15 October 2016, Volume 42 Issue 10
-
Select all|
-
Computer Engineering. 2016, 42(10): 1-5. https://doi.org/10.3969/j.issn.1000-3428.2016.10.001
3D voxel model reconstructed from images often suffers from problems of noise and incomplete building structure.Layered contour fitting method provides a way to solve this kind of problem.However,the cross planar contours are very complicated in the realistic scene.In view of this,this paper proposes a two-step contour fitting method to fit the cross-sectional contours of a building on the layer image.The hierarchical projection method is used to project the initial 3D model into a two-dimensional layer image which is then de-noised by a contextual density-based clustering algorithm.Precise fitting results of planar contours are obtained by shape classification and shape fitting.The final 3D model is composed of the results from the upper edge of the layer image.Experimental result shows that compared with the original reconstruction model,the precise building model has more regular and complete shape with less noise,and substantially reduces storage space.
-
Computer Engineering. 2016, 42(10): 6-11. https://doi.org/10.3969/j.issn.1000-3428.2016.10.002The 3D virtual tour system often provides immersive tour experience.It cannot offer information service about tour and is hard to fulfill users’ personalized requirement.To tackle this issue,an indoor positioning method based on IBeacon Bluetooth communication technology with low consumption is proposed to provide indoor positioning and automatic tour-guidance service in indoor culture tourist facilities.It is based on RSSI ranging principle and plane projection-based 3D space positioning technology,and supports receiving transmission signals from IBeacon Bluetooth base station by mobile phone to realize fast positioning.This paper applies IBeacon-based indoor positioning method in online 3D virtual tour system,and establishes an online-to-offline cultural tour guidance system combining virtuality with reality,which provides tourists with online virtual experience,offline realtime positioning and location-based information service.Application results show that the proposed system has higher speed of indoor positioning and accuracy fulfilling application requirement.It can provide users with better travel experience.
-
Computer Engineering. 2016, 42(10): 12-19. https://doi.org/10.3969/j.issn.1000-3428.2016.10.003This paper proposes an improved virtual view point rendering algorithm to solve the problems of ghosting and holes in the virtual view point rendering.This method marks the unreliable regions of depth map based on the depth information.Through multi-threshold image segmentation,the image is divided into different parts.A virtual view point is generated by making use of 3D image warping.In order to deal with the small cracks,local median filtering is used to obtain a single-view point image.Two generated virtual images are blended together to eliminate the disocclused holes in the desired view.An inpainting method is utilized to eliminate the remaining holes.Experimental results show that high quality virtual view points are generated,the Peak Signal to Noise Ratio(PSNR) of virtual view point can be 0.74 dB higher than that standard reference software of virtual view point rendering.
-
Computer Engineering. 2016, 42(10): 20-25,31. https://doi.org/10.3969/j.issn.1000-3428.2016.10.004To solve the shortage or low utilization of lack of experiment teaching equipment in some universities,this paper proposes a virtual/remote experiment system of double tank.It is based on Networked Control System Laboratory(NCSLab) architecture and accomplished by combining 3D virtual reality and remote control technology.3D modeling and rendering,mathematic modeling and system integration are included in the process of designing the experiment system.Coupled-tank experiment result shows that the system is able to share experiment equipment and bring students a strong sense of immersion.Also,the results of virtual and remote experiments can be mutually verified and the personalized monitoring interface also brings a better user experience.
-
Computer Engineering. 2016, 42(10): 26-31. https://doi.org/10.3969/j.issn.1000-3428.2016.10.005Fast and correct detection of on-road objects is a difficult problem in the field of assistant driving and automatic driving of intelligent vehicles.So,based on overall visual structure model,the visual cognitive mechanism from the local to the whole is simulated.The structural model is introduced in the object part model,as well as the statistical model,the overall visual structural model with scale invariance and posture invariance is gained.Through weak-labeled sample image,the model can automatically label the representative local areas so that this model can be more discriminable in detection.The new model is adapted to detect obstacle object in road area.Experimental results show that the object detection algorithm based on the proposed model has higher efficiency than traditional algorithms.It has high detection rate and low error rate.
-
Computer Engineering. 2016, 42(10): 32-37,44. https://doi.org/10.3969/j.issn.1000-3428.2016.10.006Aiming at the low precision of current location based applications in tourism area,considering that map matching algorithms are mainly intended for intelligent transportation systems,a score rule-based dynamic map matching algorithm is proposed.The algorithm takes road width into account and extends the line road network modeling.Based on probabilistic map matching methods,the dynamic threshold mechanism is adopted to fit the complex location environment.Closeness,topological relation as well as dynamic confidence area are applied to figure out the candidate roads.It uses cumulative score method to choice road.Experimental result indicates that,the algorithm is used in the campus navigation application,and the accuracy of roads identification reaches 89.3%,which is 18.4% higher than that of the nearest point-curve distance method and 10% higher than current map applications.The proposed algorithm can effectively improve the accuracy of road selection and location,meanwhile improves map display.
-
Computer Engineering. 2016, 42(10): 38-44. https://doi.org/10.3969/j.issn.1000-3428.2016.10.007
Aiming at the disadvantages that the existing modeling methods of human motion cannot be applied at the general nonlinear and non-Gaussian cases,based on the Markoff model,an improved modeling method of human motion is proposed.The Markov process of human motion prediction is presented by using the Auto Regressive(AR) tree,and the paper proposes extensions to AR trees and introduces the Dynamic Forest Model(DFM) and describes its training,regularization and the realization of the accurate modeling of human motion.Example research results show that performance of DFM is better than other benchmark algorithms,Hidden Markov Model(HMM) and Gaussian Process Dynamical Model(GPDM),and computing efficiency of the proposed methods is high.
