15 November 2014, Volume 40 Issue 11
-
Select all|
-
Computer Engineering. 2014, 40(11): 1-5. https://doi.org/10.3969/j.issn.1000-3428.2014.11.001
A solution of embedded endoscope video processing system based on TILE-Gx multicore processor is designed. The system supports 2-channel input and output port for 1080p60 HD video data and real-time H. 264encoding /decoding of YCbCr422 format video. FPGA is applied to realize the input / output port of video data. 4 pieces of TILE-Gxprocessors are applied to encode / decode,another TILE-Gx processor is applied to the system management,video mixing, and hard disk storage. Experimental results show that,the performance of encoding and decoding fulfills the requirement of high definition and real-time of endoscope,and the image quality reaches H. 264 High Profile level.
-
Computer Engineering. 2014, 40(11): 6-13. https://doi.org/10.3969/j.issn.1000-3428.2014.11.002
As it is common to integrate more technical image processing algorithms within remote sensing monitoring analysis system,but these systems with the problem of lack of modeling particular domain service.Moreover,these systems short on reusability and extendibility of algorithms and models. Based on the above reasons,the remote sensing monitoring analysis system is faced to multi difficulties when it is promoted. In order to address these problems,this paper proposes an integrated framework of satellite remote sensing monitoring analysis based-on Managed Extensibility Framework(MEF) that dynamically and flexibly integrates algorithms,models,and applications on the framework according to the requirements. This framework can realize coherence of processing images and extremely shortened period of processes,and improves flexibility and scalability. Moreover,this framework can be extended to other related domains because of scalability.
-
Computer Engineering. 2014, 40(11): 14-17. https://doi.org/10.3969/j.issn.1000-3428.2014.11.003
To solve the problem that a single feature leads to tracking failure easily in a complex environment,a Particle Filtering(PF) tracking algorithm based on multi-feature fusion and Mean Shift(MS) is proposed. Under the framework of PF,it is closer to the real posterior distribution by embedding MS algorithm and using color and structural as the observation model to represent the object,and the weights of particles are calculated by this integration,in order to reduce the tracking deviation. Experimental results show that the proposed algorithm has better robustness when using the same particles,and the average weight of the particle is improved and the resample times are reduced significantly,even using the less particles can achieve tracking stability.
-
Computer Engineering. 2014, 40(11): 18-25,30. https://doi.org/10.3969/j.issn.1000-3428.2014.11.004
In order to make the encryption system have the optimization performance,and solve these problems such as not achieving the global optimization and low speed of convergence,the cipher text optimization system based on the Tree Parity Interactive Learning Machine(TPILM) and discrete evolution algorithm is proposed in this paper. It defines the weight update mechanism,and couples the chaotic mappings to construct the TPILM and its mutual interference model. It introduces the cutting roulette selection mechanism into the uniform crossover operator. Meanwhile,it takes the adjacent pixels correlation coefficient and the cipher text information entropy of image block,introduces the weight theory to design the fitness function to propose a novel global discrete evolutionary algorithm for firstly applying to image encryption. At last,it produces the encryption structure of “initial optimization-cipher optimization”. Experimental results show that,compared with other encryption systems,the encryption system in this paper has the best quality and the function of global fast optimization to optimize all the iterative outcomes to make the cipher have the maximum information entropy and the lowest correlation coefficient.
-
Computer Engineering. 2014, 40(11): 26-30. https://doi.org/10.3969/j.issn.1000-3428.2014.11.005
This paper gives a detailed analysis of the performance anomaly when multiple links using different transmission rates share a common wireless channel,and the missing receiver problem in single-radio multi-channel networks. Furthermore,a Multi-channel Rate Adaptive(MCRA) Media Access Control(MAC) protocol is proposed for WLAN Mesh network. The neighboring nodes can cooperatively inform the transmitter of the channel used by the receiver in the proposed protocol. Besides,the receiver can choose the feasible transmission rate and channel and send back them to the transmitter. Simulation results show that the proposed protocol can eliminate the interference between the links using different transmission rates,and effectively solve the missing receiver problem,thus significantly improve the overall performance of the network. The proposed protocol can significantly improve the total throughput of the network compared with the existing representative protocol.
-
Computer Engineering. 2014, 40(11): 31-35，41. https://doi.org/10.3969/j.issn.1000-3428.2014.11.006
Multiple Virtual Machines(VMs) can be hosted in the same CPU core with virtualization technologies,in a fair share manner of the physical resources among the VMs. However,as the number of VMs sharing the same core / CPU increase, the CPU access latency perceived by each VM also increases, which translates into longer network I / O processing latency experienced by heterogeneous application including both network I / O and computation. To mitigate such impact, an application type driven dynamic time slice adjusting mechanism is presented. The evaluation of a prototype in Xen shows that,compared with Credit scheduler of Xen,this mechanism improves the connection rate and response time of Nginx Web server.
-
Computer Engineering. 2014, 40(11): 36-41. https://doi.org/10.3969/j.issn.1000-3428.2014.11.007
Opinions sharing or reaching consensus is a common social phenomenon. In consideration of the facts that nodes prefer to select certain nodes to communicate and they have memory for viewpoints which are different from their own,this paper tries to establish a novel opinion dynamics model by extending the Deffuant model. Priority selection strategy and the memory effect of node are adopted in the model. And it studies the influences of these two factors on network opinion formation. Experimental results show that the proposed model adopting priority selection strategy helps consensus formation in non-uniform network. But when the network adopts the priority selection strategy without considering memory effect,the formation of consensus still depends on threshold. And the joining of the memory effect not only can promote formation of network consensus, but also can make the network reach consensus at a small threshold. Research results show that with the increasing of the threshold,the smallest opinion updating time threshold to reach consensus decreases.
-
Computer Engineering. 2014, 40(11): 42-45,55. https://doi.org/10.3969/j.issn.1000-3428.2014.11.008
For the object tracking problems in computer vision,this paper proposes a tracking algorithm based on Ranking Support Vector Machine (RSVM) fused with multiple features. Firstly,RSVM is used to get rank function. Secondly,the RSVMs combined with the two different image features are learnt respectively,then the two RSVMs predict parallel. Finally,the two RSVMs are fused with the weights which are calculated by the error rates of two classifiers,then it constructs a more adaptive RSVM framework fused with multiple features. This algorithm fuses image features effectively,and gets accurate predictions using RSVM. Experimental results demonstrate that it outperforms several stateof-the-arts algorithms.
