Home      About Journal      Editorial Board      Instructions      Ethics Statement      Contact Us                   中文

Judgement and Disposal of Academic Misconduct Article Copyright Transfer Agreement Proof of Confidentiality Requirements for Electronic Version More.... Chinese Association of Automation National ResearchCenter for Intelligent Computing System Institute of Intelligent Machines,Chinese Academy of Sciences More....

2015 Vol.28 Issue.7, Published 2015-07-30 Papers and Reports    Researches and Applications    Surveys and Reviews        Papers and Reports 577 Direction-Adaptive Lifting Wavelet Transform for Image Coding WANG Xiang-Hai, XIA Chun-Yu, SONG Chuan-Ming A direction-adaptive lifting wavelet transform (DA-LWT) based on block of image is proposed in this paper. The fixed size of directional block is used for each level of transform. Directional information is retained by the first and second level transform. The direction of higher level transform is obtained by the prediction of the first two levels, and the cost of side information is reduced. According to the minimum prediction residual energy, the filtering direction of filter is selected adaptively to eliminate the redundancy between neighboring pixels effectively and reduce the energy of high-frequency coefficients. Adopting the interpolation based on fractional pixel, and the direction resolution is improved. Experimental results show that the transform coefficients of image obtained by DA-LWT have a better "zero-tree" feature. DA-LWT can obtain better coding efficiency and visual effects compared with traditional lifting wavelet transform. 2015 Vol. 28 (7): 577-585 [Abstract] ( 671 ) [HTML 1KB] [ PDF 1062KB] ( 605 ) 586 Topic Popularity Prediction of Microblog Based on Wavelet Transformation and ARIMA CHEN Yu-Zhong, FANG Ming-Yue, GUO Wen-Zhong, GUO Kun The research of topic popularity prediction problem can be an important significance for maximizing the propagation effects of network advertisements as well as guiding and controlling the network consensus. Firstly, according to user relationship and topic factor, user influence is computed and topic influence is defined. Then, based on aging theory, topic energy value is calculated by considering both topic influence and the number of microblogs related to the topic, and the heat of topic is quantified. Finally, an method named topic popularity prediction of microblog based on wavelet transformation and autoregressive integrated moving average model is proposed to predict topic popularity of microblog and forecast when a topic will hit the peak. Experimental results show that the proposed method can effectively predict the topic popularity and the peak of a topic with lower residual error and omission rate. 2015 Vol. 28 (7): 586-594 [Abstract] ( 507 ) [HTML 1KB] [ PDF 665KB] ( 601 ) 595 Face Recognition Method of Mixed Structured SparsityBased on Coding Complexity CAI Ti-Jian, FAN Xiao-Ping, XIE Xin, XU Jun Coding complexity is utilized to represent the structural sparsity, and structural sparsity is achieved by means of reducing coding complexity. Based on the model of sparse representation classification, a structural dictionary is formed from clustering and sorting, sparsity model with mixed structure is constructed. This model combines fixed-length group structure between classes, and dynamic group structure within classes, as well as standard spare structure corresponding to error part. To reconstitute this mixed structural sparsity, an improved mixed structural greedy algorithm is proposed. Experimental results show that the clustering and sorting of the data dictionary can effectively improve the performance of face recognition. Under the same conditions, the performance of mixed structure is better than other structures, and the proposed algorithm outperforms other algorithms. 2015 Vol. 28 (7): 595-602 [Abstract] ( 541 ) [HTML 1KB] [ PDF 551KB] ( 519 ) 603 Particle Swarm Optimization with Exhaustive Disturbance Based on Exploration-Exploitation Balance Theory LI Kun, LI Ming, CHEN Hao Based on the viewpoint that the algorithm gain a good performance only because it fits the characters of the optimization problem, exhaustive disturbance mechanism is introduced in the particle swarm algorithm under the theoretical framework of the exploration-exploitation balance. Based on the thorough researches of the intensity and range for exhaustive disturbance, four kinds of method for employing exhaustive disturbance are proposed in this paper. Some groups of orthogonal experiments are designed to find the best way of employing exhaustive disturbance. By analyzing the experimental results, the following conclusions are drawn. Exhaustive disturbance has its limits while dealing with high dimensional optimization problems, the intensity of exhaustive disturbance needs to be restricted within 15%, and the triggering condition of exhaustive disturbance based on population diversity shows better performance than the other triggering conditions. Finally, on the basis of the above conclusions, adaptive particle swarm optimization with exhaustive disturbance is proposed. Comparing with other algorithms, the proposed algorithm has a better performance. 