CIE Homepage  |  Join CIE  |  Login CIE  |  中文 

Most cited

Baidu   Google scholar   CSCD   Crossref   Web of science   Sciencedirect
Published within: In last 1 yearsIn last 2 yearsIn last 3 yearsAll
Condition: CSCD + All
Please wait a minute...
  • Select all
    |
  • LUO Hui-lan, CHEN Hong-kun
    Acta Electronica Sinica. 2020, 48(6): 1230-1239. https://doi.org/10.3969/j.issn.0372-2112.2020.06.026
    CSCD(65)
    Object detection is a hot topic in the field of computer vision, and has been widely used in robot navigation, intelligent video surveillance, aerospace, and other fields. The research background, significance and challenges of object detection were introduced. Then the object detection algorithms based on deep learning were reviewed according to two categories: candidate region-based and regression-based. For the candidate region-based algorithms, we first introduced the R-CNN (Region with Convolutional Neural Network) based series of algorithms, and then the R-CNN based methods were overviewed from four dimensions: the research of feature extraction networks, the region of interesting pooling researches, improved works based on region proposal networks, and some improved approaches of non maximum suppression algorithms. Next, the regression-based algorithms were surveyed in terms of YOLO (You Only Look Once) series and SSD (Single Shot multibox Detector) series. Finally, according to the current trend of object detection algorithms that are developing more efficient and reasonable detection frameworks, the future research focuses of unsupervised and unknown category object detection directions were prospected.
  • 综述评论
    QIAN Zhi-hong;WANG Yi-jun
    Acta Electronica Sinica. 2012, 40(5): 1023-1029. https://doi.org/10.3969/j.issn.0372-2112.2012.05.026
    CSCD(63)
    Based on analyzing IoT two basic concepts,architecture of IoT is proposed,which includes underlayer network distribution,convergence gateway access,inter-connected network integration and terminal user application.In the architecture,a protocol structure of IoT is given,which consists of network protocol layers,network control platform and application terminal platform,and the key technologies for IoT have been discussed concerning hardware and software.Six development conceptions of future IoT have been presented based on summarizing the existing problems of IoT in standards,technologies,security and application.
  • 综述评论
    SU Song-zhi;LI Shao-zi;CHEN Shu-yuan;CAI Guo-rong;;WU Yun-dong
    Acta Electronica Sinica. 2012, 40(4): 814-820. https://doi.org/10.3969/j.issn.0372-2112.2012.04.031
    CSCD(57)
    Pedestrian detection is an active area of research with challenge in computer vision.This study conducts a detailed survey on state-of-the-art pedestrian detection methods from 2005 to 2011,focusing on the two most important problems:feature extraction,the classification and localization.We divided these methods into different categories;pedestrian features are divided into three subcategories:low-level feature,learning-based feature and hybrid feature.On the other hand,classification and localization is also divided into two sub-categories:sliding window and beyond sliding window.According to the taxonomy,the pros and cons of different approaches are discussed.Finally,some experiences of how to construct a robust pedestrian detector are presented and future research trends are proposed.
  • ZHOU Xin-yu, WU Zhi-jian, WANG Hui, LI Kang-shun, ZHANG Hao-yu
    Accepted: 2023-01-16
    Traditional particle swarm optimization(PSO)algorithm tends to suffer from premature convergence;we proposed an elite opposition-based learning strategy in which elite particles are introduced to generate their opposite solutions by opposition-based learning.This mechanism can expand the search area and is helpful to enhance the global explorative ability of PSO.Meanwhile,a differential evolutionary mutation strategy is presented to avoid the best particle being trapped into local optima,since this may cause search stagnation of the whole swarm.This strategy adopts differential evolution algorithm to search for the neighborhoods of the global best particle and is helpful to enhance the exploitation ability of PSO.We compared our algorithm with some state-of-the-art PSOs on 14 benchmarks,the results show that our algorithm obtains better solution accuracy and quicker convergence speed.
  • CHEN Hao-wen, LI Xiang, ZHUANG Zhao-wen
    Accepted: 2023-01-18
    Multiple-input multiple-output (MIMO) radar as a new radar system is proposed at the beginning of 21st century,which has been attracting much attention of researchers and institutions all over the world.In this paper,MIMO radar theory is summarized based on the current literature firstly.Then,the potential capabilities of system configuration,signal design,target detection,parameter estimation and high resolution performance are introduced.At the same time,the restricting factors in practical application are pointed out.Finally,some further work and the technical difficulties in MIMO radar are predicted.
