Wen-Sheng Chen

Wen-Sheng Chen
Shenzhen University · College of Mathematics and Statistics

PhD

About

140
Publications
20,415
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
1,672
Citations
Citations since 2016
49 Research Items
955 Citations
2016201720182019202020212022050100150
2016201720182019202020212022050100150
2016201720182019202020212022050100150
2016201720182019202020212022050100150
Introduction
Wen-Sheng Chen currently works at the College of Mathematics and Statistics, Shenzhen University. Wen-Sheng does research in Artificial Intelligence and Algorithms. Their most recent publication is 'Preface: Special Issue — Feature extraction for pattern recognition: Generation and applications'.
Additional affiliations
September 2009 - August 2010
Hong Kong Baptist University
Position
  • Visiting Professor
December 2006 - July 2015
Shenzhen University
Position
  • Professor
January 2006 - February 2016
Shenzhen University
Position
  • Professor

Publications

Publications (140)
Article
Non-negative matrix factorization (NMF) is a powerful tool for image data analysis and has been utilized in a range of applications involving data classification, clustering, and image processing. Nevertheless, NMF and most of its variants are unsupervised learning algorithms and do not take account of the projection property of the basis image mat...
Article
Nonnegative matrix factorization (NMF), distinguished from the approaches for holistic feature representation, is able to acquire meaningful basis images for parts-based representation. However, NMF does not utilize the data-label information and usually achieves undesirable performance in classification. To address the above-mentioned problem of N...
Article
Deep Nonnegative Matrix Factorization (Deep NMF) is an effective strategy for feature extraction in recent years. By decomposing the matrix recurrently on account of the NMF algorithms, we obtain a hierarchical neural network structure as well as exploring more interpretable representations of the data. This paper mainly focuses on some theoretical...
Chapter
Non-negative matrix factorization (NMF) is a single-layer decomposition algorithm for image data analysis and cannot reveal the intrinsic hierarchical-structure information hidden under the data. Moreover, it extracts features with weak discriminant power due to merely using unlabeled data. These flaws cause the undesirable performance of NMF in im...
Article
Full-text available
Kernel-based non-negative matrix factorisation (KNMF) is a promising nonlinear approach for image data representation using non-negative features. However, most of the KNMF algorithms are developed via a specific kernel function and thus fail to adopt other kinds of kernels. Also, they have to learn pre-image inaccurately that may influence the rel...
Chapter
Non-negative matrix factorization (NMF) has the ability for non-negative feature extraction and is successfully exploited for parts-based image representation. Most NMF-based algorithms utilize loss function with \(l_2\)-norm or Kullback-Leibler divergence to evaluate the quality of factorization. However, these measurements are sensitive to noise...
Chapter
Deep non-negative matrix factorization (DNMF) is a promising method for non-negativity multi-layer feature extraction. Most of DNMF algorithms are repeatedly to run single-layer NMF to build the hierarchical structure. They have to eliminate the accumulated error via fine-tuning strategy, which is, however, too time-consuming. To deal with the draw...
Chapter
Nonnegative matrix factorization (NMF) is a low-rank decomposition based image representation method under the nonnegativity constraint. However, a lot of NMF based approaches utilize Frobenius-norm or KL-divergence as the metrics to model the loss functions. These metrics are not dilation-invariant and thus sensitive to the scale-change illuminati...
Article
In this paper, a novel energy functional minimization model is proposed for ultrasound images denoising. A controllable regularized term and the variational method are employed in the process of speckle noise. This model not only improves the plasticity of the model, but also improves the effect and efficiency of noise removal. The new model has di...
Article
Clustering aims at naturally grouping the data according to the underlying data distribution. The data distribution is often estimated using a parametric or nonparametric model, e.g., Gaussian mixture or kernel density estimation. Compared with nonparametric models, parametric models are statistically stable, i.e., a small perturbation of data poin...
Chapter
Full-text available
Choosing a proper classifier for one specific data set is important in practical application. Automatic classifier selection (CS) aims to recommend the most suitable classifiers to a new data set based on the similarity with the historical data sets. The key step of CS is the extraction of data set feature. This paper proposes a novel data set feat...
Article
Kernel nonnegative matrix factorization (KNMF) algorithms have been widely used to extract features for face recognition. The choice of kernel function is vital to facial feature extraction. The polynomial kernel function has been commonly used in KNMF. The power of the polynomial kernel is required to be a positive integer, thereby ensuring that t...
Article
Nonnegative matrix factorization (NMF) is a linear approach for extracting localized feature of facial image. However, NMF may fail to process the data points that are nonlinearly separable. The kernel extension of NMF, named kernel NMF (KNMF), can model the nonlinear relationship among data points and extract nonlinear features of facial images. K...
