Started in January,1974(Monthly)
Supervised and Sponsored by Chongqing Southwest Information Co., Ltd.
ISSN 1002-137X
CN 50-1075/TP
CODEN JKIEBK
Editors
Current Issue
Volume 42 Issue 3, 14 November 2018
  
Review of Time Series Representation and Classification Techniques
YUAN Ji-dong and WANG Zhi-hai
Computer Science. 2015, 42 (3): 1-7.  doi:10.11896/j.issn.1002-137X.2015.03.001
Abstract PDF(644KB) ( 2600 )   
References | Related Articles | Metrics
Time series is a set of random variables ordered in timestamp.It is often the observation of an underlying process,in which values are collected from uniformly spaced time instants,according to a given sampling rate.Since time series data exist widely in various application domains,such as finance,agriculture,meteorology,biological science,eco-logy and so on,discovering knowledge from time series has become one of the mainly research fields of data mining.In this paper,a comprehensive review on the existing time series representation and classification research was given.In the term of time series representation,three different categories named non-data adaptive,data adaptive and model based were summarized.A summary of several time series classification method,namely similarity in time,similarity in shape and similarity in change was also provided.
Summary and Prospect on Entity Resolution
ZHU Can and CAO Jian
Computer Science. 2015, 42 (3): 8-12.  doi:10.11896/j.issn.1002-137X.2015.03.002
Abstract PDF(544KB) ( 1241 )   
References | Related Articles | Metrics
Entity Resolution(ER) is a key step in data cleaning,data integration,data mining and the insurance of data quality.This paper listed and explained some classic algorithms in the development of entity resolution,including pair-wise entity resolution,collective entity resolution,entity resolution on big data,and entity resolution on complex data et al.We also introduced the characteristics and limitation of these algorithms and shared some state-of-the-art algorithms derived from new application environment according to different requirements.Finally,the research hotspots and the development direction of this field were discussed.
RPL:A Robot Programming Language Based on Reactive Agent
TIAN Chang-hai, YANG Shuo, CHEN Yin and MAO Xin-jun
Computer Science. 2015, 42 (3): 13-18.  doi:10.11896/j.issn.1002-137X.2015.03.003
Abstract PDF(492KB) ( 553 )   
References | Related Articles | Metrics
Robots situated in open environment are expected to be context-aware and autonomous,to support concurrent behaviours,and to react to external stimuli in a timely manner.To program such robots,a robot programming language is required to explicitly represent environmental factors,support decision-making and concurrency,specify temporal,special and physical relations among robot behaviours,and prioritize concurrent robot behaviours to avoid collision.Agent-oriented programming (AOP) provides a new solution to robot programming by taking software units as autonomous agents.Based on the requirements for robot programminges language,agent-oriented programming model—RECA and programming language—RPL based on reactive agent were proposed.RECA programming model which takes robot software as a reactive agent consists of three elements.SensorEvent shows environmental changes;ScenarioBehaviour are the different behaviour specifications for robots,and EventRule defines the dynamic mapping relations from environmental inputs to behavioral outputs.RPL was designed to meet the needs of robot programming,by providing various mechanisms supporting event-based programming,multi-thread programming,prioritization of robot behaviours,and dynamic binding of robot behaviours.We designed and implemented a programming and runtime environment for the RPL,and demonstrated the expressiveness of RPL and the effectiveness of its runtime environment through a case study of elder assistant robot.
Research of Neural Cognitive Computing Model for Visual and Auditory Cross-media Retrieval
LIU Yang, TU Chun-long and ZHENG Feng-bin
Computer Science. 2015, 42 (3): 19-25.  doi:10.11896/j.issn.1002-137X.2015.03.004
Abstract PDF(1217KB) ( 573 )   
References | Related Articles | Metrics
Cross-media semantic mapping and cross-media semantic retrieval are key problems of the search engine.In this paper,we analyzed the functionality,the hierarchy and the structure of the brain’s neurocognitive,and established a in-like neural cognitive computing model for visual and auditory cross-media application after taking into account the idea of deep belief network and hierarchical temporal memory.According to the mechanism of information processing in central nervous system and framework of functional approach in cognitive theories,we designed computational model,and discussed systematically information integration mechanism in cortical column and cooperative cognitive processing of visual and auditory.This model provides a reference to resolve application of cross-media semantic mapping and retrieval,and is significant exploration for brain-like cognitive computation of non-von Neumann structure.
Dynamic Pointer Alias Analysis Framework for Vectorization
LIU Peng, ZHAO Rong-cai and LI Peng-yuan
Computer Science. 2015, 42 (3): 26-30.  doi:10.11896/j.issn.1002-137X.2015.03.005
Abstract PDF(455KB) ( 915 )   
References | Related Articles | Metrics
Pointer alias analysis is a key technology in data-flow analysis,whose results are the basis of compiler optimization and program transformation.Based on the related research of vectorization method and dynamic pointer alias analysis,a dynamic pointer alias analysis framework oriented to vectorization was designed.The dynamic pointer alias information is extracted by dynamic instrument and test run,and the vectorization code generation is guided by the feedback information.The whole framework was studied from three aspects,which are candidate alias analysis set extraction,instrument and test run.The framework was implemented on Open64 and evaluated in benchmark GCC-VECT and typical applications.The experimental results show that the framework has the more precise alias analysis results compared with static pointer alias analysis,and can significantly improve the speedup of vectorization program.
Low-cost Access Point Detection Algorithm Based on Periodicity Identification
LI Ai-jing, DONG Chao, TAO Bing-yang, TIAN Chang, WANG Hai and GAO Wei
Computer Science. 2015, 42 (3): 31-34.  doi:10.11896/j.issn.1002-137X.2015.03.006
Abstract PDF(347KB) ( 534 )   
References | Related Articles | Metrics
Continuous search for access point may consume enormous energy,especially for battery-powered mobile devices like mobile phones.To solve this problem,second device was recommended to search for APs and evaluate their power by detecting the periodic beacons transmitted by APs.Only when WiFi signals are detected can wireless adapter be woken up.To detect periodic signals,PSFA (precision-stable folding algorithm) was proposed.Experiment results show that PSFA can achieve high accuracy under both single AP and multiple APs circumstances with low complexity.
Routing Prediction Algorithm for Aeronautical GPSR-TAMP
ZHANG Wei-long, LV Na JIA, Hang-chuan and LI Teng
Computer Science. 2015, 42 (3): 35-38.  doi:10.11896/j.issn.1002-137X.2015.03.007
Abstract PDF(689KB) ( 436 )   
References | Related Articles | Metrics
Considering FACS algorithm can solve the problem that ACSM is not capable of describing weak strength maneuvering,an advanced adaptive filtering algorithm MFACS was introduced and on this basis a mobility prediction algorithm was raised,which was used in HELLO mechanism of GPSR creatively.Combining with a two hops adaptive beacon and relevant neighbor maintaining mechanism,according to whether adopt mobility prediction,two modified routing protocol GPSR-TAMP and GPSR-TA based on GPSR were derived to modify periodical beacon.NS2 simulation result demonstrates that with a better neighbor discovery mechanism,GPSR-TAMP outperforms GPSR-TA.
Analysis of Influence of Antenna Simple Harmonic Motion on Channel Cyclostationarity under Case of Receiver Motion
WEI Li-xia and CAO Shi-ke
Computer Science. 2015, 42 (3): 39-41.  doi:10.11896/j.issn.1002-137X.2015.03.008
Abstract PDF(305KB) ( 466 )   
References | Related Articles | Metrics
Simple harmonic motion (SHM) can induce cyclostationarity (CS) in channel autocorrelation under isotropic scattering (ISC) environments.With analysis of generation mechanism of cyclostationarity(CS) in channel,cyclostationa-rity of the channel can be measured by defining degrees of cyclostationarity (DCS).The autocorrelation characteristics of the channel,including cyclostationarity,were studied when the receiver is moving at a constant speed along a straight line and the antenna is doing simple harmonic motion at the same time.Then related curves were simulated by Matlab,showing the cyclostationarity of channels more clearly.The significance of cyclostationarity and a discussion on the future research topics were refered at last.
