15 May 2017, Volume 43 Issue 5
-
Select all|
-
Computer Engineering. 2017, 43(5): 1-7. https://doi.org/10.3969/j.issn.1000-3428.2017.05.001Aiming at the problems of low efficiency and single function of the existing illegal forensics analysis system for public security,this paper proposes an illegal forensics analysis system based on cloud platform and big data.It uses MongoDB to store original and mass vehicle trajectory data and MapReduce distributed computing to mine and screen trajectory data.The system scheduler on big data platform is used to schedule the resources of the system and monitor the running state of the system.Based on the extraction of discrete trajectory data for violation of the provisions of time travel,illegal provisions of the route,and fatigue driving for forensic analysis to find illegal vehicles.Experimental result shows that,when the vehicle trajectory data scale is large,the system still has high efficiency and stability.
-
Optimization Strategy of Information Interaction Maximum Flow Transmission in Vehicle Ad Hoc NetworkComputer Engineering. 2017, 43(5): 8-15，22. https://doi.org/10.3969/j.issn.1000-3428.2017.05.002Aiming at the transmission instability problem when multiple users request a large number of data resources in intelligent vehicle network,this paper proposes a Set Covering Maximum Flow(SCMF) algorithm in mobile vehicle node environment.It uses set covering method to consider more request details and analyzes optimal stochastic demand model,which optimizes information resource sharing and data transmission process,guarantees vehicle information interaction between vehicle and wireless Access Point(AP) nodes with maximum flow efficiency transmission,and processes the customer’s big data resource request quickly and steadily.Experimental results show that the SCMF algorithm can obtain higher transmission efficiency with lower network overhead and transmission delay.It improves the utilization of network resources and comprehensive performance.
-
Computer Engineering. 2017, 43(5): 16-22. https://doi.org/10.3969/j.issn.1000-3428.2017.05.003
According to the passenger movement pattern and the hot pick-up and drop-off areas extracted from taxi driving passenger data,this paper proposes a functions discovery method of taxi hot areas.Firstly,it uses taxi driving data clustering algorithm based on the temporal and spatial characteristics of traffic data to realize hot region division.Then,the passengers travel character discovery model of passengers in hot region based on Latent Dirichlet Allocation(LDA) is built to realize clustering hot taxi region with similar passenger travel mode.Finally,by summarizing the specific function of each area,it can find the relationship between area function and passenger movement patterns at different period of passenger flow.The experimental results show the method can effectively discovery the function characteristics of hot areas.
-
Computer Engineering. 2017, 43(5): 23-27. https://doi.org/10.3969/j.issn.1000-3428.2017.05.004Focusing on the evaluation requirements of mobility support for the IEEE 802.11p protocol,this paper constructs a typical Internet of vehicles application scenario based on IEEE 802.11p and uses Poisson process to describe the vehicle arrival model in Internet of vehicles system.It uses Veins platform to build the simulation scene and sets the simulation parameters.From the end to end delay,packet loss rate and throughput,the influence of traffic density and vehicle running speed on the performance of IEEE 802.11p and the applicability of IEEE 802.11p in the Internet of vehicles environment are tested and analyzed.Test result shows that the vehicle running speed has a great influence on the Internet of vehicles performance of IEEE 802.11p,but the traffic density has little influence on the performance of it.At the same time,the IEEE 802.11p protocol is applicable in active road safety and it helps improve the cooperative transport efficiency.
-
Computer Engineering. 2017, 43(5): 28-34. https://doi.org/10.3969/j.issn.1000-3428.2017.05.005
Aiming at the problems in traffic sign detection process,such as that traffic sign shows dimensional changes,rotation distortion,projection distortion or the sign is partially occluded,in this paper,a traffic sign detection algorithm based on saliency map and Fourier Descriptor(FD) is proposed.Firstly,the Frequency Tuning(FT) method is used to get the saliency map.Secondly,binary operation is utilized on saliency map to achieve binary image and get the regions of traffic sign.By extracting outer contour of the regions,features of contour length and aspect ratio are utilized to filter interference information.Thirdly,the convex hull processing is applied for the eligible contours and FDs of convex hull are extracted and normalized.Finally,detection results are obtained according to the comparison with standard data.Experimental results demonstrate that the detection rate of the proposed algorithm is more than 95%.This algorithm meets the real-time performance requirement of traffic sign detection.
-
Computer Engineering. 2017, 43(5): 35-39,46. https://doi.org/10.3969/j.issn.1000-3428.2017.05.006
In the rail transit clusters dispatch system,the parameters of database connection pool is one-time set and can not be modified.Aiming at this shortage,a dynamic allocation strategy of multi-client database connection pool is designed.It uses a dynamic allocation algorithm to allocate the optimal number of connections according to the different access frequency of each client for current client,so as to achieve the purpose of improving the utilization of system resources.Experimental results show that the dynamic allocation strategy proposed in this paper can shorten the response time of connection pool and improve the efficiency of system.
