Search

The equivalent transformation based steganographic algorithm for PDF files shows good performance, both in capacity and invisibility. In addition, data embedding by this algorithm need not change the size of a cover-file. However, a loophole exists in the steganographic scheme. Through analyzing the structure of a PDF file, we can reconstruct a PDF file by using the software tool according to the obtained "Creator" key, and reveal the presence of secret data. Furthermore, by making use of the fact that the residue sequence distributes uniformly, this paper's steganalysis scheme can also estimate the length of hidden data. To enhance security, in this paper, it is proposed to introduce a "random moduli" equivalent transformation based steganographic algorithm. It not only has more security, but also has larger capacity. Adopted practical PDF files, computing results show that this paper's steganalysis scheme and "random moduli" equivalent transformation based steganographic algorithm are all effective. Moreover, it is possible to extend this paper's methods to some other data sets, e.g., webpage structure text files. [ABSTRACT FROM AUTHOR]

It describes each component element of Chinese ink painting in detail. Firstly, this paper presents brush emulation based on physical calculation. Then, it proposes the ink painting decomposition research technique, and proposes the simulation model system of the Chinese ink painting effect. This model had considered the writing brush model, as well as rice paper model of the water ink proliferation mode minutely. In addition, it proposed the transfusion mechanical mode which considered the gluewater's influence. So it had finished the ink proliferated interactive technology mode fusion physical and ink characteristic. It conformed to the Chinese ink painting unique style, and might carry on the effect according to the parameter the control, which caused the ink painting more vivid. [ABSTRACT FROM AUTHOR]

In this paper, we propose an improved parallelizing algorithm for serial programs without reducing the relevancy to the previous algorithm called PDMA. We add eliminating false dependences to PDMA for higher parallelism, as well as giving the description of the improved algorithm. Finally, this paper shows that the improved algorithm achieves better parallelism than PDMA through example comparison. [ABSTRACT FROM AUTHOR]

Task scheduling is a NP puzzle. Its algorithm is an important research direction. This paper proposes a task scheduling algorithm based on real-coded immune algorithm by studying and analyzing task scheduling models and immune algorithms existed. This paper discusses the coding method, the generation and update of population, the update of memory cells and the values of partial parameters. This paper explores the affinity function and the concentration function. In the end, this new algorithm is implemented in the software VC++, and is proven validity and feasibility by comparing and analyzing examples [ABSTRACT FROM AUTHOR]

With the rapid increasing of Web data volume, it has been a hot issue to find competitive intelligence from Web. However, previous studies in this direction focused on the extracting algorithms for Web-based competitive intelligence, and little work has been done in the credibility evaluation of Web-based competitive intelligence. The credibility of Web-based competitive intelligence plays critical roles in the application of competitive intelligence and can determine the effectiveness of competitive intelligence. In this paper, we focus on the credibility evaluation issue of Web-based competitive intelligence, and present a framework to model and evaluate the credibility of Web-based competitive intelligence. A networked credibility model and some new methods for credibility evaluation are proposed in the paper. [ABSTRACT FROM AUTHOR]

In this paper, we present a new method of numeral recognition that applies character features of the power meter images taken by a handheld terminal camera. The main difficulties for this recognition arise from pictures of low acquisition resolution, noise, heavy or light priting, and numerial distortion. Besides, different dials have different figures in sizes and font. Based on above difficulties, a new kind of 8 Direction Tracking Find-Hole method is proposed. Numbers are recognized by extracting holes and gaps of each character. The systerm gets binary image firstly by means of preprocessing and then extracts holes and gaps of target figures based on structural characteristics. The Find-Hole method in this paper dosen't have to normalize irregular characters. Experiments show that this method performs well in anti-interference, reducing the complexity of the algorithm effectively and enhancing the processing speed and recognition rate. [ABSTRACT FROM AUTHOR]

A new method is used to generate three-dimensional terrain maps from discrete points in this paper. The process of generating terrain generates triangulation using the discrete points, by interpolation to calculate the elevation data of regular grid, with the Diamond-Square algorithm in fractal theory to further refine the grid data. The results of programming experiments show that realistic terrain maps can be got from a small amount of terrain feature points, which is suitable for rendering the virtual scene. [ABSTRACT FROM AUTHOR]

