Author Login Editor-in-Chief Peer Review Editor Work Office Work

15 December 2016, Volume 42 Issue 12
    

  • Select all
    |
  • FENG Xiaojie,PENG Yong,WU Xisheng
    Computer Engineering. 2016, 42(12): 1-4. https://doi.org/10.3969/j.issn.1000-3428.2016.12.001
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    In the process of Wireless Sensor Network(WSN) node positioning,it is more important to improve the x-axis and y-axis positioning accuracy of an unknown node than z-axis positioning accuracy,so this paper proposes a positioning algorithm based on planar projection on the basis of RSSI Gaussian Mixed(RGM) algorithm.It improves unknown node positioning accuracy on the x-axis and y-axis directions by reducing the projected area of spatial positioning error region.Simulation results show that,compared with RGM algorithm,the proposed algorithm can improve the x-axis and y-axis positioning accuracy without any inscrease in time complexity and network traffic.

  • WEI Juyi,ZHANG Jianjun,WEI Zhenchun
    Computer Engineering. 2016, 42(12): 5-10,15. https://doi.org/10.3969/j.issn.1000-3428.2016.12.002
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at the problem of high overhead and unbalanced energy consumption caused by exchanging majority of messages for higher accuracy in typical algorithms,an energy-balanced and low-overhead time synchronization algorithm for wireless sensor networks is proposed.Based on a hierarchical topology,cluster heads accomplish the process by pair-wise message exchange mechanism with the Sink node,and adopt maximum likelihood estimation for the clock offset and skew correction,thus improving precision and decreasing the times of the synchronization process.Considing broadcasting properties and remaining energy,cluster members synchronizes to their cluster head by adopting pair-wise packet exchange and reference broadcast mechanism,thus reducing numbers of transmission packets and balancing the energy in the network.Simulation results show that the proposed algorithm relatively has less packet cost,balances the energy consumption and extends the network lifetion.while guaranteeing the precision.
  • HUANG Qingyu,LIU Xinhua
    Computer Engineering. 2016, 42(12): 11-15. https://doi.org/10.3969/j.issn.1000-3428.2016.12.003
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at the problem that the traditional three-dimensional localization algorithm for sensor network nodes produce a singular matrix and high complexity,an improved localization algorithm based on three-edge measurement is proposed to realize node localization for Wireless Sensor Network(WSN),which is suitable for 3D localization in wireless sensor networks.The proposed algorithm obtains offset vector volume by tetrahedron formula.The target location of the 2D Linear Least Squares Estimation(LLSE) transformation is applied on the plane vector rotation.The target location is calculated by target position and offset vector.Simulation results show that compared with the traditional 3D Levenberg Marquardt (LM) localization algorithm,the proposed algorithm improves the accuracy by about 10% and not singular matrix,and the complexity is lower.
  • LUO Yihua,ZHANG Yu,PEI Xuming,CAO Wei
    Computer Engineering. 2016, 42(12): 16-20,25. https://doi.org/10.3969/j.issn.1000-3428.2016.12.004
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Intelligent meter reading system based on Wireless Sensor Network(WSN) can improve efficiency,but it has higher power consumption.Therefore,its nodes need to work under listen-sleep mode to reduce the power consumption.For this working mode,a rapid node wake-up method for WSN is proposed for hand-held meter reading and the system is designed.With the proposed method,higher communication rate is improved between handsets and sensor nodes.As a result,the average current in sensor nodes can be reduced significantly.Performance analysis result shows that the proposed method can reduce the average current.Furthermore,the proposed method yields obvious advantage in saving power consumption when the listen-sleep period is short.Hence,the proposed method is suitable for hand-held meter reading,which generally requires short responding time from sensor nodes.

  • YAN Jingjing,ZHANG Tengfei
    Computer Engineering. 2016, 42(12): 21-25. https://doi.org/10.3969/j.issn.1000-3428.2016.12.005
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The cluster head nodes selected by Low Energy Adaptive Clustering Hierarchy(LEACH)protocol suffer from imbalance ofdistribution and energy consumption.In view of these problems,this paper proposes an improved LEACH algorithm based on Rough C-means(RCM) clustering.Firstly,using location technology,it gets the position coordinates of the nodes,and then the nodes’ locations of Wireless Sensor Network(WSN)are clustered by the RCM clustering algorithm,which can guarantee uniform distribution of cluster heads in the location.Secondly,considering the residual energy of nodes,it selects cluster head nodes by the LEACH algorithm.Calculation results show that,when choosing cluster head nodes in each round,the improved algorithm can guarantee the cluster head nodes be evenly distributed in the network,and prolongs the network life cycle.
  • WANG Liutao,XIA Dongliang,WANG Jianxi,MA Fei
    Computer Engineering. 2016, 42(12): 26-31,38. https://doi.org/10.3969/j.issn.1000-3428.2016.12.006
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to exchange information of partial likelihood function in a distributed target tracking,several common distributed target tracking methods are studied,and a Distributed Particle Filter method based on Belief Propagation(DPF-BP) is proposed.The maximum diameter of the graph is calculated in a limited number of iterations.In order to avoid difference in network assessment,consistency maximization is used before assessing and nodes and the number of iterations are expressed as a function.After standardization and valuation calculations,the replacement is re-sampled.Simulation experimental results show that,compared with Standard Belief Consensus(SBC),Randomized Gossip(RG)and Metropolis Belief Consensus(MBC),under the condition of the same configuration,DPF-BP is excellent at RootMean Square Error(RMSE).In addition,DPF-MBC is best in the circular network,and DPF-BP is best in the tree network.
