Showing papers by "Yuan Yan Tang published in 2006"

Wavelet-hough transform with applications in edge and target detections

Abstract: In computer vision and image processing, edge detection concerns the localization of significant variations of the grey level image and the identification of the physical phenomena that originated them. It is difficult to design a general edge detection algorithm which performs well in many contexts and captures the requirements of subsequent processing stages. Consequently, over the history of digital image processing, a variety of edge detectors have been devised which differ in their mathematical and algorithmic properties. In this paper, we briefly describe some existing edge detectors. We investigate the wavelet transform with direction which is related to the Canny and Marr–Hildreth's edge detectors. This wavelet transform has a remarkable feature that plays an important role in image processing applications. We discuss its applications to edge and target detections. The direction effect enables the wavelets to analyze the directional features of images. Algorithms and experiments for edge and target detections have been developed based on the theory. And, very promising results have been shown in image processing.

Thinning character using modulus minima of wavelet transform

Abstract: An essential step in character recognition is to extract the skeleton characteristics of the character. In this paper, an efficient algorithm is proposed to extract visually satisfactory skeleton f...

Handwritten Character Recognition Using Combined Gradient and Wavelet Feature

Abstract: A novel feature based on the combination of gradient feature and coefficients of wavelet transform is developed in this paper. In handwritten character recognition, the gradient feature represents local characteristic properly, but it is sensitive to the deformation of handwritten character. Meanwhile, wavelet transform represents the character image in multiresolution analysis and keeps adequate global characteristic in different scales. In order to improve the discrimination power, we composed local and global characteristic in a combined feature. Three combination schemes are described in this paper. Experiments are conduced on two chinese character databases, ETL8B subset (197 categories) and HKBU-SC110 (110 categories), to test the performance of proposed feature. The recognition accuracies of our feature achieve 95.53% and 93.77% for ETL8B subset and HKBU-SC110 by 1-NN classifier, respectively, which are higher than those of gradient feature.

Handwriting-based personal identification

Abstract: Handwriting-based personal identification, which is also called handwriting-based writer identification, is an active research topic in pattern recognition. Despite continuous effort, offline handwriting-based writer identification still remains as a challenging problem because writing features can only be extracted from the handwriting image. As a result, plenty of dynamic writing information, which is very valuable for writer identification, is unavailable for offline writer identification. In this paper, we present a novel wavelet-based Generalized Gaussian Density (GGD) method for offline writer identification. Compared with the 2-D Gabor model, which is currently widely acknowledged as a good method for offline handwriting identification, GGD method not only achieves a better identification accuracy but also greatly reduces the elapsed time on calculation in our experiments.

Texture Image Retrieval Using Novel Non-separable Filter Banks Based on Centrally Symmetric Matrices

Abstract: Though millions of images are stored in a large digital image library today, the user can not access or make full use of these image information unless the digital image library is well organized in order to allow efficient browsing, searching and retrieval. Thus, research in image retrieval has been an active discipline since 70’s last century. Image retrieval is a typical problem of pattern recognition, consisting of two parts: extracting features (EF) and similarity measurement (SM). In this paper, we develop new non-separable filter banks based on the centrally symmetric matrixes, and apply them to extract the features of texture images. Compared to tensor product wavelets, our new filter banks can capture more directional texture information, which is helpful for texture image retrieval. Experiments show that our novel non-separable filter banks are satisfiable and achieve a better retrieval effectiveness than Daubechies wavelets.

Wavelet natural boundary element method for the neumann exterior problem of stokes equations

Abstract: Natural boundary element approach is a promising method to solve boundary value problems of partial differential equations. This paper addresses the Neumann exterior problem of Stokes equations using the wavelet natural boundary element method. The Stokes exterior problem is reduced into an equivalent Hadamard-singular Natural Integral Equation (NIE). By virtue of the wavelet-Galerkin algorithm, the simple and accurate computational formulae of stiffness matrix are obtained. The 2J+3 × 2J+3 stiffness matrix is sparse and determined only by its 2J + 3J + 1 entries. It greatly decreases the computational complexity. Also, the condition number of stiffness matrix is , where N is the discrete node number. This indicates that the proposed algorithm is more stable than that of classical finite element method. The error estimates are established for the wavelet-Galerkin approximate solution. Several numerical examples are given to evaluate the performance of our method with encouraging results.

Guaranteed attractivity of equilibrium points in a class of delayed neural networks

Abstract: This paper addresses qualitative properties of equilibrium points in a class of delayed neural networks. We derive a sufficient condition for the local exponential stability of equilibrium points, and give an estimate on the domains of attraction of locally exponentially stable equilibrium points. Our condition and estimate are formulated in terms of the network parameters, the neurons' activation functions and the associated equilibrium point; hence, they are easily checkable. Another advantage of our results is that they neither depend on monotonicity of the activation functions nor on symmetry of the interconnection matrix. Our work has practical importance in evaluating the performance of the related associative memory. To our knowledge, this is the first time to present an estimate on the domains of attraction of equilibrium points for delayed neural networks.

Wavelet Thinning Algorithm Based Similarity Evaluation for Offline Signature Verification

Abstract: Structure distortion evaluation is able to allow us directly measure similarity between signature patterns without classification using feature vectors which usually suffers from limited training samples. In this paper, we incorporate merits of both global and local alignment algorithms to define structure distortion using signature skeletons identified by a robust wavelet thinning technique. A weak affine model is employed to globally register two signature skeletons and structure distortion between two signature patterns are determined by applying an elastic local alignment algorithm. Similarity measurement is evaluated in the form of Euclidean distance of all found corresponding feature points. Experimental results showed that the proposed similarity measurement was able to provide sufficient discriminatory information in terms of equal error rate being 18.6% with four training samples.

Clustering algorithm research based on self-organizing feature maps networks

Abstract: Self-organizing feature maps (SOFM) can learn both the distribution and topology of the input vectors they are trained on. According to this characteristic, we construct neural networks with a family of self-organizing feature maps to cluster the input data space. The proposed algorithm in this paper defines a novel similarity measure, topological similarity, and employs some new concepts, such as SOFM family, UsageFactor. The clustering algorithm handles the clusters with arbitrary shapes and avoid the limitations of the conventional clustering algorithms. We conclude our paper by several experiments with synthetic and standard data set of different characteristics, which show good performance of the proposed algorithm.

Biorthogonal Wavelets Associated with Two-Dimensional Interpolatory Function

Abstract: To construct biorthogonal wavelets from two-dimensional interpolatory function, a large amount of computation is involved in traditional method In this paper, a method is developed for constructing the biorthogonal wavelets Masks of the biorthogonal wavelets are given explicitly Neither the Gram-Schmidt processing nor the inverse of a nonsingular polynomial matrix is needed