-
Computer Engineering. 2017, 43(5): 40-46. https://doi.org/10.3969/j.issn.1000-3428.2017.05.007The management and control of power has already become a hot issue in the area of management of High Performance Computing(HPC) system and distributed data center.In order to adapt the limited power to the dynamic change of the power supply capacity when the supply of energy is limited in the computer room,it is necessary to control the upper power limit of cluster system.Aiming at this problem,this paper designs and realizes a power capping control system based on RAPL.By constructing the power model of cluster system,utilizing RAPL’s capability of controlling the power consumption limit of CPU and combining the method of measuring the difference of power,it sets the upper limit of energy consumption of cluster system within the previously set power cap.On this basis,it tries to reduce the losses of performance as much as possible.The result of experiment shows that this system can reduce the peak power effectively with slight performance and keep it below the power cap stably.
-
Computer Engineering. 2017, 43(5): 47-54，59. https://doi.org/10.3969/j.issn.1000-3428.2017.05.008Aiming at the problems that clone code tracking is inaccurate and evolution pattern recognition is tedious,especially clone class merging is difficult,this paper proposes a method of clone code evolution traces building and pattern recognition based on graph model.Clone class mapping is implemented preliminarily between adjacent versions using theme probability model,location overlap rate of clone code and cloned fragment text similarity are calculated to map clone fragments,clone class mapping is repaired to get accurate mapping results between adjacent versions.Short-term evolution pattern is recognized and figure model is established on the basis of relationship of clone class mapping quantity and the degree of cross.Considering clone class as points,mapping relationship as edges,and according to the form of production,short-term evolution model of clone class is indicated.The clone genealogy is extracted using breadth-first search algorithm,and the long-term evolution model is identified according to types of clone class included in clone genealogy and whether there is loop among them.Experiment is conducted on 70 versions of 5 open source software.Results show that about 95% of the clone code is stable in evolution,about 1% of clone code undergoes merging or recombination,and about 80% of clone code′s life cycle does not exceed half of the total quantity of released version.
-
Computer Engineering. 2017, 43(5): 55-59. https://doi.org/10.3969/j.issn.1000-3428.2017.05.009Under the condition of heterogeneous multi-core processor,the Min-Min scheduling algorithm has better performance,but it has a problem in the real-time response of the system.The Least Slack First(LSF) scheduling algorithm,Earliest Deadline First(EDF) scheduling algorithm and Highest Value First (HVF) scheduling algorithm although perform well in real-time system task scheduling response real-time performance,but they can not be used in the heterogeneous multi-core processor environment.Therefore,this paper proposes a task scheduling algorithm with high real-time performance,named HRSA.It is based on the Min-Min scheduling algorithm and the LSF,EDF,HVF algorithms’ scheduling strategy.In this algorithm,the tasks’ energy consumption,tasks’ completion value and response ratio are combined.The heterogeneous multi-core processor tasks are dynamically scheduled and the system response time of real-time tasks is reduced.Experimental results show that,compared with the EDF algorithm and Min-Min algorithm,HRSA algorithm brings higher value when consuming unit energy and has shorter response time for high real-time task processing.
-
Computer Engineering. 2017, 43(5): 60-66,74. https://doi.org/10.3969/j.issn.1000-3428.2017.05.010Remote image data parallel processing system basically relies on foreign commercial products,while domestic independent parallel processing system the task process support capability and parallel processing cannot meet the need of scale production.Therefore,based on HDFS,MapReduce cluster parallel architecture,CPU/GPU cooperative parallel processing,memory mapping,BMP and so on,this paper proposes the architecture and realization of a high-performance distributed parallel processing platform with process driven execution.Experimental results show that the multi-granularity mixing parallel processing architecture of cluster and workstation magnificently increases the parallel performance of the platform,which proposes an independent resolution to the scale production of massive remote image data products.
-
Computer Engineering. 2017, 43(5): 67-74. https://doi.org/10.3969/j.issn.1000-3428.2017.05.011This paper discusses the principle,mechanism and strategy of hard real-time task scheduling in mobile termind devices.The mobile terminal has not only the timeliness requirements of the task,but also management requirements on energy consumption.This paper attempts in the two constraint conditions of task scheduling,completes optimal energy consumption the two main objectives.The paper researches the regularity of scheduling under these two constraints,and proposes a packet scheduling algorithm based on the nearest deadline,the algorithm uses dynamic optimal strategy to decide.Experimental results show that the algorithm can ensure real-time scheduling and least energy consumption.
-
Computer Engineering. 2017, 43(5): 75-82. https://doi.org/10.3969/j.issn.1000-3428.2017.05.012In order to make the message transmit in a delay tolerant network more efficiently,the node location,moving direction and other geographic information and the degree of nodes frequently encounter history are combined,this paper proposes a delay tolerant network geographic routing algorithm is proposed.The algorithm uses nodes encounter history information as news distribution copy,thereby it reduces the copy number.The angle of the moving direction of the neighbor nodes for the relay selection is further optimized,so as to achieve high delivery ratio and low average end-to-end delay.Simulation results show that when nodes move relatively slow,and news life cycle is short,the message delivery rate of the proposed algorithm is better than that of Spray & Wait,Epidemic and Prophet routing algorithms.When message life cycle is very short and the cache resource of nodes is the bottleneck factors limiting the performance of the routing algorithm,the network load rate of the proposed algorithm is less than 40% of Epidemic,less than 25% of Prophet,and average number of hops and average delay remain less.
