Table of Content

01 April 2007, Volume 27 Issue 4

Previous Issue    Next Issue

Network and communications

Service-oriented method for grid resources allocation based on sealedbid auction

2007, 27(4):  773-775. 

Asbtract ( )   PDF (774KB) ( )  

Related Articles | Metrics

Considering dynamic, heterogeneous and autonomous characteristics of resources in the computational grid environment and the advantages of economics mechanism were applied to solve the problem of resource allocation, a service-oriented framework of grid service markets based on OGSA, named SBAGRM, was put forward for resources allocation. A private bid auction model was presented. Bayes equilibrium point and strategy, efficiency and utility in the Bayes equilibrium state were discussed too. Douglas utility function-based resources allocation arithmetic was presented at last.

Research and application of Web caching workload characteristics model

2007, 27(4):  776-779. 

Asbtract ( )   PDF (754KB) ( )  

Related Articles | Metrics

A synthetic Web workload generator (WebSimGen) was designed and implemented based on the mathematic model established by Web characteristics. The simulated experiments made use of two-level proxy caching and ADF (Aggregation, Disaggregation and Filtering) model to evaluate the reference characteristics and performance of synthetic Web workload, which showed that the synthetic workload had the similar performance characteristics with the real traces. The generator has the advantage of flexibility and can overcome some disadvantages of the real traces, so it can provide an important basis for further improving Web caching performance and prefetching technology.

P2P complex search based on Ontology

2007, 27(4):  780-783. 

Asbtract ( )   PDF (791KB) ( )  

Related Articles | Metrics

The traditional DHT-P2P system has its limitations. It can only perform single keyword search and cannot understand the meanings of users' requests. P2P complex search based on Ontology was studied. Vector Space Model theory was used to describe documents. And the P2P identifier space was divided to gather the similar documents at adjacent nodes. Therefore, the problem of multi-keywords complex search was solved, and in the mean time, the search process was speeded up. With the help of Ontology, the whole system could understand users' requests better and broaden the searching scope reasonably to avoid missing partial result. Simulated system partially implemented by this theory has accomplished complex search, accelerated the search process, and improved the recall. In addition, it has achieved a better load balance of nodes.

Analysis of Peer-to-Peer network resilience

2007, 27(4):  784-787. 

Asbtract ( )   PDF (957KB) ( )  

Related Articles | Metrics

Study on network resilience is about the characteristics of robustness and vulnerability in random failing of nodes and intended attacking on nodes. The resilience of Gnutella network was analyzed, and the result was compared with that of ER and EBA models. The result shows that Gnutella network exhibits very good fault-tolerance characteristic, but is very fragile under intended attacks. Finally，the network resilience was analyzed theoretically，and a closed-form solution for average size of clusters was reached before the largest cluster of network appeared.

Research of a TCP/IP network management model based on P2P

2007, 27(4):  788-790. 

Asbtract ( )   PDF (594KB) ( )  

Related Articles | Metrics

On the basis of introduction to traditional TCP/IP network management model and peer-to-peer technology, network management model based on P2P named NMMBP was proposed. It can satisfy inter-domain cooperative management, improve the connectivity between management entities, and balance the task load in management network. Thus it made up for the shortage of the traditional technology in inter-domain cooperative management.

Research of Peer-to-Peer network congestion

2007, 27(4):  791-794. 

Asbtract ( )   PDF (996KB) ( )  

Related Articles | Metrics

The wide use of Peer-to-Peer networks causes network congestion. Different from traditional research methods of network congestion, the topology of Peer-to-Peer networks was analyzed to understand network congestion. First, real traffic of Gnutella network was studied, and that the relationship between node's betweenness and the network congestion was found. A closed-form solution for network critical load was got by using queuing theory as the theoretical model. Finally, several schemes for relieving network congestion by augmenting the capacity of nodes with large betweenness and by adding assortative connections between those nodes were also proposed.

Parallel fast Fourier transform algorithm based on binary fat tree network

2007, 27(4):  795-797. 

Asbtract ( )   PDF (548KB) ( )  

Related Articles | Metrics

The binary fat tree is a network topology which is prone to accomplish butterfly computing. According to this feature, a logical model for binary fat tree was constructed at first, and then a parallel Fast Fourier Transform algorithm based on it was developed. In this algorithm, balance of load was achieved in the process, and time complexity was reduced compared with serial algorithm. At last the algorithm was implemented in cluster and MPI, and experimental data was analyzed.

HMIPv6 over MPLS——a mobility management scheme for 3G RAN

2007, 27(4):  798-800. 

Asbtract ( )   PDF (725KB) ( )  

Related Articles | Metrics

With the development of 3G networks, the performance of mobility management and quality of service in the networks are highly expected. For this purpose, a new mobility management scheme for 3G Radio Access Networks, which combines MPLS and HMIPv6 in a loosely coupling way, was presented. The scheme would not only support the fast-handover algorithm in HMIPv6, but also provide the guarantee of QoS when handover happens.

New restricted multicast routing algorithm in triplet-based hierarchical interconnection network

2007, 27(4):  801-804. 

Asbtract ( )   PDF (749KB) ( )  

Related Articles | Metrics

Multicast routing algorithm is extremely important to the communication performance of interconnection network and multiprocessor systems. Taking into account the performance, cost and implementation difficulty, a new tree-based restricted multicast routing algorithm named TRMA was presented for Triplet-based Hierarchical Interconnection Network (THIN). TRMA fully used the hierarchy of network and the topology information included in the nodes codes to decide the message routing, which can make the design of algorithm very simple and hardware implementation easy. Compared with other tree-based multicast routing algorithms, TRMA need not to construct multicast tree in source node before transferring a multicast destination message and can improve the system performance. A simulation has been conducted that compares the performance of TRMA and unicast-based multicast algorithms in THIN. The results indicate that TRMA is more suitable for multicast routing in THIN with lower latency and traffic.

New rate control algorithm

2007, 27(4):  805-807. 

Asbtract ( )   PDF (591KB) ( )  

Related Articles | Metrics

Traditional rate control algorithm is not fit to be used in transcoder in wireless environment. An adaptive rate control scheme based on the theory of stochastic optimal control was proposed. It can balance real-time transmission with continuity of video playing and attempt to acquire the best overall performance. In addition, it can be adapted to the change of bandwidth and make the best of the bandwidth.