This paper presents a new active calibration algorithm to calibrate uniform linear array with amplitude and phase errors by placing an auxiliary signal known with energy and azimuth in the far-field. This algorithm only needs some easy disposal of the signal covariance matrix to estimate the amplitude-phase errors without matrix eigendecomposition as the traditional active-calibration algorithm does. It's greatly reduces the computation. When the array is calibrated, the paper apply one-dimension noise subspace algorithm to estimate the signal DOA. This algorithm can be applied to electronic countermeasure. Simulation results show that the algorithm can effectively estimate the array amplitude-phase errors, and it is robust. [ABSTRACT FROM AUTHOR]

This paper analyzes the current situation of the intrusion detection system, which is the basis to put forward that data mining technology is to be applied to the intrusion detection system in terms of the problems of the traditional intrusion detection system. Meanwhile, the paper designs the intrusion detection model of data mining. With the study on intrusion detection and data mining, the algorithm of classical relation with clustering in the characteristics on intrusion detection system is improved and optimized. [ABSTRACT FROM AUTHOR]

The accuracy of traditional monitoring methods using remote sensing was lower, because of the limited number of monitoring points on the Tai lake. This paper proposed to use the Least Squares Support Vector Machine (LS-SVM) theory to improve the accuracy of water quality retrieval, which is suitable for the small-sample fitting. The LS-SVM model was used to monitor concentration of suspended matter. In this paper, the Radial Basic Function (RBF) was chosen as the kernel function of the retrieval model, and the grid searching and k-cross validation were used to choose and optimize the parameters. From the results of experiment, it showed that the proposed method had good performance and at the same time, the complexity is lower and the speed of the modeling was rapid. [ABSTRACT FROM AUTHOR]

The artificial neural network provides us a new theoretical method on the stability classification of roadway surrounding rock. The surrounding rock is a complex dielectrics, it exhibits a nonlinear relationship between influencing factors and the stability of roadway surrounding rock, for such problems, BP neural networks have higher modeling capabilities, can truly portray nonlinear relationship between the question and influencing factors, therefore, the paper uses BP neural networks to build the prediction model for the stability classification of roadway surrounding rock, and uses VC++ 6.0 to design and implementation of system. Through the training and testing of samples, the results show that the system is reliable. [ABSTRACT FROM AUTHOR]

Since there is a defect that the surface of mine area has been covered with black dust and the spectral character is inconspicuous, this paper developed a new way to find the pile-like object. First, the clinometric map is set up based on the DSM which has been collected from the data that the airborne LiDAR generated in mine area. After the high resolution image ortho-rectified according to DSM, the object orientated information extraction method is adopted to divide the fused multisource remote sensing information into some homogeneous areas. This paper builds the pile-like object extraction model by means of analyzing the elevation information, slope, dispersion, spectral, and topology relationship. Furthermore, the experiment has proved it is a high-efficient and precise way to find the pile-like object in mine area. [ABSTRACT FROM AUTHOR]

A new algorithm based on intensity curve of High-resolution images is proposed in this paper, which is used in vehicle flow measuring system. According to different characters between moving objects and background, it can detect the presence of vehicles, and complete the real-time detection consuming with less time than traditional algorithm, such as background Model and Bayes Model. It can also be used in the statistics of complete multi-lanes vehicles flow. In the 20-minutes' video test of high-resolution, the accuracy is up to 98%. [ABSTRACT FROM AUTHOR]

With the growing number of web services, the importance of matching and discovery web services is increasing. Using domain ontology to describe the semantic of web services and matching the web services on the semantic level is a hot research pot. In this paper, the problem of matching web services is transformed to the computation of semantic similarity between concepts in domain ontology. We propose the semantic similarity can be measured from the semantic distance and consider the factors of pathlengh, depth, local density and number of down direction in the algorithm. We establish that such a measure is an effective means of discriminating services at a level of granularity that is able to enhance the matching process in semantic web services. [ABSTRACT FROM AUTHOR]