Method of Constructing Spread-spectrum Code Based on Chaos and Self-coded
ZHANG Xiao-rong, WU Cheng-mao and LI Wen-xue
Computer Science. 2015, 42 (3): 42-46.  doi:10.11896/j.issn.1002-137X.2015.03.009
Abstract PDF(504KB) ( 554 )   
References | Related Articles | Metrics
Aiming at traditional spreading codes with short period and poor security,a method of spreading codes constructed by means of combination chaotic maps with self-coded was proposed.The method firstly makes a homomorphic uniform mapping to a sequence generated by Logistic chaotic mapping to obtain a high degree of balance pseudo-random sequence,and then repeates iteration with the sub-sequence of different length to obtain the complex symbol sequence of which length increases geometrically by the time of iterations.Finally,it is combined with the self-coded sequence to generate a high quality composite random sequence through the dimensional Henon map.To analyze the effectiveness of the method based on Chaos and self-coded,we compared it by simulation test.The result shows the new spreading code has better correlation and a higher degree of complexity,and obtains a lower spread bit error rate.
Real-time Analysis of Prisoner’s Abnormal Behavior Based on Wireless Body Area Network
YANG Lu-lu, CHEN Jian-xin, ZHOU Liang and WEI Xin
Computer Science. 2015, 42 (3): 47-50.  doi:10.11896/j.issn.1002-137X.2015.03.010
Abstract PDF(373KB) ( 543 )   
References | Related Articles | Metrics
With the rapid development of wireless sensor technology,the application of wireless body area network in telemedicine and smart home gradually becomes the research hotspot.Prison is a special place,and the prisoner’s daily behavior monitoring is essential.Accurate and effective monitoring system can alarm in time when abnormal behavior occurs,and it contributes to the management of prison and prevent dangerous accidents.In the prison environment,a method for prisoner’s abnormal behavior recognition based on wireless body area network was presented. The three axis acceleration data are collected during the prisoner’s movement through a wrist-worn acceleration sensor,and then the classification algorithms are used to recognize the activities to assess whether there are abnormal behaviors.The experimental results show that the recognition accuracy of the abnormal behaviors can reach 95%.
Empirical Research on Book-crossing Network Model
MA Jie-liang, SONG Yan, PAN Zhen-zhen and HAN Lu
Computer Science. 2015, 42 (3): 51-54.  doi:10.11896/j.issn.1002-137X.2015.03.011
Abstract PDF(356KB) ( 407 )   
References | Related Articles | Metrics
By collecting the data of bookcrossing website as data source in one month,this paper constructed the database model of books and the users,and then structured the bipartite graphs to describe the relationship between them.From the perspective of complex network,this paper analyzed and calculated the network related parameters,such as degree distribution,clustering coefficient,the average path length,node strength distribution,act degree distribution,act size distribution and node interest,and the conclusion is that the book-crossing network model has scale-free characte-ristics and small-world network characteristics in the same time.
New Method for User’s QoS Requirement Network Selection in Heterogeneous Wireless Networks
ZHANG Yuan-yuan, XIAO Chuang-bai and WANG Jian
Computer Science. 2015, 42 (3): 55-59.  doi:10.11896/j.issn.1002-137X.2015.03.012
Abstract PDF(396KB) ( 579 )   
References | Related Articles | Metrics
With the advance in heterogeneous wireless networks and smart mobile terminals,user’s wireless services become more and more common.The challenge is the network selection based on QoS requirement.Based on multi-attribute decision making theory and fuzzy logic,a normalization algorithm for network selection in heterogeneous wireless network was presented.It includes six main steps:network pre-selection,constructing normalized decision-making matrix,calculating analytical hierarchy weights,determining the positive-ideal and negative-ideal solution,network eva-luation and network selection.Experimental results show that the proposed algorithm can trigger the handoff precisely and reduce the ping-pong effect.Moreover,when the mobile station’s speed and network’s load are changed,the system can select the network effectively and solve the problem of call dropping and load balancing.
Data Distribution Algorithm with Out-of-order Feedback for CMT over Diversity Network
DU Wen-feng and WU Zhen
Computer Science. 2015, 42 (3): 60-64.  doi:10.11896/j.issn.1002-137X.2015.03.013
Abstract PDF(441KB) ( 437 )   
References | Related Articles | Metrics
The performance of CMT degrades remarkably when a large quantity of out-of-order packets are blocked in the receive buffer due to the discrepancy of paths in terms of performance.Based on the analysis of different network configurations,a performance evaluation model was proposed.Meanwhile,a data distribution algorithm was proposed to improve the performance of CMT over diversity paths.This algorithm dynamically adjusts the transmission ratio with the out of order feedback of receiver buffer and reduces the impact of path with bad performance.The result of analysis and simulation reveals that the performance of our scheme outperforms the original round-robin scheme.
Community Detection for Micro-blog Network Based on WB-MMSB Model
XU Jian-min, WU Xiao-bo, WU Shu-fang and SU Wu-lin
Computer Science. 2015, 42 (3): 65-70.  doi:10.11896/j.issn.1002-137X.2015.03.014
Abstract PDF(510KB) ( 606 )   
References | Related Articles | Metrics
Considering the nodes of Mico-blog network have single direction relations,a new model WB-MMSB was put forward for community detection,which uses directed edges to embody the direction relations of nodes,and two aspects link-in and link-out are used to quantify the community membership of nodes.Exponential family distribution and mean-field variational inference method were used to inference the representations of variables in this model,and SVI algorithm was used to compute relating parameters.Experiments adopted Sina-Weibo dataset and NMI to testify the performance of WB-MMSB.The results indicate that the community detection ability of WB-MMSB model is better than aMMSB model,and the convergence rate of WB-MMSB model is faster than aMMSB model.
Clustering Protocol Based on Genetic Algorithm and Probabilistic Forwarding
CHEN Hai-nan, LIU Guang-cong, WU Xiao-ling, HUANG Ting-ting and LI Cong
Computer Science. 2015, 42 (3): 71-73.  doi:10.11896/j.issn.1002-137X.2015.03.015
Abstract PDF(277KB) ( 468 )   
References | Related Articles | Metrics
Aiming at the insufficient of the wireless sensor networks’ cluster protocol based on LEACH,a new cluster protocol based on the genetic algorithm and probabilistic forwarding method was presented.With this new protocol,the course of the cluster header’s choosing and the communication method between the header and base station were optimized.To verify the performance of our protocol,the new algorithm was compared with LEACH and LEACH-C protocol.The experimental results show that the new protocol performs better in the energy balance and more stable.
Hierarchical Routing Protocol Based on Link Quality in Wireless Sensor Network
MAO Ying-chi, WANG Jiu-long, WANG Kang and REN Dao-ning
Computer Science. 2015, 42 (3): 74-80.  doi:10.11896/j.issn.1002-137X.2015.03.016
Abstract PDF(613KB) ( 429 )   
References | Related Articles | Metrics
In the application of wireless sensor network,the quality and energy consumption of each communication between the nodes are influenced by the factor of link quality.In this paper,a hierarchical routing protocol EBCLQ was proposed based on link quality.EBCLQ protocol is composed of three parts:network initialization,cluster formation,dataforwarding.The network initialization,based on LQEWAL link quality prediction method,is used to obtain link quality between nodes and the information of neighbor nodes and prepare for the follow-up.The cluster formation phase is composed of three parts:CCELE algorithm based on link quality and energy is used to select candidate cluster head and FCECC algorithm is used to select formal cluster head.The data forwarding is composed of two parts:intra-cluster algorithm SAL is used to allocate slot to every cluster member and inter-cluster algorithm MOCC which combines the one-hop and multi-hop is used to transmit the data among the cluster head.At last,EBCLQ was achieved on MATLAB.It is confirmed that EBCLQ is effective and balanced by comparing and analyzing its efficiency of the implementation in the network.