  • LIN Zhigui,ZHANG Xiaohui,LIU Yingping,LI Min
    Computer Engineering. 2016, 42(12): 32-38. https://doi.org/10.3969/j.issn.1000-3428.2016.12.007
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    When events distribute nonuniformly in the wireless sensor monitoring network,hot spots will appear in data-centric storage algorithms.Aiming at this problem,this paper introduces storage thresholds and event priorities,considers nodes’ remaining memory when selecting storage nodes,and proposes a new snake-like slot time data storage algorithm.It determines the distance from event storage location to query node grid according to the event priority,to reduce node’s energy consumption in process of data storage and query.Defining a node storage threshold,nodes reaching the threshold are no longer involved in the next round of time slot allocation.When all nodes in a grid reach the storage threshold,it stores data into neighbor grids with the same priority level.When all nodes in all grids of the same priority reach one storage threshold,it reallocates work slot,which effectively reduces the nonuniform distribution.Simulation results show that the proposed algorithm has better performance than the Snake-like Slot Time Power-saving Data Stroage Algorithm based on Event Priority(P-SLPS) in residual energy of nodes and the network life cycle.

  • LIANG Qing,JIAO Feng
    Computer Engineering. 2016, 42(12): 39-43. https://doi.org/10.3969/j.issn.1000-3428.2016.12.008
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to improve the survival period of wireless sensor networks,a data collection protocol for Wireless Sensor Network(WSN) based on dichotomy and mobile Sink called PBDM is proposed.The network is divided into two equal subdomains by the dichotomy,the intersection of two domains is the trajectory of mobile Sink.With the increase of mortality rate of sensors,and the inner subdomain is divided into two equal areas by the dichotomy,which can identify and change the trajectory of the mobile Sink.The entire network is controlled by topology,where static and mobile Sink nodes coexist.Cluster heads collect interests in the cluster and send the data to the nearest Sink with minimum hop counts.Simulation results show that PBDM is better than LEACH in the base station to obtain information,the remaining energy of the network and the number of surviving nodes.It can improve the efficiency of data collection and prolong the lifetime of WSN.
  • LIANG Jun,LI Wei,XIAO Lin,XU Xinkai
    Computer Engineering. 2016, 42(12): 44-49. https://doi.org/10.3969/j.issn.1000-3428.2016.12.009
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    During the transplantation and optimization of the heterogeneous computing platform,memory access performance of digital image data algorithm becomes the main factor.In order to solve the problem,this paper combines with the NVIDIA Tegra K1 hardware architecture’s characteristics and the specific algorithm’s characteristics,reserches the implementation and memory access performance optimization of matrix transpose and Laplace filtering algorithms on the NVIDIA Tegra K1 heterogeneous computing platform from memory access optimization of consolidation and vectorization,eliminating global memory access’s bank and channel conflict etc.Experimental result shows that the performance of matrix transpose and Laplace filtering algorithms on the NVIDIA Tegra K1 heterogeneous computing platform has an obvious improvement,and has good real-time performance.
  • DU Ran,HUANG Qiulan,CHENG Yaodong,CHEN Gang
    Computer Engineering. 2016, 42(12): 50-59. https://doi.org/10.3969/j.issn.1000-3428.2016.12.010
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    HazelNut is a block-based hierarchical storage system,which requires to manage multiple types of storage devices and different ways of I/O trace and block migration.As a result,a diversity equipment management mechanism(HNDisk) and diversity trace migration mechanism(HNType) are designed and implemented to meet these requirements.HNDisk is able to decentralized organize multiple storage devices in tiers with no constraints on device types and device amount.Meanwhile,HNType can add the type module according to the common interface,and it is easy to configure and select different trace metrics and migration algorithm.At the same time,it tests the influence of different trace metrics on the performance of btier average migration algorithm on the base of HNDisk and HNType.Experimental result show that trace metrics of access times and access sectors are not linearly correlated,as a result,these two metrics bring different migration results.Besides,I/O requests completion time used as trace metrics is able to improve sequential I/O by at least 10%.
  • CHEN Weijian,GUO Yong,ZHU Ying,YIN Fei
    Computer Engineering. 2016, 42(12): 60-65,72. https://doi.org/10.3969/j.issn.1000-3428.2016.12.011
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A domestic processor Micro Architecture Simulator(MASim) cannot support Simultaneous Multithreading(SMT) technology.It is difficult to meet the research and engineering requirements.Therefore,a new micro architecture simulator is proposed.The combination of MASim and SMT technology is adopted to design and realize the technology of modularization and queue optimization from the aspects of architecture simulation,simulation speed,simulation precision and flexible controllability.Experimental results show that the simulation accuracy and the simulation speed are almost the same as those of the original single-thread micro architecture simulator.SMT technology has the function of enhancing the performance of the series of domestic processors.The number of Instructions Per Clock Cycle(IPC) is increased by an average of 29.61%.
  • WANG Yong,LI Bingqian,LIU Da
    Computer Engineering. 2016, 42(12): 66-72. https://doi.org/10.3969/j.issn.1000-3428.2016.12.012
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To improve the information interaction ability of the Integrated Core Processor(ICP) in Integrated Modular Avionics(IMA) system and achieve the goal of Fibre Channel(FC) unified avionics network,ICP’s FC network interface card based on Field Programmable Gate Array(FPGA) is designed.In FPGA,it designs modules of high speed transmitting/receiving serdes and FC interface logic are designed and implemented.Meanwhile,modules including clock synchronization,power,internal memory and others are set to form a completely embedded system.ICP arithmetic element point to point connection is made to do Fast Fourier Transform(FFT) test,and to estimate the affect of shared computing resources for FC interface test slabs to ICP computing unit.Experimental results show that FC interface board transfer rate is faster,compared with the single ICP computing unit independent operation,ICP computing unit parallel operation can effectively reduce the operation time.
  • WANG Wenjing,FENG Rui
    Computer Engineering. 2016, 42(12): 73-77,83. https://doi.org/10.3969/j.issn.1000-3428.2016.12.013
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Test suite reduction is an important research questions in software test,its purpose is to test as little as possible to achieve the test objective.This paper presents a test suite reduction method based on bisecting K-means.It uses the bisecting K-means clustering algorithm to reduce regression test suite.Regarding the path coverage of white box test as a criterion,each test case is quantified,so that each case becomes a point,the number of black box test functional needs is taken as the number of clusters,each cluster is sorted in the clustering results according to the distance from the center,the test cases from each cluster are selected,until all testing requirements are met and reduced test suite is got.Simulation experimental results show that this method can effectively reduce the size of the test suite,and effectively reduce the impact on the use case set error detection rate.