-
Computer Engineering. 2016, 42(10): 45-50. https://doi.org/10.3969/j.issn.1000-3428.2016.10.008Recently,researchers in the field of information retrieval have proposed many ranking algorithms supporting search result diversification,but there is no literature on the performance analysis and comparison of these algorithms.Based on a traditional data fusion ranking algorithm CombSum,this paper proposes a new ranking algorithm CombSumDiv,which considers relevance and diversity of the documents at the same time.Besides,a comparison among CombSumDiv and two explicit ranking algorithms xQuAD and PM2 is conducted.Experiments are carried out with query datasets and ClueWeb09B datasets from the TREC diversity task.The result shows that the query performance of CombSumDiv algorithm is better than that of xQuAD and PM2 algorithm,and these three algorithms have strong stability and anti-interference ability.
-
Computer Engineering. 2016, 42(10): 51-56. https://doi.org/10.3969/j.issn.1000-3428.2016.10.009The sparsity of rating data is a common problem in collaborative filtering recommendation systems.This paper proposes a collaborative filtering recommendation algorithm based on positive and negative similarities which presents a calculation method of the positive and negative similarities between users by improving the Jaccard similarity model.Thus,user groups based on positive and negative similarities can be picked out for predicting ratings according to the known score of the user group.Experimental results show that the algorithm can achieve better prediction accuracy for recommendation than collaborative filtering recommendation algorithms based on similar user group.Furthermore,the algorithm can alleviate the dataset sparsity problem.
-
Computer Engineering. 2016, 42(10): 57-63. https://doi.org/10.3969/j.issn.1000-3428.2016.10.010Traditional Quality of Service(QoS) sequence description method only considers the mean(preference) which cannot reflect the QoS fluctuation(risk).To solve the problem,a parallel service selection algorithm combining risk and preference is proposed.The QoS value sequence is converted to the QoS model based on standard error of mean by using the static QoS model converter and the distributed QoS model converter.Based on the QoS model,an adaptive adjustment mechanism is designed to adapt dynamic change of QoS.The service selection algorithm based on Technique for Order Preference by Similarity to Ideal Solution(TOPSIS) is designed to obtain the optimal service reflecting users’ QoS needs.Experimental results show that the proposed algorithm improves the reliability of service selection and can solve the uncertainty problem of big QoS data sequence.
-
Collaborative Filtering Algorithm Based on Weighted Multi-fusion Preference and Structure SimilarityComputer Engineering. 2016, 42(10): 64-68. https://doi.org/10.3969/j.issn.1000-3428.2016.10.011If the rating data set of users in the recommendation system consists of sparse data,it is not ideal to use the traditional collaborative filtering algorithm based on rating similarity or structure similarity,because of the large error in nearest neighbor selection.Aiming at the shortcomings of traditional collaborative filtering algorithms,this paper proposes a collaborative filtering algorithm fusing rating and class preference similarity.In order to measure the similarity between users more accurately,the measurement method of weighted multi-fusion preference and structure similarity is the further improvement of the measurement method of fusing score and class preference similarity.Through the experimental comparison,it is showed that the proposed algorithm can reduce the Mean Absolute Error(MAE) and improve the quality of recommendation.
-
Computer Engineering. 2016, 42(10): 69-74,79. https://doi.org/10.3969/j.issn.1000-3428.2016.10.012Development of ways for data acquisition leads to limit of traditional association rule mining by I/O and memory.Aiming at this problem,this paper puts forward an improved method,which uses advantages of the Hadoop platform,reduces the time complexity of the algorithm by incremental iterative method,and makes full use of the features of MapReduce programming.It completes the frequent itemset mining through traverse and MapReduce task scheduling,which improves the efficiency of processing.In the mining of strong association,with the help of Sqoop,the external tables are migrated from Hive to Redis,which makes the data read more efficient.Experimental results show that the proposed method can improve processing efficiency.With the data increasing,the advance becomes more obvious,and improved algorithm also has better speedup and scalability,which is able to quickly mine the association rules in large data.
-
Computer Engineering. 2016, 42(10): 75-79. https://doi.org/10.3969/j.issn.1000-3428.2016.10.013People usually want to submit queries to only a few data sources to obtain high quality search results,so data source selection becomes a key issue in Deep Web integrated retrieval.To enhance the efficiency of entity data integrated retrieval,this paper designs a data source selection method based on relevance and repeatability.Firstly,it proposes a summary construction method based on subject and emotional words.The above method identifies subject category of entity information based on user feedback and calculates the data repeatability between two Deep Webs based on emotional words.Then,it proposes a Deep Web data source scoring strategy based on query subject relevance and repetition of content.Experimental result shows that the proposed method has higher accuracy and recall,although using a small data summary.It can provide an effective support to entity information integrated retrieval.
-
Computer Engineering. 2016, 42(10): 80-85,90. https://doi.org/10.3969/j.issn.1000-3428.2016.10.014According to the high real-time performance and multitasking features of the handheld terminal,some real-time mechanisms are added to the kernel,such as real-time mutual exclusion lock,threaded interruption,high precision timer and dynamic clock.Based on the idea of the RM algorithm combined with SCHED_FIFO algorithm,a taskset classified scheduling strategy is proposed.Considering the deadline scheduling algorithm has the feature of high hard real-time performance,deadline scheduling algorithm is used to schedule hard real-time task and real-time group scheduling algorithm is used to schedule soft real-time task.Theoretical derivation and experimental results show that the real-time performance of the improved Linux kernel meets the microsecond requirements of hard real-time tasks,and it also can reduce the tasks’ deadline missing rate effectively when the system is overloaded.The average deadline missing rate is dropped by 18% when CPU load range is between 1.2 and 2.8.