Energy-efficient Virtual Machine Placement for Heterogeneous Cloud Platform
ZHOU Dong-qing and SI Qing-qian
Computer Science. 2015, 42 (3): 81-84.  doi:10.11896/j.issn.1002-137X.2015.03.017
Abstract PDF(446KB) ( 454 )   
References | Related Articles | Metrics
Energy consumption has become an important part of the operational cost of data center,and virtualization technology is one of the effective methods to reduce the energy consumption of data center.In order to reduce the high energy consumption of data center,we used the technology of virtualization,combining the heterogeneity of the physical machine and the multi-dimensional nature of resources that the virtual machine requires in the data center.A measurement model for the performance of different physical machines and another one for the multi-dimensional resource utilization rate were proposed,and then,on the premise of that,a deployment algorithm for virtual machine based on heterogeneous cloud platform was proposed.Simulation results show that the algorithm,compared with the MBFD and BFD,can reduce the energy consumption of system effectively,besides,it improves the utilization rate of resources and reduces the waste of resource.
Power Control-oriented Spectrum Pricing and Allocation in OFDMA Cognitive Radio Networks
ZHANG Chi, ZENG Bi-qing, YANG Jin-song and XIE Xiao-hong
Computer Science. 2015, 42 (3): 85-90.  doi:10.11896/j.issn.1002-137X.2015.03.018
Abstract PDF(485KB) ( 528 )   
References | Related Articles | Metrics
This paper proposed a Stackelberg game-based model for spectrum pricing and allocation in orthogonal frequency division multiple access(OFDMA) cognitive radio networks.With this model,the primary base station(PBS) can obtain the optimal pricing solution in the scenario where secondary base station(SBS) controls the transmission power of the secondary network to protect the primary network transmission.We redesigned the utility function of secondary user (SU) with the consideration of power control,and formulated the trade behaviors in spectrum leasing market,in which a single PBS acts as a seller and multi-SUs act as buyers,by a Stackelberg game model.Using backward induction,we solved the optimal pricing at market equilibrium with which the PBS can maximize its profit under QoS constraints.Besides,considering that limited information is available locally at the PBS,we presented a distributed dynamic Stackelberg game-based spectrum pricing and allocation model.Simulation results demonstrate that this model can obtain the optimal spectrum pricing and allocation scheme with controlling the secondary network transmission power under the interface threshold of the primary network.
Perceptive-direction-based Data Gathering in Opportunistic Mobile Sensor Networks
LIANG Quan, HE Wen-wu and ZHANG Yong-hui
Computer Science. 2015, 42 (3): 91-95.  doi:10.11896/j.issn.1002-137X.2015.03.019
Abstract PDF(447KB) ( 455 )   
References | Related Articles | Metrics
In opportunistic mobile sensor networks,it is necessary to improve the success rate of data transmission and reduce the network costs.At the same time,to minimize the energy consumption of sensors is also needed so as to prolong the lifetime of the network.For simplicity and practicability,a strategy named data gathering based on perceptive direction(DGPD) was presented in the paper.According to the descriptions of the strategy,both two sensors meeting in the network will select the nearest Sink nodes as the reference and calculate their perceptive directions respectively.Subsequently,the routing of the message forwarding is determined in the light of the perceptive direction which is an important parameter,and one sensor sends the message to the other which better facilitates approaching the Sink node so as to improve the data gathering rate and also reduce excessive messages forwarding.The simulation results indicate that this strategy can effectively complete the data gathering and obtain a higher performance of the network.
Attribute-based Access Control Method Supporting Policies Ontology Reasoning
NI Chuan, HUANG Zhi-qiu, WANG Shan-shan and HUANG Chuan-lin
Computer Science. 2015, 42 (3): 96-101.  doi:10.11896/j.issn.1002-137X.2015.03.020
Abstract PDF(860KB) ( 548 )   
References | Related Articles | Metrics
In large-scale and distributed systems,attribute-based access control(ABAC) proves its appropriateness out of the ordinary.However,the management of policies turns out to be complex and error-prone for the heterogeneity of network environment,the complexity of policy control and policy sets of large-scale and lack-of-semantic.In order to solve the problem,this paper presented an approach based on the established XACML standard to extend current ABAC authorization architecture with ontology consistency reasoning.First,it carries out a quantitative analysis on several important access control models under distributed environment.Second,it determines the consistency of policies in accor-dance with the result of the consistency checking on the ontology knowledge base.Third,it designs an experimental scheme in order to verify the validity and correctness of our method.
Formal Modeling of Complex Network Security Based on MAS
WEI Mei-lin, ZHANG Ming-qing, TANG Jun and KONG Hong-shan
Computer Science. 2015, 42 (3): 102-105.  doi:10.11896/j.issn.1002-137X.2015.03.021
Abstract PDF(833KB) ( 456 )   
References | Related Articles | Metrics
For the low fidelity and non-normative description of formal modeling of network attack and defense,a micro-macro combining formal modeling method with good scalability was proposed based on multi-Agent,which describes the static properties and dynamic behavior of individual agent from the microcosmic and describes the methods of role allocation and contacts between each agent from the macroscopic.Then,taking the DDoS attack and defense as example,we gave specific implementation process of the above method.Finally,a simulation of DDoS attack and defense was implemented to verify the model.
Ring Signcryption Broadcasting Scheme Based on Multilinear Maps
YU Zhi-min, JING Zheng-jun and GU Chun-sheng
Computer Science. 2015, 42 (3): 106-110.  doi:10.11896/j.issn.1002-137X.2015.03.022
Abstract PDF(434KB) ( 521 )   
References | Related Articles | Metrics
We proposed a ring signcryption scheme based on multilinear maps.Each ring member can represent ring group to generate anonymous signcryption and broadcast it to multiple recipients.Ring signcryption can be communicated between two ring groups.The scheme meets the security requirements of ring signcryption broadcasting like the confidentiality of messages,unforgeability and anonymity.In the random oracle model,the security of the scheme is reduced to grading decisional Diffie-Hellman problem (GDDH) to solve.
Non-repudiable Billing Protocol Based on Self-updating Hash Chain for Heterogeneous Wireless Networks
CHEN Shou-guo, FU An-min and QIN Ning-yuan
Computer Science. 2015, 42 (3): 111-116.  doi:10.11896/j.issn.1002-137X.2015.03.023
Abstract PDF(767KB) ( 367 )   
References | Related Articles | Metrics
The integration of heterogeneous wireless networks (HWNs) is an inevitable trend in the development of next-generation networks.UMTS,LTE,WiMAX,WiFi and other wireless networks compete and cooperate mutually.Secure billing is the primary challenge faced by the commercialization of HWNs.We proposed a non-repudiable billing protocol based on the self-updating hash chain for HWNs.By using the novel self-updating hash chain technology,mobile stations (MS) can quickly update available hash chain which ensures continuous fast authentication.The proposed protocol provides the evidence to solve billing disputes,therefore,it achieves non-repudiation,confidentiality and accuracy in billing.Moreover,the theoretical analysis and performance simulation demonstrate that the proposed protocol has some advantages in aspects of low computational cost,lightweight delay and can meet the performance requirements of HWNs.
Security Principles for RBAC-based Authorization Management
XIONG Hou-ren, CHEN Xing-yuan, ZHANG Bin and YANG Yan
Computer Science. 2015, 42 (3): 117-123.  doi:10.11896/j.issn.1002-137X.2015.03.024
Abstract PDF(664KB) ( 639 )   
References | Related Articles | Metrics
Security principles are greatly significant to security analysis of authorization management model,but they are given little attention and are open problems.This paper proposed many security principles for RBAC-based authorization model with the aim at the security of the model.The security properties of RBAC were presented,including simple safety,simple availability,bounded safety,liveness and containment.Based on deep anatomy of security requirement in authorization management,the problems including data consistency,authorization without redundancy,controllable privi-lege diffusing,controllable management privilege delegating,satisfaction of separation of duty and privilege availability were discussed.The proposed security principles include consistency,security and availability principles.Analysis result indicates that the security principles are consistent with the security properties of RBAC,which can support the security requirements of authorization management efficiently and provide criterions for evaluating the security of RBAC-based authorization model.