-
Computer Engineering. 2014, 40(11): 46-49. https://doi.org/10.3969/j.issn.1000-3428.2014.11.009In embedded systems, Flash is an often-used way of read-only memory. For all kinds of Flash online programming interface in the present industry,a general Flash online programming system is implemented in this paper by using the embedded debugging techniques. This system encapsulates the different Flash online programming details by defining specific Flash online programming driver. An Integrated Development Environment(IDE) based on Eclipse is at the top of the system,and in the background,the existing embedded debug tool is reused to achieve Flash online programming with the help of specific Flash driver running on the embedded target. In terms of use,this system presents a simple operation,strong visual software interface for Flash online programming,and in the aspect of function,major Flash in the industry can be online programming in this system,which implements the generality of Flash online programming.
-
Computer Engineering. 2014, 40(11): 50-55. https://doi.org/10.3969/j.issn.1000-3428.2014.11.010A variety of visual elements are used to comprehensively express historical events,which can demonstrate vividly historical pictures. This paper proposes a Historical Events Oriented Spatio-temporal Data Model (HEOSDM) through combing event-based spatio-temporal data model and object-oriented idea. Considering the event,change and status,this model can describe and reproduce spatio-temporal information and changes in the major historical events. and a temporal tree data structure of historical events combining event list and sub-event tree is proposed. This structure adopts dual data structures which effectively organize spatio-temporal objects and related events so as to dynamically backtrack, reproduce and search the development process and status of historical events. A temporal tree data structure of events is used to realize spatio-temporal processes visualization of historical events. Adobe Flex is selected to develop and realize the corresponding algorithm and it achieves good effects.
-
Computer Engineering. 2014, 40(11): 56-59. https://doi.org/10.3969/j.issn.1000-3428.2014.11.011Aiming at the problem of eXtensible Markup Language(XML) data query and update,this paper proposes a dynamic XML coding scheme based on left son and right sibling node chain association. According to the chain related information about brother node,only by doing some simple changes in local,it can realize unlimited updates of XML data and the ancestor descendant relationships,parent-child relationship and sibling relationships and other axis operation. The results show that the proposed coding scheme not only efficiently supports structure query,but also has less coding time and insert node time. It can fast and accurately determine the relationship between any XML document structure tree nodes,so as to avoid a lot of code adjustment problems the update brings,and it efficiently supports the query and update of XML documents.
-
Computer Engineering. 2014, 40(11): 60-64. https://doi.org/10.3969/j.issn.1000-3428.2014.11.012Graph applications rise with the emerging of social network and semantic Web,and generate many graph data processing products,including Neo4j,HyperGraphDB,etc. However,current solutions fail to take into consideration graph applications’ higher requirements on data availability and scalability. This paper proposes a modeling and storage solution based on distributed memory cloud. It takes advantage of the prior work to build a key-value system over the memory cloud,then builds data modeling and read-write based on it. Experimental results on large scaled datasets show that this solution has a good figure random access performance,and it can support massive graph applications efficiently.
-
Computer Engineering. 2014, 40(11): 65-70,76. https://doi.org/10.3969/j.issn.1000-3428.2014.11.013For the problem of current access control strategies difficultly guaranteeing the flexibility of authorization of complex E-government system for Web service,this paper proposes an organization-based access control model for Web services on the basis of the research of the organization-based 4 level access control model. The model takes organization as the core and studies the issue of access control and authorization management from the perspective of management. Through importing the position agent and authorization unit in the model,the authorization can be adjusted according to the change of the environment context information to implement the dynamic authorization,while taking advantage of the state migration of authorization units,provides support for workflow patterns. Furthermore,the model divides permissions into service permissions and service attribute permissions,and achieves fine-grained resource protection. Application examples show that the model can commendably fit the complex organization structure in E-government system. Moreover,it can make authorization more efficient and flexible meanwhile protecting the Web service resources.
-
Computer Engineering. 2014, 40(11): 71-76. https://doi.org/10.3969/j.issn.1000-3428.2014.11.014As the software requirement becoming more and more complicated, developers have an increasing dependency on the code recommendation tools to assist their development tasks,while current code recommendation tools can not provide efficient recommendation for static methods. In order to recommend code snippet including static methods in a faster and more accurate mean,this paper proposes code snippet recommendation method based on program context combing with the reachability analysis, which solves the space explosion problem during the static method recommendation well,and it can also describe the program context accurately and effectively. Based on the proposed method,a code recommendation Eclipse plugin is implemented,and a corresponding experiment on Tomcat source code is conducted. Experimental results show that,the proposed method not only owns the capability of recommending code snippet with static method,but also has a higher accuracy compared with the recommendation method of Eclipse Code Recommenders.
-
Computer Engineering. 2014, 40(11): 77-82. https://doi.org/10.3969/j.issn.1000-3428.2014.11.015To solve the problems of equipment safety and sensitive data leakage in mobile terminal,this paper proposes a Mobile Terminal Security Management System ( MTSMS) based on Android,which allows the users to remotely manage the mobile terminal. MTSMS adopts C / S architecture,the client-side of MTSMS permits users to register mobile terminal on the server,and also achieves local user identification of three factors based on Trusted Platform Module (TMP) in mobile terminal. The server-side of MTSMS permits users to login Web platform to remotely encrypt and decrypt data in registered mobile terminal,locate and lock the registered mobile terminal. Further,to prevent users’ remote order from being faked,MTSMS achieves server-side authentication based on elliptic curve cryptography and zero knowledge proof. Experimental result shows that the system can ensure the security of mobile terminal devices,and prevent the fake of remote user instructions.
-
Computer Engineering. 2014, 40(11): 83-86. https://doi.org/10.3969/j.issn.1000-3428.2014.11.016
A new interference alignment method is proposed to improve the channel capacity of the multi-user Multiple Input Multiple-Output(MIMO) channel. Based upon the assumption that the antenna configuration is unchangeable,the antennas at transmit node are divided into two parts and with differentiated parameters,it conduces to eliminate the correlation between the two subparts and contributes to form two independent sub-channels. The signals received are consolidated to extend the dimensions of the signal space. According to the equivalent channel transmission matrixes,the pre-coding matrices and combining matrices are separately designed by optimizing the angle between the pre-coding vectors and using orthogonal projection method. The new scheme can reduce the computation pervasively existed in the previous iterative algorithms. Simulation indicates that the proposed method can effectively promote the anti-interference ability of the MIMO channel with the unchanged antenna configuration.