-
Computer Engineering. 2016, 42(10): 86-90. https://doi.org/10.3969/j.issn.1000-3428.2016.10.015As unique internal properties,software birthmark is often used to detect the similarity between different software.On the basis of researching software birthmark,this paper proposes a new method combining two dynamic birthmarks with different granularity to be a new software birthmark.The coarse-graired birthmark is the system function called by the executing program and the call frequency,and the fine-grained birthmark is the set of basic block instruction when user-defined function is executing.The combination of the two birthmarks is the final software birthmark.Simulation experimental results show that compared with the method using single-granularity birthmark,the proposed method improves the credibility and the reliability of software birthmark,and it has stronger ability to resist attacks.
-
Computer Engineering. 2016, 42(10): 91-95. https://doi.org/10.3969/j.issn.1000-3428.2016.10.016The traditional Digital Signal Processor(DSP) target board realizes communication with DSP through calling hardware driver lay commands by the bus.However,for this way,the application environment change,service program interruption and many other factors need to be considered,which seriously affects the efficiency of DSP and increases the development costs.Aiming at these problems,on the basis of Object Request Broker(ORB) architecture,this paper cuts and modifies the CORBA/e specification to form lightweight ORB specification.It writes the framework code and stub code on the basis of this lightweight specification and realizes lightweight design of framework,adapter,public data,encoder and decoder.Test result shows that the lightweight ORB can realize cross-platform and across-system communication of DSP,and it has higher communication efficiency.
-
Computer Engineering. 2016, 42(10): 96-100. https://doi.org/10.3969/j.issn.1000-3428.2016.10.017According to the solution characteristics of large-scale tridiagonal linear equations,based on distributed shared memory environment of mixing Message Passing Interface(MPI) with Open Multi-processing(OpenMP),this paper designs a parallel GaBP algorithm to solve large-scale tridiagonal linear equations.Tests are done through parallel computing cluster environment of Intel Xeon E5-2650,and the results show that the proposed algorithm has higher speed up ratio and better scalability,compared with parallel GaBP algorithm based on MPI.It can take fuller advantage of the overall computing performance of cluster system and improve the speed of solving the large-scale tridiagonal linear equations.
-
Computer Engineering. 2016, 42(10): 101-106,112. https://doi.org/10.3969/j.issn.1000-3428.2016.10.018In order to effectively manage storage resources of the intelligent chip and improve the utilization rate of storage space,a storage management algorithm which has fast processing and high storage utilization is proposed.The algorithm divides the whole storage space into multiple units with the same size,and uses block array and free array to manage them.The block contains multiple continuous units.There are two kinds of blocks,the occupied block and the free block.The block array records the information of blocks,such as it is occupied or not and the size of it.The algorithm uses the subscript of block array to organize the free blocks into multiple double direction link tables,which makes the management of blocks easy.The head of free block link table is stored in the free array where the subscript of array is the same as the size of the free block.According to the size of the storage required,the free block with the size can be find quickly in the free array.Experimental results show the proposed realizes the function of storage allocating,releasing and merging,so it has high storage utilization,and the overall performance of the proposed algorithm is better than that of the buddy algorithm.
-
Computer Engineering. 2016, 42(10): 107-112. https://doi.org/10.3969/j.issn.1000-3428.2016.10.019Since Loongson adopts Microprocessor without Interlocked Piped Stages(MIPS) instruction set and main supports the Linux operating system,application software on Loongson is still relatively scarce.Many traditional Windows software cannot run directly on the platform.To solve this problem,this paper proposes a method to simulate Windows system call by utilizing Wine emulator and uses QEMU user mode realize the dynamic translation from x86 instructions to MIPS instructions.It achieves software binary migration from Intel+Windows architecture to Loongson+Linux architecture.Experimental results show that this scheme has higher efficiency compared with Bochs+Wine and QEMU full system schemes.The rate of CPU calculation with this method is 8.04 times of Bochs+Wine scheme and 1.44 times of QEMU full system scheme.I/O calculation rate with this scheme is 2.44 times of Bochs+Wine scheme and 7.38 times of QEMU full system scheme.
-
Computer Engineering. 2016, 42(10): 113-117,123. https://doi.org/10.3969/j.issn.1000-3428.2016.10.020Nowadays,Transfer Control Protocol(TCP) packet reassembly technologies based on stand-alone system are getting hard to satisfy the need of processing mass data,as the larger the data amount,the lower their efficiency.Aiming at this problem,a Hadoop-based TCP packet reassembly system is proposed.It takes advantages of Hadoop Distributed File System(HDFS) and the parallel processing framework of MapReduce to parallelize data reading and result outputting in TCP packet reassembly.It also takes advantages of the secondary sort technology in MapReduce to achieve TCP traffic classification.Test result shows that the Hadoop-based packet reassembly technology works more effectively than the packet reassembly tool based on stand-alone system.The operating efficiency can be improved by 84% when the size of the data reaches 100 GB.