Research on Model Based Safety Analysis Technology for Avionics System
GU Qing-fan, WANG Guo-qing, ZHANG Li-hua and ZHAI Ming
Computer Science. 2015, 42 (3): 124-127.  doi:10.11896/j.issn.1002-137X.2015.03.025
Abstract PDF(696KB) ( 636 )   
References | Related Articles | Metrics
This paper introduced a new model based method for safety analysis to address the problem of failure modes integrity,dynamic failure and data consistency currently encountered in safety assessments for integrated avionics system.The method models integrated avionics system hierarchically with layers of application operation,function and resource.It simplifies a large part of the analysis,the development of fault trees,and can guarantee the consistency of results.Event-B language is used to model application layer to check the integrity of operations modes and AltaRica is used to model dysfunction of system to solve the problem of dynamic failure.The efficiency and practice of the method are illustrated by analyzing safety of auto pilot system through Rodin tool which is used for analyzing operational modes of application and Simfia tool which is used for safety analysis.
Understanding Spread of Worms with Multi-area and Selectivity
ZHANG Jian-feng, CHEN Gou-xi and YANG Qiu-xiang
Computer Science. 2015, 42 (3): 128-131.  doi:10.11896/j.issn.1002-137X.2015.03.026
Abstract PDF(329KB) ( 423 )   
References | Related Articles | Metrics
Today worm not only spread more quicklying,but also can implement selective infection based on the diffe-rent characteristics of the regions.Firstly,on the characteristic,quantifying certain factors of the average scan rate based on AAWP model by the probability distribution of the vulnerability of different regions,we raised a multi-area and selective worm propagation model named Areas-AAWP with discrete time.Then,under this model,we analyzed the correlation between the scanning strategies adopted by each subarea,judged the degree of relevance to the whole infectious process between subareas,and analyzed the important impact on the overall efficiency of infection by this degree of relevance.Finally,the experiments testify that the whole worm infection rate increases with the degree of correlation among multi-zone scanning increases,and forms the obvious regional differences in infection.
Certificateless Strong Designated Verifier Signature Scheme
ZHANG Yi-chen, LI Ji-guo and QIAN Na
Computer Science. 2015, 42 (3): 132-135.  doi:10.11896/j.issn.1002-137X.2015.03.027
Abstract PDF(413KB) ( 532 )   
References | Related Articles | Metrics
It has great academic significance to research certificateless strong designated verifier signature(CLSDVS) that can solve the certificate management complexity problem in traditional public key cryptosystem and key escrow problem in identity-based public key cryptosystem.We constructed a CLSDVS scheme and proved that our scheme is existentially unforgeable against adaptive chosen message attack under the assumption of the computational bilinear Diffie-Hellman problem and computational Diffie-Hellman problem in the random oracle model.Then,we analyzed the computation cost and communication cost of our scheme.
Research on Fuzzy Search over Encrypted Cloud Data Based on Keywords
FANG Zhong-jin, ZHOU Shu and XIA Zhi-hua
Computer Science. 2015, 42 (3): 136-139.  doi:10.11896/j.issn.1002-137X.2015.03.028
Abstract PDF(434KB) ( 520 )   
References | Related Articles | Metrics
With the increasing amount of data generated by individuals and business users,the advantages of cloud sto-rage such as lower price and flexible use of storage space are obvious.As a large amount of data are outsourced to the cloud server,encryption methods are used to achieve protection and limitations of sensitive data generally.It makes traditional search scheme based on plaintext no longer applicable.How to achieve efficient file search on the basis of privacy protection is the primary consideration.On the basis of the existing encrypted data search schemes,Chinese characteristics of homophone and polysemy were analyzed,and synonyms of keyword were constructed using Chinese and English,and the sets were constructed respectively.Fuzzy search scheme over encrypted cloud data based on keywords was proposed.Search of Chinese fuzzy pinyin and synonyms is achieved,and the private key is protected by pseudo-random function.High security,good practicability and high searching success rate of the system were verified by security analysis and system experiment.
Research on Exploiting DoS Attack Against DNS Based on Information Entropy
YAN Fen, DING Chao and YIN Xin-chun
Computer Science. 2015, 42 (3): 140-143.  doi:10.11896/j.issn.1002-137X.2015.03.029
Abstract PDF(376KB) ( 658 )   
References | Related Articles | Metrics
DNS server has a vital role in the Internet,and it will affect the network to provide normal services to users if DNS is attacked.DNS Query Flood attack sends a lot of fake DNS request to the DNS server,consumes the DNS server resources and causes denial of service.So it is very important to detect timely the attack.Based on the study of the DNS resolution process,we summed up the characteristics of the DNS Query Flood attack.According to the characteristics of attack,we combined the information entropy to determine whether a network abnormalities,and then used sliding window mechanism to determine whether there is any attack.
Security Risk Monitor System of Mobile Sensor Network
NI Ming, ZHANG Hong and LI Qian-mu
Computer Science. 2015, 42 (3): 144-147.  doi:10.11896/j.issn.1002-137X.2015.03.030
Abstract PDF(315KB) ( 377 )   
References | Related Articles | Metrics
This paper proposed an approach based on stability,security,cluster size selection and rational cluster head node choice,which divides the entire network into clusters logically and selects the best head node.With a intermittent anomaly identification based on the auto-regressive(AR) model,it achieves automatic detection of network traffic anomalies on the head node of the cluster and automatic alarm.Experiments show the effectiveness of the proposed method.
Fast Implementation of KLEIN for Resisting Timing and Cache Side-channel Attacks on AVR
WEN Ya-min, LI Feng-xia, GONG Zheng and TANG Shao-hua
Computer Science. 2015, 42 (3): 148-152.  doi:10.11896/j.issn.1002-137X.2015.03.031
Abstract PDF(617KB) ( 414 )   
References | Related Articles | Metrics
With the rapid development of IoT (Internet of Things) applications,lightweight block ciphers are widely focused in the applications of resource-constrained environments.In IoT applications,attackers often use side-channel information to recover secret keys.At RFIDSec 2011,Gong et al.proposed a new lightweight block cipher named KLEIN for the software implementation in resource-constrained environments.We proposed a bitslicing implementation of the KLEIN block cipher based on AVR ASM.In the implementation,look-up tables and logical operations are combined for reducing the computational costs in the MixNibbles step,which leads to a better balance between the algorithm’s speed and storage.Our experiments on AVR show the bitslicing implementation of KLEIN is feasible for practical applications.
On Relationship of Algebraic Degree,Correlation Immunity and Algebraic Immunity for a Class of H Boolean Functions
HUANG Jing-lian, WANG Zhuo and LI Juan
Computer Science. 2015, 42 (3): 153-157.  doi:10.11896/j.issn.1002-137X.2015.03.032
Abstract PDF(330KB) ( 541 )   
References | Related Articles | Metrics
Using the derivative of the Boolean function and the e-derivative defined by ourselves as research tools,we studied the relationship of algebraic degree,algebraic immunity and correlation immunity for H Boolean functions with a specific Hamming weight.We obtained the algebraic degree of the e-derivative which is a component of H Boolean functions deciding the algebraic degree of H Boolean functions.Besides,we determined the e-derivative of H Boolean functions which is closely related to the order of the algebraic immunity of H Boolean functions.We also checked the e-derivative of H Boolean functions which can put algebraic immunity,annihilators,correlation immunity and algebraic degree of H Boolean functions together.Meanwhile,we also deduced two kinds of methods which are formula method and cascade method.By using these two methods we could solve annihilators of the lowest algebraic degree of H Boolean functions.
Location Privacy Protection Method Based on Incremental Nearest Neighbor Query
WANG Peng-fei, LI Qian-mu and ZHU Bao-ping
Computer Science. 2015, 42 (3): 158-161.  doi:10.11896/j.issn.1002-137X.2015.03.033
Abstract PDF(429KB) ( 407 )   
References | Related Articles | Metrics
With the growing popularity of location-based services in people’s daily life,personal location privacy is facing a serious threat.This paper proposed a new location privacy protection method based on incremental nearest neighbor query,which considers network environment reflecting the population distribution.Using the P2P system structure,the method gets rid of the limitations of traditional central server architecture and solves the vulnerability of single point.Meanwhile,this method can guarantee the user’s privacy in the case that proxy user in P2P system architecture is non-credible.Finally,the proposed method was evaluated in simulated data sets of different road density and the result verifies the validity of the method.
