Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Conclusions
References

Pattern Analysis and Applications

Profile image of Roger Lambert IIIRoger Lambert III

Abstract

Feature extraction from images, which are typical of high dimensionality, is crucial to the recognition performance. To explore the discriminative information while depressing the intra-class variations due to variable illumination and view conditions, we propose a factor analysis framework for separate "content" from "style", identifying a familiar face seen under unfamiliar viewing conditions, classifying familiar poses presented in an unfamiliar face, estimating age across unfamiliar faces. The framework applies efficient algorithms derived from objective factor separating functions and space mapping functions, which can produce sufficiently expressive representations of feature extraction and dimensionality reduction. We report promising results on three different tasks in the high-dimensional image perceptual domains: face identification with two benchmark face databases, facial pose classification with a benchmark facial pose database, extrapolation of age to unseen facial image. Experimental results show that our approach produced higher classification performance when compared to classical LDA, WLDA, LPP, MFA and DLA algorithms.

Figures (10)
In tl his paper, we present a general framewo linearization, kermelization, and tensorization, which explaining many of the popular dimensionality red above. T simultaneously projection. Imp constructs begin a partitio factor an algorithm alysis n and weight ma minimizi ning with framewo ng style fac ementing the facto he main objective of factor analysis is to fi rk as a pla tform, we deve tor while maximizin r analysis framework requ the formulation of a factor separating fu trix. The second construct produces the space mapping f oped three novel dimensi s for facial identification, facial age grouping, and facial pose classificati offers a uni uction algori nd the essen g conte rk called fac i iW ti n i ires two essen tor analysis, together with its ied view for understanding and hms, such as those mention al low-dimensional features t factor of the object throu tial prelimin N des construc nction. Using onality reduc on. ction and inclu U
In tl his paper, we present a general framewo linearization, kermelization, and tensorization, which explaining many of the popular dimensionality red above. T simultaneously projection. Imp constructs begin a partitio factor an algorithm alysis n and weight ma minimizi ning with framewo ng style fac ementing the facto he main objective of factor analysis is to fi rk as a pla tform, we deve tor while maximizin r analysis framework requ the formulation of a factor separating fu trix. The second construct produces the space mapping f oped three novel dimensi s for facial identification, facial age grouping, and facial pose classificati offers a uni uction algori nd the essen g conte rk called fac i iW ti n i ires two essen tor analysis, together with its ied view for understanding and hms, such as those mention al low-dimensional features t factor of the object throu tial prelimin N des construc nction. Using onality reduc on. ction and inclu U
at most two age groups, therefore we elaborated the age membership functions as follows: that the age ambiguity always occurs at two adjacent age group’s boundary, each age can belong to
at most two age groups, therefore we elaborated the age membership functions as follows: that the age ambiguity always occurs at two adjacent age group’s boundary, each age can belong to
produced by our algorithm. The LDA and LPP algorith ms only utilize the class or nearest neighborhood for partition, while MFA, DLA and our algorithms consider not only the nearest neighborhood but also the class information. Thus, the classi our algorithms are higher than LDA and LPP algorithms. taken into account in WLDA and our algorithms, but is not and our algorithms produce a classification accuracy that is h fication accuracies of M1 Furthermore, the weighti FA, DLA and ng scheme is considered in the LDA .Thus, WLDA igher than LDA.
produced by our algorithm. The LDA and LPP algorith ms only utilize the class or nearest neighborhood for partition, while MFA, DLA and our algorithms consider not only the nearest neighborhood but also the class information. Thus, the classi our algorithms are higher than LDA and LPP algorithms. taken into account in WLDA and our algorithms, but is not and our algorithms produce a classification accuracy that is h fication accuracies of M1 Furthermore, the weighti FA, DLA and ng scheme is considered in the LDA .Thus, WLDA igher than LDA.
Table3 Comparisons of the performance of different approaches on PIE pose database 4.3 facial age grouping experiment the best performance. In the experiments conducted in this study, the images for age group classification are collected from multiple data sources like 1002 facial images from FG-NET database [26], 500 images from Google database and 600 images from the scanned photographs leading to a total of 2102 sample facial images. A few of them are shown in Fig 5. FG-NET consists of 1,002 images of 82
Table3 Comparisons of the performance of different approaches on PIE pose database 4.3 facial age grouping experiment the best performance. In the experiments conducted in this study, the images for age group classification are collected from multiple data sources like 1002 facial images from FG-NET database [26], 500 images from Google database and 600 images from the scanned photographs leading to a total of 2102 sample facial images. A few of them are shown in Fig 5. FG-NET consists of 1,002 images of 82
7 Table4 Comparisons of the performance of different approaches on facial age database
7 Table4 Comparisons of the performance of different approaches on facial age database

