Zhaowei Shang

Face Recognition Under Varying Illumination Using Gradientfaces

In this correspondence, we propose a novel method to extract illumination insensitive features for face recognition under varying lighting called the gradient faces. Theoretical analysis shows gradient faces is an illumination insensitive measure, and robust to different illumination, including uncontrolled, natural lighting. In addition, gradient faces is derived from the image gradient domain such that it can discover underlying inherent structure of face images since the gradient domain explicitly considers the relationships between neighboring pixel points. Therefore, gradient faces has more discriminating power than the illumination insensitive measure extracted from the pixel domain. Recognition rates of 99.83% achieved on PIE database of 68 subjects, 98.96% achieved on Yale B of ten subjects, and 95.61% achieved on Outdoor database of 132 subjects under uncontrolled natural lighting conditions show that gradient faces is an effective method for face recognition under varying illumination. Furthermore, the experimental results on Yale database validate that gradient faces is also insensitive to image noise and object artifacts (such as facial expressions).

Multiscale facial structure representation for face recognition under varying illumination

Facial structure of face image under lighting lies in multiscale space. In order to detect and eliminate illumination effect, a wavelet-based face recognition method is proposed in this paper. In this work, the effect of illuminations is effectively reduced by wavelet-based denoising techniques, and meanwhile the multiscale facial structure is generated. Among others, the proposed method has the following advantages: (1) it can be directly applied to single face image, without any prior information of 3D shape or light sources, nor many training samples; (2) due to the multiscale nature of wavelet transform, it has better edge-preserving ability in low frequency illumination fields; and (3) the parameter selection process is computationally feasible and fast. Experiments are carried out upon the Yale B and CMU PIE face databases, and the results demonstrate that the proposed method achieves satisfactory recognition rates under varying illumination conditions.

Generalized Discriminant Analysis: A Matrix Exponential Approach

Linear discriminant analysis (LDA) is well known as a powerful tool for discriminant analysis. In the case of a small training data set, however, it cannot directly be applied to high-dimensional data. This case is the so-called small-sample-size or undersampled problem. In this paper, we propose an exponential discriminant analysis (EDA) technique to overcome the undersampled problem. The advantages of EDA are that, compared with principal component analysis (PCA) + LDA, the EDA method can extract the most discriminant information that was contained in the null space of a within-class scatter matrix, and compared with another LDA extension, i.e., null-space LDA (NLDA), the discriminant information that was contained in the non-null space of the within-class scatter matrix is not discarded. Furthermore, EDA is equivalent to transforming original data into a new space by distance diffusion mapping, and then, LDA is applied in such a new space. As a result of diffusion mapping, the margin between different classes is enlarged, which is helpful in improving classification accuracy. Comparisons of experimental results on different data sets are given with respect to existing LDA extensions, including PCA + LDA, LDA via generalized singular value decomposition, regularized LDA, NLDA, and LDA via QR decomposition, which demonstrate the effectiveness of the proposed EDA method.

Negative samples reduction in cross-company software defects prediction

Abstract Context Software defect prediction has been widely studied based on various machine-learning algorithms. Previous studies usually focus on within-company defects prediction (WCDP), but lack of training data in the early stages of software testing limits the efficiency of WCDP in practice. Thus, recent research has largely examined the cross-company defects prediction (CCDP) as an alternative solution. Objective However, the gap of different distributions between cross-company (CC) data and within-company (WC) data usually makes it difficult to build a high-quality CCDP model. In this paper, a novel algorithm named Double Transfer Boosting (DTB) is introduced to narrow this gap and improve the performance of CCDP by reducing negative samples in CC data. Method The proposed DTB model integrates two levels of data transfer: first, the data gravitation method reshapes the whole distribution of CC data to fit WC data. Second, the transfer boosting method employs a small ratio of labeled WC data to eliminate negative instances in CC data. Results The empirical evaluation was conducted based on 15 publicly available datasets. CCDP experiment results indicated that the proposed model achieved better overall performance than compared CCDP models. DTB was also compared to WCDP in two different situations. Statistical analysis suggested that DTB performed significantly better than WCDP models trained by limited samples and produced comparable results to WCDP with sufficient training data. Conclusions DTB reforms the distribution of CC data from different levels to improve the performance of CCDP, and experimental results and analysis demonstrate that it could be an effective model for early software defects detection.

A Visual-Attention Model Using Earth Mover's Distance-Based Saliency Measurement and Nonlinear Feature Combination

This paper introduces a new computational visual-attention model for static and dynamic saliency maps. First, we use the Earth Movers Distance (EMD) to measure the center-surround difference in the receptive field, instead of using the Difference-of-Gaussian filter that is widely used in many previous visual-attention models. Second, we propose to take two steps of biologically inspired nonlinear operations for combining different features: combining subsets of basic features into a set of super features using the Lm-norm and then combining the super features using the Winner-Take-All mechanism. Third, we extend the proposed model to construct dynamic saliency maps from videos by using EMD for computing the center-surround difference in the spatiotemporal receptive field. We evaluate the performance of the proposed model on both static image data and video data. Comparison results show that the proposed model outperforms several existing models under a unified evaluation setting.