  • LIU Ying, LIU Hong-yan, FAN Jiu-lun, GONG Yan-chao, LI Ying-hua, WANG Fu-ping, LU Jin
    Acta Electronica Sinica. 2020, 48(3): 590-601. https://doi.org/10.3969/j.issn.0372-2112.2020.03.024
    CSCD(46)
    Object detection has been well studied based on the traditional manual features and deep learning algorithm. However, the research on small object detection has just begun in recent years and there are little research outcome available. Furthermore, most of the methods proposed are based on the traditional object detection algorithms with certain modifications so as to improve the accuracy of small object detection. Small object contains fewer pixels and has less features, and it is even harder to extract object features after down sampling. Hence, small object detection is a challenging task. Small object detection has a wide range of application requirements in the fields of automatic driving, remote sensing image detection, and criminal investigation. It has important practical value for the research of small object detection technology. In this paper, the existing research results of small object detection are summarized. Firstly, the existing algorithms are classified into one stage, two stages and multi-stages according to the number of stages for detection. The principles of RetinaNet、CornerNet-Lite、feature pyramid network (FPN) and other algorithms are described and compared. Secondly, this paper describes the application of small object detection technology in different fields, and summarizes the data sets such as MS COCO、PASCAL VOC、DOTA、KITTI and algorithm performance evaluation indicators. Finally, the challenges faced by small object detection are concluded, and the future research directions are prospected.
  • YU Neng-hai, HAO Zhuo, XU Jia-jia, ZHANG Wei-ming, ZHANG Chi
    Acta Electronica Sinica. 2013, 41(2): 371-381. https://doi.org/10.3969/j.issn.0372-2112.2013.02.026
    CSCD(45)
    With the development of cloud computing in the academia and industry,it is inevitable that many security problems arise.This paper summarizes the security requirements of cloud computing,which not only cover the traditional security requirements like confidentiality,data integrity,access control and identity authentication,but also introduce new security requirements in the credibility,configuration and virtual machinery.We make conclusions about the security situations on two typical cloud computing products:Amazon Web Services and Windows Azure and elaborate two attack mechanisms against cloud computing:Denial of service attack and Side channel attack.Based on the security requirements and attacks against cloud computing,we systematically summarize the current security protection mechanisms and further make a comparison among them.
  • ZHOU Yong-quan, HUANG Zheng-xin, LIU Hong-xia
    Accepted: 2023-01-18
    A discrete glowworm swarm optimization (DGSO) algorithm is designed to tackle the travelling salesman problem.A new encoding schema and decoding schema are given with the characteristics of the TSP problem,and a new distance formula and encoding update formula for the new algorithm are given.In order to enhance the capability of the algorithm local searching,and to speed up the algorithm convergence speed,the 2-opt local search scheme is integrated into the new algorithm for solving TSP problem.The proposed algorithm was evaluated on 10 TSP test problems.The numerical experiments show that the proposed algorithm can find the global optimal solution with less computation and evolving time.In case of large scale TSP algorithm can achieve optimal solution of the theory and the error of the optimal solution is also less than 1%.
  • 综述评论
    LI Feng-hua;SU Mang;SHI Guo-zhen;MA Jian-feng
    Acta Electronica Sinica. 2012, 40(4): 805-813. https://doi.org/10.3969/j.issn.0372-2112.2012.04.030
    CSCD(41)
    The main task of access control is to prevent unauthorized accesses to information resources.Conflict detection and resolution mainly solves problems caused by various security policies among different information systems.With the development of computer and communication technology,several access control models have appeared such as discretionary access control,mandatory access control,role based access control,task-based access control,access control for distributed environment and cross-domain,spatiotemporal attribute based access control and security attribute based access control,etc.The paper analyzes and summarizes the existing domestic and international research situation in the field of access control and conflict detection and resolution from the theoretical research and application aspects,indicates exiting problems in ubiquitous networks for the cyber-physical society,and points out some development trends of fine-grained and multi-level security access control model and scalable method for its policy.
  • LI Xiang, FAN Mei-mei
    Acta Electronica Sinica. 2012, 40(9): 1863-1870. https://doi.org/10.3969/j.issn.0372-2112.2012.09.025
    CSCD(41)
    Cognitive radar can adaptively and intelligently reconfigure its transmission and reception system based on the knowledge about the current environment to improve its performance.This paper reviews the development of cognitive radar first,and then reveals its essential character and framework;after that,the key technologies are presented in detail and research advances are discussed;at last,the prospects of cognitive radar are pointed out.