-
Computer Engineering. 2014, 40(11): 87-91. https://doi.org/10.3969/j.issn.1000-3428.2014.11.017A new energy efficient routing protocol of MDG-LEACH is proposed in the paper,which is based on the question of numerous redundancies with data collecting in nodes and imbalance of energy consumption in Low Energy Adaptive Clustering Hierarchy ( LEACH ) protocol. It selects active nodes based on virtual grids and multiple dimensions grids. Further,it adopts cluster head choice mechanism in overall consideration of residual energy and spatial distribution of nodes and establishes the shortest transmission route tree between cluster heads and base station by the dynamic programming method for date transmission. Simulation results show that MDG-LEACH protocol improves LEACH protocol and the improved LEACH protocol ( Computer Technology and Development,2013, No. 2 ) in balancing energy consumption of nodes and prolonging the life time of wireless sensor network effectively.
-
Computer Engineering. 2014, 40(11): 92-96. https://doi.org/10.3969/j.issn.1000-3428.2014.11.018Aiming at energy constrained problems in Wireless Sensor Network(WSN),in order to achieve a balanced energy consumption of nodes,balance cluster heads distribution,and maximize the network lifetime,this paper proposes a WSN energy balanced clustering protocol which is based on the Cell Membrane Optimization(CMO) algorithms. The CMO algorithm has good ability of global optimization,and fast convergence capability. The new clustering protocol divides the nodes through concentration and energy factors. Combined with the distance factor for global clustering balance,it can be a good solution to the uneven distribution for cluster head,imbalanced global energy and other issues. Experiments show that the protocol has the capability of balanced global clustering quickly. Compared with LEACH algorithm and LEACH-C algorithm,it has better performance in terms of balancing power consumption and prolonging the network lifetime.
-
Computer Engineering. 2014, 40(11): 97-100,105. https://doi.org/10.3969/j.issn.1000-3428.2014.11.019As high free degree of amount and uneven distribution of the cluster-head nodes cause high energy consumption,this paper presents a novel improved LEACH protocol algorithm based on cluster optimization and energy balance. In the process of selecting cluster-heads according to the conventional energy threshold,the improved selection mechanism selects cluster-head with the optimal cluster-radius control strategy and optimizes the distribution of cluster-head nodes. In the meantime,the coverage rate is put forward to avoid extra cluster-heads and keep the amount of cluster-head nodes. It uses CH-VCH alternate circulated strategy to simplify the calculation. The simulation reveals that the improved LEACH protocol can solve the problem that the traditional one can not do. It can average the energy consumption and prolong the survival period of the wireless sensor networks.
-
Computer Engineering. 2014, 40(11): 101-105. https://doi.org/10.3969/j.issn.1000-3428.2014.11.020Broadcasting scheduling problem is a hot research problem in Cognitive Radio Networks(CRN),and existing works for the broadcast scheduling issue in CRN are either heuristic solutions without performance guarantee or with performance far from the optimal solution. This paper proposes a Broadcasting Scheduling Algorithm Based on Unit Disk Graph Model(BS-UDGM). It constructs a Connected Dominating Set(CDS) based broadcasting tree,which serves as the scheduling infrastructure,and subsequently,the broadcasting tree is optimized by using the tessellation and coloring of a plane. The broadcast task is finished by employing mixed unicast and broadcast communication modes. Simulation results show that compared with the existing algorithms,the performance of the proposed algorithm is improved with respect to both latency and redundancy.
-
Computer Engineering. 2014, 40(11): 106-112. https://doi.org/10.3969/j.issn.1000-3428.2014.11.021For the status quo that the Botnet detection of a priori knowledge to get the matching and protocol-related are unable to be suitable for new or mutated Botnet detection,this paper proposes a dynamic Botnet detection model based on network traffic. By using clustering,it analyzes traffic and completes the correlation analysis to identify similar between bot communication and malicious behavior patterns. The test architecture has nothing to do with the agreement and Botnet prior knowledge. The model has three dynamic characteristics,such as the characteristics of library updated,detection model generation,and handling the network traffic from the dynamic Botnet. Finally,the effectiveness and the accuracy are verified by the experimental data.
-
Computer Engineering. 2014, 40(11): 113-117. https://doi.org/10.3969/j.issn.1000-3428.2014.11.022Based on the difficulty of the discrete logarithm assumption,H?lbl et al(Computer Standards & Interfaces, 2009,No. 6) presented two identity-based authenticated key exchange protocols. The first protocol,denoted by HW1, improved Hsieh et al’ s protocol which makes it immune against Tseng et al’ s attack (Journal of Computers,2002, No. 3),while the second protocol,denoted by HW2,improves the efficiency of Tseng’s protocol. Shim et al analyzes these two protocols,and then shows that the HW1 can not resist the man-in the-middle attack and the impersonation attack,and the HW2 can not resist the impersonation attack(IEEE Communications Letters,2012,No. 4). This paper conducts a detailed analysis on the flaw,and finds the reason of the protocols are tampered,making use of the Hash function,authenticates the information to prevent the information is tampered,it proposes improved protocols based on these two protocols,and analyzes the security of improved protocols. The results suggest that the improved protocols can resist the man-in-the-middle-attack and the impersonation attacks,they are safe and feasible.
-
Computer Engineering. 2014, 40(11): 118-120,125. https://doi.org/10.3969/j.issn.1000-3428.2014.11.023As an attractive cryptographic primitive,the public key searchable encryption enables users to search on encrypted data,and hence is applicable to the setting of cloud computing. But most of the existing schemes have to adopt the bilinear pairing and fail to verify search results from the server. Accordingly,these schemes suffer drawbacks in terms of efficiency and security. Aiming at this problem,based on the ElGamal encryption algorithm,a new verifiable scheme is proposed. It has more desirable computation efficiency and is easy to implement in because it replaces the bilinear pairing with the ElGamal encryption. Especially,during the generation of encrypted keywords and encrypted files,the new scheme can generate the digital signature of the hash value of keywords based on the ElGamal signature algorithm. Upon receiving the search results from the server,users can obtain the public key of the sender,and then verify the ElGamal signature, which effectively prevents the server from returning wrong results.
-
Computer Engineering. 2014, 40(11): 121-125. https://doi.org/10.3969/j.issn.1000-3428.2014.11.024Delay-tolerant Network(DTN) is an Ad Hoc network,in which there is not end-to-end path between the source and the destination nodes in most of the time,and DTN relies on the mechanism of store and forwards to realize the asynchronous communication. Aiming at the blackhole attack in DTN,this paper proposes a security routing protocol which can detect blackhole nodes. It analyzes the model of DTN,gives a scheme to detect malicious nodes,and merges this scheme with the routing protocol. Compared with existing routing protocols through the simulation,it verifies that the security routing protocol can accurately detect the malicious nodes and it has better performance in DTN environment. When being used in water environment monitoring or city traffic control and other fields,the security routing protocol can avoid the network environment from malicious nodes attack,and ensure the reliability and stability of network.