-
Computer Engineering. 2017, 43(5): 83-87. https://doi.org/10.3969/j.issn.1000-3428.2017.05.013Long Term Evolution(LTE) network performance heavily depends on the network planning.To implement the self-planning and optimization of LTE network coverage,network run is on optimal status,this paper proposes a new balance network coverage and interference mathematics model and evaluating function.According to the characteristic of LTE system,the model reconstructs evaluating method and evaluating function.It uses simulation annealing algorithm to solve NP problems,and minimize evaluating function.Simulation results show that the model and algorithm can improve network coverage rate,reduce unnecessary overlap coverage to decrease small region interference,and increase network performance.
-
Computer Engineering. 2017, 43(5): 88-91. https://doi.org/10.3969/j.issn.1000-3428.2017.05.014Aiming at the problem of low utilization rate of user resources and user Quality of Service(QoS) guaranteeing,this paper proposes a distributed D2D power allocation method by combining the Cognitive Radio(CR) technology and D2D technology for the wireless cellular networks.Based on the CR technology,the method can perceive the available spectrum resources and price it.Then,by changing the price to achieve the new Nash equilibrium,the iterative optimization makes the Nash equilibrium maximize the profit of the D2D pair,so as to effectively use the available spectrum resources.Experimental results show that compared with the traditional D2D communication model,the proposed method improves significantly in the data transmission efficiency and the utilization ratio of the resources in the communication system.
-
Computer Engineering. 2017, 43(5): 92-96. https://doi.org/10.3969/j.issn.1000-3428.2017.05.015To meet the high quality communication needs,the AltBOC modulation mechanism is proposed.However,due to the characteristics of complex multi-peaks for AltBOC signal,the problem of main peak estimation difficulty is caused in signal synchronization.Thus,according to the characteristics of AltBOC signal,the local co-direction and orthogonal correlation functions are constructed.And then through the leading and lagging processing of correlation function,an auxiliary correlation reconstruction algorithm is established,which to recovery the main peak,and overcome the side peaks.Simulation results demonstrate that Auxiliary Correlation Reconstruction(ACR) algorithm can recovery the main peak,and it can adapt to lower Signal to Noise Ratio(SNR) requirements.Meanwhile,ACR algorithm is better than BPSK-like algorithm,SCPC algorithm and Filtered algorithm in main peak detection and side peaks inhibition.
-
Computer Engineering. 2017, 43(5): 97-102,107. https://doi.org/10.3969/j.issn.1000-3428.2017.05.016In IPv6 over Low-power Wireless Personal Area Networks(6LoWPAN) based on Mesh-under routing,the missing fragmentations are retransmitted with the end-to-end method to ensure the reliable reassembly of packets.But the end-to-end retransmission can result in the large data transmission delay and energy consumption,decrease the destination’s reassembly successful rate and influence other data’s transmission.In order to solve the problem,a cache and retransmission mechanism based on the forwarding intermediate nodes is proposed.The intermediate nodes can cache and manage the forwarding fragmentations,and the destination node’s missing fragmentations are retransmitted by the intermediate nodes to improve the performance of the entire network.Simulation result shows that the performance of the proposed mechanism is better than the traditional RIF_SN,RPF_SN mechanism,the energy utility value can be drastically increased by 36.7%,the delay is decreased by 28.7% and the packet’s reassembly rate increased by 34.9%.
-
Computer Engineering. 2017, 43(5): 103-107. https://doi.org/10.3969/j.issn.1000-3428.2017.05.017In order to guarantee the security of signature scheme under quantum attack,most of the existing lattice-based signature schemes are based on the difficulty hypothesis under general lattice rather than ideal lattice,and it caused lower efficiency of these schemes.To overcome the problem,a new digital signature scheme based on ideal lattice is proposed.By reducing the problem to the PLWE,it is proved that an attacker can not forge a legitimate signature for a new message even if he gets multiple existing messages and the corresponding signature.Analysis result shows that the signature length and the key length of the scheme are reduced by O(n),compared with the existing digital signature scheme based on general lattice.
-
Computer Engineering. 2017, 43(5): 108-114. https://doi.org/10.3969/j.issn.1000-3428.2017.05.018In order to protect the user’s identity and location information in sparse environment,this paper proposes a privacy-preserving method of Location Based Service(LBS),by determining whether or not the accessed LBS service is anonymous to add the virtual user or virtual location.According to the characteristics of time delay,the real user location information and distribution are simulated to ensure the fidelity of virtual information.Virtual user information or location information and their relevance mapping complexity are added to achieve user anonymity.Experimental results show that compared with privacy-preserving method of randomly generated virtual information,the virtual information generated by this method is more consistent with the real scene and has better privacy-preserving effect.
-
Computer Engineering. 2017, 43(5): 115-120,128. https://doi.org/10.3969/j.issn.1000-3428.2017.05.019
Currently HTTPS protocolused by Web applications provides a good message encryption mechanism for the message requests between the client and the server,but it cannot selectively process the application layer message,thus the whole process message request interaction does not encrypt,which cannot prevent replay attacks in the whole process.So,this paper proposes an anti-replay attack scheme based on counterand dynamic calibration,which uses the current time on the server as the dynamic calibration,and sets the counter mechanisms feedback client data packet loss.Experimental results show the scheme compared with the sheme based on the single serial number mechanism can relieve dependence on the database server.Compared with thesheme based on a single time stamp mechanism,it can avoid clock synchronization and feedback data packet loss.
