ORL Datasets Research Articles

GOLD Features with Machine Learning for Face Recognition

Face detection and recognition is a technology in which machines intelligently recognize a human from his face. Face recognition is widely used for many applications such as tracking attendance, fraud detection for passport, in banks, Facebook’s Deep face for automatic tagging of people in photos etc. The initial steps in face recognition is feature extraction. Most of the commonly used feature extraction techniques are SIFT, SURF, Linear Discriminant Analysis. In this paper we have experimented with the use of Gaussian of local descriptors (GOLD) for face recognition. An optimal size of 8384 GOLD descriptors are obtained for every image. Classification models such as Naive Bayes, K nearest neighbor, Regression trees and multi-class ECOC are used to train with the descriptors and test the same. The dataset is partitioned in 70:30 ratio for every subject for training and testing respectively. We have used the ORL, Faces-95 and Grimace dataset to test our experimentation. From this, it was found that K nearest neighbor outperforms the others with the recall and precision of 99%.

Discriminant Analysis with Local Gaussian Similarity Preserving for Feature Extraction

In this paper, we propose a novel discriminant analysis with local Gaussian similarity preserving (DA-LGSP) method for feature extraction. DA-LGSP can be viewed as a linear approximation of manifold learning based approach which seeks to find a linear projection that maximizes the between-class dissimilarities under the constraint of locality preserving. The local geometry of each point is preserved by the Gaussian coefficients of its neighbors, meanwhile the between-class dissimilarities are represented by Euclidean distances. Experiments are conducted on USPA data, COIL-20 dataset, ORL dataset and FERET dataset. The performance of the proposed method demonstrates that DA-LGSP is effective in feature extraction.

Face recognition via fast dense correspondence

Face recognition plays a significant role in computer vision. It is well know that facial images are complex stimuli signals that suffer from non-rigid deformations, including misalignment, orientation, pose changes, and variations of facial expression, etc. In order to address these variations, this paper introduces an improved sparse-representation based face recognition method, which constructs dense pixel correspondences between training and testing facial samples. Specifically, we first construct a deformable spatial pyramid graph model that simultaneously regularizes matching consistency at multiple spatial extents - ranging from an entire image, though coarse grid cells, to every single pixel. Secondly, a matching energy function is designed to perform face alignment based on dense pixel correspondence, which is very effective to address the issue of non-rigid deformations. Finally, a novel coarse-to-fine matching scheme is designed so that we are able to speed up the optimization of the matching energy function. After the training samples are aligned with respect to testing samples, an improved sparse representation model is employed to perform face recognition. The experimental results demonstrate the superiority of the proposed method over other methods on ORL, AR, and LFWCrop datasets. Especially, the proposed approach improves nearly 4.4 % in terms of recognition accuracy and runs nearly 10 times faster than previous sparse approximation methods.

Graph-Regularized Local Coordinate Concept Factorization for Image Representation

Existing matrix factorization based techniques, such as nonnegative matrix factorization and concept factorization, have been widely applied for data representation. In order to make the obtained concepts to be as close to the original data points as possible, one state-of-the-art method called locality constraint concept factorization is put forward, which represent the data by a linear combination of only a few nearby basis concepts. But its locality constraint does not well reveal the intrinsic data structure since it only requires the concept to be as close to the original data points as possible. To address these problems, by considering the manifold geometrical structure in local concept factorization via graph-based learning, we propose a novel algorithm, called graph-regularized local coordinate concept factorization (GRLCF). By constructing a parameter-free graph using constrained Laplacian rank (CLR) algorithm, we also present an extension of GRLCF algorithm as \(\hbox {GRLCF}_{\mathrm{CLR}}\). Moreover, we develop the iterative updating optimization schemes, and provide the convergence proof of our optimization scheme. Since GRLCF simultaneously considers the geometric structures of the data manifold and the locality conditions as additional constraints, it can obtain more compact and better structured data representation. Experimental results on ORL, Yale and Mnist image datasets demonstrate the effectiveness of our proposed algorithm.