-
Computer Engineering. 2014, 40(11): 126-129,134. https://doi.org/10.3969/j.issn.1000-3428.2014.11.025Recursive diffusion layer is a new type of cryptographic linear diffusion layer. It has good structure characteristics,and can achieve the optimal effect of diffusion layer. It is more complex in the concrete implementation process because its function structure parameters in the search space is large. It seriously impacts on its cryptographic function in the actual application ability. After analyzing the structure and the lower order based on the structure of the diffusion layer,and combining with the optimal diffusion layer of relevant theoretical basis,this paper gets some general conclusions of recursive diffusion layer,and based on this,it gives a method to design the recursive diffusion layer and puts forward a scheme to improve the coefficient’s implementation of recursive diffusion layer. By the simulation realization,it gets the results of recursive diffusion layer’s structure in low order. The diffusion layer only needs a few XOR operation, rotating operations and some simple complementation operations,and it has a better security character.
-
Computer Engineering. 2014, 40(11): 130-134. https://doi.org/10.3969/j.issn.1000-3428.2014.11.026Most existing proxy re-signature schemes are existential unforgeability,where an adversary will be able to forge a signature on a new message rather than on a message that has already been signed. However,strong unforgeability can protect the existing message-signature pairs from being forged. By using TCR hash function,a bidirectional proxy resignature scheme is proposed. Based on collision-resistant Target Collision Resistant ( TCR ) hash function and computational Diffie-Hellman assumption,the proposed scheme is proved to be strongly unforgeable under adaptive chosen message attacks. The results show that the proposed scheme in computational efficiency is superior to the available proxy re-signature schemes with strong unforgeability. Compared with these schemes,the new scheme has short system parameters,short signature,short re-signature and more security properties.
-
Computer Engineering. 2014, 40(11): 135-138. https://doi.org/10.3969/j.issn.1000-3428.2014.11.027Aiming at the problem of incomplete target segmentation and low classification accuracy of traditional pedestrian-vehicle detection and classification algorithm in intelligent monitoring,this paper presents a pedestrian-vehicle detection and classification algorithm based on video. The algorithm dynamically adjusts confidence intervals for constructing Gaussian mixture model using neighborhood information,and uses the Kalman filter to predict the position of the target in the next frame. It extracts the target aspect ratio and area through adaptive EM clustering as a feature,then divides target into pedestrians and vehicles. Assume that target makes the uniform linear motion in adjacent frame and derive the target area to meet the linear relationship change. Thus target tracking and classification can be modified to improve the detection accuracy in the end. Experimental result show that the algorithm detection rate is over 90% and classification rate is over 80% .
-
Computer Engineering. 2014, 40(11): 139-142,148. https://doi.org/10.3969/j.issn.1000-3428.2014.11.028According to the characteristic of fault data of high-speed train,a dynamic feature selecting algorithm is proposed to research the measured data of the running gear(referring mainly to bogie) of high-speed train. The approach combines the advantages of Fisher ratio and fuzzy entropy dynamically, which manages to evaluate features more accurately and removes the redundant features effectively to obtain superior feature subset by weighted average method. The new approach can improve classification accuracy. Experimental results for fault data of high-speed train show that the proposed approach not only improves the classification accuracies significantly,but also strengthens the stability in low speed. The overall-precise improvement is 5. 2% after extracting the optimal feature space in average compared with that of the original feature space.
-
Computer Engineering. 2014, 40(11): 143-148. https://doi.org/10.3969/j.issn.1000-3428.2014.11.029To solve the problem that the traditional ant colony algorithm is slow to converge and easy to conflict in the collaborative trajectory planning, considering the features of network graph consist of trajectory segments, a aircraft collaborative trajectory planning algorithm is proposed based on multi-subgroup ant colony coevolution. It divides the ant colony into subgroups with the same number of the aircrafts. Heterogeneous pheromone is introduced to simulate the competition among subgroups,reference length coevolution is adopted to guide the subgroups generating trajectory satisfying the temporal constraints,and the strategy of lost ants pheromone update is added to accelerate the convergence speed. Experimental results demonstrate that this algorithm can generate collaborative flight trajectorys satisfying the constraints of time and space in complex environments for different planning tasks. Compared with the traditional ant colony algorithm,it can generate better collaborative trajectorys,while the planning speed can be improved by 2 ~3 times.
-
Computer Engineering. 2014, 40(11): 149-154. https://doi.org/10.3969/j.issn.1000-3428.2014.11.030The number of hidden layer and hidden layer node in neural network determines the size of the network and has a great influence on the performance of the network. Therefore,when the network contains the least hidden layer node number,pruning algorithm can be used to delete some redundant node,then the network is more simple. The pruning algorithm adds a penalty function to the target function,and the penalty function regards the weights of network as variable. It adds a variable parameter to the weights of network,so the simple penalty function can be generalized to a kind of new penalty function that changes as the parameter. The initial function can be treated as a special condition after the generalization of penalty function. Experiment tests the XOR data based on the BP neural network and sums up the effect of the generalization of penalty function on the pruning of the hidden layer node with neural network and the structure of the neural network. Then the parameters which can lead to better pruning effect and more optimal network structure are obtained from data in experiment.
-
Computer Engineering. 2014, 40(11): 155-159,166. https://doi.org/10.3969/j.issn.1000-3428.2014.11.031In Differential Evolution ( DE ) algorithm, the population initialization is updated by using opposition optimization rule,so as to maintain the population diversity. However,the reverse individuals are easy to deviate from the global optimal solution,has slow convergence speed and easy to fall into local optimum in function optimization. This paper proposes a fast Opposition Differential Evolution(ODE) algorithm based on M-H(Metropolis-Hastings) sampling method. M-H sampling is used in the mutation operation of ODE. M-H sampling satisfies Markov Chain reversible conditions. One step transition probability of Markov Chain is calculated according to the selecting probability of individual ranking-assignment,not only chooses the best individual,but also searches for the optimal direction and keeps the population diversity. Simulation results show these individuals from M-H sampling have Markov stationary distribution property. The algorithm can accelerate the speed of convergence in unimodal functions and multimodal functions,balance the performance of global searching and local searching,and has higher precision and better robustness.