Artificial Image Security Degradation Algorithm Based on Invertible Information Hiding in Spatial Domain
LEI Zheng-qiao and XIAO Di
Computer Science. 2015, 42 (3): 162-166.  doi:10.11896/j.issn.1002-137X.2015.03.034
Abstract PDF(933KB) ( 382 )   
References | Related Articles | Metrics
In order to meet the requirement of the digital image business model of try with option to buy,we proposed an artificial image security degradation algorithm based on invertible information hiding in spatial domain.The algorithm uses histogram shifting method of lossless hiding to embed the compensation matrix for recovering the original image precisely.By controlling the embedding depth,the algorithm can generate publishing image which has large distortion but also keeps the main information of the original image.The scrambling and encryption methods are used in the algorithm so that the invaders can not recover the original image correctly by brute-force method without authorized files.The experiment results demonstrate that the algorithm can not only obtain the publishing image which has big differences with the original one,but also recover the exact image when security permission is released.
Quantitative Evaluation for Effectiveness of Code Obfuscation Based on Multi-level Weighted Attributes
XIE Xin, LIU Fen-lin, LU Bin and GONG Dao-fu
Computer Science. 2015, 42 (3): 167-173.  doi:10.11896/j.issn.1002-137X.2015.03.035
Abstract PDF(503KB) ( 490 )   
References | Related Articles | Metrics
In order to overcome randomness and blindness for choosing code obfuscation algorithms in the process of software protection,in view of the problem that quantitative comparison and evaluation of code obfuscation are difficult,a quantitative evaluation method of obfuscation based on multi-level weighted attributes was proposed.From the aspect of attacker,it uses static and dynamic reverse analysis means to analyze the original and obfuscated programs,and quantifies evaluation index based on program attributes.Three-level hierarchical analysis model is constructed,and expert evaluation method is used to compare the importance of program attributes and determine the weights of program attributes.Based on the evaluation index quantitative values and weights of attributes,analytic hierarchy process is used to evaluate different obfuscation methods.Experiment and analysis show that the method can quantitatively compare the effectiveness of different obfuscation algorithms.
Modeling and Analysis of Improved Stochastic Petri Net Based on Delay Characteristics
LIU Jun-qiang, ZUO Hong-fu and PENG Zhi-yong
Computer Science. 2015, 42 (3): 174-177.  doi:10.11896/j.issn.1002-137X.2015.03.036
Abstract PDF(303KB) ( 394 )   
References | Related Articles | Metrics
Random characteristics can be analyzed by stochastic Petri net,but delay time and delay cost problem cannot be computed by stochastic Petri net.In order to solve above problem,a stochastic Petri net model based on delay characteristics was proposed.The delay time and delay cost were analyzed using the model proposed in this paper.The serial and parallel simplification method about delay time and delay cost were studied.Finally,an example of the capital airport shows that the method can represent the emergency rescue time and cost characteristics better than traditional one.
Index of Indoor Moving Objects Based on Semantics and Access Permission
BEN Ting-ting, QIN Xiao-lin and WANG Li
Computer Science. 2015, 42 (3): 178-184.  doi:10.11896/j.issn.1002-137X.2015.03.037
Abstract PDF(593KB) ( 406 )   
References | Related Articles | Metrics
With the development of wireless communication and positioning technology,the index techniques for indoor moving objects are more and more important in location-based services (LBSs).The structure of the indoor scenes structure is complex and diverse,while previous studies regard rooms,corridors,stairways and other indoor entities as the same cells and model a moving object by a moving point,which do not distinguish different semantic meanings between them and do not consider access issues between the objects and cells.To solve this problem,this paper studied a new method of indoor index technique based on semantics and presented an efficient trajectory query processing algorithm based on semantics and access permission.In addition,this paper also proposed an indoor semantic-based model,which gives the formal description of semantics and accesses permission of indoor cells and moving objects.Extensive experiments demonstrate that the proposed index structure is effective,robust and more efficient than the ACII and RTR-tree in several aspects.
Data Storage and Query in Internet of Things
HE Yan-xiang, YU Tao, CHEN Yan-zhao, LI Qing-an and FAN Tong-rang
Computer Science. 2015, 42 (3): 185-190.  doi:10.11896/j.issn.1002-137X.2015.03.038
Abstract PDF(465KB) ( 581 )   
References | Related Articles | Metrics
Internet of Things includes radio frequency identification devices,sensors,smart embedded devices and other heterogeneous devices,in which the services used to mark,perceive,treat,and transmit information are running.So there are a large amount of non-continuous and time-sensitive data.These kinds of data are very important to ensure the normal operation of the industrial chain.However,the existing data management technologies can not fit the need of storage and query of these data.For multidimensional and polymorphic data and data with other features in the Internet of Things,we proposed a multi-level metadata standard with learning support based on service-oriented data management framework of Internet of Things and Dublin metadata standard.Then we designed a process from XML metadata description to data stored in relational databases.Optimized data storage and query programs were proposed.Finally we verified the proposed storage and query schemes.
Fatigue Recognition Algorithm Based on Deep Learning
ZHOU Hui, ZHOU Liang and DING Qiu-lin
Computer Science. 2015, 42 (3): 191-194.  doi:10.11896/j.issn.1002-137X.2015.03.039
Abstract PDF(956KB) ( 1252 )   
References | Related Articles | Metrics
Current domestic and overseas fatigue recognition algorithms are implemented using fatigue features which are mostly singular and man-made.Most of those algorithms have complex structure,low efficiency and weak adaptability for drivers’ individual behavior habit.To this end,the paper put forward a fatigue recognition algorithm based on deep learning.It introduces deep belief network (DBN) to simulate the data distribution of input images,extracts fatigue features automatically layer by layer,and then recognizes state of fatigue from video images based on time window.The algorithm adjusts the learning rate of the net adaptively to reduce pre-training time,uses feedback mechanism to let the net evolve by itself and as a consequence improves its adaptability for user personalized fatigue features.The experimental result shows that our algorithm acquires good fatigue features,and its misjudgment rate reduces gradually along with incremental time.
Affinity Propagation Hierarchical Optimization Algorithm
NI Zhi-wei, JING Ting-ting and NI Li-ping
Computer Science. 2015, 42 (3): 195-200.  doi:10.11896/j.issn.1002-137X.2015.03.040
Abstract PDF(499KB) ( 476 )   
References | Related Articles | Metrics
Affinity propagation (AP) clustering algorithm is a new clustering algorithm,and it is used in many fields well.Affinity propagation clustering algorithm tends to generate more classes than the real data sets.P has a great influence on the result.So this paper proposed an effective affinity propagation clustering’s hierarchical optimization algorithm called as CAP.CAP algorithm uses the CURE algorithm to optimize the result of AP algorithm,and CAP is a semi-supervised clustering algorithm.The result of experiment on five UCI data sets shows that CAP algorithm achieves higher quality than AP algorithm and the number of classes is much closer to the real number.At the same time,CAP also achieves much better clustering result than K-means,Spectral and CURE.
Clustering Algorithm CARDBK Improved from K-means Algorithm
ZHU Ye-hang, LI Yan-ling, CUI Meng-tian and YANG Xian-wen
Computer Science. 2015, 42 (3): 201-205.  doi:10.11896/j.issn.1002-137X.2015.03.041
Abstract PDF(440KB) ( 602 )   
References | Related Articles | Metrics
The difference between our clustering algorithm and batch K-means algorithm is that in our algorithm each point is not only attributable to one cluster,instead affects multiple cluster centroid values,and the degree of influence of a point on a cluster centroid depends on the distance values between this point and the other more near cluster centroids.Our algorithm and a number of different algorithms on a number of different data sets were clustered respectively from the point of view of their clustering result’s five performance index values such as entropy,purity,F1 value,Rand Index and normalized mutual information,and the results show our algorithm has a better clustering results.Our algorithm and a number of different algorithms were clustered respectively on one same data set but under many different initialization conditions,and clustering results of our algorithm are preferably more stable and better.Cluster on different size data sets by our algorithm has a linear scalability and is faster.