  • LI Qiongying,XIA Yinshui,ZHANG Junli
    Computer Engineering. 2016, 42(12): 78-83. https://doi.org/10.3969/j.issn.1000-3428.2016.12.014
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    When the power consumption estimated by the probability power estimation method is used as the cost function for power optimization,the limitations of the methods themselves or ignoring the characteristics of the circuit node lead to lower accuracy of the estimation results and then affect the power optimization results.Aiming at these problems,a new power estimation method is proposed.It estimates the power effectively by giving the signal probability and transition density of input signals,and then classifying the characteristics of Reduced Ordered Binary Decision Diagram(ROBDD) nodes based on ROBDD representation.Experimental results show that this method can better predict the power consumption of the circuit,and the power consumption estimation precision is superior to that of the probability and conditional probability estimation methods.
  • ZHAO Siyi,CHEN Jing,GONG Shuiqing
    Computer Engineering. 2016, 42(12): 84-90. https://doi.org/10.3969/j.issn.1000-3428.2016.12.015
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at Software Defined Network(SDN) virtualization environment,an efficient virtual SDN network mapping algorithm is proposed to minimize network mapping overhead.Based on the Particle Swarm Optimization(PSO) algorithm,it redefines the position,velocity and related operations of the particles according to the model of the virtual SDN network,and the mapping problem is solved by iterative optimization of particles.Simulation results indicate that the algorithm improves the resource utilization and the acceptance ratio,and reduces the average controller-to-switch delay.
  • TAO Yang,LIU Xiaohu,TAO Ling
    Computer Engineering. 2016, 42(12): 91-96. https://doi.org/10.3969/j.issn.1000-3428.2016.12.016
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Most vertical handoff algorithms do not think about the actual demands of network and the mobility of user,but take network properties as the standards of judgment.In order to solve the above problems,the switching decision algorithm based on Fuzzy Analytic Hierarchy Process(FAHP) is proposed.By analyzing the influencing factors of business,ituses FAHP for different business to select the optimal network to switch.Simulation results show that the algorithm is able to make the switching decision more reasonably,can ensure the Quality of Service(QoS)of application requirements,improve the stability of the switching decision,reduce the error switching times and the network blocking probability,and improve the communication satisfaction of users.
  • LI Li,WU Zhenqiang
    Computer Engineering. 2016, 42(12): 97-101. https://doi.org/10.3969/j.issn.1000-3428.2016.12.017
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Network coding is an efficient way of implementing anonymous communication,but it cannot improve the throughput in multiple unicast networks.Aiming at this problem,this paper proposes an anonymous communication scheme under multicast network,combined with the butterfly network model.It gives the visual implementation in Matlab GUI by using the principle of linear space,information division and finite field construction in Matlab GUI platform.Analysis result shows that,this scheme not only improves the throughput,but also resists flow analysis attack.It can realize encoding confusion.
  • HUANG Peng,HUANG Chen,LIAO Xinqi,TANG Lun
    Computer Engineering. 2016, 42(12): 102-107,111. https://doi.org/10.3969/j.issn.1000-3428.2016.12.018
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Inter-cell interference coordination management scheme cannot satisfy the bandwidth rate and Quality of Service(QoS) requirement of interference source and interference network in heterogeneous cellular network.Aiming at the characteristics of heterogeneous cellular network and wireless communication mechanism,a dynamic enhanced inter-cell interference coordination scheme is proposed.According to the actual situation of the current network load,the optimal ratio of Lower Power-Almost Blank Subframe(LP-ABS) can be determined.The global interference situation of area users is extended through the residential area,and the optimal transmission power of the macro base station in the LP-ABS state is obtained,which is used in transmit power of macro base station.The maximum interference value of user tolerant macro base station is used to decide whether to perform LP-ABS ratio optimization or not to the next scheduling cycle.Experimental results show that on the basis of ensuring the overall throughput of the system is not too large,the proposed scheme can effectively guarantee the rate of the entire network users and QoS need,solve the problem that the high proportion of ABS leads to low performance for the user of interference source network,and improve overall communication network performance from multiple angles.
  • YUAN Peng,FU Jielin
    Computer Engineering. 2016, 42(12): 108-111. https://doi.org/10.3969/j.issn.1000-3428.2016.12.019
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The short wave channel has characteristics of multipath,fading and time variant.In the equilibrium process,the convergence rate of Liner Equalizer(LE) is slow,the convergence precision is low,and the Fractional Spaced Equalizer(FSE) has better balance effect.So according to the waveform structure used in short wave burst communication,against the characteristics of short wave channel,this paper proposes a composite structure of T/4 FSE by employing Least Mean Square error(LMS) algorithm,Recursive Least Square(RLS) algorithm and square root Kalman algorithm.Simulation results show that in the short wave burst communication process,compared with traditional LE,the proposed FSE has faster balance speed and higher convergence precision.At the same time,the error characteristics of the communication system are significantly improved,which is more suitable for short wave channel.
  • ZHOU Jihua,PENG Hu,WU Jian,DUAN Rongchang,ZHANG Hanwen
    Computer Engineering. 2016, 42(12): 112-117,123. https://doi.org/10.3969/j.issn.1000-3428.2016.12.020
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Because of the characteristics of strong time variability channels and dynamic changed topology in wireless multi-hop network,power control mechanism should be simple and efficient.The theory proves that the transmit power affects the data send rate.There is a feasible solution to the power control mechanism based on the length of the send queue.According to the fact that intermediate nodes of wireless multi-hop network need help other nodes to carry out the data forwarding,based on the length of the send queue,this paper proposes a power control mechanism.The mechanism considers the problem of buffer overflow packet loss of send queue when intermediate nodes forward information,dynamicly adjusts power through extension of control message carrying transmit power,information of tolerant noise and so on,thus improves the reusability of space,optimizes the network throughput and the end-to-end delay.Simulation results on NS2 show that,compared with POWMAC mechanism and ATPMAC mechanism,the throughput of the proposed mechanism is higher by 20%~30%,and the end-to-end delay is lower by 10%~20%.
