Jiangsu Sunboon Information Technology Co.

Wuxi, China

Jiangsu Sunboon Information Technology Co.

Wuxi, China
SEARCH FILTERS
Time filter
Source Type

Song X.,Nanjing University of Science and Technology | Song X.,Jiangsu Sunboon Information Technology Co. | Song X.,Jiangsu University of Science and Technology | Xue Y.,Nanjing University of Posts and Telecommunications | And 6 more authors.
Nanjing Li Gong Daxue Xuebao/Journal of Nanjing University of Science and Technology | Year: 2013

To solve the underdetermined linear problem in the signal recovery from high-dimensional data, a fuzzy adaptive method for optimizing recovery of compressive sensing matrix is proposed for image reconstruction and recognition. By this means, each high dimensional input sample is firstly partitioned into the several local blocks, and those local blocks are combined to represent the sample as a third-order tensor. Moreover, the compressive measurement is performed on the dimensionality-reduced source signal, which is able to find the properties of statistical independence and linear singular by using multi-scale structural analysis and independent component analysis. Finally, a new fuzzy cost function for optimization of sensing matrix is proposed, in which the update of atoms from sensing matrix are fuzzily handled, and the low coherence is obtained between the properties of observation matrix and dictionary matrix. The merit of the method is that the sparse signal has desirable properties for the number of measurements and representation qualities under the same reconstruction conditions. Extensive experimental studies conducted on ORL, Yale face images and 91 natural images databases show that the effectiveness of the proposed method.


Yang X.,Jiangsu Sunboon Information Technology Co.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

In this paper, the dominance-based rough set approach is introduced into multigranulation environment. Two different dominance-based multigranulation rough sets models: dominance-based optimistic multigranulation rough set and dominance-based pessimistic multigranulation rough set are constructed, respectively. Not only the properties of these two dominance-based multigranulation rough sets are discussed, but also the relationships among dominance-based optimistic multigranulation rough set, dominance-based pessimistic multigranulation rough set and the classical dominance-based rough set are investigated. © 2012 Springer-Verlag.


Yang X.,Jiangsu University of Science and Technology | Yang X.,Nanjing University of Science and Technology | Yang X.,Jiangsu Sunboon Information Technology Co. | Zhang M.,Jiangsu University of Science and Technology | And 3 more authors.
Knowledge-Based Systems | Year: 2011

Neighborhood system formalized the ancient intuition, infinitesimals, which led to the invention of calculus, topology and non-standard analysis. In this paper, the neighborhood system is researched from the view point of knowledge engineering and then each neighborhood is considered as a basic unit with knowledge. By using these knowledge in neighborhood system, the rough approximations and the corresponding properties are discussed. It is shown that in the incomplete information system, the smaller upper approximations can be obtained by neighborhood system based rough sets than by the methods in [Y. Leung, D.Y. Li, Maximal consistent block technique for rule acquisition in incomplete information systems, Information Sciences 115 (2003) 85-106] and [Y. Leung, W.Z. Wu, W.X. Zhang, Knowledge acquisition in incomplete information systems: a rough set approach, European Journal of Operational Research 168 (2006) 164-180]. Furthermore, a new knowledge operation is discussed in the neighborhood system, from which more knowledge can be derived from the initial neighborhood system. By such operations, the regions of lower and upper approximations are further expanded and narrowed, respectively. Some numerical examples are employed to substantiate the conceptual arguments. © 2011 Elsevier B.V. All rights reserved.