Performance analysis and improvement of TCP over OBS

2007, 27(4):  808-810. 

Asbtract ( )   PDF (582KB) ( )  

Related Articles | Metrics

In Optical Burst Switching（OBS） network, the FAP assembly strategy of edge router tends to slow down the sending rate of TCP traffic. As a result, the correlation gain of TCP traffic is reduced, which in turn makes the throughput drops down. There is a strategy to tackle this issue by increasing the ACK assembly time. Simulation results show that the strategy proposed here can significantly improve the performance of TCP over OBS in terms of throughput.

Self-adaptive Lagrange relaxation algorithm for aggregated multicast

2007, 27(4):  811-813. 

Asbtract ( )   PDF (766KB) ( )  

Related Articles | Metrics

Multicast has great advantages in data forwarding. But the number of forwarding states becomes huge in routers when there are a large number of multicast groups in the network, which may cause explosions of state information and control information. Aggregated multicast is a new approach to reduce the number of multicast state. It enables multicast groups to share a single distribution tree so that the tree management overhead at core routers can be reduced. Aggregated Multicast can actually be attributed to minimal set cover problem, which is an NP-complete problem. A self-adaptive Lagrange Relaxation Algorithm that can achieve global optimal solution was used to solve it. Simulation results show that this algorithm is better than the conventional greedy algorithm in that it improves aggregation degree and reduces multicast state number.

Design and implementation of time synchronization algorithm based on frequency adjustment in distributed system

2007, 27(4):  814-817. 

Asbtract ( )   PDF (745KB) ( )  

Related Articles | Metrics

In order to reduce access frequency to the NTP time server, and efficiently relieve the over-load situation of server. A time synchronization algorithm based on frequency adjustment in distributed system was proposed. Under the same accuracy requirement, the experimental results show that the algorithm is more effective than the traditional algorithm based on the method of phase adjusting time.

On-demand synchronization CDMA——a new multiple access protocol for the mobile Ad hoc network

2007, 27(4):  818-820. 

Asbtract ( )   PDF (734KB) ( )  

Related Articles | Metrics

An On-demand Synchronization Code Division Multiple Access (ODSCDMA) protocol for the mobile Ad hoc network was proposed. The protocol established the temporary synchronization between the sending node and receiving node only when they are communicating, which combines the strong points of the synchronization and asynchronization MAC protocol. The slot reservation and power control were integrated with CDMA to improve the network performance. Finally, chaotic sequences and adjustment factor were used to reduce the Multiple Access Interference (MAI) and increase the capacity of system.

A node clustering and information retrieval algorithm on P2P system

2007, 27(4):  821-824. 

Asbtract ( )   PDF (743KB) ( )  

Related Articles | Metrics

An efficient node clustering and information retrieval algorithm named NCSearch was introduced. By using the Hilbert curve's good locality characteristic of preserving ability, NCSearch clustered the nodes that have similar content and form some clusters. The search algorithm can fast locate the cluster which is most relevant to query and then flood within the cluster, and all returned results were sorted by relevant degree. Simulation results show that, compared to the Gnutella algorithm, NCSearch improves search efficiency.

New globally adaptive loadbalanced routing algorithm for k-ary n-cubes

2007, 27(4):  828-831. 

Asbtract ( )   PDF (985KB) ( )  

Related Articles | Metrics

Routing is one of the key issues in the researches of k-ary n-cubes, and globally adaptive load-balanced routing can provide high throughput on various traffic patterns. A new detection-based scheme was proposed to handle routing deadlock, which was efficient and easy to implement. A globally adaptive load-balanced routing algorithm adopting this scheme was also put forward, which has higher routing adaptability than previously published algorithms. Simulation results demonstrate that the throughput of this algorithm exceeds that of the previously published algorithms by 18% on benign traffic pattern, and by 10%~24% on adversarial traffic patterns.

Articles

Application of clustering architecture in intrusion detection for MANET

2007, 27(4):  832-834. 

Asbtract ( )   PDF (580KB) ( )  

Related Articles | Metrics

To optimize the security problem of MANET, a cooperative intrusion detection mechanism was proposed based on clustering architecture. The cooperative and communicable measures between clusters were discussed, and then the algorithm for cluster formation was put forward. The analysis of the example indicates that the clustering architecture is superior to the traditional methods in the accuracy and the cost of intrusion detection.

Cross-domain authentication alliance protocol based on lattice

2007, 27(4):  835-837. 

Asbtract ( )   PDF (745KB) ( )  

Related Articles | Metrics

A lattice-based authentication protocol model for multi-domain was proposed. The model supports mutual entity between different trust domains and unites authentication of large-scale cross-domain, which avoids the potential security hazard, network bottleneck and single-point collapse problems by certain independent prerogative organization authentication in the traditional method. Gigantic communications scale based on vicegerent cross-domain authentication and unadapted modern network environmental problem owing to lack of flexibility were simultaneously solved. The model has higher security by emulation experiment and performance analysis, and it is more flexible, lightweight and efficient in the united authentication of large-scale cross-domain.

SVM algorithm based on sample density and its application in network intrusion detection

2007, 27(4):  838-840. 

Asbtract ( )   PDF (543KB) ( )  

Related Articles | Metrics

When the network dataset is very large, conventional Support Vector Machine (SVM) learning algorithm is remarkably slow. By contrast，the proposed algorithm based on space block and sample density is fast. It was applied in intrusion detection in this paper. The algorithm selects training samples by local sample density, to reduce the training samples and thus to improve the speed of learning. Simulation shows that the algorithm is faster than the techniques of intrusion detection based on conventional SVM while it guarantees the high classification precision.

An improved attack algorithm for discrete logarithm problem

2007, 27(4):  843-845. 

Asbtract ( )   PDF (559KB) ( )  

Related Articles | Metrics

Baby-step-giant-step attack algorithm is universally suitable to solve all discrete logarithm problems, but its relatively larger storage cost is an obvious defect. An improved attack algorithm which can cut down half of space spending and cancel the inversing-computation on multiplicative group was presented. By using hash function, this new algorithm abolishes list-sorting process and drops the time complexity of list-searching into O(1). The performance analysis shows that the new algorithm obviously reduces the time and space use. Furthermore, how to play down the input size to shorten computation journey of attack algorithms was discussed and an easy way to parity sieve for discrete logarithm was given.

