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....

2017 Vol.30 Issue.8, Published 2017-08-31 Papers and Reports    Researches and Applications        Papers and Reports 673 Parallel Data: From Big Data to Data Intelligence LIU Xin, WANG Xiao, ZHANG Weishan, WANG Jianji, WANG Feiyue For many real world applications, data available from actual scenes are generally incomplete and conditional. Therefore, a mechanism of generating big data from small data and then producing small but precise knowledge for specific problems is extremely useful. To this end, the concept and the framework of parallel data are proposed and discussed. Parallel data consist of virtual data from experimental computing and real data collected for actual problems. Actual and virtual data interact and co-evolute in parallel, making virtual and actual complement, thus enabling the process of transferring big data to data intelligence for general problem solving. Parallel data is not only a new data representation method, but also a new mechanism for data generalization and evolution. The dynamic trajectories of all data constitute a data dynamic system, and provide a new paradigm for data processing, representation, mining, and applications. 2017 Vol. 30 (8): 673-681 [Abstract] ( 944 ) [HTML 1KB] [ PDF 977KB] ( 854 ) 682 Image Quality Assessment Based on Complementary Pooling of Deeply Visual Feature and Strategy LIN Zhijie, FENG Mingkun In methods of visual multiple feature pooling, visual complementary between different image features and different assessment algorithms is not taken into account. A method for complementary pooling of deeply visual feature(CPDVF) is proposed based on the integration of physiological perception of front human visual system(HVS) and psychological processing of back HVS in this paper. Firstly, two kinds of complementary features, the histogram statistics and gradient structure for visual multi-channel, are extracted and deeply processed based on visual characteristics. Secondly, the local distortion algorithm for visual histogram pooled contrast(VHPC) assessment and the local complementary similarity algorithm for visual gradient pooled contrast(VGPC) assessment are proposed. Finally, the distorted image quality is obtained with pooling of VHPC and VGPC based on psychological characteristics and regression function. The experimental results show that the CPDVF is superior to feature similarity and visual saliency in accuracy, stability and monotonicity. 2017 Vol. 30 (8): 682-691 [Abstract] ( 554 ) [HTML 1KB] [ PDF 1474KB] ( 677 ) 692 Variable Precision Intuitionistic Fuzzy Rough Set Based on θ Operator XUE Zhan′ao, YUAN Yilin, XIN Xianwei, HAN Danjie In the fuzzy approximation space, combing the membership degree and non membership degree of the intuitionistic fuzzy sets with fuzzy implication operator, the concept of intuitionistic fuzzy and its membership degree and non membership degree based on θ operator and θ* operator are presented and their properties are proved. Then, integrating the intuitionistic fuzzy set and the variable precision rough set, a variable precision intuitionistic fuzzy rough set is defined based on θ operator.A method to solve the threshold parameter β of variable precision rough set is put forward. Finally, an example for analyzing the method is provided. 2017 Vol. 30 (8): 692-701 [Abstract] ( 584 ) [HTML 1KB] [ PDF 754KB] ( 396 ) 702 Improved Power Spectrum Based Magnetotactic Bacteria Algorithm ZHANG Xuexue, LIU Sanyang A global search operator is designed to prevent the magnetic bacteria algorithm from falling into local minimum easily and an improved power spectrum-based magnetotactic bacteria algorithm(IPSMBA) is proposed. The algorithm is based on the power spectrum of the magnetic particles in bacteria bodies. The bacteria moment regulation operator and bacteria replacement operator are improved in the process of simulating magnetic bacteria moment regulation. To make the best of the power spectrum information and increase the diversity of the population, a new moment replacement operator combining chaotic mapping and power spectrum replacement operator is designed. The experiments indicate that IPSMBA achieves better convergence and robustness on low dimensional benchmark functions, and it has better performance on high dimensional functions. 2017 Vol. 30 (8): 702-709 [Abstract] ( 485 ) [HTML 1KB] [ PDF 644KB] ( 428 ) 710 Classification Model Based on Variable Multi-granulation Probabilistic Rough Set WANG Jiaqi, MIAO Duoqian, ZHANG Hongyun Based on the multi-granulation rough set theory, a variable multi-granulation probabilistic rough set (VMGPRS) model combining the ideas of variable multi-granulation and misclassification rate is proposed. A granulation reduction algorithm is put forward grounded on the concept of attribute reduction in rough set theory, and the granulation redundancy caused by parameter setting in the variable multi-granulation rough set model is found and solved. The data before and after the reduction are applied to classical classification algorithms such as support vector machine, k-nearest neighbor, Naive Bayes, and it is verified that the classification ability of data is hardly influenced by the reduction. With the combination of the rule and the proposed algorithm, a rule-based classification algorithm is designed. Furthermore, two adjustment parameters, α and β, in the VMGPRS model are analyzed for classification effect of the classifier. 