On the effect of hyperedge weights on hypergraph learning

Hypergraph is a powerful representation for several computer vision, machine learning, and pattern recognition problems. In the last decade, many researchers have been keen to develop different hypergraph models. In contrast, no much attention has been paid to the design of hyperedge weighting schemes. However, many studies on pairwise graphs showed that the choice of edge weight can significantly influence the performances of such graph algorithms. We argue that this also applies to hypergraphs. In this paper, we empirically study the influence of hyperedge weights on hypergraph learning via proposing three novel hyperedge weighting schemes from the perspectives of geometry, multivariate statistical analysis, and linear regression. Extensive experiments on ORL, COIL20, JAFFE, Sheffield, Scene15 and Caltech256 datasets verified our hypothesis for both classification and clustering problems. For each of these classes of problems, our empirical study concludes with suggesting a suitable hypergraph weighting scheme. Moreover, the experiments also demonstrate that the combinations of such weighting schemes and conventional hypergraph models can achieve competitive classification and clustering performances in comparison with some recent state-of-the-art algorithms.

Elastic Bunch Graph Matching Based Face Recognition Under Varying Lighting, Pose, and Expression Conditions

In this paper performance of elastic bunch graph matching (EBGM) for face recognition under variation in facial expression, variation in lighting condition and variation in poses are given. In this approach faces are represented by labelled graphs. Experimental results of EBGM on ORL, Yale B and FERET datasets are provided. Strong and weak features of EBGM algorithm are discussed.

Frequency Domain Approach for Face Recognition Using Optical Vanderlugt Filters

In this paper human face machine identification is experienced using optical correlation techniques in spatial frequency domain. This approach is tested on ORL dataset of faces which includes face images of 40 subjects, each in 10 different positions. The examined optical setup relies on optical correlation based on developing optical Vanderlugt filters and its basics are described in this article. With the limitation of face database of 40 persons, the recognition is examined successfully with nearly 100% of accuracy in matching the input images with their respective Vanderlugt synthesized filters. Software simulation is implemented by using MATLAB for face identification.

TBSTM: A Novel and Fast Nonlinear Classification Method for Image Data

A new classifier for image data classification named as linear twin bounded support tensor machine (linear TBSTM) is proposed by adding regularization terms in objective functions, which results in the realization of structural risk minimization avoids of the singularity of matrices. We know that up to now nonlinear classifiers based on STM for image data classification are not seen more. In order to remedy this limitation, a new matrix kernel function is introduced and based on which the nonlinear version of TBSTM is studied with a detailed theoretical derivation, and then a nonlinear classifier called as nonlinear TBSTM is suggested. In order to examine the effectiveness of the proposed classifiers, a series of comparative experiments with three linear classifiers STM, TSTM and PSTM are performed on 15 binary image classification problems taken from ORL, YALE and AR datasets. Experiment results show that the proposed classifiers are effective and efficient.

Multi-pose face ensemble classification aided by Gabor features and deep belief nets

A face image with pose variations may be encoded with different representations, which may severely degrade classification performance. In this paper, we consider the problem of classification in the multi-pose face setting using 2D-Gabor features with the deep belief nets approach. We cast the classification as one form of deep learning problem where our goal is to construct the Gabor feature maps in nonlinear characterization. By extracting the 2D-Gabor features of multi-pose faces, then combines Gabor features and the advantages of LTP features; and using of local spatial histogram to describe the face image; then taking X-means algorithm for data processing to further improve the mapping space and dimensionality reduction which can enhances the difference between the data sub-sets. In this way, to learn the complex data space will be automatically divide into multiple sample subspace. We improve the features learning with more discriminating power to benefit the classification problems. We adopt the combination neighborhood component analysis method to mapping this data in a better way. Linear variation of training samples, which can find a more favorable category of linear subspace. We formulate this problem using Gabor feature maps as the input data in deep belief nets. In addition, experimental results on ORL, Yale and LFW datasets show that our proposed algorithm can have better discriminating power and significantly enhance the classification performance, which shows that our algorithm is almost robust to the training samples both on ORL and Yale datasets. Comparisons with eight algorithms (PCA, 2DPCA, Gabor+2DPCA, SIFT, LBP, LTP, LGBP and LGTP) show that the proposed method is better in recognizing multi-pose faces without large volumes of data. And the experimental results on image classification have verified the effectiveness of the proposed approach.

Unsupervised discriminant canonical correlation analysis based on spectral clustering

Canonical correlation analysis (CCA) has been widely applied to information fusion. However, it only considers the correlated information between the paired data and ignores the correlated information between the samples in the same class. Furthermore, class information is helpful for CCA to extract the discriminant feature, but there is no class information available in application of clustering. Thus, it is difficult to utilize the correlated information between the samples in the same class. In order to utilize this correlated information, we propose a method named Unsupervised Discriminant Canonical Correlation Analysis based on Spectral Clustering (UDCCASC). Class membership of the samples is calculated using the normalized spectral clustering, while the mappings for feature fusion are computed by using the generalized eigenvalue method. These two algorithms are executed alternately before the desired result is obtained. Two extensions of UDCCASC are proposed also to deal with multi-view data and nonlinear data. The experimental results on MFD dataset, ORL dataset, MSRC-v1 dataset show that our methods outperform traditional CCA and part of state-of-art methods for feature fusion.