Article
Nonnegative matrix factorization (NMF) is a promising method to represent facial images using nonnegative features under a low-rank nonnegative basis-image matrix. The facial images usually reside on a low-dimensional manifold due to the variations of illumination, pose and facial expression. However, NMF has no ability to uncover the manifold stru...
Article
Full-text available
Total variation (TV) minimization for the sparse-view x-ray computer tomography (CT) reconstruction has been widely explored to reduce radiation dose. However, due to the piecewise constant assumption for the TV model, the reconstructed images often suffer from over-smoothness on the image edges. To mitigate this drawback of TV minimization, we pre...
Article
Full-text available
Dimensionality reduction is an important topic in machine learning community, which is widely used in the areas of face recognition, visual detection and tracking. Preserving local and global structures simultaneously is crucial for dimensionality reduction. In this paper, local and global approaches are generalized, respectively, and then a unifie...
Article
Full-text available
Rician noise removal for Magnetic Resonance Imaging (MRI) is very important because the MRI has been widely used in various clinical applications and the associated Rician noise deteriorates the image quality and causes errors in interpreting the images. Great efforts have recently been devoted to develop the corresponding noise-removal algorithms,...
Article
Non-negative Matrix Factorization (NMF), as a promising image-data representation approach, encounters the problems of slow convergence and weak classification ability. To overcome these limitations, this paper, based on different error measurements, proposes two kinds of NMF algorithms with fast gradient descent and high discriminant performance....
Conference Paper
Full-text available
Kernel based Non-Negative Matrix Factorizations (KNMFs) are one of the most important methods for non-negative nonlinear feature extractions and have achieved good performance in pattern classifications. However, most existing KNMF algorithms are merely valid for one special kernel function. Also, they model the pre-images inaccurately. In this pap...
Article
Full-text available
Unlike the visual object tracking, thermal infrared object tracking can track a target object in total darkness. Therefore, it has broad applications, such as in rescue and video surveillance at night. However, there are few studies in this field mainly because thermal infrared images have several unwanted attributes, which make it difficult to obt...
Article
In order to improve the interpretation of principal components, many sparse principal component analysis (PCA) methods have been proposed by in the form of self-contained regression-type. In this paper, we generalize the steps needed to move from PCA-like methods to its self-contained regression-type, and propose a joint sparse pixel weighted PCA m...
Conference Paper
Nonnegative matrix factorization (NMF) is a promising approach to extract the sparse features of facial images. It is known that the facial images usually reside on multi-manifold due to the variations of illumination, pose and facial expression. However, NMF lacks the ability of modeling the structure of data manifold. To improve the performance o...
Article
A method for adaptive background modeling based on the incremental non-negative matrix factorization (INMF) is proposed. INMF is used to update new background models effectively when new data streams arrive. The experimental results show that, compared with non-negative matrix factorization (NMF), INMF not only takes less running time but also can...
Article
Full-text available
Robustness and efficiency are the two main goals of existing trackers. Most robust trackers are implemented with combined features or models accompanied with a high computational cost. To achieve a robust and efficient tracking performance, we propose a multi-view correlation tracker to do tracking. On one hand, the robustness of the tracker is enh...
Conference Paper
Automatic detection of epileptic seizure plays an important role in the diagnosis of epilepsy for it can obtain invisible information of epileptic electroencephalogram (EEG) signals exactly and reduce the heavy burdens of doctors efficiently. Current automatic detection technologies are almost shallow learning models that are insufficient to learn...
Article
This paper presents our study of the problems associated with learning supervised kernels from a large amount of side information. We propose a new loss function derived from the Laplacian matrix of a special complete graph that is generated from the side information. We analyze the relationship between the proposed loss function and the kernel ali...
Conference Paper
Nonnegative matrix factorization (NMF) is a promising method for local feature extraction in face recognition. However, NMF is time-consuming when performing on a large matrix. Another limitation of NMF is that it cannot update the factors incrementally as new training data are available. To overcome these limitations, this paper proposes a block s...
Article
Nonnegative Matrix Factorization (NMF) is a promising algorithm for dimensionality reduction and local feature extraction. However, NMF is a linear and unsupervised method. The performance of NMF would be degraded when dealing with the complicated nonlinear distributed data, such as face images with variations of pose, illumination and facial expre...
Article
Full-text available
Visual tracking remains a challenging problem in computer vision due to the intricate variation of target appearances. Some progress made in recent years has revealed that correlation filters, which formulate the tracking process by creating a regressor in the frequency domain, have achieved remarkable experimental results on a large amount of vide...
Article
Choosing suitable kernels plays an important role in the performance of kernel methods. Recently, a number of studies were devoted to developing nonparametric kernels. Without assuming any parametric form of the target kernel, nonparametric kernel learning offers a flexible scheme to utilize the information of the data, which may potentially charac...