Related papers

New Feature Extraction Approaches for Face Recognition
Vo Nhat

Lecture Notes in Computer Science, 2005

All the traditional PCA-based and LDA-based methods are based on the analysis of vectors. So, it is difficult to evaluate the covariance matrices in such a high-dimensional vector space. Recently, two-dimensional PCA (2DPCA) and two-dimensional LDA (2DLDA) have been proposed in which image covariance matrices can be constructed directly using original image matrices. In contrast to the covariance matrices of traditional 1D approaches (PCA and LDA), the size of the image covariance matrices using 2D approaches (2DPCA and 2DLDA) are much smaller. As a result, it is easier to evaluate the covariance matrices accurately and computation cost is reduced. However, a drawback of 2D approaches is that it needs more coefficients than traditional approaches for image representation. Thus, 2D approach needs more memory to store its features and costs more time to calculate distance (similarity) in classification phase. In this paper, we develop a new image feature extraction methods called two-stage 2D subspace approaches to overcome the disadvantage of 2DPCA and 2DLDA. The initial idea of twostage 2D subspace approaches which consist of two-stage 2DPCA and twostage 2DLDA is to perform 2DPCA or 2DLDA twice: the first one is in horizontal direction and the second is in vertical direction. After the two sequential 2D transforms, the discriminant information is compacted into the up-left corner of the image. Experiment results show our methods achieve better performance in comparison with the other approaches with the lower computation cost.

View PDFchevron_right
A Novel Efficient Discriminant Component Analysis Algorithm for Face Recognition
M.K. Jeyakumar Jeyakumar

2013

Face recognition is the process of categorizing a p erson in an image by evaluating with a known face i mage library. The pose and illumination variations are t wo main practical confronts for an automatic face recognition system. This paper proposes a novel fac e re ognition algorithm known as EDCA for face recognition under varying poses and illumination co ditions. The main aim of this paper is to reduce t h feature vector size. The intensity of a face image is normalized to get an illumination normalized ima ge. This EDCA algorithm overcomes the high dimensionali ty problem in the feature space by extracting features from the low dimensional frequency band of the image. It combines the features of both LDA an d PCA algorithms. The experiments were performed on b oth the Extended Yale B datasets. The experimental results show that the proposed algorithm produces a higher recognition rate than the existing LDA and PCA based face recognition techniques.

View PDFchevron_right
Data dimensionality reduction for face recognition
Kukharev Georgy

In the process of image recognition in most of the applications there is a problem of gathering, processing and storing large amounts of data. A possible solution for reducing this amount and speeding-up the computations is to use some sort of data reduction. Efficient reduction of stored data without losing any important part requires an adaptive method, which works without any supervision. In this article we discuss a few variants of a two-step approach which involves Karhunen-Loeve Transform (KLT) and Linear Discriminant Analysis (LDA). The KLT gives a good approximation of input data,however it requires a large number of eigenvalues. The second step reduces data dimensionality further using LDA. The efficiency of KLT depends on quality and quantity of input data. In the case, when only one image in a class is given as input, its features are not stable in comparison with other images in other classes. In the article we present a few methods of solving this problem, which is an improvement of the ideas presented in .

View PDFchevron_right
Feature extraction based face recognition, gender and age classification
L M Patnaik

2010

Abstract-The face recognition system with large sets of training sets for personal identification normally attains good accuracy. In this paper, we proposed Feature Extraction based Face Recognition, Gender and Age Classification (FEBFRGAC) algorithm with only ...