Locally principal component analysis based on L1‐norm maximisation

Locally principal component analysis (LPCA) is a popular method of dimensionality reduction, which takes locality of data points into account. In this study, by using the L1-norm instead of the L2-norm in LPCA, the authors introduce a new formulation of LPCA based on the L1-norm maximisation, referred to as LPCA-L1. Compared with the conventional L2-norm LPCA, the proposed LPCA-L1 approach is more robust to outliers. Experiments of classification and recognition on the UCI, Yale and ORL data sets confirm the effectiveness of the proposed method.

Extended semi-supervised fuzzy learning method for nonlinear outliers via pattern discovery

This article presents an extended Parameterized Fuzzy Semi-supervised learning (PFSL) method, in which the key innovation is the capability of separating a sample set into two independent subsets: outlier sample subset and regular sample subset. In our proposed PFSL, we first develop an improved parameterized Fuzzy Linear Discriminant Analysis (F-LDA) algorithm to classify regular samples, in which the distribution information of each sample in terms of fuzzy membership degree is incorporated with the redefined within-class and between-class scatter matrices. To achieve good parameter estimation for this improved F-LDA, we advocate the use of Hopfield Neural Networks (HNN) due to its efficiency. Second, a new semi-supervised Fuzzy C-Means (S-FCM) algorithm is designed using pre-computed cluster number and cluster centers in the supervised pattern discovery stage. It is applied to classify the remaining outlier samples and generate the final classification result. Third, since Kernel Fisher Discriminant (KFD) is an efficient way to extract nonlinear discriminant features, a kernel version of the proposed PFSL (K-PFSL) is discussed. Extensive experiments on the ORL, NUST603, FERET and Yale face datasets show the effectiveness and the superiority of the proposed algorithm.

Affine Normalized Krawtchouk Moments Based Face Recognition

Abstract In this paper, we have developed a new local descriptor based on Krawtchouk polynomial moments. The interest points are initially detected using the Canny edge detector and made the region around each interest point scale and affine normalized. The region is then represented using Krawtchouk polynomial and hence formed the descriptor. Experiments have been conducted keeping the face recognition problem in focus. By using the sparse representation concept, classification of face images is done. Experimental results on the ORL dataset and a subset of pose and illumination variant FERET dataset have shown the classification capability of our descriptor for face recognition applications.

Improving Shape Retrieval by Integrating AIR and Modified MutualkNN Graph

In computer vision, image retrieval remained a significant problem and recent resurgent of image retrieval also relies on other postprocessing methods to improve the accuracy instead of solely relying on good feature representation. Our method addressed the shape retrieval of binary images. This paper proposes a new integration scheme to best utilize feature representation along with contextual information. For feature representation we used articulation invariant representation; dynamic programming is then utilized for better shape matching followed by manifold learning based postprocessing modified mutualkNN graph to further improve the similarity score. We conducted extensive experiments on widely used MPEG-7 database of shape images by so-called bulls-eye score with and without normalization of modified mutualkNN graph which clearly indicates the importance of normalization. Finally, our method demonstrated better results compared to other methods. We also computed the computational time with another graph transduction method which clearly shows that our method is computationally very fast. Furthermore, to show consistency of postprocessing method, we also performed experiments on challenging ORL and YALE face datasets and improved baseline results.

Performance Analysis of Pose Invariant Face Recognition Approaches in Unconstrained Environments

Face recognition and authentication are two significant and dynamic research issues in computer vision applications. There are many factors that should be accounted for face recognition; among them pose variation is a major challenge which severely influence in the performance of face recognition. In order to improve the performance, several research methods have been developed to perform the face recognition process with pose invariant conditions in constrained and unconstrained environments. In this paper, the authors analyzed the performance of a popular texture descriptors viz., Local Binary Pattern, Local Derivative Pattern and Histograms of Oriented Gradients for pose invariant problem. State of the art preprocessing techniques such as Discrete Cosine Transform, Difference of Gaussian, Multi Scale Retinex and Gradient face have also been applied before feature extraction. In the recognition phase K- nearest neighbor classifier is used to accomplish the classification task. To evaluate the efficiency of pose invariant face recognition algorithm three publicly available databases viz. UMIST, ORL and LFW datasets have been used. The above said databases have very wide pose variations and it is proved that the state of the art method is efficient only in constrained situations.