-
Computer Engineering. 2016, 42(10): 118-123. https://doi.org/10.3969/j.issn.1000-3428.2016.10.021In order to improve the applicability of the airborne sensor network deployment algorithm and realize the deployment of complex environments,a Weighted Virtual Force Decomposition Algorithm(WVFDA) is proposed.The air diffusion model and the surface diffusion model are established.The air diffusion model can reduce the energy consumption of nodes on the surface and make most of the diffusion process are carried out in the air.The surface diffusion model can reduce the time of exposure to the air,so as to avoid the impact of the environment and unexpected events.Simulation experimental results show that compared with 3D virtual force deployment algorithm,the proposed algorithm can adjust the deployment process in real time.The deployment time is short and the applicability is strong,which has a good deployment effect.
-
Channel Contention Delay Optimization Mechanism for 802.15.4 Network MAC Layer with Unsaturated LoadComputer Engineering. 2016, 42(10): 124-128. https://doi.org/10.3969/j.issn.1000-3428.2016.10.022To reduce the delay of channel contention at the Medium Access Control(MAC) layer for IEEE 802.15.4 networks,a channel contention mechanism for nodes,with unsaturated load is designed.A mathematical model is built and a derivation is carried on to obtain the steady-state probability of the main state.According to the result of the model,the formula of channel access delay is derived based on the modeling of channel state conversion process.The impact of protocol parameters such as λ,NB and minBE on the performance of node access delay is researched,and it is compared with the delay performance of the saturated load network.Analysis results show that the new model can describe and evaluate the mechanism of channel contention access very well,and the competition optimization mechanism can reduce effectively the deterioration degree of channel.Moreover,settings of protocol parameters can improve the performance of channel contention delay.
-
Computer Engineering. 2016, 42(10): 129-134. https://doi.org/10.3969/j.issn.1000-3428.2016.10.023In order to improve the reliability and energy efficiency of the networks,a Hop-count Gradient field based Energy-balanced Routing protocol(HGEBR) is proposed.In the protocol,each sensor node has a Neighbor Node Residual Energy Table(NNRET) which records the residual energy of the N nodes with the most residual energy in the neighbor nodes.The messages generated by the source node in a gradient carry the gradient value and the minimum value in the NNRET of the source node.According to the information carried by the message,the neighbor nodes decide whether to forward this message,thereby controlling the number of forwarding nodes,thus ensuring the reliability of the network and avoiding message flooding and too much energy consumption.Theoretical analysis and simulation results show that the protocol has better performance in the transmission reliability and network delay than GRAB and SPBT protocol.When the NNRET is 3,the performance of the proposed protocol is the best.
-
Computer Engineering. 2016, 42(10): 135-139. https://doi.org/10.3969/j.issn.1000-3428.2016.10.024An improved interference cancellation method is proposed for MIMO interference channel with three users each of which equipped with four antennas.4×4 space-time codeword with Alamouti code structure is designed.The multi-user interference at the first receiver is cancellated by the unidirectional operational link.Then,the interfering codeword are aligned at the second and the third receivers through interference alignment pre-coding.Multi-user interference at these two receivers is cancellated by the orthogonal properity of the effective channel matrices for Alamouti code.Experimental result shows that the diversity gain of the proposed method is 8,higher than that of the existing interferenc acncellation method for the same scene.And it has higher system capacity and reliability.
-
Computer Engineering. 2016, 42(10): 140-145. https://doi.org/10.3969/j.issn.1000-3428.2016.10.025Existing Attribute-based Encryption(ABE) schemes with hidden access structures are only partially hidden,which will lead to privacy leaks.Therefore,this paper proposes a Ciphertext-policy Attribute-based Encryption(CP-ABE) scheme with fully hidden access structure in prime order group.In the proposed scheme,each attribute not only has a number of possible values,but also has a mark to describe the demanding status,which enriches the description of access structure.In addition,any user learns nothing about an access policy even if an authorized user can decrypt encrypted data successfully because he is forced to use all of his attribute components for decryption.The scheme is proven secure under sequences of games based on the Decisional Bilinear Diffie-Hellman(DBDH) assumption and Decision Linear(D-Linear) assumption.Finally,comparisons of the proposed scheme and the existing schemes are offered.The results show that the proposed scheme is more suitable for the application of high confidentiality requirements.
-
Computer Engineering. 2016, 42(10): 146-150. https://doi.org/10.3969/j.issn.1000-3428.2016.10.026For the information leakage problem in the first-order mask encryption operation of vertically-reused mask in Advanced Encryption Standard(AES),this paper proposes an attack method based on power spectrum of second-order Correlation Power Analysis(CPA) in frequency domain.With the soft realization of AES first-order mask encryption algorithm in smartcard as the research object,the paper analyzes the weak points of the algorithm and conducts experiments,in order to get a more efficient high-order attack method.This paper compares the second-order CPA experiment in time domain with the one in frequency domain.The result shows that the second-order attack in frequency domain presents a better alignment effect,and has more obvious advantages in data processing complexity,data storage space,and attacks than in time domain.Furthermore,this attack method is applicable to other first-order mask cryptographic algorithms.
-
Computer Engineering. 2016, 42(10): 151-157,163. https://doi.org/10.3969/j.issn.1000-3428.2016.10.027Wireless Sensor Network(WSN) is often deployed in hostile environments,in which malicious attackers can easily disrupt the nodes’ localization procedure to challenge the WSN-based applications.A new method of secure unknown node localization is proposed on the basis of the consistency of Received Signal Strength Indicator(RSSI) by using mobile beacon node.The proposed secure localization scheme can counteract the impacts of wormhole attack by eliminating the group of beacon information with a smaller variance of RSSI.Simulation results show that this method can effectively eliminate the influence of wormhole attack on node location.