Aiming at the problem that the tradition background extraction method could not precisely extract background when moving objects have high density or appear too frequently. This paper presents an algorithm of background reconstruction based on morphology and pixel intensity classification (PIC). The algorithm classifies the pixels in each frame into background area and moving objection area using three frame differences and Morphological analysis. Then filtered foreground objection when reconstructed background with PIC that improves the frequency of background, so the background could be rebuilded properly. The experiment result shows that the proposed method in this paper could extract accurate background in high density of vehicle. [ABSTRACT FROM AUTHOR]

based on the traditional mean-shift tracking algorithm, this paper applies the method of Collins which combined the mean-shift tracking algorithm with the scale-space to settle the handicap of tracking objects while scale changes. Aimed at the character of the infrared image, a improved target modeling method is presented and the kalman filter is also applyed to improve the system performance. [ABSTRACT FROM AUTHOR]

To the security threats existed in Neighbor Discovery Protocol(NDP) of IPv6, the paper provides a distributed intrusion detection model based on Multi- Agent(MAIDM). The algorithm and Mechanism of agent's studying and detecting is described in detail based on a mechanism of Abnormity Presenting and Abnormity Triggering. Experimental results demonstrate the efficiency of the proposed MAIDM. [ABSTRACT FROM AUTHOR]

Support Vector Machine (SVM ) is widely applied to fault diagnosis of machines. However, this classification method has some weaknesses. For example, it can not separate fuzzy information, particularly sensitive to the interference and the isolated points of the training samples. Besides, it has great demand for memory in calculation. In view of the problems mentioned above, a binary tree-based fuzzy SVM multi-classification algorithm (BTFSVM) has been put forward. This paper focuses on the study of the application of the theory BTFSVM to fault diagnosis on the gearbox of ships. Simulation experiments show that the algorithm has better anti-interference ability and classification effects than others. Consideration should be taken into account that it can be further applicable to the diagnosis on other mechanical faults of ships. [ABSTRACT FROM AUTHOR]

Skeleton extraction is a very challenging subject, and has an important application value. Because of the ambiguity and complexity of cerebrovascular image, the skeleton gained by conventional skeleton algorithms is discontinuous. This paper proposes a cerebrovascular image skeleton extraction algorithm based on topological nodes. This algorithm first determines the important topological nodes from the starting point of the skeleton, then calls the extraction algorithm for a single skeleton from every topological node until the arrival of the source point. This algorithm avoids the locating and classifying of the skeleton connection points which guide the skeleton extraction. Because all its parameters are gotten by the analysis and reasoning, no artificial interference is needed. [ABSTRACT FROM AUTHOR]

It is the key step of the cerebral aneurysm recognition system to locate the cerebral aneurysm accurately and fast onto the image. A new detection method of cerebral aneurysm, which is based on the improved thinning algorithm, is proposed after analyzing the morphological characteristics of cerebral aneurysm fully in the paper. In this new detection method, the improved OPTA algorithm is used to get the skeleton tree of blood vessel firstly, and then cerebral aneurysms are detected by searching the skeleton tree. After doing lots of experiments, the cerebral aneurysm can be detected well by using this new method, which provides a premise for cerebral aneurysm recognition. [ABSTRACT FROM AUTHOR]

Based on the software platform of MATLAB simulation, the paper collects the real GPS intermediate frequency (IF) signal by the IF signal sampler, and uses it to carry out the time-domain correlation acquisition and the fast Fourier transform (FFT) acquisition on the C/A code of GPS signal, and also carry out the decision of the singlebinary search and the M/N search on the personal computer. The experimental results show that comparing to the algorithm of time-domain correlation acquisition, the algorithm of FFT acquisition can consumedly reduce the computing times, and accordingly reduce the acquisition time. In addition, comparing to the decision algorithm of single-binary search, the M/N search can consumedly reduce the false-alarm probability and improve the detection probability. So, the M/N search can both control acquisition time effectively and meet the high performance requirement of acquisition. [ABSTRACT FROM AUTHOR]