  • HE Zhi-yong, SUN Li-ning, CHEN Li-guo
    Acta Electronica Sinica. 2013, 41(2): 267-272. https://doi.org/10.3969/j.issn.0372-2112.2013.02.010
    CSCD(39)
    The traditional Otsu algorithm has to exhaustively compute all between-class variances.Based on one characteristic of Otsu threshold,this paperwork proposes a new fast algorithm.The new algorithm finds out every threshold which is equal to the integer part of the average of the mean levels of two classes,and then selects one threshold which is in accord with Otsu criterion.The traditional Otsu algorithm cannot work well when it extracts small object from gradient image,so an improved thresholding algorithm is proposed.Based on the fast Otsu algorithm provided,the improved thresholding algorithm recursively computes threshold.Experimental results show that the fast Otsu algorithm is faster than the traditional Otsu algorithm. Experimental results also show that the improved thresholding algorithm is effective to segment small object of gradient image.
  • QIAO Shao-jie, HAN Nan, ZHU Xin-wen, SHU Hong-ping, ZHENG Jiao-ling, YUAN Chang-an
    Acta Electronica Sinica. 2018, 46(2): 418-423. https://doi.org/10.3969/j.issn.0372-2112.2018.02.022
    CSCD(37)
    Traditional fitting-based trajectory prediction algorithms cannot meet the requirements of high accuracy and real-time prediction. A dynamic Kalman filter based TP approach was proposed, which performs state estimation of dynamic behavior with regard to moving objects, updates the state variable estimation value based on the estimation of the previous and current observation states, in order to infer the next location of moving objects. Extensive experiments are conducted on real datasets of moving objects and the results demonstrate that the average prediction error (root mean square error between the predicted location and the actual location) of the TP algorithm based on Kalman filter is around 12.5 meters on the GeoLife datasets. The prediction error is reduced by about 555.4 meters by compared to the fitting-based TP algorithms, and the prediction accuracy is increased by 7.1% on the T-Drive datasets as well. The dynamic TP approach based on Kalman filter can handle the problem of low prediction accuracy with the guarantee of efficient time performance.
  • XIAO Zi-ya, LIU Sheng
    Acta Electronica Sinica. 2019, 47(10): 2177-2186. https://doi.org/10.3969/j.issn.0372-2112.2019.10.020
    CSCD(37)
    In order to improve the slow convergence rate and low stability of WOA, elite opposition-based golden-sine whale optimization algorithm is proposed. Elite opposition-based learning strategy is used to improve the diversity and quality of the population so that the convergence rate can be promoted. At the same time,golden ratio is introduced to improve the optimal method of WOA, so as to ocoordinate the global exploration and local exploitation. Twenty unimodal and multimodal benchmark functions are tested and compared with that of other algorithms,and the experimental results show that EGolden-SWOA has a better performance in convergence rate and stability. The high dimensional function test shows that EGolden-SWOA perform well in solving large scale optimization problem. Finally, EGolden-SWOA is applied to the optimization design of the pressure vessel and tension/compression spring, the result shows that its performance in project optimization is better than RCSA and CPSO, it can be effectively applied to project optimization.
  • XIE Hui, HUANG Zhi-tao, WANG Feng-hua
    Acta Electronica Sinica. 2013, 41(6): 1166-1176. https://doi.org/10.3969/j.issn.0372-2112.2013.06.019
    CSCD(36)
    Blind recognition of channel coding plays an important role in the field of non-cooperative signal processing,which has been extended from the signal level to the information level.Blind recognition of channel coding is widely used in the fields of intelligence communication,information interception and information countermeasure.Firstly,the recognition algorithms of convolution code,BCH code,RS code,Turbo code and scramble code which are commonly used in modern digital communication systems were summarized and classified.Then the theories of the algorithms were described,and the computational complexity and performances in noisy environment of the algorithms were analyzed.Finally,the future of blind recognition of channel coding was pointed out based on the shortcoming of current algorithms and practical need.
  • XIA Nan, QIU Tian-shuang, LI Jing-chun, LI Shu-fang
    Acta Electronica Sinica. 2013, 41(1): 148-152. https://doi.org/10.3969/j.issn.0372-2112.2013.01.026
    CSCD(35)
    A nonlinear filtering algorithm is proposed based on the Kalman filter and the particle filter.The method can provide significant performance for dynamic nonlinear system which is consist of linear state equation and nonlinear measurement equation.Firstly,the particle filter is utilized for initial estimation of the state variables,and then the Kalman filter is performed.The Cramer-Rao Bound is derived for the nonlinear model.Computation complexity analysis and numerical simulations demonstrate that the proposed algorithm has the same complexity as the standard particle filter,but the estimation accuracy is higher than the standard particle filter and the extended Kalman filter.The estimation error is even lower than the Cramer-Rao Bound of the system model.