-
Computer Engineering. 2016, 42(10): 158-163. https://doi.org/10.3969/j.issn.1000-3428.2016.10.028Traditional botnet detection technologies mainly detect botnet nodes in specified area network on the hosts or the border of gateway export,which have small scale and low detection efficiency.To efficiently execute Peer-to-Peer(P2P) network botnet node detection in a larger range,this paper proposes a strategy to detect malicious botnet node parasitism in P2P network based on Challenge.Simulation results show that in KAD network including more than 4 million nodes,the detection technology can detect 3 000 to 9 000 parasitic P2P botnet nodes.It can estimate the number of possible parasitic botnet nodes in KAD network,which is of great guiding significance for accurate defense object positioning and defensive measures design.
-
Computer Engineering. 2016, 42(10): 164-168. https://doi.org/10.3969/j.issn.1000-3428.2016.10.029In order to implement biological keys reliablility coding,on the basis of online hand-writing signature,this paper proposes a key generation method using on-line handwritten signature,which is based on the method of equal probability mass quantization.It utilizes the user-dependent fisher ratio criterion to get the user characteristics,which are the most representative of the user,and applies the thought of equal probability mass in the static bit allocation method based on likelihood ratio,but the quantitative probability mass of the registered sample is determined by the characteristic interval control coefficient,which improves the time performance of the registration phase.Factoring in time like definite integral in the computation of the quantitative probability mass,this paper raises an optimization strategy which replaces the definite integral with the trapezoidal area of curved section.The identity authentication experiment and the time complexity contrast experiment are carried out on the SVC2004.Results show that the method gains False Acceptance Rate(FAR) of 2.54% and False Rejection Rate(FRR) of 28.63%.The registration time of the trapezoidal improvement is 9 s,which is about 1/10 of the original Gauss integral method.
-
Computer Engineering. 2016, 42(10): 169-175. https://doi.org/10.3969/j.issn.1000-3428.2016.10.030Interaction between Android application program and WebView will lead to privacy leakage.Aiming at the problem,this paper analyzes two privacy leakage channel of WebView,addJavascriptInterface and loadUrl,and proposes a data flow tracking framework of extended WebView,named WebView Tainted Detection(WTD).WTD extends the addJavascriptInterface and sensitive API in the Android source code.In sensitive API,Java function stack executes trace mechanism to record function hierarchy call relation in the system to detect whether the addJavascriptInterface registering objects accessed Android privacy API.Besides,WTD increases the taint detection of loadUrl loading page parameters.Experimental result shows that WTD can effectively detect the private data leakage from WebView.
-
Computer Engineering. 2016, 42(10): 176-180,186. https://doi.org/10.3969/j.issn.1000-3428.2016.10.031In text orientation classification,there is a problem of adjunct words’ polarity misjudgment and missing hidden view caused by unknown emotional direction.This paper proposes an orientation identification method based on evaluation modification distribution difference.It builds modification relationship bipartite graph and modification distribution vector,computes the modification distribution difference of evaluation object in positive and negative training corpus,extracts the features which have obvious difference,and integrates the positive and negative difference information into the process of characteristic value computation.Experimental results show that the proposed method can improve the precision and recall by 4.6% and 5.6% respectively for text orientation identification compared with other approaches which tend to pick the strong subjective emotional items as features and classify by Support Vector Machine(SVM) algorithm.The method has domain adaptability that it can reduce the decline of precision and recall by 6.6% and 6.4% in the cross-domain situation,when compared with the approaches which pick the emotional items as features and classify by SVM algorithm.
-
Computer Engineering. 2016, 42(10): 181-186. https://doi.org/10.3969/j.issn.1000-3428.2016.10.032An entity may belong to several different concepts.At the same time,it may belong to some varigrained similar concepts.This paper proposes a method that handles the entity conceptualization based on the given entities and attribute sets.By establishing concept’s attribute template and the typicalness of attributes in the concept,this method deduces an entity’s concept when combined with different attributes.It also proposes a Bayesian similar candidate entity ranking model which takes four characteristics into consideration.Experimental results demonstrate that the proposed method can effectively improve the similar entity recommendation effect.
-
Computer Engineering. 2016, 42(10): 187-191. https://doi.org/10.3969/j.issn.1000-3428.2016.10.033Since bipartite graph contains two different types of nodes and links only exist between different types of nodes,most link prediction methods for common single graph cannot be applied to bipartite graphs directly.In addition,the community information does not draw enough attention though it is important to improve the accuracy of link prediction.So a bipartite link prediction method using community information is developed.The method regards link prediction as a classification problem in machine learning.The local structural properties of node pair instances in a bipartite graph are extracted by the projection of the bipartite graph,together with community properties of instances by exploiting community detection techniques.Then local structural properties and community properties are used as similarity measurements of node pair instances.Training and prediction are conducted in a supervised learning process.Experimental results in a real dataset MovieLens show that the use of community information improves the link prediction accuracy and recommendation performance.
-
Computer Engineering. 2016, 42(10): 192-200. https://doi.org/10.3969/j.issn.1000-3428.2016.10.034Green supply chain network system is of great significance to improve the competitiveness of enterprises,to protect the environment and to implement the strategy of sustainable development.This paper introduces the background of the supply chain and puts forward an optimization method for the traditional supply chain network design.Meanwhile,considering the environmental factors such as water pollution and carbon emissions in the beginning of the supply chain network construction,it establishes the mathematical model.Combining with Genetic Algorithm(GA) and optimization algorithm,it designs hybrid GA based on autonomous learning to solve the model,and completes the construction of green supply chain network,so as to realize green combination selection of cooperative partner of the supply chain and provide some decision support for the enterprise managers.Experimental results show that proposed approach can objectively evaluate the service,environmental protection and other factors of the supply chain partner.It is feasible and applicable in the construction of green supply chain network system.