Multimedia authentication system in the P2P collaboration environment

2007, 27(4):  846-848. 

Asbtract ( )   PDF (801KB) ( )  

Related Articles | Metrics

In order to meet the security demand for P2P collaboration environment and the particularity of the multimedia data, a solution based on the digital watermarking and digital signature was put forward. This solution utilizes the security feature of JXTA platform and doubles entity authentication based on digital signature and fragile watermarking technique. It may prevent the illegal user from deceiving other users, at the same time also maintain the integrity of resources. Analysis shows that the proposed scheme has high security and trust.

Key agreement scheme based on GDH for virtual dynamic subgroup group

2007, 27(4):  849-851. 

Asbtract ( )   PDF (490KB) ( )  

Related Articles | Metrics

Key agreement is usually mentioned in general static group. Based on virtual dynamic subgroup, Agreed Virtual Dynamic Subgroup Scheme (AVDSG) was proposed, which can manage the virtual dynamic subgroup by using the key information of the group. Compared with GDH, it can largely reduce the average cost of rekeying and the number of keys held by each user.

New S-boxes for hardware implementation of AES

2007, 27(4):  852-853. 

Asbtract ( )   PDF (590KB) ( )  

Related Articles | Metrics

To promote the performance of AES in hardware, a new S box for AES was constructed thorough analyzing the structure of AES S box. Compared to the AES S box, the most important feature of new S box is that it can implement with a small amount of gates and get faster processing, so it is more applicable in resource constrained device. Meanwhile, the characteristics of the S box and the safety of the modified AES were analyzed, which proves the safety to AES with new S box.

Attack traceback method based on collaborative mechanism

2007, 27(4):  854-856. 

Asbtract ( )   PDF (639KB) ( )  

Related Articles | Metrics

In order to coordinate monitoring Agents with each other in tracing the source of network attack by sharing the tracing data, a traceback method based on collaborative mechanism was proposed. In this method an adaptive-collaborative-aggregative tracing protocol was designed which could exchange simple and safe query messages reciprocally between monitoring Agents. By this means, distributed traceback system could achieve collaborative tracing and reconstruct attack path. Experimental results show that this method can make use of the tracing information effectively and improve the tracing precision.

Acquiring testing factors in component software integration based on DFA model

2007, 27(4):  857-859. 

Asbtract ( )   PDF (624KB) ( )  

Related Articles | Metrics

Functions of the interface, event, context dependence relationship and content dependence relationship in componentbased software were analyzed with regard to integration testing. The difficulty in acquiring these four factors was also analyzed. A DFA (Deterministic Finite Automaton) model based on the characteristics of component-based software was proposed so that these four factors could be acquired from the model. An example was given to prove the validity of the model.

Multiple COTS selection process based on extension to FCD

2007, 27(4):  860-862. 

Asbtract ( )   PDF (854KB) ( )  

Related Articles | Metrics

In large software system, components depend on each other, which leads to the difficult evaluation of individual component. The present methods for single COTS selection do not fit multiple COTS selection very well. Then a multiple COTS selection process was proposed based on extension of a proven system decomposition technique named FCD. During the process of decomposition, local requirements and crosscutting requirements were identified. Local requirements were allocated into modules as local evaluation criteria for candidates COTS while crosscutting requirements were considered in global evaluation. The process supports requirements adaptation to capabilities inherent in COTS products and takes into consideration varying granularity of COTS as well. The global selection was described as a nonlinearly constrained optimization problem with the purpose of determining an optimal combination of COTS products with maximal global fitness under certain constraints.

Description language supporting workflow services integration and collaboration: PSXPDL

2007, 27(4):  863-867. 

Asbtract ( )   PDF (970KB) ( )  

Related Articles | Metrics

With the rapid development of service-oriented architecture and Web services technology, supporting integration and collaboration of workflows encapsulated into services become a tendency of workflow language development. After the analysis of the levels and scenes of workflows' collaboration, an extended meta-model based on WfMC's was proposed. According to this extended meta-model, a XPDL-based workflow language named PS-XPDL, which has ability to describe the collaboration and data transportation between workflows accurately, was put forward. Based on the standard protocol of WfMC, PS-XPDL can be easily implemented.

Research of pattern language for Agent-based application

2007, 27(4):  868-871. 

Asbtract ( )   PDF (805KB) ( )  

Related Articles | Metrics

To develop robust, maintainable Multi-agent System (MAS), an Agent Pattern Language (APL) has been proposed in the light of Generative programming and engineering disciplines for Grammarware, in company with a single Agent architecture based on blackboard. With a Business Process Management (BPM) application scenario, how to use the APL language to describe the data type and components like knowledge sources and so on was introduced, as well as a parser for APL parsed APL files and generated executable code for Agent which ran on JADE. In this way, systems based on Agent can be developed fast and MAS can be constructed in a top-down manner. Then not only the productivity for the development agent-base system is improved, but flexibility and reusability of the system are also enhanced.

Technology of recovering function return type in binary translation

2007, 27(4):  872-874. 

Asbtract ( )   PDF (804KB) ( )  

Related Articles | Metrics

Function return type recovery includes two types,one is library function return type; the other is user function return type. By analyzing the characteristics of call and return for function in IA-64 architecture, we know that the work could be done not only in the caller's position, but also in the callee's position. A new method used to recover function return type was introduced combining with the relative merits of them. The technology of recover function return type in ITA binary translator, which is based on the control-flow and data-flow, was introduced, and the corresponding algorithm was also proposed. The instance verification under IA-64 architecture proves that the method is a practical way to recover function return type.

Development for a generic In-Circuit Emulator based on SOPC IPcore

2007, 27(4):  875-876. 

Asbtract ( )   PDF (583KB) ( )  

Related Articles | Metrics

A generic In-Circuit Emulator (ICE) architecture based on SOPC was put forward. New fetures can be downloaded more easily by using IPcore. The ICE of different embedded CPU can be completed in a piece of FPGA. According to the architecture, a generic ICE has been realized with the developing kits from ALTERA company, an IPcore independently designed for ICE, and some commercial IPcores were used too. The design and implementation were introduced.