The Collaboration of Multi-Agent is one of the most important issues in the study of Multi-Agent Systems, and agent union has become the priority and hot spot as a form of collaboration when studying multi-Agent system theory and technology. Though, there exists a variety of coordination algorithms, it is difficult to compare and verify these algorithms. So, the simulation platform called MAZE is developed, that users can use it to compare and verify various algorithms about Agent. Although there have been a variety of Agent infrastructure platforms, such as JADE, which has been widely used. However, it is difficult to extend and customize the customer-centric application based on these platforms. This paper will summarize some common patterns (e.g. coalitions )to support the development of the MAS system by the design and implementation of MAZE which is a multi-Agent system simulation platform. Users can test and compare various Agent Alliance algorithms. [ABSTRACT FROM AUTHOR]

The volume of junk emails on the Internet has grown tremendously in the past few years and is causing serious problems. Content-based filtering is one of mainstream technologies used so far. This paper has had a deep study in the content of emails and come up with a better idea to get the features which make it even convenient to e-mail classify as well. This paper uses the classification algorithm by essential emerging patterns CeEP to the junk email examine, and carries out a new categorization and filtering algorithm ECFEP of emails based on the EP. The experiments show, the new feature extraction methods and the combination CeEP classification is a very efficient method of classification, and The classification efficiency of the algorithm ECFEP is higher than currently several better classification algorithm. [ABSTRACT FROM AUTHOR]

This paper introduces an improved clustering algorithm GCA(Gravitational Clustering Algorithm ), it is extended in such a way that the Gravitational Law is not the only law that can be applied. This algorithm can decide automatically the number of clusters in the target data set, and find any clusters with arbitrary forms and filter the noisy data. The experimental results show that GCA algorithm creates high quality greatly. [ABSTRACT FROM AUTHOR]

Due to the variations of the network in actual distributed system, failure detectors without adaptive mechanism cannot meet the requirements of QOS of applications. Adaptive failure detectors should dynamically adjust the detecting quality according to the real-time state of the network. Assuming that delay of the message and loss of the messages is a random probability, the failure detection model based on the predicted message delay is proposed in this paper. A PAC-AFD adaptive failure detection algorithm is realized based on the above model based on the prediction from historical message delay and it contains checking idea. Experimental results show that the algorithm can relieve the effect of delay of the message and loss of the message on the failure detection while ensuring the accuracy and completeness of detection. [ABSTRACT FROM AUTHOR]

In large scale server systems, Quality of Service (QoS) performance cannot be guaranteed if only the objective is to minimize the total power consumptions separately. In this paper a power aware feedback control based power aware job scheduling algorithm (PFM) is proposed to minimize power consumption in computing system and to provide QoS guarantees. Simulations and experiments show that the proposed algorithm can reduce power consumptions significantly while still providing QoS guarantees and the performance degradation is acceptable. The results also show that conventional hardware based per-component and system-wide power management methods can save more power consumptions if they are in assistance with job-level adaptation. [ABSTRACT FROM AUTHOR]

The research on intelligent vehicle mainly includes safety monitoring, intelligent anticollision, aided driving, auto driving, behavior planning decision-making, system structure, and synthetical integration, etc. Sensor and control algorithm are major factors influencing the development of intelligent vehicle. This paper introduces an intelligent vehicle system with chip of FRSSCALE MC9SDG128. By adding fuzzy algorithm into this research on turning control angle of intelligent vehicle, it makes decision according to the lateral error and orientation error. This system, which includes automatic recognition and finished special function, is simple and useful, lower requirement to hardware and has capability of adapting existing structured road environment. [ABSTRACT FROM AUTHOR]