  • CHU Ding-li, CHEN Hong, WANG Xu-guang
    Acta Electronica Sinica. 2019, 47(5): 992-999. https://doi.org/10.3969/j.issn.0372-2112.2019.05.003
    CSCD(34)
    Aiming at the problem that whale optimization algorithm is easy to fall into local extreme value and slow convergence speed,this paper proposes a whale optimization algorithm based on adaptive weight and simulated annealing.The improved convergence weight strategy is used to adjust the convergence speed of the algorithm,and the global optimization ability of the whale optimization algorithm is enhanced by simulated annealing.In the simulation experiment,18 test functions were calculated and the genetic algorithm,the particle swarm optimization algorithm and the standard whale algorithm were compared and statistically analyzed.At the same time,the influence of the adaptive weight and simulated annealing on the whale optimization is compared.The results show that the improved algorithm has a significant improvement in the calculation of the extremum of the test function,and the effectiveness of the improved algorithm is verified.
  • YU Pei-dong, LI Jing, PENG Hua
    Acta Electronica Sinica. 2013, 41(2): 301-306. https://doi.org/10.3969/j.issn.0372-2112.2013.02.015
    CSCD(34)
    The existing methods for channel coding recognition usually use hard-decision of the demodulator output sequence,and their robustness against error bits is to be improved.Focusing on situations of low signal-to-noise ratio,this paper presents a novel recognition algorithm which uses soft-decision.The algorithm is based on the error-containing equation model,and it solves the equation,thus accomplishes the recognition,through regarding the probability that the equation holds right as the measurement of the performance of a solution vector.The use of log-likelihood ratio(LLR) algebra makes the algorithm greatly simplified.Results of simulation experiments show that the new algorithm improves recognition performance,especially in lower signal-to-noise ratio cases,compared to the existing algorithm based on Walsh-Hadamard transform.
  • WU Cheng-mao
    Accepted: 2023-01-17
    From the viewpoint of optimal image contrast measure,a kind of optimization mathematical models and its enhanced methods are proposed for image enhancement by histogram equalization.For the shortcoming of low operation efficiency of optimal contrast image enhancement method by linear programming,the gray level mapping optimization mathematical model for classical histogram equalization is established to improve its mathematical theory.The improved gray level mapping optimization models for histogram equalization are constructed by means of weighted geometry averaging in order to avoid the shortcoming of classical histogram equalization.The good properties of its optimal solution for gray level mapping optimization models are discussed,and the classical and adjustable histogram equalization methods are regarded as special cases of the proposed optimization mathematical models for image enhancement.Experimental results show that the optimization model of histogram equalization is reasonable and can obtain satisfactory image enhancement effect.To some extent,it is more universal than classical and adjustable histogram equalization methods.
  • LI Shun-dong, WANG Dao-shun
    Acta Electronica Sinica. 2013, 41(4): 798-803. https://doi.org/10.3969/j.issn.0372-2112.2013.04.029
    CSCD(32)
    Secure multiparty computation is a key privacy-preserving technology in cyberspaces and a research focus in the international cryptographic community.We first present a new encoding scheme to encode private data.By using this encoding scheme together with homomorphic encryption scheme,we construct a new scheme for Yao's millionaires' problem and prove its privacy-preserving property.This new scheme is more concise,more general and can be applied to compare any two objects on which a total order can be defined.We finally utilize the new scheme to propose a solution to the coprime problem and prove the privacy-preserving properties of the solution.
  • 科研通信
    ZHANG Xue-jun;CAI Wen-qi;WANG Suo-ping
    Acta Electronica Sinica. 2012, 40(1): 193-198. https://doi.org/10.3969/j.issn.0372-2112.2012.01.032
    CSCD(31)
    Based on the adaptive multi-tree search anti-collision algorithm,we proposed an improved adaptive multi-tree search anti-collision algorithm (IAMS) by optimizing the prefix sent by readers.Our algorithm reduces idle time slots by choosing the number of search tree branches adaptively according to the collision factor and optimizing the query prefix of quadtree.By mathematical analysis,we accurately predict the total number of time slots required for the tag identification in IAMS algorithm.Simulation results show that the IAMS algorithm has faster identification speed and higher system throughput.
  • WANG Qiang, LI Jia, SHEN Yi
    Acta Electronica Sinica. 2013, 41(10): 2041-2050. https://doi.org/10.3969/j.issn.0372-2112.2013.10.027
    CSCD(30)
    Measurement matrix,whose performance can affect the compression and reconstruction of original signal,plays a key role in compressive sensing.Most of the existing measurement matrices are random ones,which have shortcomings in practical application,such as large storage capacity,low efficiency and difficulty when implemented in the hardware.Therefore,it is of important practical significance to construct deterministic measurement matrix for the promotion and application of the compressive sensing theory.In this paper,the existing construction algorithms for deterministic measurement matrix are reviewed,introduced and classified in detail.Finally the performances of all algorithms are summarized in terms of common indicators.