Database Technology

Mining method of credit evaluation model based on gene expression programming

2007, 27(4):  877-880. 

Asbtract ( )   PDF (757KB) ( )  

Related Articles | Metrics

A new mining method of credit evaluation model based on Gene Expression Programming (GEP), named GEP-CREDIT, was presented. The GEP-CREDIT model can extract relative attributions and fuse them automatically. Based on these attributions, GEP-CREDIT model can train and build the credit evaluation model. Experiments and performance analysis on Germany credit database were given. The results show that, compared with Naive Bayes algorithm, SVM algorithm, and KNN(K=17) algorithm, the prediction precision is increased by 3%, 1.6% and 6.83% respectively by using GEP-CREDIT model.

Data stream clustering algorithm based on probability density

2007, 27(4):  881-883. 

Asbtract ( )   PDF (587KB) ( )  

Related Articles | Metrics

Data stream is characterized by infinite data and quick stream speed, so traditional clustering algorithm cannot be applied to data stream clustering directly. In view of above questions, a probability-density-based data stream clustering algorithm was proposed. It requires only newly arrived data, not the entire historical data, to be saved in memory. It applies EM algorithm on the newly arrived data and updates probability-density function by incremental Gaussian mixture model. Experimental results show that the algorithm is very effective to solve data stream clustering.

Rules extraction method of decision tree based on new conditional entropy

2007, 27(4):  884-887. 

Asbtract ( )   PDF (755KB) ( )  

Related Articles | Metrics

The disadvantages of the current information entropy for estimating decision ability were analyzed deeply. To eliminate the limitations, a new conditional entropy was defined. The attribute selection metric of traditional heuristic algorithm was modified, so the new improved significance of an attribute was proposed. Finally, a heuristic algorithm for rules extraction of decision tree was designed. This reduction method does not need attribute reduction before extracting decision rules, and its computation is direct and efficient, and its time complexity is less than the others. The experiment and comparison show that the algorithm provides more precise and simple decision rules.

Classification algorithm for self-learning Naive Bayes based on conditional information entropy

2007, 27(4):  888-891. 

Asbtract ( )   PDF (773KB) ( )  

Related Articles | Metrics

Naive Bayes algorithm is an effective simple classification algorithm. But two central assumptions made by the Naive Bayes approach are that the attributes are independent within each class and the importance of the attributes is equal, which can harm the classification process to some extent. It is a very difficult problem in machine learning to carry out self-learning knowledge according to the characteristic of source data without prior domain knowledge. Based on the theory of rough set, a new Naive Bayes method named Conditional Information Entropy-based Algorithm for Selflearning Naive Bayes (CIEBASLNB) was proposed, which combined the merits of selective Naive Bayes (SNB) and Weighted Naive Bayes (WNB). Simulation results on a variety of UCI data sets illustrate the efficiency of this method.

Measurement of association rules correlation

2007, 27(4):  892-893. 

Asbtract ( )   PDF (501KB) ( )  

Related Articles | Metrics

Association rules generated by Apriori algorithm includes some useless and even misleading rules.To gain more effective rules, a statistical criteria,Chi-Squared was used to measure the associations, furthermore the quantitative relations between the Chi-Squred test and correlation coefficient was found out, and two measures was united to generate association rules with correlation coefficient.

Intelligence perception

Modified nearest feature classifiers and their application in radar target recognition

2007, 27(4):  894-896. 

Asbtract ( )   PDF (565KB) ( )  

Related Articles | Metrics

The classifiers of Nearest Feature Line (NFL) and Nearest Feature Plane (NFP) share the same drawback in terms of the computation complexity under large data sample size and high dimensionality. Therefore, a new search strategy based on locally nearest neighborhood rule was proposed to modify the two classifiers. Compared to the traditional NFL and NFP, the modified ones can not only improve the real-time performance significantly, but also achieve competitive recognition rate. Experimental results on three measured airplanes data have confirmed the effectiveness of the proposed methods.

Human gait recognition based on bilinear modeling and HMM

2007, 27(4):  897-900. 

Asbtract ( )   PDF (1102KB) ( )  

Related Articles | Metrics

Motivated by bi-factor-invariant human gait recognition problem, a new gait recognition algorithm was proposed in which two factors, generically called "style" and "content" were analyzed and manipulated. First, image sequences were clustered into a fixed number of content with fixed dynamics HMM-EM algorithm. Then the observation data were generated according to an asymmetric bilinear model. After that, SVD and NN were used to classify new sequences characterized by a different style label. Body width between vertical line through centroid and outer contour was used as the feature. Test on the CASIA datasets shows the proposed method's advantage in increasing the recognition rate and adapting to new styles or content. Some other facts affecting ID identification were also discussed.

Low overhead multipath routing protocol for wireless sensor networks

2007, 27(4):  901-904. 

Asbtract ( )   PDF (836KB) ( )  

Related Articles | Metrics

A low overhead routing protocol for many-to-one data collection in wireless sensor networks named LOMP was proposed. The real-time link quality estimation is combined with periodical link quality estimation to provide reliable routing selection in face of the poor link quality in sensor networks. Instead of providing end-to-end node-disjoint multipath routing, LOMP only ensures data forwarding for hop-by-hop multipath. The simulation results indicate that LOMP can greatly reduce routing overhead and improve routing reliability at the same time.

Dynamic routing algorithm based on application rules and probability in wireless sensor networks

2007, 27(4):  905-908. 

Asbtract ( )   PDF (992KB) ( )  

Related Articles | Metrics

A dynamic routing algorithm based on application rules and probability in Wireless Sensor Networks (WSN) was proposed. Making use of nodes neighbors, the algorithm periodically interacted with application rules to construct a breath-first data-centric tree after two stages, and then routing path was posed. On the first stage, the data-centric tree was initiated after computing the node's probability to act as a routing one. Second, some appropriate nodes were added to the tree to be a complete connectivity one. Simulation results show its superiority to TinyOS Beaconing algorithm.

Three-dimensional sensor networks: spatial structure and coverage property

2007, 27(4):  909-912. 