Conference Paper
This paper attempts to develop a novel Non-negative Matrix Factorization (NMF) algorithm to improve traditional NMF approach. Based on gradient descent method, we appropriately choose a larger step-length than that of traditional NMF and obtain efficient NMF update rules with fast convergence rate and high performance. The step-length is determined...
Conference Paper
The existing Kernel Nonnegative Matrix Factorization (KNMF) cannot ensure the non-negativity of the mapped data in the kernel feature space. This is called the nonnegative in-compatible problem of KNMF. To tackle this problem, this paper presents a new methodology to construct Nonnegative Compatible Kernel (NC-Kernel) for face recognition. We obtai...
Article
Full-text available
In face recognition (FR), a lot of algorithms just utilize one single type of facial features namely global feature or local feature, and cannot obtain better performance under the complicated variations of the facial images. To extract robust facial features, this paper proposes a novel Semi-Supervised Discriminant Analysis (SSDA) criterion via no...
Conference Paper
The regularization plays an important role in the sparse-view x-ray computer tomography (CT) reconstruction. Based on the piecewise constant assumption, total variation (TV) regularization has been widely discussed for the sparse-view CT reconstruction. However, TV minimization often leads to some loss of the image edge information during reducing...
Article
Full-text available
The data from real world usually have nonlinear geometric structure, which are often assumed to lie on or close to a low-dimensional manifold in a high-dimensional space. How to detect this nonlinear geometric structure of the data is important for the learning algorithms. Recently, there has been a surge of interest in utilizing kernels to exploit...
Article
Feature extraction is an important problem in face recognition. There are two kinds of structural features, namely Euclidean structure and the manifold structure. However, the single-structural feature extraction methods cannot fully utilize the advantages of global feature and local feature simultaneously. Thus their performances will be degraded....
Article
Biclustering algorithm is used to find local patterns as an important tool in the analysis of gene expression data. However, most of the biclusters found by existing biclustering algorithms consist of non-continuous columns. It is not suitable for time series gene expression data, which has not been extensively studied. This paper presents an effic...
Conference Paper
This paper aims to establish a novel framework for highperformance Mercer kernel construction. Based on a given kernel matrix incorporated the class label information, a nonlinear mapping is firstly generated and well-defined on the training samples. The partial data-defined mapping can be extended and well-defined on the entire pattern space by me...
Conference Paper
To enhance the discriminant power of features in face recognition, this paper builds a novel discriminant criterion by nonlinearly combining global feature and local feature, which also incorporates the geometric distribution weight information of the training data. Two formulae are theoretically derived to determine the optimal parameters that bal...
Article
Many partial differential equation (PDE) models have been proposed for image segmentation. However, most of them cannot handle the complicated images such as inhomogeneous images. To overcome the limitation of traditional PDE-based models, a novel PDF model is proposed in this paper and then successfully applied to inhomogeneous image segmentation...
Article
Full-text available
We are grateful to the authors of the special issue for their contributions and the reviewers for their valuable comments on the submissions. Ming Li acknowledges the supports for his work in part by the National Natural Science Foundation of China under the Project Grant nos. 61272402, 61070214, and 60873264 and by the Macau Science and Technology...
Conference Paper
In this paper, we propose a regularization framework for learning geometry-aware kernels. Some existing geometry-aware kernels can be viewed as instances in our framework. Moreover, the proposed framework can be used as a general platform for developing new geometry-aware kernels. We show how multiple sources of information can be integrated in our...
Conference Paper
In this paper, a linear method to determining intrinsic parameters from two parallel line-segments is proposed. Constrains based on the length ratio of line-segments are used to solve the camera calibration problem from images of two parallel line-segments under different conditions. And for each setting, we can get linear solution for intrinsic pa...
Conference Paper
Traditional Fisher linear discriminant analysis (FLDA) method is a promising algorithm for face recognition. However, FLDA does not utilize the geometric distribution information of the training face data, which will degrade its performance. In order to enhance the discriminant power of FLDA, this paper proposes a novel Fisher criterion by using ge...
Article
Full-text available
By summarizing some classical active contour models from the view of level set representation, a simple energy function expression with the Gaussian kernel of fractional order is proposed, and then a novel region-based geometric active contour model is established. In this proposed model, the energy function with value of [−1, 1] is built, the loca...
Article
Full-text available
Fisher linear discriminant analysis (FLDA) is a classic linear feature extraction and dimensionality reduction approach for face recognition. It is known that geometric distribution weight information of image data plays an important role in machine learning approaches. However, FLDA does not employ the geometric distribution weight information of...
Chapter
This paper attempts to utilize the basis images of block non-negative matrix factorization (BNMF) to serve as the sparse learning dictionary, which is more suitable for non-negative sparse representation (NSR) because they have non-negative compatibility. Based on BNMF-basis-image dictionary, the NSR features of query facial images can be learnt di...
