method with other existing fairing schemes has been discussed. We also applied our method to arbitrarily simplified meshes for demonstration. [ABSTRACT FROM AUTHOR]

In this paper, some new and efficient algorithms are described for feature description, analysis and recognition of contours. One linearization method is introduced . The series of curvature angle, linearity, and bend angle between two neighboring linearized lines are calculated from the starting line to the end line. The series of structural points are described. The useful series of features can be used for shape analysis and recognition of binary contours. [ABSTRACT FROM AUTHOR]

Terrain data set is essential for many 3D computer graphics applications. This paper presents a general framework for real-time terrain walk-through application, where the LOD(levels of detail) construction and occlusion culling are alternately selected depending on the terrain geometry. For highly occluded areas, only the occlusion culling is enabled. For less occluded areas, only the LOD rendering is enabled. A small piece of pre-computed information helps the renderer alternate the two methods. The experiment results demonstrate significant savings in rendering time. [ABSTRACT FROM AUTHOR]

This paper presents a real-time method for synthesizing texture. The algorithm consists of two stages: a preprocess stage and a real-time synthesis stage. In the preprocess stage, jump lists are built for each texture patch in the input sample texture and all the jump lists are stored in patch jump maps. In the synthesis stage, we synthesize the output texture with the aid of stored patch jump maps. Experiments show 200-500 frames of 256 × 256 high quality textures can be produced within a second. [ABSTRACT FROM AUTHOR]

General procedure for stereo image coding is to use disparity compensated prediction methods and code residuals separately. Although the characteristics of stereo residuals are different from those of common images, JPEG-like methods which are applied to monocular images are used frequently and less research has been devoted to residual image coding. The focus of this paper is to make stereo image coding more efficient by speculating the characteristics of the residual image. By measuring the edge tendency of residual image, our method modifies the quantization matrix adaptively using discrete wavelet transform. [ABSTRACT FROM AUTHOR]

This paper focuses on boundary surface based ray casting. In general the boundary surface based ray casting is processed in two stages. In the first stage, boundary surfaces are found and stored into buffers. In the second stage, the distance between the viewpoint and the voxels of the area of interest is calculated by projecting boundary surfaces on the view plane, and then the traverse of the volume data space with the distance is started. Our approach differs from the general boundary surface ray casting in its first stage. In contrast to the typical boundary surface based ray casting where all boundary surfaces of volume data are stored into buffers, in our proposal, they are projected on the planes aligned to the axis of volume data coordinates and these projected data are stored into 6-depth buffers. Such a maneuver shortens the time for ray casting, and reduces memory usage because it can be carried out independently from the amount of the volume data. [ABSTRACT FROM AUTHOR]

This paper proposes a new image stabilization mechanism based on filtering of absolute frame positions for mobile video appliances. The proposed mechanism removes undesired motion effects in real-time, while preserving desired gross camera displacements. The well known finite impulse response (FIR) filter is adopted for the filtering. The proposed mechanism provides the filtered position and velocity that have good inherent properties. It is shown that the filtered position is not affected by the constant velocity. It is also shown that the filtered velocity is separated from the position. Via computer simulations, the performance of the proposed mechanism is shown to be superior to the existing Kalman filtering based mechanism. [ABSTRACT FROM AUTHOR]

Driven by the need for interactive entertainment, modern PCs are equipped with specialized graphics processors (GPUs) for creation and display of images. These GPUs have become increasingly programmable, to the point that they now are capable of efficiently executing a significant number of computational kernels from non-graphical applications. In this introductory paper we first present a high-level overview of modern graphics hardware's architecture, then introduce several applications in scientific computing that can be efficiently accelerated by GPUs. Finally we list programming tools available for application development on GPUs. [ABSTRACT FROM AUTHOR]