Web Content Rating Algorithm Based on Network Community Structure and its Performance Analysis
LIU Yan and WANG Tai
Computer Science. 2015, 42 (3): 206-209.  doi:10.11896/j.issn.1002-137X.2015.03.042
Abstract PDF(332KB) ( 413 )   
References | Related Articles | Metrics
The Web contents are massive,diverse and semantics-missing,which bring significant challenges to the content-rating.This new algorithm takes full advantage of the Web structure feature that the similar topic webpages aggregate into the Web community,and uses the Web community detection algorithm to rate more content-similar webpages automatically when rating one webpage.In addition,it can be used in the current third-party content-rating system.Theory analysis shows that this algorithm significantly raises the efficiency of Web contents rating.
Extraction Method of Text Summarization Based on Event Network
YANG Jun-hui, LIU Zong-tian, LIU Wei and SU Xiao-ying
Computer Science. 2015, 42 (3): 210-213.  doi:10.11896/j.issn.1002-137X.2015.03.043
Abstract PDF(681KB) ( 368 )   
References | Related Articles | Metrics
Text was expressed by the means of event,and event ontology was built by using event as the basic semantic unit.According to the relationship between events,we built event network direct diagram which can express more semantic information of the text and describe the importance of relationship between events.The importance degree of event of the event network corresponding to each node was calculated and ranked by using the PAGERANK algorithm.According to the time sequence of events,event corresponding primitives were exported as abstract.The experimental results show that automatic summary based on the event network method has better performance.
Bayesian Network Structure Learning Algorithm Based on Conditional Mutual Information and Probabilistic Jumping Mechanism
WEI Zhong-qiang, XU Hong-zhe, LI Wen and GUI Xiao-lin
Computer Science. 2015, 42 (3): 214-217.  doi:10.11896/j.issn.1002-137X.2015.03.044
Abstract PDF(327KB) ( 485 )   
References | Related Articles | Metrics
Precise construction of Bayesian network classifier is an NP-hard problem.K2 algorithm can reduce search space effectively and improve learning efficiency,but it requires the initial node ordering as input,which is very limited by the absence of the priori information.On the other hand,K2 algorithm uses a greedy search strategy and is easy to fall into local optimization solutions.This paper presented a new Bayesian network structure learning algorithm based on conditional mutual information and probabilistic jumping mechanism.Firstly,conditional mutual information is used to determine the initial node ordering as input of K2 algorithm.Then probabilistic jumping mechanism is introduced into K2 algorithm to improve the search process and the ability of global optimization,and learn more reasonable network structure.Experimental results over two benchmark networks Asia and Alarm show that this new improved algorithm has higher classification accuracy and better degree of data matching.
New Vis-Meta Graph Knowledge Representation for Association Rules
CHEN Min, ZHAO Shu-liang, GUO Xiao-bo, LI Xiao-chao and LIU Meng-meng
Computer Science. 2015, 42 (3): 218-223.  doi:10.11896/j.issn.1002-137X.2015.03.045
Abstract PDF(993KB) ( 376 )   
References | Related Articles | Metrics
Considering the problems aroused by the traditional association rules presentation formalizing approaches which are powerless to demonstrate the domain knowledge,lack of displaying multi-schema association rules of one to one,one to many,many to one,many-to-many,and especially ignoring the sharing knowledge of discovering results,this paper proposed a novel knowledge representation method for showing multi-mode association rules based on Vis-Meta graph.Firstly,it gave the relevant definitions of Vis-Meta graph and Vis-Meta graph presentation method of association rules,then introduced the conceptual relationship in Vis-Meta graph for knowledge representation,and presented associa-tion rule’s conceptual relationship knowledge representation algorithm,association rule’s instance compared algorithm,as well as association rule’s knowledge representation optimizing algorithm.Finally,with the help of experimental data obtained from demographic data of a province,we finished the visualizing analysis for association rules information.Experimental results turn out that the knowledge representation algorithm proposed has better display effect and knowledge-sharing.
Mining Algorithm of Frequency Domain Migration Intrusion Feature Based on Information Fusion Transfer
MI Xiao-ping and LI Xue-mei
Computer Science. 2015, 42 (3): 224-227.  doi:10.11896/j.issn.1002-137X.2015.03.046
Abstract PDF(691KB) ( 443 )   
References | Related Articles | Metrics
In the power self incentive networks, the difference property of routing phase group characteristics produces resonance signal,therefore frequency domain migration feature needs to be mined for intrusion signal interception.Traditional methods use shuffled frog leaping algorithm for data mining,and the clustering center vector is close to fuzzy edge,resulting in low search and mining accuracy.An improved mining algorithm of frequency domain migration intrusion feature was proposed based on shuffled frog leaping optimal mode information fusion transfer.The power self combination network system model and mathematical model of intrusion signal are constructed.On the basis of frequency resonant slow fading amplitude equalization principle,the multi-source network attack source signals in the coherent point integrated power accumulation scale coordinate are obtained.The Doppler frequency shift fuzzy search algorithm is used for intrusion signal smoothing processing.The intrusion signal state space modal function of Doppler frequency shift is calculated.Amplitude estimation value is obtained.IIR filtering algorithm is used for signal filtering processing to improve the signal purity.The shuffled frog leaping intrusion detection algorithm based on information fusion of transfer is obtained.Feature mining results are optimized.The frequency domain migration intrusion signal feature mining algorithm is completed.The simulation results show that the algorithm can accurately mine the frequency domain migration feature of intrusion signal.The wave ridge highlight is obvious,and it can improve the detection performance of the intrusion signal in low SNR.
Randomized Power Tree Method for Shortest Addition Chains
JIANG Shun-liang, XU Qing-yong, HUANG Wei, YE Fa-mao and XU Shao-ping
Computer Science. 2015, 42 (3): 228-232.  doi:10.11896/j.issn.1002-137X.2015.03.047
Abstract PDF(411KB) ( 1014 )   
References | Related Articles | Metrics
The power tree method is a simple addition chain approximation method for shortest addition chains with its high computing efficiency and ability to generate multi-results by single running,and unfortunately its accuracy is poor.Randomized power tree method improves calculation accuracy significantly while maintaining high efficiency.The methodextends nodes randomly one layer by one layer,and updates the better results by multi-running.For all numbers of n less than 24924,accuracy rate above 95% with an average of 97% was achieved by nine times running,while guarante-eing that the result is suboptimal.The solved problem is up to 155691199 scale by ordinary desktop PC.
Bipartite Graph Scheduling Method for Grid Task Atomized
LI Jian-xun, GUO Jian-hua, LI Wei-qian and CAO Mao-sheng
Computer Science. 2015, 42 (3): 233-236.  doi:10.11896/j.issn.1002-137X.2015.03.048
Abstract PDF(691KB) ( 395 )   
References | Related Articles | Metrics
Focusing on the scheduling problems in grid system,adopting directed acyclic task graph and undirected node graph,the paper proposed a grid scheduling algorithm based on bipartite graph (BGS) by using atom parallel set,and made the dynamic grid scheduling gradually optimized by introducing punishment strategies,load balancing and revived mechanism.The results show that the BGS algorithm can adapt to the changes in grid resources better,reduce the load of tasks,improve the degree of parallel operations and reasonably use the node resources in accordance with system load.
Protein Conformational Space Optimization Algorithm Based on Fragment-assembly
HAO Xiao-hu, ZHANG Gui-jun, ZHOU Xiao-gen, CHENG Zheng-hua and ZHANG Qi-peng
Computer Science. 2015, 42 (3): 237-240.  doi:10.11896/j.issn.1002-137X.2015.03.049
Abstract PDF(631KB) ( 668 )   
References | Related Articles | Metrics
An optimization algorithm based on fragment-assembly was proposed for the optimization problems of protein conformational space.The algorithm employs Rosetta energy model based on the knowledge and coarse-grained to improve the convergence rate.Simultaneously,fragment-assembly techniques are able to compensate the defect of prediction accuracies caused by the inaccuracy of energy functions.The introduction of differential evolution algorithm successfully improves the global searching capability of the algorithm as well.The experiments on five test proteins verify the superior searching performance and prediction accuracy of the proposed algorithm.