  • WU Guangxian,LIU Nianyi,LIU Boya
    Computer Engineering. 2016, 42(12): 118-123. https://doi.org/10.3969/j.issn.1000-3428.2016.12.021
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Aiming at the awful encryption expansion rate while encrypting a binary plaintext matrix in GHV scheme,a BGN-type public cryptosystem is proposed for p plaintext matrix.It gives the security proof and analyzes the Somewhat homomorphism of Indistinguishability under Chosen Plaintext Attack(IND-CPA),and the new scheme is applied to the construction of privacy information retrieval protocol.Compared with the original GHV scheme,analysis results show that the proposed scheme not only supports multiple homomorphic addition and a single homomorphic multiplication,but also makes some encoding modification in the plaintext space.The ciphertext extension rate is reduced to logq/logp from logq.
  • FANG Xuefeng,WANG Xiaoming
    Computer Engineering. 2016, 42(12): 124-128,132. https://doi.org/10.3969/j.issn.1000-3428.2016.12.022
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to reduce the computational costs of Attribute-based Encryption(ABE) scheme,this paper proposes an outsourced encryption and decryption Ciphertext-policy ABE(CP-ABE) scheme with user revocation.It delegates the complex computation of the encryption and decryption process to the cloud server to reduce the user’s encryption and decryption computational cost.Moreover,the proposed scheme implements user revocation and ciphertext update by using the Chinese remainder theorem,and makes the unrevoked users do not need to update their decrypted key.Analysis results show that,compared with the scheme proposed by Zhou et al.,the proposed scheme can improve computational performance,and it can be proved secure under general group model.
  • ZHONG Xianfeng,CAO Lei,LIU Weiyue,WANG Chaoze,LI Fengzhi,JIANG Zhidi,PENG Chengzhi
    Computer Engineering. 2016, 42(12): 129-132. https://doi.org/10.3969/j.issn.1000-3428.2016.12.023
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Quantum key distribution uses polarization coding.When using conventional single-mode fiber as the channel for quantum state transfer,optical fiber has a slight birefringence and cannot maintain the quantum state of polarization.In this case,this paper proposes a method to make polarization compensation for quantum state.Using the polarization compensation method of successive approximation,it is a promising scheme to complete the polarization compensation by rotating the combination of 2 quarter-wave plate and 1 half-wave plate successively.By measuring the values of polarization contrast ratio,it rotates the three wave plates according to the results of the measurement,until the polarization performance of the system finally meets the requirements.Experimental results show that compared with the polarization compensation method of traversal mode,the method of successive approximation can realize polarization compensation more quickly.
  • ZHONG Jianyou,CHANG Shan,LIU Xiaoqiang,SONG Hui
    Computer Engineering. 2016, 42(12): 133-138. https://doi.org/10.3969/j.issn.1000-3428.2016.12.024
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To protect the trace privacy of mobile objects,pseudonym is used to the anonymous processing of trace before the release of the trace dataset.However,the anonymous trace of pseudonym users still faces the risk of privacy leakage.This paper proposes a new de-anonymization attack method.If an attacker obtains several track segments of his attack target at present or in any future period,comparing the traces with the anonymous historical trace dataset,the historical traces of the attack target are identified.The characteristics of the real moving track data of the two groups are analyzed,and a de-anonymization method based on characteristic similarity is presented.The feature vectors of history trace are extracted based on improved Term Frequency-Inverse Document Frequency(TF-IDF) method.The dimension is reduced by Principal Component Analysis(PCA),and the feature matching is performed on the track segments obtained by the historical track and the attacker,to recognize the historical trace with the highest degree of similarity with the trace characteristics of attackers.Experimental results show that the proposed method can obtain higher accuracy.
  • WANG Caifen,KANG Burong
    Computer Engineering. 2016, 42(12): 139-144. https://doi.org/10.3969/j.issn.1000-3428.2016.12.025
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to improve the efficiency of the existing ID-based signcryption algorithm,taking into consideration the characteristic of low computing power of low-end devices,introducing the idea of server-aided verification,this paper presents a server-aided verification signcryption scheme.Complex operations in the verification process are carried out through the server,which effectively reduces the amount of computation and running time in the validation phase of the algorithm,and makes the application of the algorithm in the low-end device possible.Based on the difficult hypothesis of Decision Bilinear Diffie-Hellman(DBDH) problem and Calculation of Diffie-Hellman(CDH) problem,it proves the scheme meets the unforgeability and confidentiality in random oracle model.
  • WANG Ziwang,XU Yang,XIE Xiaoyao,ZHANG Shuai,CHEN Yi
    Computer Engineering. 2016, 42(12): 145-150. https://doi.org/10.3969/j.issn.1000-3428.2016.12.026
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Two types of schemes based on pseudonym authentication and group signature authentication are compared and analyzed.This paper proposes a hybrid authentication scheme.The model builds an exchange protocol for exchanging pseudonym between adjacent nodes.The node can apply for the exchange key from the Trusted Authority (TA) with the nearby nodes.TA uses asymmetric key to realize the unforgeability of exchange identity and pseudonym.The proposed protocol can effectively resist the collusion attack.The exchanged pseudonym also can be used to signature and authentication between trusted nodes.And the group signature is introduced as the identity attribute label.The group signature label ensures the message unforgeable and auditable.The theoretical and efficiency analysis shows that the proposed scheme introduces the white list mechanism for pseudonym.Under this mechanism,the verification efficiency of a single message is improved obviously.At the same time,the white list mechanism can effectively resist the replay attacks using pseudonym.

