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....

2016 Vol.29 Issue.10, Published 2016-10-31 Papers and Reports    Researches and Applications    Research and Review        Papers and Reports 865 Disparity Optimization Algorithm on Sub-pixel Accuracy for Stereo Matching Using Segmentation Guided Filtering SHI Hua, ZHU Hong , YU Shunyuan To solve the problems of low accuracy and staircase effect in slope and weak texture regions for local stereo matching, a disparity optimization algorithm on sub-pixel accuracy using segmentation guided filtering is proposed. Firstly, the mismatch pixels are checked on the initial stereo matching disparity map according to left-right consistency criterion, and the mismatch disparity is corrected by average filtering. Then, the guide image is obtained by segmenting the corrected disparity map, and the sub-pixel accuracy dense disparity map can be achieved via disparity optimization based on the segmentation guided filtering method. Experimental results show that by the proposed algorithm the smoothness of the disparity map in slope regions is improved effectively, the mismatch rates of the initial stereo matching disparity are reduced, and the higher precision of dense disparity is obtained. 2016 Vol. 29 (10): 865-875 [Abstract] ( 857 ) [HTML 1KB] [ PDF 1689KB] ( 954 ) 876 Software-Defined Crime Scene Analysis Process and Its Knowledge Automation Scheme LIU Shuo, WANG Shuai, FU Huanzhang, WANG Feiyue Crime scene analysis is a backtracking process based on the consequence from traces to psychology (physiology). It is the starting point and foundation of criminal investigation and the key part for solving criminal cases. Learning from the successful application and good performance of the relative theories and methods about complex systems in various domains, the crime scene analysis model based on complex systems is constructed in this paper. From the view of software-defined systems and processes, the basic idea, framework and main methods of the construction of parallel crime scene system model are introduced based on artificial societies, computational experiments, and parallel execution (ACP) method, and the important role played by knowledge automation in parallel systems for the control and management of crime scene analysis is discussed. 2016 Vol. 29 (10): 876-883 [Abstract] ( 489 ) [HTML 1KB] [ PDF 796KB] ( 916 ) 884 Hierarchical Algorithm of Self-Organizing Background Subtraction with Memory Storage LIN Dewei, CHEN Zhaojiong, KE Xiao, YE Dongyi In the algorithm of self-organizing background subtraction, the real-time performance is poor and the background model is easy to offset under complicated environment due to non-periodic changes. Aiming at these problems, a hierarchical algorithm of self-organizing background subtraction with memory storage is proposed. Firstly, a border-shared background model is established to reduce the time and space complexity. And a cache strategy is introduced on the basis of the original matrix model to store the past and the current background data, respectively. Then, during the object detection, a decision mechanism for different granularity levels is designed to determine whether a pixel is an object or not. Experimental results show that the proposed algorithm can overcome the shortcomings of the original one and achieve higher detection accuracy and better real-time performance. 2016 Vol. 29 (10): 884-893 [Abstract] ( 490 ) [HTML 1KB] [ PDF 1331KB] ( 524 ) 894 Frequent Pattern Mining from Biological Sequences Based on Score Matrix YUAN Ermao, GUO Dan, HU Xuegang, WU Xindong Mining significant frequent patterns from biological sequences is an important task in bioinformatics. An algorithm of mining approximate frequent pattern based on score matrix (MAPS) is proposed. Firstly, approximate matching score matrix (MSM) is constructed to handle insertion, replacement and deletion operations with interval constraints. Secondly, the approximate pattern matching based on score matrix (S-APM) scheme is designed to obtain counts of approximate occurrences of each pattern. Finally, a data driven pattern generation method and an Apriori-like rule are adopted to avoid unnecessary candidate patterns. Experiments on protein and DNA sequences show that the MAPS produces better performance, and it can be used to discover co-occurrence frequent patterns among different sequences. 2016 Vol. 29 (10): 894-906 [Abstract] ( 442 ) [HTML 1KB] [ PDF 631KB] ( 460 ) Research and Review 907 Survey of Sparse Structure Learning of Bayesian Networks GUO Min, SHI Hongbo, JI Suqin Sparse structure learning of Bayesian networks can simplify network structure without losing important information of the original network structure. In this paper, the necessity of the sparse structure learning of Bayesian networks and the definition of the sparsity of those are firstly discussed. Based on the general structure learning of Bayesian networks, the existing problems for high-dimensional data are analyzed, and then it is found that score-based structure learning is suitable for sparse structure learning. Therefore, the objective functions and their optimization algorithms are mainly described. Finally, some meaningful research trends are discussed. 