-
Computer Engineering. 2017, 43(5): 121-128. https://doi.org/10.3969/j.issn.1000-3428.2017.05.020In order to meet the high level security requirement of specific application areas,based on domestic basic software and hardware platform,a user behavior measurement method is proposed.Through the combination of USBKey dual factor authentication mechanism and Trusted Cryptography Module(TCM) authentication and authorization,the security authorization of user identity is realized.User behavior control chain is created according to different user identity,and a trust measure of fine grain to specify the user and process and operate the specified system resource is done,so as to realize the hierarchical and dynamic access control of the system resource according to the user’s behavior.Dynamic encryption and decryption for different user identity is provided to ensure the security of high level users while operating system resources.Experimental results show that the method can effectively improve the security of the system.
-
Computer Engineering. 2017, 43(5): 129-133,142. https://doi.org/10.3969/j.issn.1000-3428.2017.05.021In order to adapt the change of network topology and enhance the security of routing protocol,a heuristic strategy of malicious node discovery and isolation is proposed for security applications of Mobile Ad-hoc Network(MANET).It calculates the difference value between two destination sequence numbers from the received packet and routing table respectively,compares it with the heuristic calculated difference threshold to distinguish whether a node is suspicious,and executes bait detection for suspicious nodes to find and identify the malicious nodes.And then,the malicious nodes are isolated to build a secure and reliable route for data transmission.Simulation results show that,compared with two secure routing protocols including SAODV and ARAN,the AODV routing protocol modified by the new strategy has higher packet delivery rate and more stable transmission performance in the case of different number of nodes,node movement speed and the proportion of malicious nodes.
-
Computer Engineering. 2017, 43(5): 134-142. https://doi.org/10.3969/j.issn.1000-3428.2017.05.022Based on the architecture of General Reconfigurable Processor(GReP),this paper proposes an Advanced Encryption Standard(AES)-128 implementation scheme,which is optimized in algorithm structure,Processing Element(PE) utilization ratio and storage cost.This scheme uses a 256 bit×32 bit look-up table to complete the byte instead of mixing with the column on encryption time,improves PE utilization ratio through data decomposition and data stream decomposition,and reduces storage overhead through local register and reasonable layout of PE.Experimental result indicates that,PE utilization ratio reaches 60% and memory access cost reduces 74% than totally using shared memory.Compared with the serial method on Intel Atom230,this scheme increases throughput rate by about 100%.
-
Computer Engineering. 2017, 43(5): 143-148. https://doi.org/10.3969/j.issn.1000-3428.2017.05.023In Word2vec framework,aiming at the feature of Microblog text,this paper proposes the method using word vector or sentence vector of high dimension word database mapping calculation,and constructs sentence vector using three algorithms:Using Word2vec to expand text,and TF-IDF to obtain sentence vector.Adding the word vectors into sentence vector;Building a word bank to obtain high dimension vector space for the sentence.Through comparing the three methods,it selects the model fitting Microblog field.Experimental results show that the sentence vector calculation method using high dimension word database mapping is the best.
-
Computer Engineering. 2017, 43(5): 149-155. https://doi.org/10.3969/j.issn.1000-3428.2017.05.024
-
Computer Engineering. 2017, 43(5): 156-162. https://doi.org/10.3969/j.issn.1000-3428.2017.05.025This paper analyzes the characteristics of the unsupervised and supervised fault diagnosis methods,and presents an Imbalanced Support Vector Data Description(ISVDD) algorithm,which combines both advantages.The algorithm can find the most representative feature based on the description of the boundary distribution of samples under normal operating conditions. It takes in the supervised fault diagnosis method,and describes the true boundary of samples under normal operating conditions more correctly by introducing the discriminant information from fault operating conditions. It is optimized for fault detection where the imbalance data is common. The empirical error which is sensitive to the sample distribution in traditional SVDD is replaced by the Area Under Curve(AUC) which is robust to the sample distribution.Numerical simulation and industrial cases are presented to verify the effectiveness of the proposed algorithm.
-
Computer Engineering. 2017, 43(5): 163-168,173. https://doi.org/10.3969/j.issn.1000-3428.2017.05.026As traditional Adaboost algorithm is sensitive to noisy sample and the linear combination of base classifiers is irrational,a piecewise nonlinear Adaboost algorithm based on noise self-detection called NDK Adaboost is proposed.NDK Adaboost,drawing on traditional Adaboost algorithm whose error rate in training set decreases with iteration times exponentially,establishes directly a noise detection model to recognize noise,and maps the prediction samples to the relative positions of the training set.According to the neighbor samples’ distribution,it determines the weight of the base classifier.A higher classification accuracy rate of the algorithm can be drawn out among the different sample distribution.Experimental results show that,compared with the traditional Adaboost algorithm and the related improved algorithms,NDK Adaboost has a higher classification accuracy rate.
-
Computer Engineering. 2017, 43(5): 169-173. https://doi.org/10.3969/j.issn.1000-3428.2017.05.027
On the basis of Histogram of Oriented Gradient with Support Vector Machine(HOG-SVM) algorithm and LeNet network model, a pedestrian detection algorithm which is the combination of HOG and Convolutional Neural Network(CNN) is proposed. Firstly, multi-scale sliding windows are used to extract the HOG features which are then sent to SVM classifier to find the candidate regions.The regions are judged according to the posterior probability. And the CNN algorithm is used to eliminate the false detection window. In order to solve the problem that a single target is framed by multiple candidate regions, the Non-maximum Suppression(NMS) algorithm is used to fuse the multi-rectangles, remaining the largest posterior probabilitywindow and suppressing the overlapped windows.In the classifying process, the candidate region is judged as pedestrian region based on the posterior probability in SVM and LeNet. Experimental results show that this algorithm can get higher recognition rate and recall rate compared with HOG-SVM and LeNet algorithms.