  • GUAN Xiuying,ZENG Weiming,WANG Nizhuan
    Computer Engineering. 2016, 42(12): 151-155. https://doi.org/10.3969/j.issn.1000-3428.2016.12.027
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    Affinity Propagation Clustering(APC) method shows its limitations in time complexity,data storage and clustering results while handling massive functional Magnetic Resonance Imaging(fMRI) data.Aiming at these problems,this paper proposes a new method named SDAPC,which combines Sparse APC(SAPC) with similarity matrix reduction.It starts from sparse approximation on fMRI data,continues with the density analysis on sparse data by Gaussian density function and Euclidean distance,and finally realizes the detection on the functional connectivity of reduced fMRI data.The task-related data experiment gets the following results:SDAPC produces a fine ROC curve for single subject while running about three times faster than SAPC.SDAPC and SAPC both get better ROC curves for multiple subjects than single subject.The resting-state data experiment leads to the further finding that SDAPC can successfully identify nine resting-state networks.

  • DING Xiancheng
    Computer Engineering. 2016, 42(12): 156-163. https://doi.org/10.3969/j.issn.1000-3428.2016.12.028
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    In view of the decision making problems with multiplicative preference information,the order consistency,acceptable consistency and consistency index of multiplicative preference relation are introduced.An optimization model is developed with the principle of minimum non-negative deviation variables.An algorithm to derive the ranking of alternatives in decision making is investigated based on the consistency of multiplicative preference relation.The main characteristics of the proposed algorithm are that:It helps decision makers make decision quickly and efficiently by using the properties of the order consistent multiplicative preference relation.It repairs the consistency of multiplicative preference relation by using the optimal deviation values and ensures that the repaired multiplicative preference relation is acceptable consistent.The priority weight vector is obtained,and it helps the decision maker to obtain the reasonable and reliable decision making results.A practical application to selection of comprehensive replenishment ship is demonstrated to show the rationality and effectiveness of the proposed algorithm.

  • PENG Tianqiang,LI Fang
    Computer Engineering. 2016, 42(12): 164-170. https://doi.org/10.3969/j.issn.1000-3428.2016.12.029
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Constructing visual vocabulary in the Bag of Visual Word(BoVW) model is a critical step,most visual vocabulary is generated by the k-means algorithm or its improved algorithm.Because of the limitation of the k-means algorithm and the complexity and the high-dimensionality of the sample space,visual vocabulary generated by these methods have the problem of low discriminative long running time and without space information.For these problems,a BoVW model is proposed based on binary Hashing and space pyramid,which can shorten the visual vocabulary generation time sharply.It extracts the local feature points from the images,learns binary Hashing functions,which map the local feature points into visual words,filters the visual words and generates the visual vocabulary whose visual word is binary hash code.The new BoVW model is composed with the visual vocabulary and Spatial Pyramid Matching(SPM) model,which represents the images by the histogram vector of space pyramid and is applied in image classification and retrieval.Experimental results on the common datasets show that the model has higher classification accuracy and retrieval performance.
  • WU Ke,ZHAN Yinwei,LI Ying
    Computer Engineering. 2016, 42(12): 171-175. https://doi.org/10.3969/j.issn.1000-3428.2016.12.030
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To solve the problems of data sparsity and cold start on latent factor model recommendation algorithm,a recommendation algorithm fused with user attribute information is presented.After adding binary user attributes to the latent factor model,it finds similar users according to the attributes of the target user and importance degrees of other user attributes measured by classification model.Combining with the target user’s rating information,the recommendation result is obtained.The proposed method is tested by the Movielens data set,and the result shows that it outperforms collaborative filtering algorithm and latent factor model algorithm on accuracy of recommendation results even when data is extremely sparse.
  • HE Xiping,ZHANG Qionghua,LIU Bo
    Computer Engineering. 2016, 42(12): 176-180,187. https://doi.org/10.3969/j.issn.1000-3428.2016.12.031
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To improve the feature extraction timeliness and classification validity of real-time classification of visual object in low computing profile,a Histogram of Oriented Gradients(HOG)-based feature deep learning model for object classification is proposed.For the requirements of high timeliness,offline deep learning strategy is applied to the classifier model to save its online training time.In view of the requirements of network depth limitation and high recognition rate,the local feature of HOG feature of an image is extracted to be used as the input of the sparse autoencoder stack so as to output the high level feature code of the sample image.The Softmax multiple classifier is designed to classify the extracted features.During the learning process of the deep neural network model,the two-stage optimization strategy is introduced,which minimizes the structural risk of every layer and fine-tune the parameters of the whole model.Using some samples of the scene image database Caltech101 and that of the handwritten digits database MNIST as the training set and the others as the test set to perform the comparative experiment,results show that the time performance of the proposed model is better than that of one-layer only Convolutional Neural Network(CNN),and the classification accuracy of the trained model is higher than that of CNN,Stacked Autoencoder(SAE) comparative models.
  • YE Shiren,SUN Ning
    Computer Engineering. 2016, 42(12): 181-187. https://doi.org/10.3969/j.issn.1000-3428.2016.12.032
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    This paper proposes an improved Chinese text classification algorithm by Domain Ontology Graph(DOG)based on semi-supervised concept clustering.According to the DOG structure model,the ontology learning framework of Chinese classification is created,and then the HowNet dictionary is used to extract the word disambiguation,and then the Chinese term-term relationship mapping is established.Based on the weight connection between the terms of the binary classification relationship,here designs the KLSeeker ontology Chinese text classification algorithm.It realizes accurate classification of Chinese text through DOG semi-supervised learning.Experimental results show that the proposed algorithm has a higher classification accuracy compared with the Chinese text classification algorithm based on non-negative tensor decomposition ontology concepts.
  • PENG Yan,XI Liya
    Computer Engineering. 2016, 42(12): 188-195,203. https://doi.org/10.3969/j.issn.1000-3428.2016.12.033
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Community detection is an important image mining problem,and the overlapping community detection and the community evolution pattern in dynamic networks is a hot topic.However,There is no work on the integration of the local-first community detection.This paper proposes a local-first analytical approach to analyze the evolution pattern of the overlapping communities on dynamic networks.This approach detects the local Ego community via label propagation on each network snapshot.Then,the global communities are generated via merging the local Ego communities successively.In this process,the two concepts of the community similarity and community association are introduced,and it analyzes the overlapping community evolution pattern via an evolutionary clustering framework.Experimental results show that this approach not only detects the overlapping communities in dynamic networks effectively,but also analyzes the evolutionary pattern.