-
Computer Engineering. 2014, 40(11): 160-166. https://doi.org/10.3969/j.issn.1000-3428.2014.11.032The Improved Local Tangent Space Alignment(ILTSA) can obtain better low dimension feature for face recognition because it can efficiently recover the problem that the Local Tangent Space Alignment(LISA) fails to reveal the manifold structure in the case when data are sparse or non-uniformly distribute or when the data manifold has large curvatures. To solve the problem that the ILTSA cannot efficiently handle ever-increasing data set,this paper presents a Generalization method for the ILTSA(GILTSA). The nearest neighborhood set is obtained based on the distance defined according to the classes of the samples,then the low manifold of the training set is implemented using the ILTSA. Through finding the nearest sample in the training set,and the low manifold of a new sample is approximately calculated by the projection of its nearest sample. Experimental results on the ORL,the Yale and the University of Essex face image database indicate that the proposed GILTSA method increases the overall accuracy compared with Principal Component Analysis(PCA) and Linear Local Tangent Space Alignment(LLTSA) algorithm etc.
-
Computer Engineering. 2014, 40(11): 167-171. https://doi.org/10.3969/j.issn.1000-3428.2014.11.033There are many algorithms based on data distribution to effectively solve semi-supervised text categorization. However,they may perform badly when the labeled data distribution is different from the unlabeled data. This paper presents a semi-supervised text classification algorithm based on aggregation pheromone, which is used for species aggregation in real ants and other insects. The proposed method,which has no assumption regarding the data distribution, can be applied to any kind of data distribution. In light of aggregation pheromone,colonies that unlabeled ants may belong to are selected with a Top-k strategy. Then the confidence of unlabeled ants is determined by a judgment rule. Unlabeled ants with higher confidence are added into the most attractive training colony by a random selection strategy. Compared with Na?ve Bayes and EM algorithm,the experiments on benchmark dataset show that this algorithm performs better on precision,recall and Macro F1.
-
Computer Engineering. 2014, 40(11): 172-177. https://doi.org/10.3969/j.issn.1000-3428.2014.11.034Clustering result of k-means clustering algorithm is highly dependent on the choice of the initial cluster center. With regards to this,a text clustering algorithm based on improved Particle Swarm Optimization ( PSO) is presented. Features of particle swarm algorithm and k-means algorithm are analysed. Considering the disadvantages of PSO including low solving precisions, high possibilities of being trapped in local optimization and premature convergence,self-regulating mechanism of inertia weight and cloud mutation operator are designed to improve PSO. Selfregulating mechanism of inertia weight adjusts the inertia weight dynamically according to the degree of the population evolution. Cloud mutation operator is based on stable tendency and randomness property of cloud model. The global best individual is used to complete mutation on particles. Those two algorithms are combined by taking advantages of power global search ability of PSO and strong capacity of local search of k-means. A particle is a group of clustering centers,and a sum of scatter within class is fitness function. Experimental results show that this algorithm is an accurate,efficient and stable text clustering algorithm.
-
Computer Engineering. 2014, 40(11): 178-182. https://doi.org/10.3969/j.issn.1000-3428.2014.11.035With the rapid development of the human motion capture technology,large amount of motion capture data is gradually accumulated,then human motion retrieval and recognition technology becomes the essential issue for motion data management and reuse. Logically similar motions may be numerically dissimilar,so it is difficult to get feasible results if the logical similarity between two movements is measured with Euclidean distance. This paper presents a semisupervised distance learning method for measuring the logical similarity with Mahalanobis distance which is trained by labeled and unlabeled motion data. Experimental evaluation result of the method shows that the proposed method is effective for motion retrieval.
-
Computer Engineering. 2014, 40(11): 183-188. https://doi.org/10.3969/j.issn.1000-3428.2014.11.036Based on the characteristics of logistic network design problem of some special materials,a joint Locationinventory Problem(LIP) model with lead-time is built,considering stochastic demands,facility capacity constraints and the client time constraints. The goal is to minimize system cost and maximize system timeliness. A discrete nonlinear mixed integer programming model with 2 goals is built to describe the problem. An improved NSGAII based on niching technology is worked out to solve the model,in order to enrich the number of non-dominated solutions. Numerical example and control experiment indicate that the Pateto front solution set can be obtained and the improved NSGAII has obvious advantages compared with standard NSGAII. The model and algorithm can be used to make location and inventory decision of blood banks or other emergency medicine warehouses. And optimal decision schemes can be selected from a cluster of Pateto solutions according to the preferences and actual needs of decision makers.
-
Computer Engineering. 2014, 40(11): 189-193. https://doi.org/10.3969/j.issn.1000-3428.2014.11.037Artificial Bee Colony ( ABC) algorithm can be efficiently employed to solve the multimodal and high dimensional function optimization problem. However,low search speed and premature convergence frequently appear with more complex problem. In order to improve the algorithm performance,this paper proposes a new artifciall bee colony algorithm. It introduces a search equation based on multi-dimensional greedy search to enhance local search and avoid the solution to be abandoned which achieves optimum value in some dimensions but reach the maximum update limit. New algorithm also adds a disturbance mechanism to avoid obtaining partial optimal solutions when premature convergence in a few dimensions. Experimental results show the new algorithm can balance the exploitation and exploration,has more fast convergence speed and better computational precision in solving the multimodal and high dimensional function optimization problem.
-
Computer Engineering. 2014, 40(11): 194-199. https://doi.org/10.3969/j.issn.1000-3428.2014.11.038According to the existing difficulties of parameter adjustment in taking pictures and the problems of how to keep two cameras cooperating coordination and coherence in the existing 3D-camera,this paper presents an automatic acquisition method of stereo shooting parameters based on sterescopic vision. The method makes full use of the ability of panoramic 3D-camera made of four Omnidirectional Vision Sensors(ODVS) to obtain panoramic stereo images and get the desired convergence point, distance, focal length, aperture factor and other parameters automatically through Three-Dimensional (3D) image processing technology. Experimental results show that the proposed method can effectively guarantee consistency of the 3D photographic parameters of focal length,shooting direction,shooting angle,3D depth and so on,solve the coordinating difficulties of two cameras’ shooting action and the adjustment of 3D depth parameters during 3D shooting,achieve capturing 3D panoramic video and 3D close-up video and shooting 3D close-up video automatically by panoramic points controlled on the monitor at the same time.
-
Computer Engineering. 2014, 40(11): 200-204. https://doi.org/10.3969/j.issn.1000-3428.2014.11.039Particle Swarm Optimization (PSO) algorithm is inclined to fall into diversity loss and low optimizing efficiency in dynamic environment. In this paper,a PSO algorithm with symmetric displacement mapping is proposed. The main subpopulation and assistant subpopulation particle swarm work with each other to search the changing global optimum by the parallel searching. The assistant subpopulation particle swarm uses differential evolutionary mechanism to constantly explore the new environment,when the environment is changed,and a symmetrical displacement mapping strategy is introduced to improve the convergence probability to the global optimum through symmetrical particle distribution surrounding the global optimum. The simulative environment in experiments is generated by MPB and DF1 two benchmark functions,the results demonstrate that the algorithm can improve the accuracy of tracking the changing global optimum.