2015 Vol. 28 (7): 603-612 [Abstract] ( 432 ) [HTML 1KB] [ PDF 730KB] ( 545 ) Surveys and Reviews 613 A Survey of Multi-pose Face Recognition ZOU Guo-Feng, FU Gui-Xia, LI Hai-Tao, GAO Ming-Liang, WANG Ke-Jun Multi-pose face recognition has became one of important research directions of face recognition. In this paper, the research progress of face recognition is briefly reviewed. Multi-pose face recognition techniques and methods at home and abroad in recent years are introduced briefly and categorized systematically. The advantages and disadvantages of each method are analyzed, and a brief evaluation is made. The challenges of multi-pose face recognition technology are elucidated, and the development direction of the future research on multi-pose face recognition is discussed. 2015 Vol. 28 (7): 613-625 [Abstract] ( 878 ) [HTML 1KB] [ PDF 734KB] ( 1935 ) Researches and Applications 626 Heuristic Simulated Annealing Algorithm for Orthogonal Rectangle Packing Problem with Static Non-Equilibrium Constraints LIU Jing-Fa, ZHANG Zhen, XUE Yu, LIU Wen-Jie, JIANG Yu-Cong With the background of the satellite module layout, the orthogonal rectangle packing problem with static non-equilibrium constraints is studied. By drawing lessons from quasiphysical strategy and defining embedded computing formula between each two rectangles, and between the rectangle and the circular container, this problem is converted into an unconstrained optimization problem. By incorporating heuristic configuration update strategies, local search strategy based on the gradient method and the simulated annealing algorithm with global optimization, a heuristic simulated annealing algorithm for orthogonal rectangle packing problem with static non-equilibrium constraints is put forward.The heuristic configuration update strategies in this algorithm produce new configurations and jump out of trap. The gradient method is searched for lower-energy minima near newly generated configurations. In addition, in the process of layout optimization, a static non-equilibrium penalty term on the basis of the extrusive elastic energy is introduced. Subsequently, by adopting the translation of the center of mass, the static non-equilibrium constraints of the whole system can be satisfied. The experimental results show that the proposed algorithm is an effective algorithm for solving the orthogonal rectangle packing problem with static non-equilibrium constraints. 2015 Vol. 28 (7): 626-632 [Abstract] ( 556 ) [HTML 1KB] [ PDF 461KB] ( 589 ) 633 Sparse Representation with Weighted Fusion of Local Based Non-minimum Square Error and Global for Face Recognition under Occlusion Condition HU Zheng-Ping, PENG Yan, ZHAO Shu-Huan In the occlusion face recognition, some covered parts change the property of local information. It may lead to a wrong classification using the minimum residual as a decision function for sparse representation classification when the residual is approximate. In this case, proceeding from the decision rule of the classifier, the algorithm of sparse representation with weighted fusion of local based non-minimum square error and global is proposed for face recognition. The accumulation of each class of coefficient is mainly used as the decision function and the Borda votes system is introduced for sparse representation classification. Firstly, the sparse coefficient accumulation of each class is calculated for global classification. Then, for the local information, the subblocks coefficient accumulation is used to classify. Considering the different effects of subblocks, the sparsity and residual are utilized to jointly express the weight of credibility. Finally, the global and local blocks are combined to Borda vote for the final classification. The experimental results on public available database demonstrate that the proposed algorithm has good effectiveness and robustness. 