  • WU Zhonggang,Lü Zhao
    Computer Engineering. 2016, 42(12): 196-203. https://doi.org/10.3969/j.issn.1000-3428.2016.12.034
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Many existing community detection algorithms focus on topological structure or node attributes.Some attribut graph clustering algorithms consider both of them but the quality of community is not good.Shared neighbors based local similarity algorithms underestimate pairwise of node similarity.Hence,this paper proposes a new Local Similarity based Community Detection(LS-CD)algorithm.The proposed algorithm contains two main components:node local similarity calculation and node clustering.It evaluates the vertex importance using the Pagerank algorithm and calculates the similarity of pairwise vertexes by combining connetion strength and node attribute.To avoid underestimating node similarity based on shared neighbors,the similarity of vertexes is calculated by the similarity of their local neighborhoods.The K-Medoids clustering algorithm is used to identify community by measuring the local similarity of node and cluster centroid.Experimental results show that,compared with traditional SA-Cluster and k-SNAP algorithms,this algorithm can mine high quality community and has good community identification effect.
  • LIN Chen,GU Junzhong
    Computer Engineering. 2016, 42(12): 204-207,216. https://doi.org/10.3969/j.issn.1000-3428.2016.12.035
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The major performance bottlenecks of most maximal frequent itemset mining algorithms based on FP-Tree are caused by recursively traversing and constructing conditional FP-Trees and superset check.Therefore,this paper proposes a maximal frequent itemset mining algorithm based on Nodeset named MFIN.Instead of recursively traversing and constructing conditional FP-Trees,this algorithm adopts a novel data structure called Nodeset to encode the nodes of POC-Tree and uses a set-enumeration tree to represent search space.Besides,this paper proposes an early-stopping method to minimize the cost of intersection operation of Nodesets and adopts parent equivalence pruning technique and look-ahead pruning technique to reduce the search space.The efficiency of superset check method is promoted by improving the projection strategy based on MFI-Tree.Experimental results show that the performance of MFIN algorithm is superior to the classic FP-Max algorithm which is based on FP-Tree in running time and efficiency on mushroom,pumsb and webdocs datasets.
  • KOU Yanmin,WANG Junfeng,XUE Xiao
    Computer Engineering. 2016, 42(12): 208-216. https://doi.org/10.3969/j.issn.1000-3428.2016.12.036
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    Through the dynamic matching and composition of all kinds of manufacturing service resources,Cluster Supply Chain (CSC) can help manufacturing enterprises face the changing market competition.However,the current service composition strategy focuses on the selection of the optimal service composition,but neglects the common phenomenon that the service resources are idle.In view of this situation,this paper uses the method of computational experiment,with cluster collaborative manufacturing as the research object to simulate supply and demand uncertainty evolutionary process of non-collaborative service composition strategy and collaborative service composition strategy,and makes a comparative research of the two service composition strategies.Experimental results show that,under the market environment of supply and demand uncertainty,the performance of collaborative service composition strategy is optimal.
  • LI Na,WANG Hongyuan,WANG Jia
    Computer Engineering. 2016, 42(12): 217-221. https://doi.org/10.3969/j.issn.1000-3428.2016.12.037
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to solve the problem of object matching based on non-overlapping camera views,a feature extraction model based on pedestrian’s main color statistical information and spatial distribution information is put forward.The main color feature is obtained by clustering the target through the nearest clustering method in HSV color space.Each part of the pedestrian is weighted based on the pedestrian’s main color feature.And it determines whether the two targets’ are matched by setting the threshold.Experimental results show that the method can achieve reasonable similarity value and eliminate undesired color matching.It is capable of reducing false matching of different pedestrians with similar information and has good robustness to the change of shape and gesture.
  • ZHOU Yan,ZHOU Yuan,WANG Xuhui
    Computer Engineering. 2016, 42(12): 222-227. https://doi.org/10.3969/j.issn.1000-3428.2016.12.038
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    To improve the fusion effect of multi-source grayscale image,a new grayscale image fusion algorithm is proposed combining the shift invariance and good directional sensitivity of FDST.The original images after registration are decomposed by FDST,and the low frequency sub-band coefficients and high frequency sub-band coefficients of different scales and directions are obtained.The combination fusion algorithms of regional average energy and average gradient are used for the low frequency,and the combination fusion algorithms of relative region variance and average gradient are used for the high frequency.The fused image is reconstructed by inverse transform of finite discrete shear wave,and the result is evaluated by subjective vision and objective performance.Experimental results show that the algorithm can get better fusion results and image detail extraction compared with low frequency regional energy fusion and high frequency region variance fusion algorithms based on wavelet transform.

  • HUANG Xiaoxia,GU Yuzhang,ZHAN Yunlong,ZHAO Luyang
    Computer Engineering. 2016, 42(12): 228-234. https://doi.org/10.3969/j.issn.1000-3428.2016.12.039
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The traditional climbing detection method is based on monocular vision,which has specific limit and low accuracy.In order to detect the climbing behavior more accurately in video surveillance,this paper proposes a climbing detection method based on binocular stereo vision.The feature points of both rectified images are extracted and the vision disparity of the point is obtained by the sparse stereo matching method.The 3D coordinates of the points are calculated based on the standard 3D measurement theory.Targets are detected by points clustering.According to targets’ location and modified joint probabilistic data association method,targets can be tracked.Based on targets’ location,the trajectory and moving direction are analyzed to detect the behavior of climbing.Experimental result shows that the method has high accuracy in the conditions of complex background and the average accuracy is 94% with appropriate parameters.