Manifold Adaptive Label Propagation for Face Clustering.

In this paper, a novel label propagation (LP) method is presented, called the manifold adaptive label propagation (MALP) method, which is to extend original LP by integrating sparse representation constraint into regularization framework of LP method. Similar to most LP, first of all, MALP also finds graph edges from given data and gives weights to the graph edges. Our goal is to find graph weights matrix adaptively. The key advantage of our approach is that MALP simultaneously finds graph weights matrix and predicts the label of unlabeled data. This paper also derives efficient algorithm to solve the proposed problem. Extensions of our MALP in kernel space and robust version are presented. The proposed method has been applied to the problem of semi-supervised face clustering using the well-known ORL, Yale, extended YaleB, and PIE datasets. Our experimental evaluations show the effectiveness of our method.

Practical emotional neural networks

In this paper, we propose a limbic-based artificial emotional neural network (LiAENN) for a pattern recognition problem. LiAENN is a novel computational neural model of the emotional brain that models emotional situations such as anxiety and confidence in the learning process, the short paths, the forgetting processes, and inhibitory mechanisms of the emotional brain. In the model, the learning weights are adjusted by the proposed anxious confident decayed brain emotional learning rules (ACDBEL). In engineering applications, LiAENN is utilized in facial detection, and emotion recognition. According to the comparative results on ORL and Yale datasets, LiAENN shows a higher accuracy than other applied emotional networks such as brain emotional learning (BEL) and emotional back propagation (EmBP) based networks.

An Improved Face Recognition Technique Based on Modular Multi-directional Two-dimensional Principle Component Analysis Approach

In this paper, a new method named modular multi-directional two-dimensional principle component analysis (M 2 D2DPCA) is proposed for face recognition. First, the original images are rotated at some predetermined angles so that we may extract features from the images in any direction. Then we divide the rotated images into smaller sub-images and apply 2DPCA approach to each of these sub-images. Finally we propose a fusion method named modular multi-directional 2DPCA (M 2 D2DPCA) to combine a bank of preliminary results in different directions. Compared with conventional 2DPCA based algorithms, the advantage of the proposed method is that it can extract significant features from the images in any direction and avoid the effects of varying illumination and facial expression. The results of the experiments on ORL and Yale datasets show that the proposed M 2 D2DPCA method can obtain a higher recognition rate than the conventional 2DPCA based methods.

DECISION TREE BASED OCCLUSION DETECTION IN FACE RECOGNITION AND ESTIMATION OF HUMAN AGE USING BACK PROPAGATION NEURAL NETWORK

Occlusion detection in face verification is an essential problem that has not widely addressed. In this study the research is deals about occlusion detection in face recognition and estimation of human age using image processing. The objects hide from another object is called as occlusion. Occlusion conditions may vary from face wearing sunglasses, wearing of scarf in the eyes and mouth positions. The proposed work consists four stages. Initial stage is to extract the features using canny edge detection technique and to classify the occluded and non occluded region using Decision Tree Based Occlusion Detection (DTOD) classifier. Secondly the face verification and recognition is carried out using Elastic Matching Pattern (EMP) and Maximum Likelihood Classifier (MLC). Back Propagation Neural Network (BPNN) can be used to estimate the age of the human in the third stage. Our experiments are conducted on the database images for the first stage. By considering the first stage the various performance measures of the classifiers are analyzed. The correctly classified instances rate are high compared with the existing classifiers like random forest and bayes classifier. Experiments are conducted using ORL dataset for the second and the third stage. On the basis of the results obtained from the second stage we observed that the face verification was completed with 95% of accuracy. In the third stage, the age estimation using BPNN algorithm shows better performance results compared with the existing neural network algorithm.

Customized orthogonalization via deflation algorithm with applications in face recognition

The face recognition problem is a topical issue in computer vision. In this paper we propose a customized version of the orthogonalization via deflation algorithm to tackle this problem. We test the new proposed algorithm on two datasets: the well-known ORL dataset and an own face dataset, CTOVF; also, we compare our results (in terms of rate recognition and average quiery time) with the outcome of a standard algorithm in this class (dimension reduction methods using numerical linear algebra tools).