  • WANG Yi, LIU San-yang, ZHANG Wen, WANG Ya-nan
    Acta Electronica Sinica. 2014, 42(12): 2509-2514. https://doi.org/10.3969/j.issn.0372-2112.2014.12.025
    CSCD(30)

    Under the framework of IFS(Intuitionistic Fuzzy Sets),aiming at the situation that targets' attributes are interval value,the weights are totally unknown,and decision-makers have preference information,a threat assessment method with uncertain attribute weight based on intuitionistic fuzzy multi-attribute decision is proposed.Firstly,based on the above-mentioned problem,intuitionistic fuzzy interval judgment matrix is set up,and a standard interval value index processing method is put forward;secondly,by analyzing the target attribute,attribute weight,and decision-makers' authority degree in project,the preference project model in group decision making is finally constructed;thirdly,according to the attribute's variety character in interval value,the definition of intuitionistic fuzzy interval value's similarity and ideal resolve is given,and the optimization attribute weight restriction model is built up;lastly,some typical threat assessment examples are cited to verify that the method can reflect the influence of both subjective and objective information,which can avoid the deviation caused by sensor invalidation,outside environment's effect,or decision-makers' subjective experience.Finally,the superiority of the method is proved by the threat assessment of some typical targets.

  • LONG Wen, CAI Shao-hong, JIAO Jian-jun, WU Tie-bin
    Acta Electronica Sinica. 2019, 47(1): 169-175. https://doi.org/10.3969/j.issn.0372-2112.2019.01.022
    CSCD(29)
    Grey wolf optimization (GWO) algorithm is a relatively novel optimization technique which has been shown to be competitive to other population-based algorithms.However,there is still an insufficiency in canonical GWO regarding its position update equation,which is good at exploitation but poor at exploration.Inspired by differential evolution and particle swarm optimization,the personal best information and the random selected individual from population are used to construct a modified position update equation for enhancing the exploration.Inspired by particle swarm optimization,a random adjustment strategy of control parameterais proposed.In addition,to enhance the global convergence,when producing the initial population,the chaos method is employed.Simulation experiments were conducted on the 18 high-dimensional conventional test functions.The simulation results show that the proposed algorithm provides better performance than basic GWO algorithms in the same or less number of maximum fitness function evaluation in most cases.
  • KANG Bing-yi, LI Ya, DENG Yong, ZHANG Ya-juan, DENG Xin-yang
    Accepted: 2023-01-18
    One of the open issues of Dempster Shafer theory is how to determine basic probability assignment function (BPA).To solve this problem,a method to determine BPA based on interval numbers is proposed in the paper.At first,the model of interval numbers is constructed with the samples.Then the distance of interval numbers is used to represent difference among the attributes of the samples,so the similarity of them is calculated.At last,the similarity is normalized to get the value of BPA.The effectiveness of this method is proved by classifying the Iris Set.It concludes that the total recognition rate is 96%.This method is simple and practical;it can determine BPA in the case of the little number of the samples.
  • 学术论文
    LIAN Qiu-sheng;ZHANG Wei
    Acta Electronica Sinica. 2012, 40(5): 920-925. https://doi.org/10.3969/j.issn.0372-2112.2012.05.010
    CSCD(29)
    At present,super-resolution algorithms based on sparse representation of image patches exploit single dictionary to represent the image patches,which can not reflect the differences of various image patches types.In this paper,a novel method based on sparse representation of classified image patches is proposed to overcome this disadvantage.In this method,image patches are firstly divided into smooth patches,different directional edge patches and irregular structure patches by local features.Then these classified patches are applied into training the corresponding high and low resolution dictionary pairs.During the reconstruction process,simple bicubic interpolation approach is used for smooth patches while edge and irregular structure patches are reconstructed from their corresponding dictionary pairs using orthogonal matching pursuit algorithm.Experiment results show that the proposed algorithm significantly improves the visual quality of the edges and has faster speed compared with other single dictionary methods.