  • CHEN Tongqian,DAI Yong,YUAN Dibo
    Computer Engineering. 2016, 42(12): 235-240,247. https://doi.org/10.3969/j.issn.1000-3428.2016.12.040
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    For the existing related works cannot be applied to automatic teaching of hand-drawn graph with high freedom degree,this paper proposes an automatic teaching method for fuzzy evaluation of line drawing quality based on open-loop and closed-loop line graphs.This method defines a high freedom degree drawing of open-loop and closed-loop lines and their unique relationships.A formal structure of knowledge graph is put forward.It utilizes the similarity method to recognize the drawing structure roughly.Fuzzy evaluation of the drawing quality is carried out according to the key points,stroke segments,scale,angles,radians and centroid distances.The difference is identified and marked.The homework checking contrast experiments between teachers and system show that the structure identification,quality evaluation and differential label of the method can reach the practical requirement.The directive opinion is objective,which achieves practical effect.
  • CHU Heng,WANG Xing,LIU Hongbin,CHAO Shuanshe
    Computer Engineering. 2016, 42(12): 241-247. https://doi.org/10.3969/j.issn.1000-3428.2016.12.041
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    As the spectral distortion is prone to exist in the vegetation area during the process of image fusion,a new method is proposed for image fusion of resource satellite-3 remote sensing by analyzing different characteristics of several common fusion methods.Normalized Difference Vegetation Index(NDVI) is used to separate green vegetation areas and non-green vegetation areas from original images.A combined fusion strategy is built according to different fusion characteristics and requirements of the two different areas to accomplish fusion process.For non-green vegetation areas,it constructs the linear relationship between multispectral image and HP obtained after an image processing to panchromatic image,and fittings LRP by calculating the correlation coefficient to implement fusion.In green vegetation areas,it carries out image fusion by using GS method for an advantage in spectral fidelity.Experimental results show that compared with the fusion methods of FIHS,GIHS,Brovey,GS and divided region fitting method,the new method based on combined fusion strategy has the best fusion quality.It can obtain a relatively high spatial resolution while reducing spectral distortion of fused image effectively,especially in solving the problem of spectral distortion in green vegetation area.
  • ZHANG Yubing,SONG Wei
    Computer Engineering. 2016, 42(12): 248-253,261. https://doi.org/10.3969/j.issn.1000-3428.2016.12.042
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    For retrieval accuracy of Text-based Image Retrieval(TBIR) method is not high,this paper proposes a new image retrieval method.The hue and saturation components are reasonably quantified in HSV color space to make more detailed description for color difference.Local Diagonal Color Difference(LDCD) descriptor is extracted on the luminance component to reveal the spatial distribution information of color.Composite color feature and color spatial distribution information are retrieved,which solves the problem that color histogram cannot describe the color spatial distribution information.Experimental results show that this method can achieve higher retrieval accuracy and reduce the time required for image retrieval by selecting the Corel image database.
  • ZHANG Na,LIU Hui,SHANG Zhenhong,ZOU Binyi
    Computer Engineering. 2016, 42(12): 254-261. https://doi.org/10.3969/j.issn.1000-3428.2016.12.043
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to solve inflexible rotation and translation in measuring neighborhood similarity for structural characteristics in Non-local Means(NLM) denoising method,this paper presents the NLM image denoising algorithm based on Normalized Moment of Inertia(NMI).The normalized moment of inertia as a part of the similar weight is proposed to balance neighborhood patches’ similarity.Experimental results show that compared with the original nonlocal mean denoising algorithm,this algorithm can keep the detail of the image structure information preferably,and can improve the Peak Signal to Noise Ratio(PSNR) and structural similarity.Especially,when the noise intensity increases,the algorithm improves the PSNR by 1.03 dB and the structure similarity index by 0.04.
  • XU Nini,HE Xiaotong,XU Shiyin
    Computer Engineering. 2016, 42(12): 262-267. https://doi.org/10.3969/j.issn.1000-3428.2016.12.044
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to further improve the compression ratio of 12-bit static image,a static image compression scheme based on M encoder is presented.On the basis of JPEG standard coding framework,the 12-bit gray image quantization table is added in this compression scheme.The M encoder,which is the arithmetic coding engine of the CABAC is improved and used as the coder for static image compression.The compressed code stream is encapsulated in byte stream format,which is similar to those presented in H.264 and H.265.This scheme is tested on a large number of 12-bit gray images.Experimental results show that with the same quality factor,the image compression ratio of this scheme is higher than that of the Huffman coding scheme and QM coding scheme.
  • WANG Bo,YU Fengqin
    Computer Engineering. 2016, 42(12): 268-271. https://doi.org/10.3969/j.issn.1000-3428.2016.12.045
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In order to overcome the defect that sample entropy can be falsely detected due to its sensitivity to the suddenly changing noise,this paper proposes a speech endpoint detection algorithm.This algorithm does the multi-scale transform for the speech signal in the time domain.The sample entropy and threshold of different scales can be calculated.The number of the sample entropy which is greater than the threshold of corresponding scale is counted and compared with the number of total scale to realize speech endpoint detection.Experimental results show that this algorithm can eliminate the mutation noise sensitivity of the sample entropy,and the detection accuracy is well improved in the low Signal Noise Ratio(SNR) conditions,compared with approximate entropy and sample entropy detection algorithms.
  • YUAN Hui,ZHANG Dawei,ZHANG Ke,ZHAN Yongsong
    Computer Engineering. 2016, 42(12): 272-277. https://doi.org/10.3969/j.issn.1000-3428.2016.12.046
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    An audio-video fusion detection method is proposed in this paper for indicator light security incident surveillance in key areas of the electrical room.The optimal indicator range is determined by color image nonlinear transform for monitoring area and OTSU automatic threshold segmentation technique.The local gray scale histograms are taken as the video feature vector,and the Mel-frequency Cepstral Coefficien(MFCC) for multi frame is selected as the audio feature vector.The audio feature vector and the video feature vector are combined into a relatively large-scale vector.Principal Component Analysis (PCA) is employed for dimensionality reduction of the fused feature vector,and Support Vector Machine(SVM) is used to classify and detect the security incidents.Experimental result show that the proposed method obtains higher detection rate and lower false positive rate over the traditional single security incident detection(only audio-based or video-based) methods.