A real number usually cannot be exactly represented by a floating-point number in a computer. Namely, a floating-point number frequently stands for any real number in a specific interval. In this paper, we present a method for computing the sign of a dot product sum. Each initial datum that is a floating-point number is considered as an interval. With interval analysis and floating-point summation methods, an explicit formula for calculating the minimal interval of a dot product sum is presented. Error analysis and some examples are provided as well. Keywords: Floating-point arithmetic; Rounding error; Interval analysis. [ABSTRACT FROM AUTHOR]

Image thresholding is one of the main techniques for image segmentation. It has many applications in pattern recognition, computer vision, and image and video understanding. This paper formulates the thresholding as an optimization problem: finding the best thresholds that minimize a weighted sum-of-squared-error function. A fast iterative optimization algorithm is presented to reach this goal. Our algorithm is compared with a classic, most commonly-used thresholding approach. Both theoretic analysis and experiments show that the two approaches are equivalent. However, our formulation of the problem allows us to develop a much more efficient algorithm, which has more applications, especially in real-time video surveillance and tracking systems. [ABSTRACT FROM AUTHOR]

A method for constructing a cubic Pythagorean hodograph (PH) curve (called a Tschirnhausen cubic curve as well) satisfying unbalanced Hermite interpolation conditions is presented. The resultant curve interpolates two given end points, and has a given vector as the tangent vector at the starting point. The generation method is based on complex number calculation. Resultant curves are represented in a Bézier form. Our result shows that there are two Tschirnhausen cubic curves fulfilling the unbalanced Hermite interpolation conditions. An explicit formula for calculating the absolute rotation number is provided to select the better curve from the two Tschirnhausen cubic curves. Examples are given as well to illustrate the method proposed in this paper. Keywords: Hermite; Pythagorean hodograph; Absolute rotation number. [ABSTRACT FROM AUTHOR]

Model selection plays a key role in the performance of a support vector machine (SVM). In this paper, we propose two algorithms that use the Vapnik Chervonenkis (VC) bound for SVM model selection. The algorithms employ a coarse-to-fine search strategy to obtain the best parameters in some predefined ranges for a given problem. Experimental results on several benchmark datasets show that the proposed hybrid algorithm has very comparative performance with the cross validation algorithm. [ABSTRACT FROM AUTHOR]

In this paper we present a multiple bank electronic cash system based on group blind signature scheme, which offers a new kind of tracing mechanism. It provides conditional anonymity both for the customers and electronic coins under a judge. The coins can be marked using undeniable signature scheme so that the bank will recognize these coins at deposit. We also use the secret sharing scheme to trace the customer under the permission of a judge. The security of our scheme is analyzed. And compared with other works, our proposed tracing methods offer more privacy and do not need any trusted third parties. Our system is able to prevent from blackmailing, kidnapping, and bank robberies. Also we extend electronic cash system to multiple banks, it is more practical in the real life. [ABSTRACT FROM AUTHOR]

Digital signature is an important task in modern cryptography, which concerns about authenticity, integrity and non-repudiation of data on communication channel. In this paper, an arbitrated quantum digital signature scheme is proposed, in which the receiver verifies the signature with the help of an arbitrator. The most important property of the proposed scheme is that it can sign unknown quantum states. The security of the protocol relies on the quantum one-time-pad and the Greenberger-Horne-Zeilinger (GHZ) triplet particles distributed among communication entities. Security analysis shows that the proposed quantum signature is a secure scheme. [ABSTRACT FROM AUTHOR]

Keyword-based search has been popularized by Web search services. However, due to the problems associated with polysemy and synonym, users are often unable to get the exact information they are looking for. In this paper, we propose a web search framework in which ontologies and rules are used to deal with the synonym and polysemy problem to gain a better recall and precision performence than the traditional search services. A demonstration system based on Google's web search APIs is currently under implementation. [ABSTRACT FROM AUTHOR]