Asbtract ( )   PDF (808KB) ( )  

Related Articles | Metrics

The spatial structure of three-dimensional（3D）sensor networks is much more complicated than that of two-dimensional（2D）sensor networks. The random structure assumption which is adopted by most of the current researchers cannot satisfy all demands of 3D sensor networks in real life. Then, three spatial symmetric cubic lattice structures were proposed. The coverage properties of them were analyzed based on the spatial tessellations theory. In addition, since previous works on random structured wireless networks used sophisticated probabilistic tools to derive their bounds, a simpler analysis technique named bin-covering was applied to get tight threshold for coverage properties of random structured 3D sensor networks.

Wavelet based independent component analysis for palmprint identification

2007, 27(4):  913-915. 

Asbtract ( )   PDF (653KB) ( )  

Related Articles | Metrics

Independent Component Analysis (ICA) has particular advantages in image processing. Used for palmprint feature extraction, it can ensure the components to be irrelevant and statistical independent among each other, and can more roundly describe the essential characteristics in palmprint features. In order to decrease the calculation of complication, the wavelet based independent component analysis for palmprint identification method was proposed. First, the palmprint images were decomposed into wavelet subimages, in which the profile and the minutia information were reserved, and the dimensions were decreased at the same time. Then take the wavelet subimages as the input of ICA, by using FastICA algorithm, the experimental results show that the integrated method is more efficient than ICA algorithm in accuracy and speed.

Tracking of moving object based on genetic algorithm and Kalman filter

2007, 27(4):  916-918. 

Asbtract ( )   PDF (833KB) ( )  

Related Articles | Metrics

A real-time tracking of moving object method based on Genetic Algorithm (GA) and Kalman filter was proposed. The possible position of moving target center in the next frame image was predicted by Kalman filter, and a search region of target was generated around the center position. Genetic Algorithms (GA) was utilized to search the best position in the region. The GA fitness function was Bhattacharyya coefficient between the gray features of the target template and the candidate area's, and the parameter code was the coordinates of candidate center position. Finally, the best position was used as an observed value for next prediction. The experimental results show that the method is effective and robust.

Robust speaker verification in low bit rate channels

2007, 27(4):  919-921. 

Asbtract ( )   PDF (584KB) ( )  

Related Articles | Metrics

The influence of speech coding on text-independent speaker verification was studied. Two methods were proposed to improve the robustness of systems under the matched condition and mismatched one, respectively. The weighted LPCC feature based on the distortion of speech coding was explored to reduce the influence under the matched condition. Under the mismatched condition, the solution was to construct multi-models in multi-channels. In this method, a GMM speech coder detector was used to determine the pattern of the testing speech, and then the corresponding speaker verification model was chosen. Experiments show that the two methods can improve the performance of speaker verification in low bit channels.

Gait Recognition by Fitting Moment Invariants

2007, 27(4):  922-924. 

Asbtract ( )   PDF (965KB) ( )  

Related Articles | Metrics

According to the idea of combining static components and dynamic components from the walking way, as the moment invariants may represent geometrical traits in images, they were extracted as gait features from a subsequent gait series. The moment invariants of human silhouettes were represented by the Fourier series. A genetic algorithm was deployed to obtain the Fourier coefficients. The magnitudes of the coefficients were classified through the kNN classifier. The recognition results with the proposed scheme in the CMU gait database have recognition rates of more than 90% and show that it has achieved a high performance and effectively made use of two kinds of contents.

Dealing with faces illumination variation based on dynamic directional gradient vector flow Snake and gamma algorithm

2007, 27(4):  925-928. 

Asbtract ( )   PDF (1029KB) ( )  

Related Articles | Metrics

A new Dynamic Directional Gradient Vector Flow Snake (DDGVF Snake) algorithm was presented which can detect the image shadow and highlight contours of the unknown face images. According to the given standard face database, the image gray median could be computed. Then, the weight nonlinear Gamma gray correct algorithm was proposed to normalize the illumination variation in the unknown face image. Experimental results on the Yale B, CMU-PIE and CAS-PEAL face databases show that the proposed approach is more precise in recovering the unknown face images than other methods.

New preprocessing method for fingerprint images

2007, 27(4):  929-931. 

Asbtract ( )   PDF (829KB) ( )  

Related Articles | Metrics

A new full fingerprint preprocessing algorithm, which improves traditional fingerprint images segment algorithm by using residual block disposal and smoothing, was introduced, and an orientation-match method that filters orientation field to enhance images was proposed, which removes short line and spur after thinning. The performance of the algorithm has been evaluated by conducting experiments on fingerprint images in FVC2004 database. Experimental results show that the algorithm improves the quality of fingerprint image and makes it convenient for minutiae extraction.

Artificial intelligence

Optimum route sequence search in Petri net based on ant colony algorithm

2007, 27(4):  932-935. 

Asbtract ( )   PDF (769KB) ( )  

Related Articles | Metrics

A Memory Extended Timed Petri Net (METPN) whose elements can record a little information was proposed based on the ant colony optimization algorithm and the features of the Timed Petri Net (TPN). When METPN was running, enough tokens walked and left odor in METPN so that route selections of tokens could be adjusted, in this way it made lots of ant walk routes to approach the transitional sequences with less delay. At last a clear ant walk route could be found on the transitional sequence with the least delay, and the route search problem of complex TPN was solved to certain extent. The result of the simulation shows that the ant walk route is formed along the least delay route effectively by tokens, and the shortest route from initial places to every place of METPN can be gotten.

Ant colony and particle swarm optimization algorithm-based solution to multi-objective flexible job-shop scheduling problems

2007, 27(4):  936-938. 

Asbtract ( )   PDF (746KB) ( )  

Related Articles | Metrics

A hybrid of ant colony and particle swarm optimization algorithms was proposed to solve the multi-objective flexible jobshop scheduling problem based on the analysis of objectives and their relationship. The hybrid was formulated in a form of hierarchical structure. The ant colony algorithm was performed at the master level to minimize the total load and bottleneck load through selecting job-processing route, while the particle swarm optimization algorithm was carried out at the slave level to minimize the makespan through scheduling the operations with machines without violating the result from the master level. The transfer probabilities of ant between machines were designed by using heuristic information of processing time and machine load. The decoding method of particle vector was well designed in order to sequence operations of every machine based on the size relations of element priority values. The simulation and results from comparison with other algorithms demonstrate the effectiveness of the proposed algorithm.