  • ZHANG Yang-rui, LI Yun-jie, LI Man-ling, GAO Mei-guo, FU Xiong-jun
    Acta Electronica Sinica. 2016, 44(1): 46-53. https://doi.org/10.3969/j.issn.0372-2112.2016.01.008
    CSCD(28)

    A multiple false targets method based on interrupted-sampling and nonuniform periodic repeater jamming (ISNPRJ) and against to the linear frequency modulation (LFM) pulsed radar with mean level (ML) CFAR detector is proposed.Firstly, the principle is illuminated which takes advantage of the interrupted sampling and repeater jamming (ISRJ) to realize the distribution of multiple false targets.After that several key parameters both the number and the SNR of the false targets are derived.Then jamming effectiveness of the ISNPRJ is analyzed on the basis of the mathematics principle of the interrupted-sampling and periodic repeater jamming (ISPRJ), meanwhile, numerical value of the sampling interval, the sampling pulse width, transmitted pulse width and the transmitted power are calculated.Simulation results show that ISNPRJ can reduce the detection probability of target greatly and jam the radar's detector effectively.

  • LIU Jiang, ZHANG Hong-qi, LIU Yi
    Acta Electronica Sinica. 2018, 46(1): 82-89. https://doi.org/10.3969/j.issn.0372-2112.2018.01.012
    CSCD(28)
    This paper is focused on the optimal policy selection for moving target defense. Attack-defense confrontation in moving target defense environment is analyzed. Reward quantization method of moving target defense policy is proposed. Single-stage and multi-stage moving target defense game models are constructed based on the dynamic game with incomplete information. The algorithm to obtain perfect Bayesian equilibrium and the method to revise prior belief are proposed. Optimal moving target defense policies are obtained under different security situations. Finally, not only the feasibility and effectiveness of the proposed model and method are illustrated and verified in a representative example but also general rules of network defense using moving target defense policies are summarized.
  • LI Wei-gang, YE Xin, ZHAO Yun-tao, WANG Wen-bo
    Acta Electronica Sinica. 2020, 48(7): 1284-1292. https://doi.org/10.3969/j.issn.0372-2112.2020.07.006
    CSCD(27)
    To solve the problem of slow speed and low accuracy in the surface defect detection of hot rolled strips, an improved YOLOv3 algorithm is proposed. Firstly, the weighting K-means clustering algorithm is put forward to optimize priors anchor's parameters, which can improve the match between priors anchor and feature map. Secondly, the improved network structure of the YOLOv3 algorithm is proposed to improve the detection accuracy, whose shallow features and deep features are combined to form the new large-scale inspection layer. The experiments are carried out on the NEU-DET dataset, the results show that the average accuracy of the improved YOLOv3 algorithm is 80%, which is 11% higher than that of the original algorithm; the detection speed is 50fps, which is faster than other strip surface defect detection algorithms based on deep learning.
  • DONG Li-li, DING Chang, XU Wen-hai
    Acta Electronica Sinica. 2018, 46(10): 2367-2375. https://doi.org/10.3969/j.issn.0372-2112.2018.10.009
    CSCD(26)
    HE (Histogram Equalization) is a fundamental method in the field of image enhancement, the research and improvement about which is very significant. First, this paper analyzes the disadvantages of the classical HE algorithm and summarizes five kinds of image enhancement techniques based on HE. Then, two kinds of improved methods are proposed aiming at the disadvantages of the classical HE algorithm, the techniques of the peak clipping and the edge information fusion are introduced. Finally, underexposure and overexposure images are selected to verify the algorithms' properties, the standard of efficient image objective quality assessment is selected to evaluate the experimental results. The assessment of image subjective and objective quality shows the algorithms this paper proposes have the characteristics of better results, less input parameters and so on.
  • YANG Zhao-cheng, LI Xiang, WANG Hong-qiang
    Acta Electronica Sinica. 2014, 42(6): 1194-1204. https://doi.org/10.3969/j.issn.0372-2112.2014.06.024
    CSCD(26)

    With the development of compressive sensing theory,the space-time adaptive processing (STAP) technology based on sparsity of space-time power spectrum (STPS) receives a growing interest.This paper firstly reviews the traditional STAP algorithms and shows analysis of sparsity of STPS from three different points of view and potential advantages of STAP technology based on sparsity of STPS.Then,the current developed STAP algorithms based on sparsity of STPS are categorized into three classes,such as STAP based on prior knowledge of array manifold,STAP based on sparse recovery of STPS and STAP based on both prior knowledge of array manifold and sparse recovery of STPS.It also performs an overview of those algorithms.Finally,based on the progress of the existing research,some key issues to enhance the performance of clutter suppression and moving target detection are introduced,which include intrinsic mechanism analysis of sparsity,space-time steering dictionary design,easy parameters setting,fast and low complexity algorithms design,robust algorithm design,STAP algorithms based on sparse recovery of STPS design by exploiting different types of prior knowledge,and constant false alarm rate detector design of STAP based on sparsity of STPS and validation using measurement data.