Affinity Propagation Clustering Based Ensemble Feature Selection Method
MENG Jun and YU Shuang-yun
Computer Science. 2015, 42 (3): 241-244.  doi:10.11896/j.issn.1002-137X.2015.03.050
Abstract PDF(452KB) ( 396 )   
References | Related Articles | Metrics
Aiming at the problem that only a small part of features are associated with the sample classification in high-dimensional data containing thousands of features,a filtering and grouping based feature random selection ensemble learning method was proposed.Rank aggregation technique was used to select the relevant features,and we grouped them by affinity propagation clustering algorithm using bicor correlation coefficient as distance measure.The feature clusters were produced and the feature pairs from any two different clusters are not correlated.A feature from each cluster was selected randomly,and then a relevant and discriminative feature subspace was generated.In this way,many feature subspaces can be generated.Base classifiers were trained in the produced feature subspaces and fused together using a majority voting method.The experiments on 7 gene expression data sets show that the proposed method can effectively reduce the classification error.Meanwhile,it also has more stable performance,and good expansibility.
On Edge-balanced Index Sets of Two Classes of Nested Network Graph
LIU Jin-meng, HOU Tao and ZHENG Yu-ge
Computer Science. 2015, 42 (3): 245-251.  doi:10.11896/j.issn.1002-137X.2015.03.051
Abstract PDF(769KB) ( 410 )   
References | Related Articles | Metrics
On the basis of smaller power-cycle nested network graph,the edge-balanced index sets of ten-power-cycle nested network graph were investigated.It reduces the difficulty of ten-power-cycle nested network graph labeling using the novel design of the basic graph,nested-cycle subgraph with gear and single-point sector subgraph.When n is an even number,a new method of changing index was provided,simplifying the proving process.The edge-balanced index sets of ten-power-cycle nested graph were determined when m≡1,3(mod 6) and m≥2.This paper proved the existence of the edge-balanced index sets of two classes of nested network graph.The computational formulas and the construction of the corresponding graphs were also provided.
Collaborative Filtering Recommendation Method Based on Dynamic Social Behavior and Users’ Background Information
JIANG Sheng, WANG Zhong-qun, XIU Yu, HUANG Su-bing and WANG Qian-song
Computer Science. 2015, 42 (3): 252-255.  doi:10.11896/j.issn.1002-137X.2015.03.052
Abstract PDF(413KB) ( 373 )   
References | Related Articles | Metrics
To address the difficulty of data sparsity and lower recommendation precision in the traditional collaborative filtering recommendation algorithm,a new collaborative filtering recommendation method was presented based on dynamic social behavior and users’ background information.As the result of user annotation behavior,variable social tags can reflect the changes of user social behavior.Firstly,the similarities of users’ dynamic preferences are calculated based on users’ social tags.Secondly,the similarities between users are calculated based on users’ background information.Finally,the similarities of user rating are calculated based on time weight,and the above three similarities are integrated to get the nearest neighbor set for targeted users to provide more accurate individual recommendation.The experimental results show that the new method can not only improve the accuracy of recommendation,but also solve the problems of data sparsity and cold-start.
Using Bipartite Network for Enhancement of Collaborative Filtering
LENG Ya-jun, LU Qing and ZHANG Jun-ling
Computer Science. 2015, 42 (3): 256-260.  doi:10.11896/j.issn.1002-137X.2015.03.053
Abstract PDF(692KB) ( 376 )   
References | Related Articles | Metrics
Collaborative filtering is one of the most successful and widely used techniques among recommender systems.However,it suffers from serious problem in sparsity.Sparsity in ratings makes the formation of neighborhood inaccurate,thereby resulting in poor recommendations.In this paper,bipartite network was used to alleviate the sparsity problem in collaborative filtering.Users and items are mapped to nodes in bipartite network,and resources on items are redistributed.Resource approach degree between items is computed,and the original rating matrix is converted to complete matrix based on the resource approach degree.Then affinity propagation clustering was applied to cluster the ra-ting matrix to improve the scalability of our approach.Finally,two different recommendation methods were presented.One is generating recommendations according to neighbors in the cluster which active user belongs to (BNAPC1),and the other is generating recommendations according to clusters’ preferences (BNAPC2).Experiments on MovieLens and Netflix datasets show that BNAPC1 is more accurate than BNAPC2,and is also superior to existing alternatives.
Clustering with Mixed Condition Attributes Based on Average Mutual Information
LIU Jin-sheng
Computer Science. 2015, 42 (3): 261-265.  doi:10.11896/j.issn.1002-137X.2015.03.054
Abstract PDF(940KB) ( 429 )   
References | Related Articles | Metrics
There is a great difference between the distances of mixed condition attributes parameter.The numeric condition attributes object with larger and law magnitude tends to be clustered only.With small and chaos magnitude,the cate-gorical condition attributes object which has obvious category characteristics will be ignored.A clustering algorithm based on average mutual information was proposed.First,the size of parameter category characteristics is quantified through entropy.Then,the similarity and the difference between category characteristics are measured according ave-rage mutual information of entropy.The magnitude between distances of numeric and categorical condition attributes parameter is unified.At last,the final clustering result is got by optimizing iterative adaptive process.The experimental results show that the proposed algorithm was high clustering quality and good adaptability.
Fast Video Detection Scheme Based on Multi-core Processor and GPU
YANG Juan, ZENG Miao-xiang, XU Jing and XU Wei
Computer Science. 2015, 42 (3): 266-270.  doi:10.11896/j.issn.1002-137X.2015.03.055
Abstract PDF(526KB) ( 526 )   
References | Related Articles | Metrics
At present,the speed of video detection based on general structure is very slow,and it is difficult to meet the requirement of real-time network video monitoring.This paper showed a new video detection method based on multi-core processor and graphic processing unit (GPU).This method uses multi-core processor to realize video decoding,and uses the GPU to realize the SURF (Speed Up Robust Features) and SVM (Support Vector Machines) algorithm to detect the image.Compared with video detection scheme based on general PC architecture,the performance of the method based on multi-core processor and GPU can be improved over 10 times.
Novel Image Zooming Method Based on Sparse Decomposition
LI Qiu-ju, ZHU Xuan, ZHANG Xu-feng and WANG Ning
Computer Science. 2015, 42 (3): 271-273.  doi:10.11896/j.issn.1002-137X.2015.03.056
Abstract PDF(592KB) ( 420 )   
References | Related Articles | Metrics
Two new dictionaries,RDWT and WAT,were proposed in this paper,and we used them to sparsely decompose one image into cartoon component and texture component.Based on the fact that the cartoon and texture in one ima-ge have different morphological characteristics,we zoomed the cartoon by self-snake model with the characteristics of curvature motion,edge shock and smooth denoising,and zoomed the texture by bicubic interpolation.Through superpo-sing the zoomed cartoon and texture,the zoomed image will be obtained.The experiment results show,compared with the traditional zooming methods processing the whole image,the new zooming model based on morphological component decomposition has good performance for enhancing edge,protecting small curvature and large gradient,and ensuring the texture clear and completion.
Modular Multilinear Principal Component Analysis and Application in Face Recognition
XIE Pei and WU Xiao-jun
Computer Science. 2015, 42 (3): 274-279.  doi:10.11896/j.issn.1002-137X.2015.03.057
Abstract PDF(477KB) ( 595 )   
References | Related Articles | Metrics
Though principal component analysis (PCA) is a classical method for face recognition,the PCA method extracts global features of the original images,and it does not consider the local discriminant features.In contrast,Modular PCA method extracts the important local discriminant features,and it achieves better performance than the PCA method in face recognition.However,vectorization in PCA or modular PCA often causes "curse of dimensionality".In order to extract features from matrix or higher-order tensor objects directly,multilinear principal component analysis (Multilinear PCA) is developed.Multilinear PCA can avoid "curse of dimensionality",meanwhile it will not destroy the original data structure.Inspired by Modular PCA and Multilinear PCA,we proposed a new method called modular multilinear principal component analysis (M2PCA) for face recognition.Experiments were conducted on the Yale,XM2VTS and JAFFE databases respectively,and experimental results indicate that,under the same condition of sub-blocks,the proposed method is obviously superior to the general Modular PCA.