-
Computer Engineering. 2017, 43(5): 174-178,184. https://doi.org/10.3969/j.issn.1000-3428.2017.05.028Image segmentation algorithm ignores the image patch which contains the important local facial feature should have bigger weights of voting in the final classification.To solve this problem,this paper proposes an automatic weighting sparse representation algorithm.Through using sliding windows with overlapping,the mean square error value is calculated,and an automatic weighting strategy is given to measure the weight of each patch in the final classification.Experimental results on the public data set show that,compared with the commonly used classification algorithm and block algorithm,the proposed algorithm can get higher recognition rate in face recognition wherever it gives weights to the minimum error or the biggest votes.
-
Computer Engineering. 2017, 43(5): 179-184. https://doi.org/10.3969/j.issn.1000-3428.2017.05.029
Aiming at the theme drifting and the page weight splitting of traditional PageRank algorithm,an improved PageRank algorithm is proposed.In order to improve the user query efficiency and search quality,combined with the time feedback factor,it makes a comprehensive analysis on user forwarding,user comments and micro-blog mentions.Statistical analysis is used to measure the contribution of user behavior in the ranking of micro-blog user influence.By using the improved TF-IDF algorithm to calculate the similarity weight of the topic,the user can select the Web page with high relevance to obtain the corresponding PageRank weight.Experimental results show that compared with common microblog ranking algorithms,the improved PageRank algorithm has better user influence ranking effect.
-
Computer Engineering. 2017, 43(5): 185-190,196. https://doi.org/10.3969/j.issn.1000-3428.2017.05.030The quality of the recommendation system is affected by the computing of the similarity between users in the collaborative filtering algorithm. Especially in the case of data sparsity,compared with the actual situation, the resulting relationship between users deviates too much, affecting the accuracy of recommendation. Aiming at the problem, a new similarity calculation method is proposed.Using the stability of a user’s score structural features, combined with the preference distance between the score structure, the characteristics of the two parts are fused to recalculate the similarity between user preferences. On the whole, this method makes full use of the user’s score features and score data space. Experimental results on the MovieLens dataset show that compared with the traditional User-based Similarity(UBS)collaborative filtering algorithm and Cosine(COS)similarity algorithm,the recommendation accuracy of the proposed algorithm has an average increase of 3.94% and 2.99%.
-
Computer Engineering. 2017, 43(5): 191-196. https://doi.org/10.3969/j.issn.1000-3428.2017.05.031Aiming at the situations of poor adaptive learning effect and dynamically changed negotiation environment of current Agent production and marketing negotiation,the influence of conflict level,cooperation possibility and negotiation remaining time on negotiation in the dynamic negotiation environment are considered,and the entropy evaluation method is used to determine the weights of the three factors and perform linear weighting.According to the difference of the current negotiation issues,this paper constructs a model for predicting the concession amplitude based on dynamic selective ensemble learning,and proposes a negotiation optimization strategy of supply chain production and marketing.The experimental results show that compared with the single machine learning negotiation strategy,the proposed strategy improves the adaptive learning success rate and joint utility of Agent.It can ensure mutual benefit of both sides of the supply chain production and marketing to achieve win-win situation of cooperation between the two sides.
-
Computer Engineering. 2017, 43(5): 197-203,209. https://doi.org/10.3969/j.issn.1000-3428.2017.05.032Ontology mapping is aimed to establish the semantic relationship of the elements between the two ontologies, and the key to affect the ontology mapping is the method of calculating the similarity. In order to solve the imprecise semantic relation problem of the similarity calculation method, this paper proposes an ontology mapping method, which transforms the ontology mapping problem into the problem of solving the maximum common sub-graph. Ontology with graph structure can better reflect the latent semantic relations between the structure of the body and the maximum common sub-graph is introduced to extract the common parts of the ontology. The properties of the maximum common sub-graph are used to measure the similarity between two ontology elements.Then the mapping relationship between the two ontologies is obtained. Compared with CtxMatch and COMA,the proposed method is proved to have a certain improvement in the recall rate and the precision.
-
Computer Engineering. 2017, 43(5): 204-209. https://doi.org/10.3969/j.issn.1000-3428.2017.05.033In view of the problem that the accuracy of abnormal traffic identification is low and fast identification is dependent on the threshold,an abnormal traffic identification method based on BoW(Bag of Words) model clustering is proposed.By means of K-means mean clustering for existing abnormal traffic and normal traffic,it finds the key points of network traffic.The original traffic is tranformed and mapped to the corresponding traffic critical points and then histogram is established.Abnormal traffic is detected by using semi-supervised learning.The experimental results show that this method has better recognition effect of abnormal traffic compared with identification method based on Naive Bayes(NB),Support Vector Machine(SVM) and others.