Song X.-N.,Nanjing University of Science and Technology | Song X.-N.,Jiangsu Sunboon Information Technology Co. | Song X.-N.,Jiangsu University of Science and Technology | Liu Z.,Nanjing University of Science and Technology | And 4 more authors.
Nanjing Li Gong Daxue Xuebao/Journal of Nanjing University of Science and Technology | Year: 2012

To solve the problem of recognition for handwritten tick symbols filled in form bill images, a hybrid detection approach is presented for form bill image processing. A chain code algorithm based on the sparse representation is proposed by using the morphological method to describe the external contour of form bills. An optimal-axis projection estimation algorithm is proposed for angular coordinate detection of frame lines and angle correction. By means of the frame line characteristics of standard template, the form areas of special characters are determined by template matching, and the judgment criterion of tick symbol is obtained by using the spatial convolution algorithm. This algorithm has generality for form bill image processing, and can decrease the difficulty of location and identification of large-scale multiple bill images. Detection results from the practical form bill images from banks demonstrate the effectiveness of the proposed method.


Yang X.,Jiangsu University of Science and Technology | Yang X.,Nanjing University of Science and Technology | Yang X.,Jiangsu Sunboon Information Technology Co. | Dou H.,Jiangsu University of Science and Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

In this paper, we present an explorative research focusing on dominance-based rough set approach to the incomplete information systems. In most of the rough set literatures, an incomplete information system indicates an information system with unknown values. By assuming that the unknown value can be compared with any other values in the domain of the corresponding attributes, the concept of the valued dominance relation is proposed to show the probability that an object is dominating another one. The fuzzy rough approximations in terms of the valued dominance relation are then constructed. It is shown that by the valued dominance-based fuzzy rough set, we can obtain greater lower approximations and smaller upper approximations than the old dominance-based rough set in the incomplete information systems. Further on the problem of inducing "at least" and "at most" decision rules from incomplete decision system is also addressed. Some numerical examples are employed to substantiate the conceptual arguments. © 2011 Springer-Verlag Berlin Heidelberg.


Song X.,Jiangsu University of Science and Technology | Song X.,Nanjing University of Science and Technology | Song X.,Jiangsu Sunboon Information Technology Co. | Liu Z.,Nanjing University of Science and Technology | And 5 more authors.
Neurocomputing | Year: 2014

Sparse representations using over complete dictionaries has concentrated mainly on the study of pursuit algorithms that decompose signals with respect to a given dictionary. Designing dictionaries to better fit the above model can be done by either selecting one from a pre-specified set of linear transforms or adapting the dictionary to a set of training signals. The K-SVD algorithm is an iterative method that alternates between sparse coding of the examples based on the current dictionary and a process of updating the dictionary atoms to better fit the data. However, the existing K-SVD algorithm is employed to dwell on the concept of a binary class assignment, which means that the multi-classes samples are assigned to the given classes definitely. The work proposed in this paper provides a novel parameterized fuzzy adaptive way to adapting dictionaries. In order to achieve the fuzzy sparse signal representations, the update of the dictionary columns is combined with an update of the sparse representations by embedding a new mechanism of fuzzy set, which is called parameterized fuzzy K-SVD. Experimental results conducted on the ORL, Yale and FERET face databases demonstrate the effectiveness of the proposed method. © 2013 Elsevier B.V.


Song X.,Jiangsu Sunboon Information Technology Co. | Liu Z.,Jiangsu Sunboon Information Technology Co.
Proceedings - 2013 2nd International Conference on Robot, Vision and Signal Processing, RVSP 2013 | Year: 2013

In this paper, we develop a hybrid fuzzy semisupervised learning algorithm (HFSA) for face recognition, which is based on the segregation of distinctive regions that include outlier instances and its counterparts. First, it achieves the distribution information of each sample that represented with fuzzy membership degree, and then the membership grade is incorporated into the redefinition of scatter matrices, as a result, the initial fuzzy classification of whole regular feature space is obtained. Second, a new semi-supervised fuzzy clustering algorithm is presented on the basis of the precise number of clusters and initial pattern centers that have been previously obtained in the pattern discovery stage, and then applied in order to perform the outlier instances classification, yielding the final pattern recognition. Experimental results conducted on the ORL and XM2VTS face databases demonstrate the effectiveness of the proposed method. © 2013 IEEE.

Loading Jiangsu Sunboon Information Technology Co. collaborators
Loading Jiangsu Sunboon Information Technology Co. collaborators