Entity

Time filter

Source Type


Yonghui P.,Jiangyin Polytechnic College | Yonghui P.,Jiangsu Engineering R and nter for Information Fusion Software | Rui F.,Jiangyin Polytechnic College
Journal of Computers | Year: 2012

Using wavelet transform to handle auto-mobile image with complex background for license localization, then preprocess license characters on vehicle licenses, and extracting the textural features of license characters in wavelet space, this paper proposed a novel algorithm for vehicle license localization and character recognition which is based on adaptive wavelet neural networks. Firstly, it uses the wavelet transform to preprocess color vehicle image into index image which undergoes wavelet transform to obtain wavelet feature coefficients. Secondly, license position could be located through morphological operation. Thirdly, it extracts the features of localized license characters in wavelet space which is presented to the wavelet neural network as inputs. At last, an adaptive wavelet neural network based on wavelet transform is constructed to recognize license characters. Experimental results demonstrate that the proposed approach could efficiently be used as a vehicle license characters recognition system with high convergence, which is robust for license-size, licensecolor and background complexity. © 2012 ACADEMY PUBLISHER. Source


Deng Z.,Jiangnan University | Deng Z.,Soochow University of China | Shitong Wang,Jiangnan University | Shitong Wang,Soochow University of China | And 2 more authors.
Proceedings - 2011 8th International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2011 | Year: 2011

A key technique for protein analysis is the geometric alignment of two-dimensional polyacrylamide gel electrophoresis (2-D PAGE), i.e. 2-D PAGE image registration. In this study, according to the characteristics of 2-D gel image registration, a fuzzy-inference-rule based flexible model (FIM-FM) is proposed to model the complex geometric transformation and intensity variation between 2-D gel image pairs. Then, by introducing the concept of motion estimation and the corresponding sum-of-squared-difference (SSD) objective function, the parameter learning rules of the proposed model are derived for registration. Accordingly, an adaptive FIR-FM-based automatic elastic image registration algorithm is presented for 2-D gel image registration here. It is distinguished by its strong ability in approximating complex nonlinear transformation inherited from fuzzy inference and its efficiency and adaptability in obtaining precise model parameters through effective parameter learning rules. Our experiments show that the proposed algorithm is highly effective for registration of 2-D gel images. © 2011 IEEE. Source


Tu L.,Jiangsu Engineering R and nter for Information Fusion Software | Tu L.,Jiangyin Polytechnic College
Advances in Intelligent and Soft Computing | Year: 2012

This paper presents a multiple data streams clustering algorithm CA-cluster based on Kendall correlation analysis. CA-cluster instantaneously adjusts the number of clusters and detects the development of the data stream. Since the high velocity and the large number of the data stream, it is not feasible to retain the raw data to calculate the correlation coefficient. We propose a compression mechanism based on AU statistics to support the only once-scanned algorithm to calculate the Kendall correlation coefficient. Experimental results show that our algorithms are more superior to other methods in the aspect of clustering quality, speed and scalability. © 2012 Springer-Verlag GmbH. Source


Pan Y.,Jiangsu Engineering R and nter for Information Fusion Software | Deng Z.,Jiangsu Engineering R and nter for Information Fusion Software | Deng Z.,Jiangnan University | Deng Z.,Soochow University of China | And 3 more authors.
Information Technology Journal | Year: 2012

A key technique for protein analysis is the geometric alignment of two-dimensional polyacrylamide gel electrophoresis (2-D PAGE), i.e., 2-D PAGE image registration. In this study, the adaptability in elastic image registration was emphasized. According to the characteristics of 2-D gel image registration, a fuzzy-inference-rule based flexible model (FIM-FM) is proposed to model the complex transformation between 2-D gel image pairs. By introducing the concept of motion estimation, the parameter learning rules of the proposed model are derived for registration. The experiments show that the proposed algorithm is highly effective for registration of 2-D gel images and is competitive to the existing state-of-the-art algorithms. © 2012 Asian Network for Scientific Information. Source


Tu L.,Jiangsu Engineering R and nter for Information Fusion Software | Tu L.,Jiangyin Polytechnic College | Cui P.,Jiangsu Engineering R and nter for Information Fusion Software | Cui P.,Jiangyin Polytechnic College
WIT Transactions on Engineering Sciences | Year: 2014

Most existing clustering algorithms on uncertain data stream cannot discover the arbitrary shapes since they are based on k-means. To address this issue, this paper proposes a density grid-based clustering algorithm (DG-UStream) over uncertain data stream. DG-UStream uses the grid structure to store the summary information of data tuples in the stream which could be easily updated periodically. Clusters are formed by merging the adjacent grids. Furthermore, an efficient technique is developed to detect and delete the isolated grids which could greatly reduce the time and space costs. The experimental results show that DG-UStream has superior clustering performance in terms of clustering quality and time efficiency. © 2013 WIT Press. Source

Discover hidden collaborations