2017 Vol. 30 (8): 710-717 [Abstract] ( 554 ) [HTML 1KB] [ PDF 705KB] ( 474 ) Researches and Applications 718 Target Tracking Based on Occlusion Detection and Spatio-Temporal Context Information CHU Jun, ZHU Tao, MIAO Jun, JIANG Landa The existing spatio-temporal context based target tracking algorithms have good performance for static occlusion due to the consideration of the spatio-temporal relationship between the object and the background. However, the large occlusion area of the object or the occluded fast-moving object still easily lead to inaccurate tracking or lost tracking. A target tracking algorithm combining target occlusion detection and context information is proposed in this paper. Firstly, the compressed illumination-invariant color features extracted from the first frame are utilized to constitute and initialize spatio-temporal context model. Then, the occlusions in the inputted video frames are judged by bidirectional trajectory error. If the bidirectional matching error of key points in object region between consecutive frames is less than a set threshold, there is no dynamic occlusion or severe static occlusion. Accordingly, the accurate tracking is conducted in virtue of spatio-temporal context model. Otherwise, the objects in the subsequent frames are detected by the combined classifiers until the objects can be detected again. Meanwhile, the context model and classifiers are updated online. The experimental results on several video frame sequences show that the proposed method can deal with severe static occlusion and dynamic occlusion in complex scenario well. 2017 Vol. 30 (8): 718-727 [Abstract] ( 848 ) [HTML 1KB] [ PDF 2549KB] ( 535 ) 728 Super Pixel Tracking Algorithm via Fusing Saliency Detection into Spatio-Temporal Context GUO Chunmei, CHEN Ken, LI Meng, LI Fei To achieve more efficient utilization of image feature information and improve the tracking accuracy as well as robustness of the target, an improved super pixel tracking algorithm via fusing salient region detection into spatio-temporal context is proposed. Firstly, super pixel segmentation is conducted in the context region of the target, then the motion relevance of target context and the regional covariance information are utilized to calculate the correlation saliency of the image super pixels. Based on Bayesian framework, the model fusing saliency detection into spatio-temporal context is built in the frequency domain. Next, the color and texture histograms of current frame and reference template are employed to calculate the Bhattacharyya coefficient and update the spatial and temporal context model. The scale pyramid model is introduced to estimate the target scale. Finally, the adaptive motion prediction module is incorporated by updating online dynamic model sample set and using ridge regression method to determine the parameters of a low pass filter. Experimental results on public database indicate the superiority of the proposed algorithm over other algorithms in illumination change, complex background, object rotation, high mobility, low resolution. 2017 Vol. 30 (8): 728-739 [Abstract] ( 485 ) [HTML 1KB] [ PDF 6305KB] ( 404 ) 740 Convex Discriminant Canonical Correlation Analysis JIANG Fan, CHEN Songcan Inspired by geometric mean metric learning(GMML), a convex discriminant canonical correlation analysis(CDCA) is proposed. The learning of two projection matrices is transformed into a geodesic convex problem of metric learning. Thereby a closed form solution is acquired and simultaneously discriminant fused features are extracted directly. The experiments on artificial and real datasets verify the effectiveness of CDCA. 2017 Vol. 30 (8): 740-746 [Abstract] ( 599 ) [HTML 1KB] [ PDF 616KB] ( 428 ) 747 Medical Image Super-Resolution Reconstruction Method Based on Non-local Autoregressive Learning XU Jun, LIU Hui, YIN Yilong In the process of medical imaging, the image resolution is limited by radiation dose constraints and imaging equipment conditions. The accuracy of late clinical diagnosis and treatment is affected by the low resolution of the medical image. To solve this problem, a medical image super-resolution reconstruction method based on non-local autoregressive learning is proposed. According to the non-local similarity characteristic inherent in medical images, the autoregressive model based on sparse representation is applied to the super-resolution reconstruction process. Furthermore, to improve the efficiency of the experiment, the clustering algorithm is utilized to acquire the classification dictionary. The experimental results demonstrate the feasibility of the proposed method in improving the resolution of medical images as well as the reconstruction efficiency and performance. 2017 Vol. 30 (8): 747-753 [Abstract] ( 781 ) [HTML 1KB] [ PDF 970KB] ( 500 ) 754 Semi-supervised Manifold Learning Algorithm Based on Neighbourhood Components Analysis LI Xueqing, WANG Jing, DU Jixiang In most of the existing manifold learning algorithms, the geometry structure of the data instances is preserved, but the label information is ignored. Therefore, the application of manifold learning algorithms in data classification is limited. In this paper, a semi-supervised manifold learning algorithm based on neighborhood components analysis is proposed. A distance metric matrix is learned by using neighbor components analysis and local neighbors of the sample points are selected by using the new distance metric. The local geometric structures of the sample points and their neighbors are constructed under the new distance metric, and the local geometric structures are preserved in the low-dimensional embedding coordinates of the sample points. The classification experiments conducted on three different datasets demonstrate the efficiency of the proposed algorithm. 2017 Vol. 30 (8): 754-760 [Abstract] ( 560 ) [HTML 1KB] [ PDF 542KB] ( 558 ) 761 Sorting Method of Multi-attribute Decision Making Based on Weighted α Dominance Relation LI Jia, LIANG Jiye, PANG Tianjie In the dominance rough set model, simple average method is exploited to weight sorting results of multiple thresholds , but the dataset information is ignored, and a greater difference in the sorting quality of different datasets is generated. Aiming at this problem, a dominance degree sorting method of weighted dominance relation is proposed in this paper. Firstly, dominance rough set method is applied to decision-making objects for detailed analysis. Then, to overcome the problem of parallel decision phenomenon in multi-attribute decision making sorting results caused by the subjective weight of parameter α, two criteria are applied to weight the parameter α by sorting results, and comprehensive dominance degree of all objects are fused by the two criteria to further refine the sorting result. Finally, through the comparison with other sorting methods on specific cases, the feasibility and the validity of the proposed sorting method are verified. 2017 Vol. 30 (8): 761-768 [Abstract] ( 503 ) [HTML 1KB] [ PDF 507KB] ( 524 )

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