-
Computer Engineering. 2016, 42(10): 201-206. https://doi.org/10.3969/j.issn.1000-3428.2016.10.035In order to mine more information relationship in information recommendation,this paper puts forward an information recommendation method based on the Deep Belief Network(DBN).It uses fuzzy clustering to do some pretreatment in order to achieve the aim of cross-category recommentation.According to the user browsing history,it works out high-mark information through the Internet,and then uses Latent Dirichlet Allocation(LDA) model combining the interests of users to do some weight value adjustments.So as to improve the accuracy of recommendation.In the process of recommendation,the Internet will adjust the weight value of recommended information according to the action of users.Experimental results show that the recommendation accuracy of this method can be improved by 5.7% than BP neural network.
-
Computer Engineering. 2016, 42(10): 207-212. https://doi.org/10.3969/j.issn.1000-3428.2016.10.036Current community detection algorithms based on modularity have shortcomings like high time-complexity and low accuracy,they suffer form resolution limit.Therefore an algorithm based on Multilayer Particle Swarm Optimization(MPSO) for detecting community is proposed.Each particle has flight direction and a fitness value f.Whether two adjacent nodes belong to the same community is dependent on whether the fitness value is increasing when the particle move from one node to another.The proposed algorithm treats community as one super node to construct an upper network and get more coarse-grained community structure and higher modularity value.Experimental results show that the proposed algorithm can rival infomap algorithm when the mixing parameter u is less than 0.7 for dealing with large-scale network.When u is greater than 0.7,the proposed algorithm is better.It can get accurate results,and alleviate resolution limit problem effectively.
-
Computer Engineering. 2016, 42(10): 213-218. https://doi.org/10.3969/j.issn.1000-3428.2016.10.037Short text is generally of short structure and semantic inadequacy which makes it difficult to model.Therefore a method for short text feature extension via improved frequent term set is proposed.By calculating support and confidence,with co-occurrence relations and same lategory tendency are extracted.Based on the extracted frequent term set correlations are defined to further extend the term set.Meanwhile,information gain is introduced based on the traditional TF-IDF,to express the category distribution information and enhance the weight of the words for each category.Finally,the word similarity matrix is constructed via the frequent term set,and the symmetric non-negative matrix factorization technique is used to extend the feature space.Experimental results show that the constructed short text model built by this method can significantly improve the performance of short text clustering.
-
Computer Engineering. 2016, 42(10): 219-225. https://doi.org/10.3969/j.issn.1000-3428.2016.10.038Traditional biometric recognition system uses single source biometric information for authentication.However,the recognition performance decreases while biometric information is contaminated by various artifacts such as illumination variations,noise,occlusion and so on.A multimodal Sparse Representation(SR) algorithm is proposed,which represents the test data by a sparse linear combination of training data,while enforcing the observations from different modalities of the test subject to share their sparse representations.The optimization problem is solved using an efficient alternative direction method.Experimental results show that the proposed method has better recognition performance than competing method based on information fusion.
-
Computer Engineering. 2016, 42(10): 226-230. https://doi.org/10.3969/j.issn.1000-3428.2016.10.039The Local Binary Fitting(LBF) model uses the local average gray level information of image to cope with gray level inhomogeneity.However,by only using the local gray means of the image,it is not possible to get ideal results of segmentation for the medical images which involve strong noise and low contrast.This paper puts forward an improved method.Taking the local average information into consideration,it uses the global constraint information to detect the large gradient amplitude at the outer region.By adopting the level set methods to minimize the energy function,the contour is able to stay at the objective boundaries accurately.Experimental results show that the improved method enhances the robustness to noise,and its accuracy of segmentation exceeds that of other similar methods.
-
Computer Engineering. 2016, 42(10): 231-235,241. https://doi.org/10.3969/j.issn.1000-3428.2016.10.040Due to the light,complex structure and other reasons,there are inevitable heavy boundary with large boundary intensity near the target contour.Using the Gradient Vector Flow(GVF) Snake model to extract the contour of the image,the initial contour wrongly converges to the heavy boundary,and this phenomenon is called distraction.From the perspective that the GVF field force changes,the influence of the number of iterations and the intensity difference on the distraction are analyzed in detail.It is found that the enhancing intensity of GVF edge map’s weak boundary makes the field force near the weak boundary free from the affect of the heavy boundary.To this end,the edge detected by the classical Canny algorithm is used to replace the traditional GVF edge map calculation method.To solve the problem of weak boundary loss of classical Canny algorithm,an improved Canny algorithm is proposed,and the gradient direction is used as the compensation weight to increase the gradient amplitude.Experimental results show that the improved Canny algorithm can preserve the weak boundary well,and the GVF edge map based on this algorithm can effectively solve the problem of distraction.Compared with the method of reducing the number of iterations,this method has better stability.
-
Computer Engineering. 2016, 42(10): 236-241. https://doi.org/10.3969/j.issn.1000-3428.2016.10.041
Image segmentation algorithm by bipartite graph considers the spatial organization relation between superpixels as well as pixel and superpixels,which is robust for ore image segmentation.This paper proposes a bipartite graph algorithm based on Collaborative Representation(CR),which is able to ensure global features and local information.It takes image segmentation as a bipartite graph partitioning problem and uses a superpixel segmentation to search for the most probable groups of superpixels.CR method can reduce the complexity of 0 normalized image segmentation algorithm.Besides it is robust for segmenting ore images which have monotonic color changes and overlapping fragments,and compares to different segmentation algorithm.Simulation results of different segmentation algorithms show the validity of the proposed algorithm.