Structure risk minimization based weighted partial least-squared method

2007, 27(4):  939-941. 

Asbtract ( )   PDF (511KB) ( )  

Related Articles | Metrics

Weighted Partial Least-Squared (WPLS) method was proposed to achieve Structure Risk Minimization (SRM) in the Partial Least-Squares (PLS) modeling process. At first, WPLS abstracted the principal components of training samples, and then it trained the weight of samples by means of Support Vector Machine (SVM) algorithm, and finally computed the regression model in the original universe discourse. WPLS not only takes the advantage of PLS to extract most explanatory variables, but also improves generalization property through the weight of samples and SRM is achieved with interpretable model. Simulation results show the effectiveness of the proposed method.

Multi-agent negotiation model with emphasis on privacy

2007, 27(4):  942-944. 

Asbtract ( )   PDF (553KB) ( )  

Related Articles | Metrics

A market-mechanism-based multi-agent negotiation model with emphasis on agents' privacy was put forward. Agent just sends a symbol of its marginal utility and needs not to send its demand of raw materials. The market agent summarizes the individual information by running pricing algorithm and allocation algorithm, and returns the summarized result to agents. The simulation results prove that the negotiation results is similar to present methods, but negotiation process is more confidential, thus satisfies the requirement of distributed autonomous agents.

Improved algorithm for ART2 neural network

2007, 27(4):  945-947. 

Asbtract ( )   PDF (532KB) ( )  

Related Articles | Metrics

Due to normalization in pretreatment stage, some important but small amplitude component will be removed as noise in traditional ART2 neural network. It will lose important information in clustering process. Lack of pattern drifting also exists and the causes for this lack were analyzed. The traditional ART2 neural network was modified based on this idea: removing unit and different distances between samples to the cluster center have different influences on excursion of the cluster center. It verifies that the improved ART2 neural network has reformed pattern drifting greatly according to the result of test on gradual changing data. The improved ART2 neural network has definite value in data classification of nuclear radiation field.

Research and application of genetic expression programming algorithm based on uniform-design

2007, 27(4):  948-951. 

Asbtract ( )   PDF (695KB) ( )  

Related Articles | Metrics

A new evolutionary algorithm named UniformDesign Genetic Expression Programming (UGEP) based on uniformdesign was proposed, which improved the way of generating initial population and also improved hybrid crossover operators. Theoretic analysis proves that UGEP converges to the global optimum and its convergence speed is faster than the classic GEP. The simulation results from function fitting and time series prediction indicate that UGEP performs better than other similar algorithms in each of experimental parameters.

Application of the solution schema in ant colony algorithm

2007, 27(4):  952-955. 

Asbtract ( )   PDF (807KB) ( )  

Related Articles | Metrics

Based on the solution space analysis of the Quadratic Assignment Problem (QAP), we tried to apply the solution schema to ant system algorithm, and made an experiment on the four classes problem of QAP. The experimental results demonstrate that the solution affects the forth class instances of QAP positively, but affects the second class negatively.

New hybrid evolution algorithm based on multi-layer chromosomes gene expression programming

2007, 27(4):  956-959. 

Asbtract ( )   PDF (720KB) ( )  

Related Articles | Metrics

A new hybrid evolution algorithm named MGEP-GA was proposed. It is based on the new concept of the multi-layer chromosomes in gene expression programming (GEP) and uses two-level evolution method, whose main idea is to embed Genetic Algorithm (GA) in GEP where GEP is employed to optimize the structure of a model, and GA is employed to optimize the parameters of a model. The results of experiments show that the algorithm has better search ability and higher stability than the traditional single gene and multi-genes GEP and GP-GA method.

Open genetic algorithms based on NICHE

2007, 27(4):  960-962. 

Asbtract ( )   PDF (666KB) ( )  

Related Articles | Metrics

Taking biosphere and adaptive mathematic models into account, a new Open Genetic Algorithm (OGA) based on NICHE was proposed, which overcome the defects of current genetic algorithm in solving constrained optimization problems. The convergence of global optimal solution of OGA was verified. Firstly, OGA does not need to confirm penalty coefficient, so it is strongly adaptable; secondly, OGA almost does not request the problems and the constraint, so it is easy to apply, which shows the advantage of GA; finally, in order to make good use of the population search characteristics of GA, ensure the diversity of population, and accelerate the search speed, OGA adopted the interaction mechanism between two NICHEs. Experiments show the algorithm is effective.

Improved transformation based part of speech tagging of Latin Mongolian

2007, 27(4):  963-965. 

Asbtract ( )   PDF (547KB) ( )  

Related Articles | Metrics

To solve the problem of rule learning time cost for traditional transformation based part of speech tagging method of Latin Mongolian, a dynamic partition algorithm was presented. It used rule conflict and rule dependency to dynamically partition the training corpus, reduced the searching space and increased the rule learning speed. In an open test of a Latin Mongolian corpus with 10000 sentences, the time that new algorithm cost was only 32% of the old one.

Graphics and image processing

Image threshold denoising algorithm based on monoscale ridgelet transform

2007, 27(4):  966-969. 

Asbtract ( )   PDF (1055KB) ( )  

Related Articles | Metrics

The statistical property of Gaussian White noise in the Monoscale Ridgelet domain was analyzed, and a new image threshold denoising algorithm based on monoscale ridgelet transform was proposed. The experimental results demonstrate that the new denoising algorithm substantially surpasses those based on wavelet transforms, both visually and in terms of PSNR.

Edge detection of infrared image based on multi-scale morphology

2007, 27(4):  970-971. 

Asbtract ( )   PDF (873KB) ( )  

Related Articles | Metrics

An edge detection algorithm based on multi-scale morphology was presented. Firstly, image edge was detected by using several typical structure elements. Then, edges in different scale were combined according to the life of local feature, which were detected by changing the size of structure elements. Compared with several conventional edge detectors, the proposed algorithm has better noise immunity and performance on edge detection of infrared image with large noise and blurry edge.