Real-time transmission of video data in wireless network environments is a challenging task, as it have higher bit error rate and packet loss ratio. An Self-Adaptive Temporal and Spatial combined Error Resiliency(SATSER) scheme based on H.264/AVC is proposed in the paper to improve the quality of the rebuilt video pictures in decoder side. Combining the spatial error resiliency method with the temporal one, a new scheme is proposed to conceal an erroneous block according to its encoded mode. The spatial error resiliency scheme exploits the directional interpolation which subdivides the block by the number and the direction of edges going through the erroneous block to recover it. The temporal error resiliency scheme adopts the weighted neighboring Boundary Matching Algorithm to recover the erroneous block. Experiments show that the proposed error resiliency scheme can restrain the diffusion of erroneous block and improve the quality of the rebuilt image. [ABSTRACT FROM AUTHOR]

Security assurance is a critical requirement for QoS or SLA satisfactions in risky grid environments because jobs may be scheduled to multiple machines across different distributed administrative domains. Unlike conventional methods using fixed-number job replications, in this paper, we propose a security-aware parallel and independent job scheduling algorithm based on adaptive job replications to make sure the job scheduling decision secure, reliable and fault tolerant. In risky and failure- prone grids, the replication number is changed according to the current security conditions and the end-user settings. Simulation results show that it is robust due to its adaptive job replications and rescheduling mechanism. The performance results also show that it is better than non-security-aware scheduling algorithms based on fixed-number job replications. [ABSTRACT FROM AUTHOR]

Currently, the demand for multimedia information publication system supporting for multiple terminals is getting more and more emergent. But the capacity of set-top box terminal is limited. In order to improve scalability and reliability, we separated its control logic from physical data transmission. Data transmission is optimized to support a large number of terminals with special algorithms. Time-sharing engine is used to improve performance of system when sources exceed the capacity of STB. Evaluation result shows that the system designed in this paper can provide high-quality information display services for terminal users. [ABSTRACT FROM AUTHOR]

Structural learning is finding the best network that fits the available data and is optimally complex. This can be accomplished by utilizing a search algorithm over the possible network structures. In this paper, a greater importance is given to the search algorithm because we have assumed that the data will be complete. That is, each element of the database is a valid state of a variable. If there are non-applicable entries in the database then the database is said to be incomplete. In the paper, we explore the search algorithms used in this research. In the algorithms, the arcs are added heuristically and exhaustively. [ABSTRACT FROM AUTHOR]

This paper analyzes the status and shortcomings of the existing queue management algorithms of the Ad Hoc networks and puts forward a queue managemen algorithm with priority self-adaptive random early detection (PSRED). The algorithm can not only avoid the network congestion effectively but also resolve the "starving to death" phenomenon of inferior priority queue. Then the algorithm can make router manage queue more accurately. The simulation shows that the PSRED algorithm can reduce queuing delay and the loss probability of the packets effectively, at the same time the jitter of queue length restricted in a relatively slight range. [ABSTRACT FROM AUTHOR]

Digital image histogram is an important global feature, in this paper, histogram will be used to embed binary watermark in the space domain. In the watermark embedding process, firstly, the corresponding histogram of image is generated according to the original images' characteristics, which shows the probability of the corresponding pixel in gray scale; Secondly, the distances between adjacent elements of histogram are calculated, and found their maximum; Then, the histogram is adjusted according to the maximum and pretreated watermark; Finally, the watermark image is generated. In the watermark extracting process, in accordance with the histogram of watermark image and the original image, and the pixels embedded watermark are found, thereby extracting watermark. The experimental results show that the algorithm has good anti-geometric nature of attacks. [ABSTRACT FROM AUTHOR]

This paper describes a new Hough transform (HT) algorithm for multi-objective detection. The core principle of HT-voting principle, which needs to find the maximal voting rate, makes HT impossible to detect many targets from single image synchronously. An improved HT algorithm based on clustering is presented. When the number of targets in the image is known, all of targets can be detected once by this algorithm. The simulation results of iris location with proposed approach reveal that this approach can detect the needed targets in the image once and has great detection effect [ABSTRACT FROM AUTHOR]