-
Computer Engineering. 2016, 42(10): 242-248. https://doi.org/10.3969/j.issn.1000-3428.2016.10.042Aiming at the fuzzy distance definition of the classical Vector Median Filtering(VMF) algorithm,a new vector median filtering algorithm based on random projection depth sorting is proposed.The algorithm uses the sorting method of center outward.The global color pixels are sorted by the depth of the random projection.The noise pixels in the local window are determined according to the sorting.The noise is filtered by vector median filtering.The filtering experiment for color images with different noise density is carried out.The improved algorithm is compared with vector dual morphological filtering,vector morphological filtering of spatial transformation,vector median filtering and a half space vector median filtering algorithm.The results show that the modified algorithm can efficiently remove salt and pepper noise.The filtering effect of the modified algorithm is more consistent with human visual characteristics and has better performance and robustness.
-
Computer Engineering. 2016, 42(10): 249-254. https://doi.org/10.3969/j.issn.1000-3428.2016.10.043A new algorithm is presented based on curve matching to detect the reflection symmetry of the image.The Intensity Order based Mean Standard Deviation descriptor(IOMSD) is adopted to obtain the symmetric curve pairs in the single image.The gradient symmetry measurement for the points lying on the symmetric curve pair is made to determine the gradient symmetry point pairs.The best symmetry point pair on one symmetric curve pair is picked out under the minimum distance constraint,and the midpoint of the best symmetry point pair can be thus located.For all the midpoints of the symmetric curve pairs in the image,Hough transform is used to detect the reflection symmetry axis.Experimental results show that the proposed algorithm can precisely locate the reflection symmetry axis in the case of illumination change,contrast change,noise contamination and image rotation.A good application on symmetry detection in reflecting image is also demonstrated.
-
Computer Engineering. 2016, 42(10): 255-260. https://doi.org/10.3969/j.issn.1000-3428.2016.10.044Since the Otsu method and most of its improved methods take between-class variance as the foundation of picking threshold,which causes great difference in segmentation for image with different histogram distribution,a new threshold segmentation algorithm is proposed in this paper.Firstly,a model of Coupled Object Similarity(COS) is introduced,which can take both the relationship of the various attributes of the object itself and the relationship between the properties into account,and can capture the relationships between the objects.Secondly,between-class variance in the Otsu method is replaced by COS to pick threshold.Each class distinguished by the selected threshold is regarded as an object in the model of COS.Each class has two attributes,the probability of class and gray mean.Similarity between classes is calculated,and the optimal threshold value is obtained according to the minimum of similarity between classes.Experimental results show that compared with Otsu algorithm,two-dimensional Otsu algorithm and maximum entropy algorithm,the algorithm can measure the difference of classes at a higher accuracy and obtains better segmentation results.
-
Computer Engineering. 2016, 42(10): 261-265,270. https://doi.org/10.3969/j.issn.1000-3428.2016.10.045The global shape retrieval is used for relic model which has the characteristics of similar overall shape and obvious detail difference.Due to the limitation of accuracy,it is difficult to obtain satisfactory retrieval results.So this paper proposes a retrieval algorthm for local geometric feature extraction.The principal curvature of each model vertex is calculated,and the shape index of the local surface is obtained from the principal curvature.The local geometric features of the model are got by using region growing and expanding search method based on the shape index.The template library is constructed with the local geometric features and prior knowledge.The similarity of the template and the detected model is compared.The Terra-cotta Warrior pieces are used as experimental data and the proposed method is applied to identify and classify the pieces.Experimental results show that the method is effective and feasible.
-
Computer Engineering. 2016, 42(10): 266-270. https://doi.org/10.3969/j.issn.1000-3428.2016.10.046In view of the real-time defects of existing defogging algorithms as well as the error and insufficient stability of gradient domain defogging algorithms,a gradient domain restoration algorithm for a single haze image is put forward in this paper.Based on the atmospheric light scattering model,gradient attenuation of haze image is directly calculated in accordance with the estimated value of transmission,and then gradient domain restoration is proceeded.For the pixels in the image boundary position,defogging is carried out in the space domain.At last,taking the processed value of boundary pixel as the initial condition,the ultimate restoration image is reconstructed from gradient domain.Experimental results show that this algorithm can get good defogging effect and it has high real-time performance.
-
Computer Engineering. 2016, 42(10): 271-276. https://doi.org/10.3969/j.issn.1000-3428.2016.10.047Aiming at the registration of laser point cloud data with no additional information,this paper proposes a new registration method based on the measure criterion for matching point and Iterative Closest Surface(ICS) algorithm.The method introduces a new Normalized Zero-mean Cross-correlation Coefficient(NZCC) to measure curvature similarity of the neighborhood of a point.The effective array of one-to-one initial matching points is built.The initial registration parameters can be computed by using the four elements and the linear least square method.The method uses local surface instead of discrete points,the efficient point sets which involve in ICS are built,and uses one-time similar distance instead of the geometric distance from point to its corresponding surface patches,the nonlinear least square optimization model and solution strategy of registration is established.Numerical example results show that compared with the iterative stitching algorithm,this method is feasible,accurate and efficient.