-
Computer Engineering. 2014, 40(11): 205-210. https://doi.org/10.3969/j.issn.1000-3428.2014.11.040Mining roles in large scale organizations are very redundant and can not reflect system security requirements in existing approaches to bottom-up role engineering. In order to discover optimal roles, this paper proposes an optimization method for role mining,which is based on the enumeration approach to role mining. The method utilizes mutually exclusive permissions constraints to optimize the role mining process by clustering users. It analyzes the correctness of secure constraints by using separation of duty,and cells’ values of mining permissions are adjusted in access control matrix. The method mines roles that cover all permissions. Experimental results show that the set of optimal roles can ensure the completeness of role mining. Compared with the enumeration method,the optimization method can reduce the redundancy of roles and ensure system security.
-
Computer Engineering. 2014, 40(11): 211-214. https://doi.org/10.3969/j.issn.1000-3428.2014.11.041Emotion thesaurus plays an important role in the text sentiment analysis,but it is particularly inadequate in the analysis of fine-grained emotions such as human emotions. To solve this problem,this paper presents a fine-grained emotion thesaurus construction method via the calculation of sememe similarity,and finishes the construction of seven sorts of thesaurus. Based on this work,this paper researches on the calculation method of the weight of fine-grained emotion words, and proposes a new weight calculation method of emotion words. Finally, this paper finishes the construction of seven sorts of thesaurus with weight value. Experimental results show that the introduction of the finegrained emotion thesaurus with weights can make the accuracy rate of the text emotional tendencies increased by about 5% .
-
Computer Engineering. 2014, 40(11): 215-219. https://doi.org/10.3969/j.issn.1000-3428.2014.11.042Zernike moment,as a shape descriptor,has been widely used in image characteristics extraction and pattern recognition. It is low information redundancy and not sensitive to noise. To improve the shape description capability of the images which are degraded by combined blur and affine transformation,a new shape descriptor based on Zernike moment is proposed. The normalization method is used to construct affine invariants of Zernike moment. The combined blur and affine moment invariants of Zernike moment is achieved by the help of the blur invariants. The combined moment invariants is used as the shape descriptor to describe the shape feature of images,and is implemented comparison with the combined affine and blur invariants based on geometric moment with relative error. Experimental results show that the combined blur and affine invariants of Zernike moment can get better shape description and invariance in combined degrades,and robustness to noise.
-
Computer Engineering. 2014, 40(11): 220-224. https://doi.org/10.3969/j.issn.1000-3428.2014.11.043The Partial Differential Equation(PDE) image processing is an advanced technology due to the ability of obtaining continuous and one-pixel edges. The improved Geodesic Active Contour(GAC) model by using the gradient vector field and a balloon force———GAC_GVF&B is an important one,because it can convergence accurately to target edges on images with multi-object or complex objects. But the model suffers from long running time which blocks its application. By appropriately increasing the time step and reducing the number of iterations,a semi-implicit additive split operator———Additive Operator Splitting (AOS) is used to speed up the computing of the GAC _GVF&B model and improve the convergence rate with the same accuracy of the segmentation. Experimental results show that the AOS algorithm is correct and effective,can reduces the number of iterations required,and lowers iteration time and cpu time. Furthermore,it speeds up the segmentation.
-
Computer Engineering. 2014, 40(11): 225-228,249. https://doi.org/10.3969/j.issn.1000-3428.2014.11.044The computation of approximate geodesics algorithm on triangle mesh can be transformed to the computation of the shortest path on triangle mesh. In order to figure out the geodesics on triangle mesh efficiently,an improved approximate geodesics algorithm is presented. A weighted graph is constructed by splitting triangle mesh along the Cartesian coordinate axes,thus the shortest path on lattice can be computed by A? algorithm and a new region of search can be defined. The neighboring edges of the shortest path are iteratively subdivided to construct a new weighted graphs to approach the geodesic. Ex perimental results show the improved algorithm can figure out the geodesics precisely and efficiently. The speed up ratio of the algorithm range is increased between 10 and 59. A region labeling method on virtual liver surgery based on the improved algorithm is also proposed,which can meet the computational real-time and quality requirements of virtual scene.
-
Computer Engineering. 2014, 40(11): 229-232,254. https://doi.org/10.3969/j.issn.1000-3428.2014.11.045This paper presents a novel Region-based Image Retrieval (RBIR) method to reduce the semantic gap between low-level visual feature and high-level semantic of images in the content-based Image retrieval area. K-means clustering algorithm is used in the LUV color space for image segmentation. The color and shape feature in each region as well as the region auto-correlation feature are extracted as the integrating features of each region. And this paper defines a new Quadratic Distance Similarity Measure(QDSM) to calculate the similarity between different images. Experimental results show that the novel RBIR method using integrating features increases the retrieval performance by 12% ~47. 8% compared with the traditional methods under Average Normalized Modified Retrieval Rank(ANMRR) metric.
-
Computer Engineering. 2014, 40(11): 233-236,259. https://doi.org/10.3969/j.issn.1000-3428.2014.11.046In the digital mine system,in order to make intricate underground roadway reconstruction effect reach the acceptance criteria of the mine,a 3D modeling method for fork entity based on measured contour using a partition or block modeling idea is proposed,which is is based on the measured contour of the roadway cross fork. The measured contour data of fork is classified by features into top,bottom and side contour point set. The convex hull algorithm is used to construct the roof and floor of a simple polygon,and triangulation. Based on the order of the top (bottom) plate polygon vertices,the adjacent relationship of the lateral broadside contours is established,and connection box algorithm is used to triangulate for adjacent contours. Experimental results show that,the algorithm is simple and effective and can solve any cross-sectional shape and any multiple fork modeling problems. The algorithm has been applied to the measurement plug-in contained in DIMINE software,and has good results.
-
Computer Engineering. 2014, 40(11): 237-240,265. https://doi.org/10.3969/j.issn.1000-3428.2014.11.047Aiming at the problem that Boundary Discriminative Noise Detection ( BDND ) performs poorly when detecting unbalanced salt-and-pepper noise or random-valued impulse noise. Based on estimated noise distribution,this paper proposes a modification of BDND. It modifies the clustering inequality by placing b2 in high intensity group. And it uses the first stage of BDND detecting all pixels to generate a noise histogram. And it redefines the upper and lower boundary values by comparing the ratio of adjacent noise with a given threshold value. It uses the new boundary values for further noise detection. Experimental results show that the proposed method is superior than BDND on detection performance,miss detection and false detection have been greatly reduced. On the visual effects,it can filter out noise while preserving image details well.