A good target tracking system depends on not only the algorithm, but also system hardware design. Speed and accuracy are two important evaluations indicators in target tracking system. This paper uses Harris algorithm to extract little part of target corner features which can improve this system's speed. Then, it uses optic flow method which has high matching accuracy to match those extracted corner features in the following video frames. Bary algorithm is used to compute the bary of those matched feature points as displacement of target. According to property of this algorithm, the author also designs a high performance mini hardware system which can be mounted in aircraft or camera. This hardware design is constructed mainly by DSP as the core processor and FPGA as logical controller. The main part of algorithm is run in high speed DSP and preprocessing algorithm is run in FPGA which has the property of parallel computing. It is proved by experiments that the designed system has not only high speed and tracking accuracy but also higher robustness to target rotation, transformation, shelter etc. [ABSTRACT FROM AUTHOR]

Our goal of this paper is to provide a new method to construct the gyroscope random drift. Based on the State-space subspace system identification algorithm, the gyroscope random drift data are constructed to a series of scale gyroscope drift model to reduce the predict errors. The results of the experiment indicate that the gyroscope drift model constructed by State-space subspace system identification algorithm for the scale data low variance for the traditional method. [ABSTRACT FROM AUTHOR]

Feature selection is an effective method for reducing the size of text feature space. So far, some effective methods for feature selection have been developed. For the purpose of acquiring the optimal number of features, these methods mainly depend on observation or experience. In this paper, by combining the overall with the local distribution of features in categories, a high performance algorithm for feature automation selection (Named FAS) is proposed. By using FAS, the feature set can be obtained automatically. Besides, it can effectively amend the distribution of features by using cloud model theory. Analysis and open experimental results show the selected feature set has fewer features and better classification performance than the existing methods. [ABSTRACT FROM AUTHOR]

To enhance the realistic experience of virtual conference, this paper proposed a distributed virtual conference stereo audio reconstruction model. The spatial audio parameters inter-aural level difference (ILD) is used to reconstruct the spatial sound field for each listener. The distributed synthesis system is designed to get a lower network payload. [ABSTRACT FROM AUTHOR]

Clustering has played a very important role in data mining. In this paper, a fast incremental clustering algorithm is proposed by changing the radius threshold value dynamically. The algorithm restricts the number of the final clusters and reads the original dataset only once. At the same time an inter-cluster dissimilarity measure taking into account the frequency information of the attribute values is introduced. It can be used for the categorical data. The experimental results on the mushroom dataset show that the proposed algorithm is feasible and effective. It can be used for the large-scale data set. [ABSTRACT FROM AUTHOR]

Logging while drilling (LWD) is an advanced drilling technology in oil/gas exploration and development, which was developed to increase drilling efficiency and decrease drilling cost. The real-time logging data transmission by the limited transmission bandwidth is a key technology in LWD. Compressing logging data by some algorithm is an evident method to enhance data transmission. This paper introduced a novel LWD data compression method on the basis of Differential Pulse Code Modulation (DPCM). Our algorithm can optimize compression parameters to improve the compression efficiency and adjust compression parameters for different application. Extensive experiments show that our method can give enough precise results with the compression ratio of 50% at least and our algorithm has the advantages of alterable compressing ratio, excellent decode quality and low algorithm complexity. [ABSTRACT FROM AUTHOR]

Analyzed and summarized the research results on the ship target detection and identification in SAR imagery. The characteristics of the ship target recognition in SAR imagery are listed. Ship target recognition techniques are introduced. And the ship detection and recognition algorithms have been reviewed and analyzed. The main problems on ship detection and recognition process have been described as well. and a possible direction of development are given. It is expected that the researchers will obtain useful information from this paper, which will help for their future works and promote the SAR image target detection and identification of ship development. [ABSTRACT FROM AUTHOR]

Security is an important performance of watermarking system. In this paper, the idea of side information is introduced into the security domain of watermarking system, and a secure fragile digital watermarking algorithm with side information is proposed associating with the idea of cryptography. In the proposed algorithm, the original image is segmented into blocks and the watermark is segmented into subsections, and then the current subsection of watermark is preprocessed by using the former subsection of watermark as side information. The embedding position of each subsection of watermark is also chosen by using the position of the former subsection of watermark as side information. Each part of watermark and embedding position is good associated, so good security of the proposed watermarking system can be obtained. Experimental results show that the proposed algorithm can improve the security of watermarking system effectually. [ABSTRACT FROM AUTHOR]