  • GE Yongkan,YU Fengqin
    Computer Engineering. 2016, 42(12): 278-281,289. https://doi.org/10.3969/j.issn.1000-3428.2016.12.047
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The excitation signal used in the traditional Hidden Markov Model(HMM)-based speech synthesis algorithm is either a pulse train or white gaussian noise,and the synthesis speech sounds buzzy.An improved speech synthesis algorithm based on harmonic plus noise excitation model is proposed to enhance the quality of speech.After inverse filtering,the harmonic signal in glottal flow is extracted and modeled by Linear Spectrum Pairs(LSP) coefficients.The LSP coefficients are sent into HMM training as the harmonic feature.In synthesis stage,the harmonic part and the noise part are reconstructed from the newly generated coefficients and mixed together as the excitation of speech signal.Experiment results demonstrate that the excitation generated by this algorithm is more accurate compared with speech synthesis algorithm based on pulsed excitation.The algorithm can effectively relieve machine noise of synthesized speech and improve the naturalness of the speech.
  • QIN Tian,ZHAO Hui
    Computer Engineering. 2016, 42(12): 282-289. https://doi.org/10.3969/j.issn.1000-3428.2016.12.048
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In 3D face synthesis,the lip movement is the most expressive.But in Uyghur lip animation,there has not been a realization method of 3D lip animation system.Therefore,this paper proposes a lip animation system.On the basis of the standard model and according to the current mainstream thinning algorithm,it uses the formula transformation to get the new transformation and model driven method,realizing fast refinement and the motion of the refined model.Experimental results show that the system can achieve the generation of facial details and smooth movement of the transition,so that the detail of the model is full,with a strong sense of reality.
  • JIA Junjie,CHEN Fei
    Computer Engineering. 2016, 42(12): 290-294,301. https://doi.org/10.3969/j.issn.1000-3428.2016.12.049
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    In the information publishing of digital library,user identity information is the important basis for data analysis.This paper builds an anonymous data publishing,makes reasonable selection for user attributes,and partitions the user identity to get the user class by anonymous way.Through connection of user classes and borrowing information,the mapping relationship between user identity and borrowing information is obtained,thus releasing the anonymous data which retains user identity.Experimental results show that under the premise of user security,the anonymity publishing which retains user identity can better realize usefulness of digital library data,and improve the quality of service of digital library as a whole.
  • WANG Yongjia,BAI Ruilin,JI Feng
    Computer Engineering. 2016, 42(12): 295-301. https://doi.org/10.3969/j.issn.1000-3428.2016.12.050
    Abstract ( ) Download PDF ( )   Knowledge map   Save

    In order to improve the Delta robot movement speed,a method of time-optimal trajectory planning is proposed with the combination of off-line optimization and online inquiry.Griding the working area on robot conveyor belt,it selects the center point of each grid as a standard,uses linear and circular interpolation in cartesian space for “door” type track to get the discrete location sequence,and calculates the discrete angle sequence in joint space through the inverse kinematics.It uses the particle swarm algorithm and gravitational search algorithm,based on fitness function of time optimal and under the premise of the joint velocity,acceleration,jerk continuous and smooth and constraints,constructs 7th order B-spline curve to get the off-line interpolation sequence of the angle-time in joint space.The optimal timing sequence is obtained by online query of 3D arrays.Experimental results show that the method is simple and easy.Delta robot in laboratory picks up object from the conveyor belt to the fixed position,and the time range is 0.676 1 s~0.786 9 s,so the method overcomes the shortcomings of low speeds of traditional planning method.

  • LIAN Yueyong,ZHANG Chao,ZHAN Yinhu
    Computer Engineering. 2016, 42(12): 302-306,313. https://doi.org/10.3969/j.issn.1000-3428.2016.12.051
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    A new method of star map simulation is proposed to test and evaluate the performance of star sensors.To start with,a star position calculation model based on the spacecraft is established,and stars are drawed in the celestial sphere according to the star point model.Then the field of view is changed on any attitude by the imaging method of star sensor.OpenGL’s perspective projection function is used to automatically image the star which can be seen in the field so as simplify the step of navigation star search and improve the simulation efficiency.The apparent position calculation model used for simulation is more precise,and the star model dispersion effect is better,which improves the centroid sub-pixel location accuracy.Simulation results show that the standard deviation of star positioning is less than 0.023 pixel and lower star angular distance error in the 32° field of view remains,which displays the high precision of star position and high imitation of real starry sky.
  • MA Jinxing,CHENG Cheng,XU Yimin,LENG Wenhao
    Computer Engineering. 2016, 42(12): 307-313. https://doi.org/10.3969/j.issn.1000-3428.2016.12.052
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    The hull temperature field is the core issue in the design of Liquefied Natural Gas(LNG) carrier.In the iteration process,large amount of data and large number of iterations lead to long computation time,unexpected MSC.Patran quit.In view of these,an efficient iterative calculation optimization method of temperature field is proposed.This method uses Patran Command Language(PCL) of MSC.Patran to obtain the load data and generate external files,and realizes the Bulk Data File(BDF) automatic generation algorithm by using the C++ Standard Template Library(STL) programming to reduce iterative calculation time.The temperature filed calculation experiment is carried out on a prismatic LNG carrier with the capacity of 220 000 m3.Results show that the new method ensures correctness and improves the efficiency of iterative calculation.
  • WANG Zhen,XU Xiaohang,WANG Jing,LI Sheng,ZHENG Hong
    Computer Engineering. 2016, 42(12): 314-320. https://doi.org/10.3969/j.issn.1000-3428.2016.12.053
    Abstract ( ) Download PDF ( )   Knowledge map   Save
    To improve the real-time performance of high-definition video stitching,a stitching method of GPU-based real-time multi-channel high-definition YUV video is proposed.Firstly,the perspective model of YUV422 image stitching is deduced,and then parallel optimization of the stitching steps including perspective warping and image blending are realized and performed on GPU by Compute Unified Device Architecture(CUDA) technology.The experimental results show that,on the condition of implementing four-channel 1080p video stitching,the method in this paper achieves 20%~40% improvement in real-time performance over RGB color model-based stitching method on different GPU,and video frame rate of 33 frame per second on GTX 780.