Ontology plays an important role on the Semantic Web. In this paper, we propose a method, AOIWD, of acquiring ontology information from Web documents. The AOIWD method employs data mining techniques combined with inference engine techniques to develop ontology. One key feature of AOIWD is that it employs association rules and inference mechanism to create association instances of ontology classes. Another key feature of AOIWD is that it employs clustering to create ontology class hierarchy. The method can discover more meaningful relationships and has less human intervention. [ABSTRACT FROM AUTHOR]

The property of forward-security can be used to reduce the potential damages caused by key exposure. However, none of the existing proxy signature schemes has taken forward-security into consideration. And most proxy signature schemes are based on discrete logarithm problem. In this paper, we first present a forward-secure proxy signature scheme based on factoring problem. Next, we present a forward-secure threshold proxy signature scheme. [ABSTRACT FROM AUTHOR]

Cryptographic support is an important dimension of securing sensitive data in databases. But this usually implies that one have to sacrifice performance for security. In this paper, we describe an novel cryptographic schemes for the problem of searching on encrypted character data and provide analysis of security for the resulting cryptographic systems.When the character data are stored in the form of cipher, we store not only the encrypted character data, but also characteristic values of character data as an additional field in database.When querying the encrypted character data,we apply the principle of two-phase query. Firstly, it implements a coarse query over encrypted data in order to filter part records not related to querying conditions. Secondly, it decrypts the rest records and implements a refined query over them. Results of a set of experiments validate the functionality and usability of this approach. [ABSTRACT FROM AUTHOR]

Existing discretization methods cannot process continuous interval-valued attributes in rough set theory. This paper extended the existing definition of discretization based on cut-splitting and gave the definition of generalized discretization using class-separability criterion function firstly. Then, a new approach was proposed to discretize continuous interval-valued attributes. The introduced approach emphasized on the class-separability in the process of discretization of continuous attributes, so the approach helped to simplify the classifier design and to enhance accurate recognition rate in pattern recognition and machine learning. In the simulation experiment, the decision table was composed of 8 features and 10 radar emitter signals, and the results obtained from discretization of continuous interval-valued attributes, reduction of attributes and automatic recognition of 10 radar emitter signals show that the reduced attribute set achieves higher accurate recognition rate than the original attribute set, which verifies that the introduced approach is valid and feasible. [ABSTRACT FROM AUTHOR]

With the development of telecom, it becomes possible to provide people with Location Based Service (LBS) according to their spatial positions by using mobile equipments. Before using such services, people should evaluate them and make suitable strategies. Demand Bus System is attracting attention as a new public transportation system that provides convenient transportation for special services while solving traffic jams in urban areas. Previous researches focused on the algorithms to the dial-a-ride problem, and few works did on evaluating the usability of the demand bus system. In this paper we proposed a Q-based multi-agent simulation framework for this purpose and implemented a prototype system. [ABSTRACT FROM AUTHOR]

Recommender systems make suggestions to users. Collaborative filtering techniques make the predictions by using the ratings on items of other users. In this paper, we have studied item-based and user-based collaborative filtering techniques. We identify the shortcomings of current filtering techniques. The performance of recommender systems was deeply affected by user's rating behavior. We propose some improvements to overcome this limitation. User evaluation has been conducted. Experiment results show that the new algorithms improve the performance of recommender systems significantly. [ABSTRACT FROM AUTHOR]

Currently there appear only few practical semantic web applications. The reason is mainly in that a large number of existed web documents contain only machine-unreadable information on which software agent can do nothing. There have been some works devoting to web document annotation manually or semi-automatically to solve this problem. This paper presents an automatic approach for web document annotation based on specific domain ontology. Because complete semantic annotation of web document is still a tough task, we simplify the problem by annotating ontology concept instances on web documents and propose an Ontology Instance Learning (OIL) method to extract instances from structure and free text of web documents. These instances of the ontology concept will be used to annotate web pages in the related domain. Our OIL method exhibits quite good performance in real life web documents as shown in our experiment. [ABSTRACT FROM AUTHOR]