Article
Locality-preserving projection (LPP) is a promising manifold-based dimensionality reduction and linear feature extraction method for face recognition. However, there exist two main issues in traditional LPP algorithm. LPP does not utilize the class label information at the training stage and its performance will be affected for classification tasks...
Conference Paper
Chan-Vese (CV) model is a promising active contour model for image segmentation. However, CV model does not utilize local region information of images and thus CV model based segmentation methods cannot achieve good segmentation results for complex image with some in-homogeneity intensities. To overcome the limitation of CV model, this paper presen...
Article
In face recognition tasks, Fisher discriminant analysis (FDA) is one of the promising methods for dimensionality reduction and discriminant feature extraction. The objective of FDA is to find an optimal projection matrix, which maximizes the between-class-distance and simultaneously minimizes within-class-distance. The main limitation of traditiona...
Article
Full-text available
The 2-D bar code possesses large capacity of data, strong ability for error correction, and high safety, which boosts the 2-D bar code recognition technology being widely used and developed fast. This paper presents a novel algorithm for locating data matrix code based on finder pattern detection and bar code border fitting. The proposed method mai...
Article
Full-text available
Multiplicative noise, also known as speckle noise, is signal dependent and difficult to remove. Based on a fourth-order PDE model, this paper proposes a novel approach to remove the multiplicative noise on images. In practice, Fourier transform and logarithm strategy are utilized on the noisy image to convert the convolutional noise into additive n...
Article
Full-text available
Abnormal running behavior frequently happen in robbery cases and other criminal cases. In order to identity these abnormal behaviors a method to detect and recognize abnormal running behavior, is presented based on spatiotemporal parameters. Meanwhile, to obtain more accurate spatiotemporal parameters and improve the real-time performance of the al...
Conference Paper
This paper presents an Automatical Parameter Determination (APD) approach for face recognition. A modified discriminant criterion is first proposed via consideration of both Euclidean structure and the manifold structure in the face pattern space. Two parameter-formulae are derived by maximizing the discriminant index. The update equations are then...
Conference Paper
This paper addresses Small Sample Size (3S) problem of Locality Preserving Projection (LPP) approach in face recognition. It is well-known that the dimension of pattern vector obtained by vectorizing a facial image is very high and usually greater than the number of training samples. Under this situation, 3S problem always occurs and direct utilizi...
Article
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 (NI...
Article
This paper presents a new regularization technique to deal with the small sample size (S3) problem in linear discriminant analysis (LDA) based face recognition. Regularization on the within-class scatter matrix Sw has been shown to be a good direction for solving the S3 problem because the solution is found in full space instead of a subspace. The...
Article
Integration of various face recognition algorithms has proved to be a feasible approach to improve the performance of a face recognition system. Different face recognition algorithms are often based on different representations of the input patterns or on extracted features and hence may complement each other. Linear and nonlinear feature based alg...
Article
Structure distortion evaluation allows us to directly measure the similarity between signature patterns without classification using feature vectors, which usually suffers from limited training samples. In this paper, we incorporate the merits of both global and local alignment algorithms to define structure distortion using signature skeletons ide...
Article
To address two problems, namely nonlinear problem and singularity problem, of linear discriminant analysis (LDA) approach in face recognition, this paper proposes a novel kernel machine-based rank-lifting regularized discriminant analysis (KRLRDA) method. A rank-lifting theorem is first proven using linear algebraic theory. Combining the rank-lifti...
Article
Generalizations ofnonnegative matrix factorization (NMF) in kernel feature space, such as projected gradient kernel NMF (PGKNMF) and polynomial Kernel NMF (PNMF), have been developed for face and facial expression recognition recently. However, these existing kernel NMF approaches cannot guarantee the nonnegativity of bases in kernel feature space...
Article
The authors propose to extract local texture features for image-based coin recognition in this study. A set of Gabor wavelets and local binary pattern (LBP) operator are employed to represent texture information. Concentric ring structure is used to divide the coin image into a number of small sections. Statistics of Gabor coefficients or LBP value...
Article
This article addresses the noisy image segmentation problems based on wavelet transform and active contour model. In order to get better results, this article proposes a new segmentation and selective smoothing algorithm. First, a new adaptive segmentation model based on grey-level image segmentation model is proposed, and this model can also be ex...
Article
In order to address the shadow, shelter problems in image segmentation, a novel intensity-based model with shape prior constraint is proposed in this pa­ per. Two dimensional principle component analysis (2D-PCA) approach and pre-image strategy are ex­ ploited to extract the shape prior . The obtained shape prior term is then incorporated into the...
Conference Paper
Dimensionality reduction technologies are very important for pattern representation and recognition. Among them, locality preserving projection (LPP) is a manifold dimensionality reduction scheme and has been successfully applied to face recognition. However, LPP is an unsupervised linear approach, its performance will degrade for classification ta...