-
Computer Engineering. 2017, 43(5): 210-216. https://doi.org/10.3969/j.issn.1000-3428.2017.05.034
In order to enhance the robustness and the instantaneity of the traditional vehicle recognition algorithms,a fine vehicle make and model recognition algorithm is proposed by incorporating Locality-constrained Linear Coding(LLC) and the weighted Spatial Pyramid Matching(SPM) model.Firstly,the Histogram of Oriented Gradient(HOG) feature is extracted to obtain the image representation vector which improves the recognition accuracy.Then the weighted SPM is applied to integrate the spatial information of vehicle images into their final representation vector.Finally,the final representation of vehicle images generated by above steps is sent to a linear Support Vector Machine(SVM) classifier for training and testing.Experiments are conducted on 56 827 vehicle images coming from 150 vehicle makes and models which are captured by real-time traffic surveillance system in various weather conditions and illuminations.The results demonstrate that the proposed algorithm has better performance both in recognition accuracy and the recognition time.
-
Computer Engineering. 2017, 43(5): 217-223. https://doi.org/10.3969/j.issn.1000-3428.2017.05.035In order to increase the retrieval accuracy of the rotated image,this paper proposes a novel rotated image retrieval method by introducing Color Motif Co-occurrence Matrix(CMCM).It designs eight kinds of rotation invariant motifs and gets Motif Excessive Matrix(MEM) and Motif Average Matrix(MAM) to depict the color and texture features of image.Based on MEM and MAM,it uses statistical methods to obtain motif sequence and the occurrence frequency of corresponding color value,and gets an 8×8×72 CMCM as the characteristic matrice of the original image and rotated image.The experiments are carried out on the standard Corel image dataset,and the results show that the proposed method has higher retrieval accuracy compared with CMM,M&D,SITF,SED and CDH method.
-
Computer Engineering. 2017, 43(5): 224-228. https://doi.org/10.3969/j.issn.1000-3428.2017.05.036For the problem of image distortion and blurring caused by the ripples in the water,a new iterative registration algorithm adopting similarity measure and optimization method to recover the real image from some stretched-out ones is proposed.Firstly,distorted sequence images are chosen with the correlation coefficient method and the real images are got from the grid changes.Grids are obtained from the iterative registration of the sequence images and their mean value ones on the basis of B-spline free transformation model.Local-normalized Mutual Information(LRMI) is used as similarity measure,and the Differential Evolution-Quasi-Newton optimization(DE-LBFGS) algorithm is adopted for optimization.Noises are removed by sparse de-noising.Experimental results show that the proposed algorithm shortens the registration time and improves the registration precision and visual effect of the image.
-
Computer Engineering. 2017, 43(5): 229-233. https://doi.org/10.3969/j.issn.1000-3428.2017.05.037For the limitations of the traditional color cast detection methods,this paper proposes a new color cast detection method based on super-pixel saliency.The video image is divided into super-pixel blocks whose luminance and color saliency are extracted.Then,the luminance saliency weight coefficient is used to realize Bayesian fusion of color saliency to get comprehensive and significant feature information.The block feature vectors are used as training samples,and Support Vector Machine(SVM) is used to reallize classification and regression training to detect color cast video and score of the degree of the video color cast.Experimental result shows that,compared with traditional color cast detection methods,the proposed method gets 95.1% accuracy and the dection result is more consistent to human eye subjective evaluations.
-
Computer Engineering. 2017, 43(5): 234-239,247. https://doi.org/10.3969/j.issn.1000-3428.2017.05.038Aiming at the problem of the uncertainty and randomness of the selection of penalty factor and radial basis function in the Support Vector Regression(SVR),based on the monocular and binocular information as well as the SVR optimization model based on Genetic Algorithm(GA),a non-reference stereoscopic image quality assessment method is proposed.The monocular and binocular visual features of the stereoscopic images are extracted.The gradient magnitude and laplacian features are used as monocular visual features.Then,in order to better combine the binocular visual features of human beings,left and right distorted images are fused to construct a cyclopean map from which the statistical characteristics of natural scenes in the spatial domain are extracted.The SVR parameter combination is optimized by using selection,crossover and mutation operations of the GA.The optimum parameter composition is introduced into SVR to predict the objective scores of the stereoscopic images.Considering the responses of human eyes to distortions of the left and right images are different,the gain control model is adopted to fuse the left and right image quality to get final quality evaluation value.Experimental results show that The Pearson Linear Correlation Coefficient(PLCC) indicators reach 0.95 and Spearman Rank-order Correlation Coefficient(SROCC) indicators reach 0.94 on stereoscopic image database established by Ningbo University.The proposed algorithm is highly consistent with the subjective perception.
-
Computer Engineering. 2017, 43(5): 240-247. https://doi.org/10.3969/j.issn.1000-3428.2017.05.039With the increase of the size of the image data and the improvement of the image resolution,the performance of the image edge detection algorithm becomes the key to the real-time processing of the image.Based on the three aspects of quantitative acess memory,data localization and conditional branch optimization,this paper studies the GPU performance optimization of Canny edge detection algorithm on NVIDIA Tegra K1 heterogeneous computing platform combined with algorithm characteristics and underlying hardware architecture characteristics.The experimental results show that compared with the Canny edge detection algorithm based on OpenCV3.0 CPU,the optimized Canny edge detection algorithm achieves 13.2 times to 17.8 times performance acceleration ratio with different graphic data size,and has better detection performance.