In this paper, semantic role labeling is addressed. We formulate the problem as a classification task, in which the words of a sentence are assigned to semantic role classes using a classifier. The maximum entropy approach is applied to train the classifier, by using a large real corpus annotated with argument structures. [ABSTRACT FROM AUTHOR]

The present paper is devoted to the pattern recognition procedure based on the set of expert rules with unprecisely formulated weights understood as appropriate probabilities. Adopting the probabilistic model the different interpretations of rule weight are discussed and the consistency conditions of set of rules are given. [ABSTRACT FROM AUTHOR]

We examined the wave pattern of the spleen by using abdominal CT images of a patient with liver cirrhosis, and found that they are different from those of a person with a normal liver. This paper suggests a method to diagnose liver cirrhosis by using the wave pattern of the spleen in abdominal CT images on the basis of the two principles. It tells us that we can judge if the liver has liver cirrhosis, without the test of the ratio of caudate lobe to right lobe, only with the spleen. [ABSTRACT FROM AUTHOR]

Case-based reasoning (CBR) has been applied to various problem-solving areas for a long time because it is suitable to complex and unstructured problems. However, the design of appropriate case retrieval mechanisms to improve the performance of CBR is still a challenging issue. In this paper, we encode the feature weighting and instance selection within the same genetic algorithm (GA) and suggest simultaneous optimization model of feature weighting and instance selection. This study applies the novel model to corporate bankruptcy prediction. Experimental results show that the proposed model outperforms other CBR models. [ABSTRACT FROM AUTHOR]

In the computer graphics field, mesh simplification, multi-resolution analysis and surface parameterization techniques have been widely investigated. This paper presents innovative applications of those techniques to the biomechanical analysis of thin elastic braces. The braces are represented in polygonal meshes, and mesh simplification and multi-resolution analysis enable fast geometric computation and processing. The multi-resolution brace mesh is parameterized for strain energy calculation. The experiment results prove that 3D geometrical analysis works quite well for assistive medical device design. [ABSTRACT FROM AUTHOR]

This paper introduces a novel paradigm of privacy preserving mining for distributed databases. The paradigm includes an agent-based approach for distributed learning of a decision tree to fully analyze data located at several distributed sites without revealing any information at each site. The distributed decision tree approach has been developed from the well-known decision tree algorithm, for the distributed and privacy preserving data mining process. It is performed on the agent based architecture dealing with distributed databases in a collaborative fashion. This approach is very useful to be applied to a variety of domains which require information security and privacy during data mining process. [ABSTRACT FROM AUTHOR]

The Internet site is encouraged to take a little time to get the preferred information and bring it to the customer through personalization of customers' information. The mobile shopping users have a strong desire to reduce the shopping hours due to the expensive telecommunication service fees. It requires much time and perseverance to input the personalization information into the server and even to input it repeatedly into a bunch of shopping malls while maintaining contact. In this paper, we develop an application to protect the general personalized information in the mobile client-side based on XML. The mobile shopping using the client-side personalization reduces the connection time and improves the personalization information management through the enhanced data transmission and data input. [ABSTRACT FROM AUTHOR]

The problem of determining optimal decision model is a difficult combinatorial task in the fields of pattern classification and machine learning. In this paper, we propose a new method to find the optimal decision model for SVM, which consists of the minimal set of highly discriminative features and the set of parameters for the kernel. To cope with this problem, we adopted genetic algorithm (GA) which provides efficient optimization tool simulating the natural evolution procedures in iterative fashion to select the optimal set of features and set of kernel parameters. In the method, the decision models generated by GA are evaluated by SVM, and GA selects the only good models and gives the selected models the chance to survive and improve by crossover and mutation operation. Combining GA and SVM, we can obtain the optimal decision model which reduces the execution time as well as improves the classification rate of SVM. We also demonstrated the feasibility of our proposed method by several experiments on the sets of clinical data such as KDD Cup 1999 intrusion detection pattern samples and stomach cancer proteome pattern samples. [ABSTRACT FROM AUTHOR]