Web service composition is most impressing method for development and deployment of e-business. Modeling and analysis of the behavior requirements of compositional Web service plays an important role in behavioral verification. Traditional methods are expressing requirements as LTL like logic specifications which are based on activities or as MSC like graph forms which are based scenarios. In this paper, we propose the concept of behavior specification based on activity chain in which its atomic granularity is between activity and scenario. Four behavioral modes such as chain existence mode, chain absence mode, chain precondition mode and chain response mode are adopted to express usual requirement specifications. Encode them on Labeled Transition System LTS and then give them operation semantics. Check compositional Web services based on LTS corresponds with behavior modes or not. Give the sufficient, necessary condition and algorithm for checking. [ABSTRACT FROM AUTHOR]

In Chinese information processing, Naive Bayes is a simple text classification method that is easily implemented. Its core is the realization of the calculating posterior probability algorithm and the effectively reducing dimension for feature words. This paper improved Naive Bayes text classification from the calculating posterior probability and the reducing dimension of feature words of text. The result of experiment indicated that the improved method is of the higher efficiency than the original algorithm. [ABSTRACT FROM AUTHOR]

This paper presents a medical image retrieval method based on the integration of corner and center local binary pattern (CLBP). The CLBP is adopted to describe the texture feature of the medical image firstly. Based on multi-scale curvature product (MSCP), the interest corners are extracted and then a new method is introduced to describe the shape feature. Then, such features are used together as the index for medical image retrieval. Experiment results show that the presented method not only reflects the texture information very well, but also has a rotating shift invariant feature, and it can improve the precision of medical image retrieval and the recall rate. [ABSTRACT FROM AUTHOR]

A new image classification method based on texture feature and improved BP neural network is introduced in the paper. By this method, image matching can be limited to the specific category, which can greatly reduce the retrieval time complexity. Comparisons are given between the improved BP neural network and the genetic BP, the experimental results show that the new method got higher efficiency in texture image classification than the traditional methods. [ABSTRACT FROM AUTHOR]

Impulse response of sparse multi-path channel (SMPC) can be recovered from a short training sequence since most entries of SMPC are zeros. Though the ordinary orthogonal matching pursuit (OMP) algorithm provides a very fast implementation of SMPC estimation, it suffers from the coherence of atoms in dictionary, especially in the case of SMPC with a large delay spread and short training sequence. In this paper, a modified OMP method is proposed and a sensing dictionary is designed adaptively to improve the performance of the OMP algorithm. Numeral experiments illustrate that the proposed algorithm based on adaptive sensing dictionary outperforms the ordinary OMP algorithm. [ABSTRACT FROM AUTHOR]

Parallel computer cluster technology is an important development direction of high-performance parallel computer system. Parallel computer system is the best choice in institution, which has the high frequent operation requirement. This paper gives the ideal scheme based on the analysis for the complex degree of common distributed algorithm, and points out the advantages of distributed algorithm applied on the high-performance parallel computer system. [ABSTRACT FROM AUTHOR]

The partial systematic resampling algorithm (PSR) classifies the particles according to the weight threshold value established in advance before the resampling. Systematic resampling (SR) is carries on the minority particles, so it increases the particle diversity and reduces the computation time. In this paper, it firstly presents PSR algorithm, then, analyzes several kinds of the weight threshold values .At last, the conclusion are drawn by comparison the performance of partial systematic resampling particle filters (PSPF) when take the different threshold values under simulation. [ABSTRACT FROM AUTHOR]

Currently, the implementation of WEB caching is mostly based on traditional cache updating algorithms. However, due to the diversity of the WEB traffic pattern, the traditional algorithm s for cache updating can not be used in WEB environment effectively. Literature (F. Bonchi, 2001) provides a Web updating algorithms based on Decision Trees, in this paper, we provide a web policy, based on Genetically Evolved Decision Trees, the result of experiment indicates this policy improves the updating efficiency of the Web pages. [ABSTRACT FROM AUTHOR]