-
Computer Engineering. 2017, 43(5): 248-254,260. https://doi.org/10.3969/j.issn.1000-3428.2017.05.040In complex scenes,salient region matching has problems such as difficult targeting positioning and too much redundant information,resulting in high error matching rate and long matching time.Aiming at these problems,this paper proposes a salient region matching algorithm fused with depth information.First of all,adopting visual attention mechanism model fused with depth information,it extracts salient region of the scene image to get rough location of the scene target.Secondly,it pinpoints relevant target in the areas of the scene using matching strategy based on local feature points.At last,it uses FLANN bilateral matching method to match and locate relevant target.Experimental result shows that the proposed method can effectively reduce the number of match points while improving matching accuracy and matching efficiency,even in complex scenes with changed light intensity,changed view angle of the scene and other changes.
-
Computer Engineering. 2017, 43(5): 255-260. https://doi.org/10.3969/j.issn.1000-3428.2017.05.041
Aiming at the problem that the interpolation point gray estimation is not accurate in the process of image reconstruction,a single-image super-resolution algorithm in frequency domain based on neighborhood feature learning is proposed in this paper.When giving a low-resolution image as input,it uses image feature to learn local covariance from low-resolution image and its corresponding high-resolution image’s geometric similar structure.For each patch in the neighborhood,four directional variances are estimated to adapt the interpolated pixels.Experimental results demonstrate that the proposed method not only can guarantee the consistency of the smooth region in the reconstructed high-resolution image,but also can retain the image details and the integrity of the edge profile.
-
Computer Engineering. 2017, 43(5): 261-267,274. https://doi.org/10.3969/j.issn.1000-3428.2017.05.042By taking the image of ethnic minority costumes as the segmentation object,based on block truncation theory,an algorithm for image segmentation based on spatial neighborhood of Fuzzy C-means(FCM) algorithm is proposed.Firstly,the Block Truncation Coding(BTC) theory is used to cut the image RGB color space into six components.Then,the six-dimensional feature vector is used to express the characteristics of the national costume image.Finally,this six-dimensional feature vector is used as the data input of the algorithm.Experimental results show that the performance of the proposed algorithm in this paper is better than FCM,FCM_S1,FCM_S2 in segmentation accuracy,partition coefficient and partition entropy.It has better performance on ethnic costume image segmentation,especially for typical elements of ethnic constume.
-
Computer Engineering. 2017, 43(5): 268-274. https://doi.org/10.3969/j.issn.1000-3428.2017.05.043Speech endpoint detection is an important step in the processing of speech signal.The accuracy and robustness of endpoint detection are low at low Signal to Noise Ratio(SNR).In order to solve this problem,this paper proposes a new endpoint detection algorithm based on wavelet energy entropy and pitch period.It analyzes the speech signal wavelet energy and wavelet entropy to build wavelet energy entropy endpoint detection parameters for different speaker,and utilizes the pronunciation characteristics of different speakers to detect the speech endpoint by wavelet energy entropy and pitch period.Experimental result shows that at different noise background the average detection accuracy is 84.375% in 5 dB SNR which is more accurate than the wavelet energy algorithm and wavelet energy entropy algorithm.
-
Computer Engineering. 2017, 43(5): 275-280. https://doi.org/10.3969/j.issn.1000-3428.2017.05.044In order to realize 3-dimensional model reconstruction of soybean, this paper treats the Yellow No.13 soybean in seedling stage as research object. It takes the advantage of cubic parametric spline curve to fit the outline of the leaf, according to the incremental insertion algorithm of Delaunay to realize the triangular mesh of irregular polygon. In research of texture energy, considering the existing methods for the information representation of texture orientation and scale is insufficient, making use of Gabor filter to extract the texture energy feature and Spatial Pyramid Matching(SPM) algorithm to analyze and render texture, combining L-system and the finite state automaton model to realize the visualization of the soybean. Compared with traditional 3D reconstruction method, experimental results show that this method can effectively describe the relationship between the texture clues and the depth of scene, better to keep the topology structure and contour features of the soybean.
-
Computer Engineering. 2017, 43(5): 281-285. https://doi.org/10.3969/j.issn.1000-3428.2017.05.045For different video content causes poor quality in the front few frames.This paper proposes an initid Quantization Parameter(QP) selection algorithm based on I frame complexity.The algorithm utilizes the three prediction modes of I frame and calculates the complexity of I frame.At the same time,combining with the target bit rate and the relationship between initial QP selects the best initial QP.Experimental results show that compared with other algorithms under the same bit rate,the algorithm avoids disadvantages of the traditional rate control of initial QP selection algorithm,it improves the quality of video and reduces the complexity of judgment I frame.The predicted initial QP is more accurate,and it greatly improves the performance of rate control.The average Peak Signal to Noise Ratio(PSNR) is improved by 0.6 dB and the maximum performance gain in PSNR is up to 1.69 dB.
-
Computer Engineering. 2017, 43(5): 286-289,293. https://doi.org/10.3969/j.issn.1000-3428.2017.05.046Child-directed Speech(CDS) has large influence on early child growth,so it is significant to recognize CDS from a speech and make full use of it.In order to improve the detection accurary,this paper constructs a Chinese CDS detection model based on Adaboost algorithm.It uses a decision tree as a weak classifier for extracting features of Chinese CDS to learn,and forms weak classifier tuple,while the classification results of this group of weak classifiers are weighted voting,to distinguish the voice category.Experimental results show that the vowel duration of CDS is longer than non-CDS;increasing the number of weak classifiers will improve the accuracy of Chinese CDS;the longer the length of test speech is,the higher the detection accuracy of Chinese CDS;compared with v-SVM algorithm,Adaboost algorithm has higher accuracy and precision in Chinese CDS detection,and it improves the robustness of the detection system.