New multi-resolution SPIHT approach with simple error detection

2007, 27(4):  972-975. 

Asbtract ( )   PDF (1009KB) ( )  

Related Articles | Metrics

SPIHT algorithm, evolving from EZW, is so far one of the most influential wavelet compression algorithms, which encodes wavelet coefficients according to their energy or significance through a unique pass. Due to the introduction of wavelet tree and the certain pass, high compression ratio was obtained as well as high quality of image or video. Multi-resolution SPIHT, being later proposed, promised resolution option based on channel condition. However, accurate path information is so demanded in SPIHT that any error of path bit will bring decoding failure of following bits. Various mechanisms were employed to protect the path information, which turn out of little help to fundamentally enhance its capability of antinoise. The algorithm proposed in the paper provides the decoder with high PSNR and error detection without additional bits by improving the original algorithm.

Image registration based on piecewise linear group parameters

2007, 27(4):  976-978. 

Asbtract ( )   PDF (923KB) ( )  

Related Articles | Metrics

An improved image registration method based on pseudoperspective model and comparametric function was presented. Firstly, computation of corner match was performed to improve the matching stability and operation speed. Secondly, an approximation of the cameras nonlinear response function was accomplished with a constrained piecewise linear one, and the joint registration model was well established. Then the more precise initial parameters could be obtained with a computationally attractive least squares method. Lastly, optimal iteration algorithm was performed to reduce the registration error. The experimental results show that the method is doing well to solve the problem of exposure difference between images and improves the speed of iteration meanwhile.

Ellipse extracting from weak edges based on GVF model and the least square method

2007, 27(4):  979-981. 

Asbtract ( )   PDF (1011KB) ( )  

Related Articles | Metrics

A new fast approach which combines improved GVF model with least square method was proposed to extract ellipse by overcoming the influence of noise and weak edges. At first, the image was preprocessed through median filtering and gradient enhancing of weak boundary, and then the priori knowledge of target region intensity was employed to gain Gaussian fuzzy closeness, which combined with pressure force and formed fussy pressure force. Under the effect of GVF field and the new force, inner control points were pushed toward boundary rapidly, points on the boundary almost bore no forces and outer points bore strong repulsive forces which pushed the points back to the boundary. After iterating several times, the control points were fitted as ellipse by least square method, which was sampled to be new initial boundary of the improved GVF model. The shapeconstrained curve deformable method was applied to footprint analysis system and extracted ellipse from blur, noise and weak boundary image precisely, while it was difficult for conventional edge detection methods. Simulation results show that the method is rapid and satisfying.

Method of thresholding using two-dimensional exponent entropy and its fast algorithm

2007, 27(4):  982-985. 

Asbtract ( )   PDF (1007KB) ( )  

Related Articles | Metrics

The two-dimensional Shannon's information entropy is a classical and commonly used image segmentation method, but there still are some disadvantages involved. Then, a two-dimensional exponent information entropy method for threshold selection was proposed here, which could overcome the disadvantages in the Shannon's entropy. The logarithm entropy was changed into the exponent entropy to get a higher speed. Meanwhile, a fast algorithm of two-dimensional exponent entropy thresholding method was also given, which changed the two-dimensional threshold into one-dimensional by using new area partition method, and the speed of the algorithm was no less than 1% of the original algorithm. The results of the experiment indicate that the proposed algorithm has high speed of calculation and good segmentation quality.

Ringing effect postprocessing and quality assessment method of blind restoration image

2007, 27(4):  986-988. 

Asbtract ( )   PDF (814KB) ( )  

Related Articles | Metrics

Because of the shortage of prior information in the blind restoration processing, the result of restoration is usually not satisfactory. Ringing effect is one of the important factors that influence restoration image quality. It is difficult to do the postprocessing and the quality assessment of restoration image because of the ringing effect. Taking the characteristics of blind restoration into account, a ringing effect postprocessing method was proposed. And according to the thought of assessing ringing effect and definition respectively, a nonreference assessment project of restoration image was put forward. Experimental results prove that the ringing postprocessing method is simple and valid, and the assessment project of restoration image reflects human vision characteristics nicely.

Method for displaying high dynamic range images based on local adaptive

2007, 27(4):  989-990. 

Asbtract ( )   PDF (930KB) ( )  

Related Articles | Metrics

The visual system is able to perceive details in an environment with a high dynamic range by means of local adaptation. A region-based method for calculating local adaptation luminance was proposed to imitate the local adaptation mechanism. It segmented image with region growing. Then local adaptation luminance was computed via region-based bilateral filter. It outputted low dynamic range images in conjunction with tone mapping operator. The resulting images preserve details without halo artifacts.

Typical applications

Algorithm of closed circle DNA computing model for time-table problem

2007, 27(4):  991-993. 

Asbtract ( )   PDF (585KB) ( )  

Related Articles | Metrics

Time-table problem is a NP-complete problem. Algorithm of DNA computing for time-table problem was obtained with introducing manifold biochemical experiments on closed circle DNA computing model. The algorithm produced initialization pool of data using two DNA encoding modes. It solved the problem of the conflict between teacher and grade and the problem of incorporate-class class by introducing group deleting experiment, solved the problem of normal incorporate-class class and problem of teacher's request of time by introducing group separate experiment, solved the problem of how to balance the scheme of time-table by introducing electrophoresis experiment, and obtained a set of whole-optimization solution by introducing mark experiment. And the process of algorithm's realization was put forward. Finally, validity of the algorithm was proved, and complexity of the algorithm was discussed.

Realization of two-intersections traffic control using improved genetic algorithm

2007, 27(4):  994-996. 

Asbtract ( )   PDF (510KB) ( )  

Related Articles | Metrics

With the rapid development of automobile and traffic management, the heavy traffic problem becomes more and more serious, and the solution by increasing new road and constraining the vehicle is not the most economical and effective measures. How to solve traffic problems in existing condition is worth considering. As the basic unit of the whole traffic network, studding on the traffic control method of isolated intersection is the foundation. But it could not reflect the infection of another intersection. That is to say the coupling relationship of intersections could not be reflected, so the method of two-intersection traffic control was proposed. The optimal model of adjacent intersection was built, and the objective function and its constraints were obtained. It was optimized by using improved genetic algorithm. The simulated software of twointersection traffic control was developed, and the control effect was satisfying.