-
Computer Engineering. 2014, 40(11): 241-244. https://doi.org/10.3969/j.issn.1000-3428.2014.11.048To solve the problem that the traditional minimum-load-priority load balancing algorithm has quite long response time of requests and low load balance degree,this paper proposes a load balancing algorithm which is the combination of static load scheduling and dynamic load migration. To reduce the response time of requests,a static load scheduling algorithm is used,which is based on the similarity of video requests. The algorithm aims to make full use of caching capacity on Video-on-Demand (VOD) by scheduling similar requests to the same VOD. The dynamic load migration algorithm,which is based on cache considering REM load migration,aims to improve the load balance degree by migrating load form high load VOD to low load VOD. By practical test,it proves that this algorithm can reduce the response time of requests by 17. 5% and cluster load variance by 53. 4% comparing with minimum-load-priority algorithm at typical scale of metro video surveillance system under the number 500 requests.
-
Computer Engineering. 2014, 40(11): 245-249. https://doi.org/10.3969/j.issn.1000-3428.2014.11.049Considering the fact that,Android Board Support Package(BSP) transplantation personnel not only needs to be familiar with the specific model chip development manual over again,but also needs to adjust the package code of some parts of the drivers and other issues,this paper reconstructs Android Linux source code to design and implement a framework of Android BSP transplantation based on the classified management of source code. The framework defines Android Linux kernel source code reconstruction rules. Based on the rules,it manages Android Linux kernel source code in accordance with the native Linux kernel code,Google Android drivers and BSP code associated with manufacturer board. In addition,an Android BSP transplantation automatic adaptation tool is designed which implements the automatic constitution of Android Linux kernel source code. Test result proves that the framework can significantly improve work efficiency of transplanting and upgrading board Android Linux kernel.
-
Computer Engineering. 2014, 40(11): 250-254. https://doi.org/10.3969/j.issn.1000-3428.2014.11.050According to feature of 16/ 32 bit mixed instruction set of CK810,assembly and link time optimization techniques based on the 16/ 32 mixed instruction set of CK810 make use of assembler output binary files to achieve instructions relax and coding instructions dynamically. The assembler gives relocations to linker for instructions that can’ t decide how to code at assembly time,the techniques fully utilizes the information only available at link time to realize of the entire program for compression and replacement of instructions to make program more efficient and small. Assembly and link optimization techniques overcome the limitations of traditional compilers by enlarging the optimizing scope from a single function or a module to the whole program,making the code density on CK810 increase an average of 7. 52% , the average performance improvement of 7. 91% .
-
Computer Engineering. 2014, 40(11): 255-259. https://doi.org/10.3969/j.issn.1000-3428.2014.11.051The paper discusses the binary self-dual code of error correcting code,looking the code as a polynomial on the field GF(2)n and factorizing it. It constructs the longer length code by the shorter length code and gives the code’s generator matrix. Owning to code’s equivalence,the possible classification of length 54 binary self-dual codes with an automorphism of order 13 is given. By implementing Matlab procedures,there are eight self-dual codes up to equivalence. By similar method,there are 16 self-dual codes with type of 13-(4,4) and there are 10 self-dual codes with type of 13-(4,6) up to equinalence. Therefore,the generator matrices and classifications of the binary self-dual code with type of 13-(4,2),13-(4,4) and 13-(4,6) are solved completely.
-
Computer Engineering. 2014, 40(11): 260-265. https://doi.org/10.3969/j.issn.1000-3428.2014.11.052CAN bus is normally used as internal bus in the large-scale Battery Energy Storage System (BESS) at present. FlexRay used as internal bus replacing CAN bus can well meet the requirements of internal communications in the large-scale battery energy storage system. The shortage of CAN bus which is used in the large-scale battery energy storage system and the superiority of FlexRay bus are analysed. Based on the simple introduction of the communication architecture of battery energy storage system,the communication architecture of energy storage battery management system is introduced by focus,and the feasibility of the application and major scheme of the FlexRay communication protocol in large-scale energy storage battery management system are discussed. Network Designer and CANoe. FlexRay are used to simulate the feasibility of FlexRay communication system used in the large-scale battery energy storage system. Simulation results prove that FlexRay bus used in energy storage battery management system can achieve good effects compared with CAN bus.
-
Computer Engineering. 2014, 40(11): 266-272. https://doi.org/10.3969/j.issn.1000-3428.2014.11.053Joint Picture Experts Group(JPEG) compression standard is widely used in static image codec for its high compression ratio and strong practicality,while it requires better performance in feature of real-time and compression speed. According to the basic vector arithmetic extracted from JPEG process,a flexible and configurable accelerator for JPEG codec is proposed to achieve the vector addition, multiplication, dot multiplication, shifting and saturation operations. Software and hardware cooperation is used in driving the accelerator to finish the Inverse Discrete Cosine Transform(IDCT),inverses quantization and color space conversion,while doing the other part in JPEG decoding with software. In the condition of increasing 0. 229 mm2 area,the process doing with accelerator costs 35% of the time costs in software,and the JPEG decoding costs only 60% of the time. This method improves the JPEG codec speed and can be used widely according to its flexibility and configurability.
-
Computer Engineering. 2014, 40(11): 273-277. https://doi.org/10.3969/j.issn.1000-3428.2014.11.054With the development of cloud accounting,the trustworthiness of its service draws more attention. Aiming at solving the problem that most existing trustworthiness evaluation methods are hard to be reasonable and effective enough to evaluate it,this paper proposes Accounting Information System (AIS) trustworthiness evaluation method based on Analytic Network Process(ANP) in cloud accounting. Considering influence factors such as user’s industry categories, dependence and feedback relationships between evaluation indexes and evolution of trustworthiness demand,it builds up trustworthiness evaluation index system which contains three dimensions:service,maintenance,reputation,establishes trustworthiness evaluation model, and based on this, through trustworthiness evaluation algorithm based on analytic network process obtained evaluation results considering the influence between indexes. Simulation experimental results show the effectiveness of this method,for it draws difference evaluation conclusion on similar service of global level, providing a strong support to users on selecting accounting information system service.