  • LUO Hui-lan, TONG Kang, KONG Fan-sheng
    Acta Electronica Sinica. 2019, 47(5): 1162-1173. https://doi.org/10.3969/j.issn.0372-2112.2019.05.025
    CSCD(25)
    Human action recognition in videos is a challenging topic in the field of computer vision.It is widely not only used in video information retrieval,daily life security,public video surveillance,but also human-computer interaction,scientific cognition and other fields.First,the research background,research significance and difficulties of action recognition are briefly introduced,and then the deep learning model based action recognition methods are comprehensively reviewed from three different aspects:the types and numbers of input signals,the combination with traditional feature extraction methods,and the pre-trained datasets.Furthermore,the performances of some typical methods on UCF101 and HMDB51 datasets are overviewed and analyzed.Last the possible future research directions are discussed from three perspectives:the video data preprocessing,the video human motion feature representation,and the model training.
  • YANG Wei, FU Yao-wen, LONG Jian-qian, LI Xiang
    Accepted: 2023-01-18
    Finite Sets Statistics (FISST) provides an "engineering friendly" theoretic tool for target tracking in clutter.An overview of the studies on the FISST-based target tracking techniques is presented here.Special attention is paid to the following areas:optimal multi-target Bayes filter and its principled approximations,multi-target filter under unknown parameters,multiple maneuvering targets tracking,track-valued estimation,Joint Target detection and Tracking Filter (JoTTF),Bayesian filtering with random finite set observations,and also the relevant applications.Finally,based on the progress of existing research in these areas,some key issues to enhance the precision and robustness of target tracking further are introduced which deserve more attention of the researchers' for solution.These include:performance evaluation of multi-target filtering,dim/small target tracking,multiple maneuvering targets tracking,multi-sensor multi-target tracking,Joint target Detection,Tracking and Classification (JDTC),and so on.
  • WEN Xiang-xi, MENG Xiang-ru, MA Zhi-qiang, ZHANG Yong-chun
    Accepted: 2023-01-18
    The chaotic performance of small-time scale network traffic was covered by noise,which made the traffic unpredictable.This paper introduces the local projection to denosie network traffic;a chaotic and predictable traffic trend is obtained.As the network traffic series is long-period and time-varying,a new method named optimal training subset online fuzzy least squares support vector machines (OTSOF-LSSVM) is proposed.Samples temporal and distance nearest to prediction sample are chosen as optimal training subset,and the subset are fuzzified.On this basis,the prediction model is established by fuzzy LSSVM.The model update computational complexity is reduced by partitioned matrix calculation.The noise reduction and trend prediction on network traffic shows the proposed method can predict the trend quickly and exactly.
  • 学术论文
    QI Xiao-zhen;WANG Qing
    Acta Electronica Sinica. 2012, 40(4): 773-779. https://doi.org/10.3969/j.issn.0372-2112.2012.04.025
    CSCD(24)
    A novel image classification method based on sparse coding and multiple kernel learning is proposed in the paper.Traditional methods of image classification used common sparse coding but lose the spatial information.We add this spatial information by dividing the image with the spatial pyramid.With the nonlinear SVM for image classification,each level of spatial pyramid has its own kernel,and we adopt machine learning for the optimal trade-off between different kernels.A much more discriminative kernel can be seen as the linear combination of base kernels corresponding to different pyramid levels.The experiments on the benchmark dataset show the effectiveness and robustness of our method.The precision on scene categories dataset can reach 83.10%,and it is the best result comparing to the state-of-the-art work.
  • WANG Wen-bo, ZHANG Xiao-dong, WANG Xiang-li
    Acta Electronica Sinica. 2013, 41(7): 1425-1430. https://doi.org/10.3969/j.issn.0372-2112.2013.07.028
    CSCD(24)
    In order to solve the problem of nonlinear and nonstationary signal de-noising,a novel de-noising method is proposed by combining the principal component analysis(PCA) and empirical mode decomposition(EMD).The method removes noise of intrinsic mode functions(IMFs) using PCA,after the noisy signal is decomposed by EMD.Firstly,the signal details of the first IMF are extracted by using 3σ criterion,and the noise energy of each level IMF is estimated.Secondly,the PCA is implemented on each IMF,and the part of principle components are selected to reconstruct the IMF according to noise energy of IMFs,then the noise of IMF is removed efficiently.Numerical simulation and real data test were carried out to evaluate the performance of the proposed method.The experimental results showed that the proposed method outperformed the Bayesian wavelet threshold de-noising algorithm and mode cell EMD de-noising algorithm.So it is an effective signal de-noising method.