-
Computer Engineering. 2016, 42(10): 277-282. https://doi.org/10.3969/j.issn.1000-3428.2016.10.048In the process of basic oxygen furnace blowing,it is required to make accurate and real-time judgment of the furnace endpoint.In order to enhance the accuracy of blowing endpoint judgment,a method for basic oxygen furnace endpoint judgment based on flame image Convolution Neural Network(CNN) recognization modelling is proposed.The related features are learned by CNN from the sample image,avoiding misleading from hand-crafted features,so as to realize the improvement in the accuracy of endpoint judgment.The OSTU algorithm is used for flame image segmentation in HSI color space to find out the best parameters of the model.The performance of the algorithm is verified with the flame data of five furnaces.Experimental results show that the recognition rate is respectively improved by 29% and 4% compared with that of the gray level co-occurrence matrix and gray scale difference statistics method.Therefore,the proposed model meets the real-time needs with high accuracy and can be used for actual basic oxygen furnace endpoint judgment.
-
Computer Engineering. 2016, 42(10): 283-288. https://doi.org/10.3969/j.issn.1000-3428.2016.10.049Sample Adaptive Offset(SAO) in High Efficient Video Coding(HEVC) uses the pixel as the unit for classification and choose the optimal value for compensation,which requires a large amount of computation..According to the dependency of offsets and Coding Unit(CU) depths,a fast SAO method is proposed.It classifies videos according to the texture complexity and skips some unnecessary samples for SAO,thereby improving video coding efficiency.Experimental results show that the proposed method reduces 60.32% SAO encoding time with negligible performance loss.
-
Computer Engineering. 2016, 42(10): 289-295. https://doi.org/10.3969/j.issn.1000-3428.2016.10.050To solve the problems such as poor side information quality,lower efficiency of compression and severe delay in Distributed Video Coding(DVC) under large Group of Picture(GOP),this paper proposes a Hybrid Side Information(HSI) generation algorithm based on extrapolation and interpolation.The proposed algorithm generates side information by interpolating after extrapolating in time domain,and then the side information is updated utilizing the multi-hypothesis prediction method.The algorithm fully utilizes the decoded frame information to suppress the prediction noise and get better compensation frames.Experimental results show that the proposed HSI can significantly increase the accuracy of side information and improves the Rate-Distortion(RD) performance steadily in DVC system under large GOP for sequences with different moving intensity without increasing the complexity of the encoder.
-
Computer Engineering. 2016, 42(10): 296-302. https://doi.org/10.3969/j.issn.1000-3428.2016.10.051A global path planning algorithm based on multistage decision is proposed for the problem that the single algorithm is not easy to coordinate multiple demands in the path planning process of mobile robot.The algorithm divides the planning process into several stages according to the different functions,and the decision of each stage is dependent on the decision of the last stage.The algorithm uses the improved location code quad-tree model to construct the model of the planning environment,uses the A* algorithm on the basis of the previous step to get the region of optimal path,which is called the path area,and then uses the improved particle filtering algorithm to get the path point.The path points are taken as the key points of the B spline interpolation to get the planning results.Experimental results show that compared with the A* algorithm and particle filtering,the time consuming,the number of path points,and the evaluation value results of the proposed algorithm are improved.Through the automatic selection of the parameters,the influence of human factors on the results of planning is reduced.
-
Computer Engineering. 2016, 42(10): 303-307. https://doi.org/10.3969/j.issn.1000-3428.2016.10.052To solve the difficulty of selecting nodes for network topology in complex environment,a selection method of random minimum spanning tree based on intuitionistic fuzzy sets is proposed.Through analyzing the uncertainty,like randomness and fuzziness,information in the network topology design,random fuzzy variables are expanded for random intuitionistic fuzzy variables,and edge weights in the topological structure are measured.The minimum spanning tree problem in network topology is transformed into random intuitionistic fuzzy spanning tree problem.Cut sets are used to remove fuzziness,and Prim algorithm to obtain the optimal solution,which will optimize the network topology.The effectiveness and rationality of the method is verified through a network experiment.
-
Computer Engineering. 2016, 42(10): 308-317. https://doi.org/10.3969/j.issn.1000-3428.2016.10.053In recent years,independent travel gradually becomes the mainstream of traveling.However,due to the Internet information flooding and search mechanism based on text matching,much attention and effort have to be paid by users to search and organize a tour route.To provide personalized and customized travel planning services for users,this paper proposes a personalized customization system of travel route based on semantic Web technology.The domain ontology model of tourist attractions is constructed and attraction instances are filled under semi-supervised conditions.User retrieval need interpretation is implemented and the retrieval results are sorted according to user’s interest model by applying Chinese word segmentation and part of speech tagging and ontology-based user modeling techniques.Finally,the system completes the time-space correlation extension by using semantic correlation between attractions.System performance test and usability evaluation results indicate that the system has high precision and recall,and it implements personalized and customized travel routes.
-
Computer Engineering. 2016, 42(10): 318-321. https://doi.org/10.3969/j.issn.1000-3428.2016.10.054In order to reduce the influence of signal drift on gas identification,an ensemble classifier model based on Deep Belief Network(DBN) is proposed.A single DBN classifier is trained by different periods of datasets and the obtained classifier is used to classify the datasets.By minimizing the classification error,the optimal integrated weights of each single classifier are obtained.Particle Swarm Optimization(PSO) is used to find the optimal weights and integrate all the classifiers to get the final gas identification results.The performance of the proposed method is compared with that of the uniform weighted DBN and the optimal Support Vector Machine(SVM) by using a sensor array composed of 4 gas sensors.Experimental results show that,this method can maintain higher classification accuracy in a long time.To a certain extent,the effect of signal drift on the classification results is suppressed.
2024
Volume 50 No 3
Volume 50 No 3
(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