2015 Vol. 28 (7): 633-640 [Abstract] ( 557 ) [HTML 1KB] [ PDF 853KB] ( 510 ) 641 Dynamic Hybrid Ant Colony Optimization Algorithm for Solving the Vehicle Routing Problem with Time Windows GE Bin, HAN Jiang-Hong, WEI Zhen, CHENG Lei, HAN Yue To solve the vehicle routing problem with time windows (VRPTW), a dynamic hybrid ant colony optimization algorithm (DHACO) is proposed, so as to avoid the disadvantages of traditional ant genetic hybrid algorithm, such as static setting, redundant iteration and slow convergence. Firstly, an initial solution is obtained through max-min ant system, and the ant colony optimization algorithm is adopted to get a basic feasible solution to VRPTW. Then, the crossing and mutation operations of genetic algorithm are employed to re-optimize local and global solutions, thus the optimal solution is obtained. Finally, based on the fusion strategy of ant genetic hybrid algorithm, and by employing ant algorithm and genetic algorithm dynamically and alternately, the parameters of ant colony algorithm is self-adaptively controlled according to cloud association rules. DHACO reduces the times of redundant iteration and speeds up the rate of the convergence. Simulation results show that DHACO is better than the other related heuristic algorithms as to the optimal solutions. 2015 Vol. 28 (7): 641-650 [Abstract] ( 507 ) [HTML 1KB] [ PDF 639KB] ( 582 ) 651 Low-Rank Matrix Recovery Based on Fisher Discriminant Criterion ZHANG Hai-Xin, ZHENG Zhong-Long, JIA Jiong, YANG Fan The original dataset is decomposed into a set of representative bases with corresponding sparse errors for modeling the raw data in standard low-rank matrix recovery algorithm. Inspired by the Fisher criterion, a low-rank matrix recovery algorithm based on Fisher discriminate criterion is presented in this paper. Low-rank matrix recovery is executed in a supervised learning mode, i.e., taking the within-class scatter and between-class scatter into account when the whole label information is available. The proposed model can be solved by the augmented Lagrange multipliers, and the additional discriminating ability is provided to the standard low-rank models for improving performance. The representative bases learned by the proposed algorithm are encouraged to be structurally coherent within the same class and be independent between classes as much as possible. Numerical simulations on face recognition tasks demonstrate that the proposed algorithm is competitive with the state-of-the-art alternatives. 2015 Vol. 28 (7): 651-656 [Abstract] ( 481 ) [HTML 1KB] [ PDF 395KB] ( 705 ) 657 Evaluation of Decision Rules Performance for Multi-source Decision Information Systems LIN Guo-Ping, LIANG Ji-Ye, LI Jin-Jin The multigranulation rough set theory is proved to be an effective method for extracting decision rules from the multi-source decision information systems. However, how to evaluate the decision rules is one of the key problems to find reasonable and accurate decision rules and predict an unknown sample in terms of decision rules. In this paper, according to the disadvantage of the existing evaluation measures of rule performance, the whole certainty measure, the whole consistency measure, and the whole support measure are proposed. These evaluation measures will be helpful for the solution of some decision problems. 2015 Vol. 28 (7): 657-664 [Abstract] ( 483 ) [HTML 1KB] [ PDF 336KB] ( 488 ) 665 Active Contour Model for Image Segmentation Based on Clustering Information LI Min, LIANG Jiu-Zhen, LIAO Cui-Cui Based on traditional Chan-Vese (CV) model, combining image clustering information, an effective active contour model for image segmentation is proposed in this paper.Firstly, the energy functional of CV model is improved, the gradient information of image is considered, and the accuracy of image segmentation is improved. Then, the coefficient K based on image clustering information is added in energy functional. And the image clustering information is used to initialize the level set curves automatically. In color image segmentation processing, weighting process on the RGB channel is proposed to improve the efficiency of segmentation. Finally, regularization term is added in energy functional to avoid re-initialization of the level set. The gray images and color images are segmented quickly and accurately. Experimental results shows the effectiveness of the proposed method. 2015 Vol. 28 (7): 665-672 [Abstract] ( 580 ) [HTML 1KB] [ PDF 1355KB] ( 962 )

Supervised by China Association for Science and Technology Sponsored by Chinese Association of Automation NationalResearchCenter for Intelligent Computing System Institute of Intelligent Machines, Chinese Academy of Sciences Published by Science Press

Copyright © 2010 Editorial Office of Pattern Recognition and Artificial Intelligence Address: No.350 Shushanhu Road, Hefei, Anhui Province, P.R. China Tel: 0551-65591176 Fax:0551-65591176 Email: bjb@iim.ac.cn Supported by Beijing Magtech 　Email:support@magtech.com.cn