Fast Image Stitching Algorithm Eliminates Seam Line and Ghosting
QU Zhong, QIAO Gao-yuan and LIN Si-peng
Computer Science. 2015, 42 (3): 280-283.  doi:10.11896/j.issn.1002-137X.2015.03.058
Abstract PDF(1104KB) ( 1054 )   
References | Related Articles | Metrics
The image stitching algorithm using SIFT feature is complicated and hard to handle the seam line as well as ghosting.We proposed a fast image stitching algorithm to eliminate the seam line and ghosting.First,we extracted SIFT features in the given area of the full image.After obtaining the feature correspondences,we used the RANSAC algorithm to compute the homograph matrix H.Then,an integral image without seam line was received by the combination of the best seam line and the improved image fusion algorithm.The experimental results show that the algorithm not only eliminates the seam line and ghosting effectively,but also improves the efficiency of image stitching.
Obstacle Detection Method for Mobile Robot Using Active Omnidirectional Vision Sensor
TANG Yi-ping, JIANG Rong-jian and LIN Lu-lu
Computer Science. 2015, 42 (3): 284-288.  doi:10.11896/j.issn.1002-137X.2015.03.059
Abstract PDF(1288KB) ( 579 )   
References | Related Articles | Metrics
To solve the problems of excessive consumption of computing resource,poor real-time performance and limited detection range of current mobile robot vision system,a fast and efficient obstacle detection method was proposed in this paper based on active omnidirectional vision sensor (AODVS).AODVS integrates ODVS with single view point and a planar laser generator composed by four red liner laser that installed on the same plane.All obstacles around the mobile robot can be detected in real time by AODVS.According to laser information on the surrounding obstacles projected by planar laser generator,the obstacles distance and direction information are obtained with robot vision method.Then,based on this information,the direction of movement and speed of the mobile robot are controlled adopting omnidirectional mobile robot obstacle avoidance algorithm.The experiment results show that the obstacle detection method based on AODVS can achieve the goals of avoiding the obstacles quickly and accurately and reducing the requirement of computing resource for the mobile robot.
Face Recognition with Occlusion Based on Removing Outliers Area
LI Dong-mei, XIONG Cheng-yi, GAO Zhi-rong, ZHOU Cheng and WANG Han-xin
Computer Science. 2015, 42 (3): 289-295.  doi:10.11896/j.issn.1002-137X.2015.03.060
Abstract PDF(1620KB) ( 603 )   
References | Related Articles | Metrics
Aiming at the issue of face recognition with partial occlusion,an improved face recognition method based on removing the outlier area was proposed in this paper.A mean face image is firstly obtained from train images,which is subtracted by the test face to form an error face image.Then the error face image is used to obtain the occlusion area of the test image by image segmentation technique,and the train images and test image are tailored by removing the corresponding occlusion area.Finally,face recognition is performed by linear regression classifier or sparse coding classifier.Compared to the similar works,the proposed method has considerable recognition performance improvement with relatively sample computational complexity.Simulation results based on the standard extended Yale B and AR face databasesshow effectiveness of the proposed method.
Feature Extraction Based on Low Rank Representation Linear Preserving Projections
YANG Guo-liang, XIE Nai-jun, YU Jia-wei and LIANG Li-ming
Computer Science. 2015, 42 (3): 296-300.  doi:10.11896/j.issn.1002-137X.2015.03.061
Abstract PDF(775KB) ( 457 )   
References | Related Articles | Metrics
For preserving the low rank properties the same,we proposed an algorithm,called linear preserving projection based on low rank representations (LLRLPP),to reduce the dimension of data.It can preserve the low rank properties of the original data space in the resulting low dimensional embedding subspace and correctly learn the low-dimensional subspace.Through constructing two different low rank representation model,the low rank weights of representing different structural characteristics are revealed.Then the low-dimensional subspace of the original high-dimensional data is obtained by preserving such low rank weight relationship.The effectiveness of the proposed method is verified on two face databases(ORL,Yale) with the traditional algorithms.
Video Smoke Detection Using Two-level Classification Algorithm
TONG Bo-bing and WANG Shi-tong
Computer Science. 2015, 42 (3): 301-306.  doi:10.11896/j.issn.1002-137X.2015.03.062
Abstract PDF(1000KB) ( 611 )   
References | Related Articles | Metrics
In order to improve the accuracy of video smoke detection,a probability-based two-level nearest neighbor classification algorithm (PTLNN) was proposed to detect smoke in video.Aiming at minimizing the mean absolute error of principle,combining the advantages of AdaBoost and KNN algorithm,and fully considering local and global sample distribution,the proposed algorithm can significantly improve the classification accuracy.The proposed algorithm adopts the discrete cosine transform (DCT) and discrete wavelet transform (DWT) two ways to extract smoke characteristics.By comparing with the traditional algorithms,the proposed PTLNN algorithm with the discrete cosine transform has better effectiveness on video smoke detection which not only meets the real-time requirements but also improves the detection accuracy.
Image Classification and Identification through SVM Based on Fuzzy Kernel Clustering
YU Wen-yong, KANG Xiao-dong, GE Wen-jie and WANG Hao
Computer Science. 2015, 42 (3): 307-310.  doi:10.11896/j.issn.1002-137X.2015.03.063
Abstract PDF(400KB) ( 510 )   
References | Related Articles | Metrics
A method of image classification and identification combined with characteristic field and SVM based on fuzzy kernel clustering was proposed in this paper.First,the structure corresponds to image color of human visual characteristics and data field of texture.For one thing,the new threshold is introduced and the image texture is established.For another,attractive pixel to pixel area is weighted and processed,and the spatial distribution of color is described by using dispersion of color spatial distribution.Second,SVM based on fuzzy kernel clustering is adopted to study a classification of image identification.On the feature space,not only the relationship of samples between its cluster centers but also the each samples are all taken into account.The relation between each samples in the cluster is measured based on fuzzy connectedness and a binary tree classifier is constructed.Experimental results show that this method can achieve a better effect of image classification.
Multiple Ship Tracking in Inland Waterway via Deformable Part Model
ZHU Lin, GUO Jian-ming, LIU Qing and LI Jing
Computer Science. 2015, 42 (3): 311-315.  doi:10.11896/j.issn.1002-137X.2015.03.064
Abstract PDF(701KB) ( 366 )   
References | Related Articles | Metrics
The closed-circuit television (CCTV) surveillance system is developing rapidly in recent years.But the intelligence level is relatively low.In this paper,a robust multiple ship tracking algorithm was proposed based on the defor-mable part model.The proposed algorithm treats every ship as a part.By incorporating the spatial constraints,the interrelations model between ships with the minimum spanning tree model can be effectively built.Then the robust multiple ship tracking is accomplished based on the deformable part model.Moreover,aiming at obtaining an accurate paramete-rized appearance model of ships,the HOG features combined with the fuzzy SVM is adopted to train those object regions.Especially,because of the ambiguity in the fuzzy SVM,every training samples are given different importance so as to obtain a more accurate appearance model.At the same time,structured learning can guarantee to update the interrelation parameters on time when ships move.Experimental results demonstrate that our proposed algorithm is suitable for inland waterway and can accomplish robust and effective multiple ship tracking.
Image Restoration Algorithm of Unknown Priori Pixel Based on Artificial Fish Swarm Decomposition
SUI Dan and GAO Guo-wei
Computer Science. 2015, 42 (3): 316-320.  doi:10.11896/j.issn.1002-137X.2015.03.065
Abstract PDF(947KB) ( 391 )   
References | Related Articles | Metrics
Because unknown pixels lack prior information,module matching and edge structure information is unknown,and the holographic reconstruction is difficult.The traditional method uses multidimensional search method of subspace feature information,which failes to achieve the fine image texture template matching of structure information,and the effect is not good.Introducing the artificial fish swarm algorithm,this paper proposed a new image holographic image restoration algorithm based on artificial fish swarm micro decomposition and brightness compensation.Sub space feature information multidimensional search method is used for unknown pixel confidence updates.In order to maintain the continuity of damaged region in image,the artificial fish swarm algorithm decomposition model is constructed,combined with the edge feature of image brightness compensation strategy,and the image restoration algorithm is obtained.The simulation result shows that it has a good visual effect in image restoration of a priori unknown pixel,resulting in less recovery time and computation costs,and the stability and convergence performance are improved.SNR error is smaller within 6%,so it has superior performance in application.