-
Computer Engineering. 2014, 40(11): 278-281. https://doi.org/10.3969/j.issn.1000-3428.2014.11.055The sparsity of the linear predictive residual of speech signal can be used in the speech compressed sensing, but needs the predictive coefficients of the signal to build sparse transformational matrix increasing the data. This paper proposes to save the linear predictive coefficients into diagonal matrix in order to build circulant matrix,and measures the speech signal in circulant way, extracts the linear predictive coefficients from circulant matrix to build sparse transformational matrix in residual domain, and reconstructs the speech with Orthogonal Matching Pursuit ( OMP ) algorithm. Simulation experimental result shows that,using circulant measure built by predictive coefficients decreases 3． 9% data more than the original linear predictive method,and has higher reconstruction signal to noise ratio per-frame than the Gaussian random matrix as measure matrix.
-
Computer Engineering. 2014, 40(11): 282-287. https://doi.org/10.3969/j.issn.1000-3428.2014.11.056The classic recommendation system is mainly based on the users for the project evaluation or the keywords similarity between the user and the item for recommendation, there is a low degree of information structure, lack of semantic and other issues. To solve these problems, this paper proposes an ontology-based recommender system model. This model uses OWL language describe the user and project information by introducing ontology to bear more semantic information and improve the degree of information structure. In the recommendation process, the results of analyzing user behavior information by rules is considered for improve the quality of recommendation. Experimental results show that the model has better effect in degree of information structure and semantics. It can effectively improve the recall and precision rates by using this model for recommendation.
-
Computer Engineering. 2014, 40(11): 288-291,296. https://doi.org/10.3969/j.issn.1000-3428.2014.11.057Aiming at high real-time demand for accurate motion detection in high-definition video surveillance system, this paper proposes an efficient motion detection algorithm based on Compute Unified Device Architecture(CUDA). By using an improved frame difference and background subtraction method of combining,background interference update is reduced and the accuracy of motion detection is enhanced. By performing video motion detection calculations in the CUDA, the traditional Graphic Processing Unit (GPU) hard decoded video data avoid storing in display memory transferred to CPU,which is the bottleneck problem. This paper uses multiple threads within a block coalesced access shared memory, reduces the amount by 52． 9% of the global memory access,solves large-scale access to global memory latency CUDA larger problem. Experimental results show that for high-definition video monitoring system,the proposed motion detection algorithm can reach 52． 6 frames per second,ensuring the accuracy as well,and can meet the real-time requirements.
-
Computer Engineering. 2014, 40(11): 292-296. https://doi.org/10.3969/j.issn.1000-3428.2014.11.058This paper designs a portable transformer condition monitoring and fault diagnosis system based on vibration analysis method. The system designs and offers pivotal technologies,including hardware and software modules such as signal sampling and conditioning,A/ D converting,database,fault diagnosis algorithms and user interface. This system can analyze the condition of transformer based on monitoring transformer signals including vibration,current,voltage and temperature. It contains fault diagnosis algorithms,besides normal signal processing method such as spectrum analysis and amplitude analysis. There are also new diagnosis methods like periodicity diagnosis algorithm and the diagnosis algorithm based on transformer vibration modeling. The methods are used for detecting the transformer abnormal and sending alarm. The test result shows that this system conveniently and effectively realizes live monitoring of transformer condition and fault.
-
Computer Engineering. 2014, 40(11): 297-302. https://doi.org/10.3969/j.issn.1000-3428.2014.11.059A novel Pulse Width Modulation(PWM) waveform generating method based on modulation waveform cycle normalization is proposed. Mapping between address storing data in digital system and pulse width of PWM waveform is the key essence. According to it,a system based on Field Programmable Gate Array(FPGA) is designed and implemented,and difficulties like analog digitization, modulation ratio M, carrier ratio N and details for modulation waveform cycle normalization digitization are solved. With setting modulation waveform Pn for PWM,a specific PWM waveform based on modulation waveform cycle normalization can be achieved successfully,which shows the effectiveness of the method.
-
Computer Engineering. 2014, 40(11): 304-309. https://doi.org/10.3969/j.issn.1000-3428.2014.11.060Autonomous component is capable of sensing and making decisions automatically according to the environment changes. The purpose of developing complex software system based on autonomous component is to reduce the complexity of software development and maintenance effectively. Self-adaptive software development requires the system and constructive component should provide a new abstract and model method to adapt to complex environment and business requirements changes. After analyzing the software entity functions and characteristics,this paper proposes an autonomous component model based on intelligent agent. The autonomous component can sense the complex environment dynamic changes,encapsulates self-adaptation logic reasonably,makes decisions in accordance with policy rules,and plans the goal-directed behaviors automatically. Autonomous component provides the efficient support for modeling and developing complex self-adaptive software.
-
Computer Engineering. 2014, 40(11): 310-317. https://doi.org/10.3969/j.issn.1000-3428.2014.11.061In order to further reduce the great computational complexity for High Efficient Video Coding(HEVC),two novel algorithms are proposed in this paper. It figures out the distribution of each LCU’s gray values in advance so that a gray histogram can be generated. By using the self-correlation of the gray histogram,some unnecessary depth levels can be skipped. Meanwhile,it also proposes three strategies to reduce the number of intra modes which are based on edge detection,intra modes correlation,texture features,respectively. Experimental results show that the proposed algorithm can save 48% of encoding time on average with negligible loss of coding efficiency compared with HM10． 1.
-
Computer Engineering. 2014, 40(11): 318-321. https://doi.org/10.3969/j.issn.1000-3428.2014.11.062Many recommendation methods based on social tagging ignore the change and repeatability of user interests,which may lead to unsatisfactory results. In order to solve these problems,a new method which efficiently combines exponential forgetting-based data weight and time windows is proposed. The method not only highlights the importance of recent interest, but also stresses the recurring early data. Based on standard tag set of the target user,the nearest neighbour set can be gained according to exponential offset tag vectors,and then calculates weight vectors via items within time windows. Recommendation values of the nearest neighbour set are computed by weight vectors and similarity. Finally,it makes recommendation of items within the top K predicted values. Simulation experimental results show that the proposed algorithm for recommendation can dynamically track the changes in user`s interest and has high quality of precision to some extent.
2024
Volume 50 No 4
Volume 50 No 4
(Started in 1975, Monthly)
Sponsored by：
East-China Institute of Computer Technology Shanghai Computer Society
Edited & Published by：
Editorial Board of Computer Engineering
Chief Editor：Jiang Bo
Executive Chief Editor:
Hu Bing
Tel：021-67092217
E-mail: ecice06@ecict.com.cn
CN: 31-1289/TP
ISSN: 1000-3428
Sponsored by：
East-China Institute of Computer Technology Shanghai Computer Society
Edited & Published by：
Editorial Board of Computer Engineering
Chief Editor：Jiang Bo
Executive Chief Editor:
Hu Bing
Tel：021-67092217
E-mail: ecice06@ecict.com.cn
CN: 31-1289/TP
ISSN: 1000-3428