Improved algorithm for finding the sample points of algebraic decomposition on plane

2007, 27(4):  997-999. 

Asbtract ( )   PDF (487KB) ( )  

Related Articles | Metrics

Based on the critical point algorithm and zero-match algorithm, an improved algorithm for finding sample points of algebraic decomposition was proposed. The proposed algorithm aims at reducing the redundant sample points. In the whole computing process, it utilizes an interval with rational endpoints to represent the exact algebraic number and avoid floating-point computation. Furthermore, compared with the existing algorithm by some examples, the effectiveness of the proposed method is verified.

New method of analyzing singularities in noise based on singularity spectrum

2007, 27(4):  1000-1002. 

Asbtract ( )   PDF (737KB) ( )  

Related Articles | Metrics

Singularity analysis is one of the most promising new approaches for extracting hidden information from noise. Singularity spectrum is a statistical method to show global singularities distributions in noise. We introduced an algorithm based on wavelet transform modulus maxima to get singularity spectrum. The experiment on emulation noises shows that the new algorithm is accurate and feasible. Also we found out that singularity spectrum had some advantage over the traditional analysis method: power spectrum.

Distributed temperature-pressure measurement and control system based on expert PID control theory and COM technology

2007, 27(4):  1003-1005. 

Asbtract ( )   PDF (615KB) ( )  

Related Articles | Metrics

Taking account of the necessary conditions of the temperature and pressure experiments, a distributed temperature-pressure measurement and control system was designed and realized, which could be used to control the temperature or pressure of environmental simulation cabin, of which the software was implemented by using COM technology. To solve the contradiction between system overshoot and speed of response, and get better performance, an expert control method was introduced into this temperature-control system based on the classical PID control theory. Comparing the simulated result with the experimental data, the expert control method has been proved with the advantages of no-overshoot, fast response, high precision and great efficiency. It is significant to the same system.

Research into the algorithm of supply chain partner selection and optimization based on the work breakdown structures

2007, 27(4):  1006-1008. 

Asbtract ( )   PDF (668KB) ( )  

Related Articles | Metrics

A multi-objective model of decision-making for partner selection and optimization in supply chain based on work breakdown structures was proposed, which took the factors of due date, quality, cost and service as evaluation indicators. The weight of each factor under the main task and the correlation of each sub-task were considered, and quantitative calculation of the time spent on task based on the model was made. An example with a selfadaptive genetic algorithm solution was included to demonstrate the validity of the model and algorithm.

Prediction research of the supper-shortterm exchange rate

2007, 27(4):  1009-1012. 

Asbtract ( )   PDF (809KB) ( )  

Related Articles | Metrics

A new method which is fit to the prediction of supper-shortterm exchange rate was proposed. The data for experimentation was got from Internet, and the model was established with the reconstructed phase space and Kalman. Compared with the neural network predicting model, the results show that the proposed model can track the instant exchange rates change well, and it is better than the neural network model both in accuracy rate and run-time length. Finally, the whole procedure of prediction on line based on .NET was introduced.

Realistic rapid 3D face modeling for film/TV animation

2007, 27(4):  1013-1016. 

Asbtract ( )   PDF (1428KB) ( )  

Related Articles | Metrics

Using LS_5000 3D laser scanner, an efficient realistic 3D face modeling algorithm was proposed to create virtual actor for film and TV animation. Inputting the point cloud data and uncalibrated photographs of the real actors face as well as little user interaction, the algorithm can generate realistic 3D face models including geometry model, texture model and deformable model for expression animation. Experiments show that the output 3D face is ideal to be directly used for film/TV animation and can enhance the speed of facial modeling.

Rapid acquisition of long PN code with large Doppler shift in low SNR environment and its DSP platform

2007, 27(4):  1017-1019. 

Asbtract ( )   PDF (820KB) ( )  

Related Articles | Metrics

According to the Long PN Code with large Doppler shift in Low SNR Environment, an improved method of carrier-frequency and code startpoint planar search is proposed. Through cycle shift in frequency space，the transformation between several frequency channels is implemented. Using the code search based on multi-phase structure, the calculation complexity was reduced. The platform device is set up with TMS320C6416, and the simulation result demonstrates that this method has better acquisition performance.

Application of image preprocessing in auto-classification of scanned hospital documents

2007, 27(4):  1020-1022. 

Asbtract ( )   PDF (700KB) ( )  

Related Articles | Metrics

In order to make recognition and classification more effective, the scanned hospital documents were preprocessed based on their characteristics. At first, the image was binarized by using Ostu method. Then the skewness of the image was dectectd by Radon transform. Finally, mathematical morphology open operator was used to reduce the disturbances and the projecting method was used to segment the header of the image.

Architecture design of NMS platform based on TMN

2007, 27(4):  1023-1026. 

Asbtract ( )   PDF (886KB) ( )  

Related Articles | Metrics

To develop the network management system in telecom network, a common NMS platform framework based on the TMN architecture was described. The system architecture of this platform was introduced, and the design of some core application frameworks such as information model and notification service was presented. A common process to develop an EMS system based on this platform was introduced too.

Study for setting up the decision-support system based on KW,MW and DW

2007, 27(4):  1027-1029. 

Asbtract ( )   PDF (861KB) ( )  

Related Articles | Metrics

Spatial Decision Support System （SDSS） is an advanced Decision-Support System (DSS) at present. But the models in the model database are unchangeable, so it is certainly limited in practice. The structure of the system based on data warehouse, knowledge warehouse and model warehouse was put forward, which can effectively solve the problem. The structure simplifies the model obtainment and the coding task, and strengthens the effectuation and intelligence of decision so as to improve the efficiency of decision making.

A DRM broadcasting system design based on PKI

2007, 27(4):  1030-1032. 

Asbtract ( )   PDF (593KB) ( )  

Related Articles | Metrics

Considering the realtime and singletrack characteristics of the broadcasting system, a new DRM broadcasting system based on Public Key Infrastructure (PKI) was designed, which integrated broadcasting network, traditional Internet and telecommunication network. The corresponding security analysis of it was also given.