  • GAO Ni, GAO Ling, HE Yi-yue, WANG Hai
    Acta Electronica Sinica. 2017, 45(3): 730-739. https://doi.org/10.3969/j.issn.0372-2112.2017.03.033
    CSCD(24)

    Owing to the constraints of time and space complexity,support vector machine (SVM) faced with the problem of ‘curse of dimensionality’ when computation happens in high-dimensional feature space.Therefore,an intrusion detection model of support vector machine based on autoencoder network (AN-SVM) is proposed.First,the multilayer unsupervised restricted boltzmann machine (RBM) in our model is employed in mapping the vector of raw dada from high-dimensional nonlinear space to low-dimensional space,and a mutual mapping autoencoder network of high-dimensional space and low-dimensional space is constructed.Then autoencoder network weights of fine-tuning algorithm based on back propagation network is employed to reconstruct the new optimal high-dimensional representation of data in low-dimensional space,and the corresponding optimal low-dimensional representation of raw data can be obtained.Furthermore,SVM classification algorithm is employed to detect intrusion from the optimal low-dimensional data.The experimental results demonstrate that AN-SVM model can effectively reduce the training time and testing time of classifier in the intrusion detection model and its classification performance outperforms those traditional methods.So,AN-SVM model is a feasible and efficient lightweight intrusion detection model.

  • XIAO Jin-sheng, GAO Wei, ZOU Bai-yu, YAO Yuan, ZHANG Yong-qin
    Acta Electronica Sinica. 2017, 45(2): 346-352. https://doi.org/10.3969/j.issn.0372-2112.2017.02.012
    CSCD(24)

    An improved image dehazing algorithm based on dark channel prior is proposed to overcome the color distortion of sky and halo effects.The guided filter is utilized to segment the sky area finely to avoid the sky color distortion,since the defect of the classical dark channel prior theory.The global atmospheric light of image in sky region is estimated accurately.In addition,the detailed edge information can be got by taking advantage of median filter technique.So more effective transmission map estimation will be achieved which effectively inhibited the halo.Last,because the brightness of image after haze removal is lower than the actual scene,histogram equalization is used for channel V of the HSV color space.The experiment results show that.the proposed method can not only restore the clean scene from hazy images effectively,but also avoid color distortion of the sky region and halo artifacts.

  • CHU Hong-li, LI Yuan-xiang, ZHOU Ze-ming, SHEN Ji
    Acta Electronica Sinica. 2013, 41(4): 791-797. https://doi.org/10.3969/j.issn.0372-2112.2013.04.028
    CSCD(23)
    The dark channel prior algorithm is effective to some extent when working on some single outdoor scene hazing images.However,it has to expend large amounts of memory storage and computational time.Besides,the result of the sky region is not accurate enough.In this paper,we estimate the transmission map through processing the edges and non-edges respectively with different neighborhoods and obtain more accurate atmospheric light value by departing the sky part or the most hazed region of the image.This method can get a similar or even a better dehazing result than the classical haze removal methods,and moreover take less memory consumption and get much faster computational speed.
  • LIU Hai-bo, WU De-wei, DAI Chuan-jin, MAO Hu
    Acta Electronica Sinica. 2013, 41(1): 8-12. https://doi.org/10.3969/j.issn.0372-2112.2013.01.002
    CSCD(23)
    It is found in numerical simulations that Duffing oscillator state transition from big-cycle motion to chaotic motion has more robust sensitivity to the amplitude of the forced periodic term than the transition from chaotic motion to big-cycle motion,as long as the reference signal initial phase and the Duffing oscillator initial value are matched.On this basis,a new weak sinusoidal signal detection scheme is advanced,whose feasibility is then analyzed.The advanced scheme shows higher detection capability by comparing with traditional methods.Besides,considering the characteristics of the Duffing oscillator phase space,a new Duffing oscillator motion state judging method is proposed,which features good real-time performance and low computational complexity,and is verified with simulations.
  • 综述评论
    WANG Qin-hui;YE Bao-liu;TIAN Yu;LI Wen-zhong;LU Sang-lu;CHEN Dao-xu
    Acta Electronica Sinica. 2012, 40(1): 147-154. https://doi.org/10.3969/j.issn.0372-2112.2012.01.024
    CSCD(23)
    With the explosion of novel wireless applications,the increasing growth of spectrum requirement has outpaced available spectrum resources.Cognitive Radio Network (CRN) has emerged as a promising solution to address the above dilemma by dynamically sharing spectrum among users.Recently,CRN technology has attracted great research interests as well as efforts.In this paper,we review the state-of-the-art of spectrum allocation techniques in CRNs.We first illustrate the technical background of CRNs,and then analyze the key issues in spectrum allocation algorithms design.Following that we review the design rationale and technical feature of typical allocation models,and further investigate the implementation mechanism of classical algorithms for each model.Finally,we envision the possible issues for future work on spectrum allocation.