2016 Vol. 29 (10): 907-923 [Abstract] ( 676 ) [HTML 1KB] [ PDF 642KB] ( 876 ) Researches and Applications 924 User Interest Related Information Diffusion Pattern Mining in Microblog HAO Zhifeng, HUANG Canjin, CAI Ruichu, WEN Wen, HUANG Yupeng, CHEN Bingfeng Information diffusion modeling is the basis of the community mining and community influence research. Based on a user interest related information diffusion model, a microscopic pattern mining method is proposed to detect the information diffusion features using frequent subtree mining in this paper. Firstly, microscopic information diffusion pattern is converted into frequent subtrees mining by formulating social network in microblog as a series of graphs with users multiple labels. In terms of the microblog social network characteristics of multiple labels on single node, an efficient frequent subtrees mining algorithm on the tree with multiple labels tree miner (MLTreeMiner) is proposed. Finally, combined with topic information extraction method, MLTreeMiner is used to mine information diffusion patterns. Experiments on synthetic data demonstrate that MLTreeMiner is efficient for frequent subtrees mining on the tree with multiple labels. Experiments are also carried out on real data from Sina Weibo, and the validity of the MLTreeMinner is verified. 2016 Vol. 29 (10): 924-935 [Abstract] ( 524 ) [HTML 1KB] [ PDF 588KB] ( 917 ) 936 Construction of Phylogenetic Tree of Flu Virus Proteins Based on Coarse Graining LI Yang, TANG Xuqing Based on the coarse graining theory, a method for constructing phylogenetic tree of flu virus proteins is proposed by combining total 127 065 hemagglutinin and neuraminidase protein sequences. Firstly, to determine the appropriate granularity, a feature vector is obtained to present a virus protein sequence and then an approach is given to construct hierarchical structure of virus system by analyzing similarity among multi-protein sequences. The suitable number of clusters is determined according to hierarchical evaluation index based on the system structure. Furthermore, on the basis of the nearest-to-center principle, the significant viruses can be selected to represent characteristics of the whole class. Finally, the phylogenetic tree is established through the distance metric. The test result indicates that the influenza viruses with same host, similar time span, close outbreak location and same names are more likely to belong to the same branch. The results are identical with that of the existing literature on flu virus. The results provide a foundation for investigating the mutation, evolution and prediction of flu viruses. 2016 Vol. 29 (10): 936-942 [Abstract] ( 420 ) [HTML 1KB] [ PDF 703KB] ( 324 ) 943 Near-Duplicate Image Retrieval Based on Multi-features Bundle SONG Ge, JIN Xin, TAN Xiaoyang The bundle feature based method is effective for near-duplicate image retrieval. However, there are some limitations due to the use of SIFT descriptor and coordinate information. Therefore, in this paper color features are integrated into traditional SIFT representation to form the multi-features bundle and two kinds of robust orders between points in bundle feature are introduced-the circle order and the main orientation order. Firstly, the color distribution features of interesting points are extracted and indexing is embedded into them, and then the potential false matchings are discarded by the Kullback-Leibler divergence of corresponding color features. Secondly, the geometric loss score is computed according to the inconsistency between orders of the distance from points to the center of bundle region or the orders of the angle between the orientation of points and main orientation of bundle. The experiment on Copydays dataset shows the improved retrieval performance by adding color features and better circle order and main orientation order geometric constraints of the bundle feature in the image retrieval task. 2016 Vol. 29 (10): 943-950 [Abstract] ( 477 ) [HTML 1KB] [ PDF 1155KB] ( 325 ) 951 Analysis and Comparison of Concept Lattices from the Perspective of Three-Way Decisions LI Leijun, LI Meizheng, XIE Bin, MI Jusheng Based on the construction of formal concepts and the constitution of formal contexts, the inherent connections between different concept lattices are explored from the perspective of three-way decisions. The comparison of the concept lattices in classical formal context and incomplete formal context, as well as in fuzzy formal context and intuitionistic fuzzy formal context, is given, respectively. Then, the important value of three-way decisions in concept lattice theory is shown. Compared with the concept lattices in classical formal context and fuzzy formal context, the concept lattices in incomplete formal context and intuitionistic fuzzy formal context can reflect the idea of three-way decisions, and they have advantages of small data storage requirement, concise attribute reduction, etc. 2016 Vol. 29 (10): 951-960 [Abstract] ( 478 ) [HTML 1KB] [ PDF 457KB] ( 439 )

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