A self-adaptive image normalization and quaternion PCA based color image watermarking algorithm

This paper proposes a novel robust digital color image watermarking algorithm which combines color image feature point extraction, shape image normalization and QPCA (quaternion principal component algorithm) based watermarking embedding (QWEMS) and extraction (QWEXS) schemes. The feature point extraction method called Mexican Hat wavelet scale interaction is used to select the points which can survive various attacks and also be used as reference points for both watermarking embedding and extraction. The normalization shape image of the local quadrangle image of which the four corners are feature points of the original image is invariant to translation, rotation, scaling and skew, by which we can obtain the relationship between the feature images of the original image and the watermarked image which has suffered with geometrical attacks. The proposed QWEMS and QWEXS schemes which denote the color pixel as a pure quaternion and the feature image as a quaternion matrix can improve the robustness and the imperceptibility of the embedding watermarking. To simplify the eigen-decomposition procedure of the quaternion matrix, we develop a calculation approach with which the eigen-values and the corresponding eigen-vectors of the quaternion matrix can be computed. A binary watermark image is embedded in the principal component coefficients of the feature image. Simulation results demonstrate that the proposed algorithm can survive a variety of geometry attacks, i.e. translation, rotation, scaling and skew, and can also resist the attacks of many signal processing procedures, for example, moderate JPEG compression, salt and pepper noise, Gaussian filtering, median filtering, and so on.

Learning the Distribution Preserving Semantic Subspace for Clustering

This paper proposes a new clustering method for images called distribution preserving indexing (DPI). It aims to find a lower dimensional semantic space approximating the original image space in the sense of preserving the distribution of the data. In the theory, the intrinsic structure of the data clusters can be described by the distribution of the data effectively. Therefore, the cluster structure of the data in a lower dimensional semantic space derived by the DPI becomes clear. Unlike these distance-based clustering methods, which reveal the intrinsic Euclidean structure of data, our method attempts to discover the intrinsic cluster structure of the data space that actually is the union of some sub-manifolds. Moreover, we propose a revised kernel density estimator for the case of high-dimensional data, which is a crucial step in DPI. In addition, we provide a theoretical analysis of the bound of our method. Finally, the extensive experiments compared with other algorithms, on COIL20, CBCL, and MNIST demonstrate the effectiveness of our proposed approach.This paper proposes a new clustering method for images called distribution preserving indexing (DPI). It aims to find a lower dimensional semantic space approximating the original image space in the sense of preserving the distribution of the data. In the theory, the intrinsic structure of the data clusters can be described by the distribution of the data effectively. Therefore, the cluster structure of the data in a lower dimensional semantic space derived by the DPI becomes clear. Unlike these distance-based clustering methods, which reveal the intrinsic Euclidean structure of data, our method attempts to discover the intrinsic cluster structure of the data space that actually is the union of some sub-manifolds. Moreover, we propose a revised kernel density estimator for the case of high-dimensional data, which is a crucial step in DPI. In addition, we provide a theoretical analysis of the bound of our method. Finally, the extensive experiments compared with other algorithms, on COIL20, CBCL, and MNIST demonstrate the effectiveness of our proposed approach.

Sift features based object tracking with discrete wavelet transform

A novel first-detect-then-identify approach with SIFT features and discrete wavelet transform for tracking object is proposed in real surveillance scenarios. For accurate and fast moving object detection, discrete wavelet transform is adopted to eliminate the noises of the frames which may cause detection errors, and then objects are detected by applying the inter-frame difference method on the low frequency parts of two consecutive frames, and then SIFT feature is used for object representation and identification due to its invariant properties. Experimental results demonstrate that the proposed strategy improves the tracking performance by comparing with the classical mean shift method, and it is also shown that the proposed algorithm can be also applied in multiple objects tracking in real scenarios.

LOCALITY PRESERVING NONNEGATIVE MATRIX FACTORIZATION WITH APPLICATION TO FACE RECOGNITION

In this paper, we propose a Locality Preserving Nonnegative Matrix Factorization (LPNMF) method to discover the manifold structure embedded in high-dimensional face space that is applied for face recognition. It is done by incorporating locality preserving constraints inside the cost function of NMF, then a new decomposition of a face with locality preserving can be obtained. As a result, the proposed LPNMF method shares some properties with the Locality Preserving Projection (LPP) such that it can effectively discover the manifold structure embedded in a high-dimensional face space. Experimental results show that LPNMF provides a better representation and achieves higher recognition rates in face recognition.

Multi-Label Learning With Fuzzy Hypergraph Regularization for Protein Subcellular Location Prediction

Protein subcellular location prediction aims to predict the location where a protein resides within a cell using computational methods. Considering the main limitations of the existing methods, we propose a hierarchical multi-label learning model FHML for both single-location proteins and multi-location proteins. The latent concepts are extracted through feature space decomposition and label space decomposition under the nonnegative data factorization framework. The extracted latent concepts are used as the codebook to indirectly connect the protein features to their annotations. We construct dual fuzzy hypergraphs to capture the intrinsic high-order relations embedded in not only feature space, but also label space. Finally, the subcellular location annotation information is propagated from the labeled proteins to the unlabeled proteins by performing dual fuzzy hypergraph Laplacian regularization. The experimental results on the six protein benchmark datasets demonstrate the superiority of our proposed method by comparing it with the state-of-the-art methods, and illustrate the benefit of exploiting both feature correlations and label correlations.