-
Computer Engineering. 2017, 43(5): 290-293. https://doi.org/10.3969/j.issn.1000-3428.2017.05.047Aiming at the problem of the high false alarm rate of moving object detection in the dynamic scenes,this paper proposes an improved moving objects detection method.Based on Dynamic Texture(DT),the video sequence is divided into many Micro Video Elements(MVE) by the sliding window technique.Each MVE is modeled by DT model.In order to reduce the frequencies of Singular Value Decomposition(SVD),the Batch Principal Component Analysis(Batch-PCA) and the Increment Principal Component Analysis(IPCA) are combined to estimate the DT parameters in a MVE group.The observability measure approach of control theory is introduced to detect the salient moving objects.Experimental results show that the method has the better detection rate,robustness and real-time performance in the dynamic scenes with the swaying branches,falling snow or waving water.
-
Computer Engineering. 2017, 43(5): 294-298. https://doi.org/10.3969/j.issn.1000-3428.2017.05.048
Current researches concerning network rumor control strategy mainly focus on the effect of high connectivity on information propagation,namely,the number of neighbor nodes,and pay less attention to the indirect effects induced by neighbor information.Therefore,this paper proposes a Self-degree and Neighbor-degree(SDND) rumor immunization strategy,which merely needs to know the local information of network.In this strategy,immunization node is selected by comprehensively considering on the out-degree of the node and the maximum out-degree of the neighbor nodes.Subsequently,based on the dataset of Sina microblog,the effects of target immunization,acquaintance immunization,important acquaintance immunization,and SDND immunization on rumor propagation are compared by simulating rumor propagation using the SEIR rumor propagation model.Simulation results show that SDND presents a superior immune effect than target immunization and acquaintance immunization strategies.It can effectively inhibit rumor propagation.
-
Computer Engineering. 2017, 43(5): 299-305,312. https://doi.org/10.3969/j.issn.1000-3428.2017.05.049
This paper proposes a computable filling method,aiming at the problem of deformable filling under certain shape constraint conditions.Quadrilateral mesh split of the target area and filling template is done.Integer planning is presented under constraint conditions such as template connecting,boundary,rotation,minimum deformation,etc.Discrete splicing in filling area is realized using filling template.Through the global optimization,model deformation is iterated,so as to achieve the desired filling effect.Experimental results show that the proposed filling method has no direct relationship with the constraints in domain coverage ratio of target area and edge fitting degree,and it can better achieve region filling under the specified constraints.
-
Computer Engineering. 2017, 43(5): 306-312. https://doi.org/10.3969/j.issn.1000-3428.2017.05.050According to the stochastic and fuzzy features of inventory control and decision-making process of the two-level supply chain system,this paper proposes a joint cost-based triangular fuzzy demand stochastic process model for two-level supply chain.Firstly,a two-level supply chain system is studied and the joint cost optimization model is presented.Secondly,according to the fuzzy characteristics of delivery time and demand rate,the backorder quantity fuzzy model is designed based on the triangular fuzzy function considering different cases.By using it,the two-level supply chain cost optimization model is improved to establish the triangular fuzzy stochastic demand process model,and then the model of solving process is given.Finally,the validity of the proposed fuzzy stochastic process model is verified by model parameters sensitivity experiment and comparative experiment.
-
Computer Engineering. 2017, 43(5): 313-316. https://doi.org/10.3969/j.issn.1000-3428.2017.05.051The security and reliability of the transmission of signals collected by the detection terminals and sensors of great importance in the high-rail track signal detection system.However,the relevant signal processing device is limited in computing power and it needs higher real-time requirements.Therefore a certificateless aggregate scheme and a new adversary model are proposed.Under the random oracle model,it is proved that the scheme cannot be provably forged by the adaptive chosen-message and adversary assumption attack.Moreover,compared with other existing aggregate signature schemes like identity-based aggregate signature or certificateless aggregate signature with constant length in the computational efficiency,the proposed scheme is more efficient as it doesn’t contain bilinear pairings.And this scheme is more suitable for the application in secure signal transmission and data authentication service of the high-rail track detection network system which has low bandwidth as well as limited storage and computing environment.
-
Computer Engineering. 2017, 43(5): 317-321. https://doi.org/10.3969/j.issn.1000-3428.2017.05.052Due to the problem that K2 algorithm requires node ordering in learning Bayesian network structure,this paper proposes a hybrid Bayesian network structure learning algorithm.In the situation of a given data set,it uses Max-min Parents and Children(MMPC) algorithm to obtain the initial network structure and utilizes the way of Breadth First Search(BFS) to search the initial network structure.It startly searchs from the node whose in-degree is zero and visits in turn the adjacent points in the figure according to level,thereby it can gain the node order and make it as the initial node order of K2 algorithm.Then,it uses K2 algorithm to search the network space to find out the global optimal solution.The experimental results show that compared with K2 algorithm and Restricted Particle Swarm Optimization(RPSO) algorithm,the new algorithm has lower probability of multi-edge,lack-edge and reverse-edge under the same sample data set.It can learn more accurate Bayesian network with faster convergence speed and higher precision.
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