View PDFchevron_right
An MPCA/LDA Based Dimensionality Reduction Algorithm for Face Recognition
Jamal El-den

Mathematical Problems in Engineering, 2014

We proposed a face recognition algorithm based on both the multilinear principal component analysis (MPCA) and linear discriminant analysis (LDA). Compared with current traditional existing face recognition methods, our approach treats face images as multidimensional tensor in order to find the optimal tensor subspace for accomplishing dimension reduction. The LDA is used to project samples to a new discriminant feature space, while theKnearest neighbor (KNN) is adopted for sample set classification. The results of our study and the developed algorithm are validated with face databases ORL, FERET, and YALE and compared with PCA, MPCA, and PCA + LDA methods, which demonstrates an improvement in face recognition accuracy.

View PDFchevron_right
A Direct LDA Algorithm for High-Dimensional Data-With Application to Face Recognition
SATISH CHANDRA PULIPAKA

Pattern Recognition, 2001

View PDFchevron_right
Factored principal components analysis, with applications to face recognition
Linlin Shen

Statistics and Computing, 2009

A dimension reduction technique is proposed for matrix data, with applications to face recognition from images. In particular, we propose a factored covariance model for the data under study, estimate the parameters using maximum likelihood, and then carry out eigendecompositions of the estimated covariance matrix. We call the resulting method factored principal components analysis. We also develop a method for classification using a likelihood ratio criterion, which has previously been used for evaluating the strength of forensic evidence. The methodology is illustrated with applications in face recognition.

View PDFchevron_right
Face Recognition in Machine Learning: A Framework for Dimensionality Reduction Algorithms
Professor Gabriel Kabanda

International Journal of Advanced Networking and Applications, 2022

A facial recognition system matches a human face from a digital image or a video frame against an authentic repository of faces or Eigenfaces subject to algorithmic performance and detection accuracy. Dimensionality reduction is a type of unsupervised learning for which input is images of higher-dimensional data and these images are represented with a lower-dimensional space. The purpose of the research paper is to evaluate the performance of Dimensionality Reduction algorithms for face recognition using different approaches of Machine Learning (ML). The research uses the Interpretivist Paradigm characterised by a subjectivist epistemology, relativist ontology, naturalist methodology, and a balanced axiology. The quantitative methodology with an experimental research design was used. The results of the experiment show that only selecting the top M eigenfaces reduces the dimensionality of the data, and that too few eigenfaces results in too much information loss, and hence less discrimination between faces. With increasing dimensionality, the amount of training instances needed rises exponentially (i.e., kd). The performance of the Dimensionality Reduction Algorithm is benchmarked against the Clustering, Bayesian, Genetic, Reinforcement Q-Learning and Reinforcement A3C Algorithms. The outcome of the research makes significant value-adding contributions to the future of advances in Big Data Analytics and ML.

View PDFchevron_right
Discriminant Feature Manifold for Facial Aging Estimation
Phil Grant

2010 20th International Conference on Pattern Recognition, 2010

Computerised facial aging estimation, which has the potential for many applications in human-computer interactions, has been investigated by many computer vision researchers in recent years. In this paper, a feature-based discriminant subspace is proposed to extract more discriminating and robust representations for aging estimation. After aligning all the faces by a piece-wise affine transform, orthogonal locality preserving projection (OLPP) is employed to project local binary patterns (LBP) from the faces into an age-discriminant subspace. The feature extracted from this manifold is more distinctive for age estimation compared with the features using in the state-of-the-art methods. Based on the public database FG-NET, the performance of the proposed feature is evaluated by using two different regression techniques, quadratic function and neural-network regression. The proposed feature subspace achieves the best performance based on both types of regression.

View PDFchevron_right
Face recognition using principle components and linear discriminant analysis
Mona Morsi

2009

Face recognition has recently received significant attention as one of the challenging and promising fields of computer vision and pattern recognition. It plays a significant role in many security and forensic applications such as person authentication in access control systems and person identification in real time video surveillance systems. This paper studies two appearance-based approaches for feature extraction and dimension reduction, namely, Principal Components Analysis (PCA) and Linear Discriminant Analysis (LDA). Numerical experiments were carried out on the ORL face database and many parameters were investigated, this included the effect of changing the number of training images, scaling factor, and the effect of feature vector length on the recognition rate. Classification is performed using the minimum Euclidean distance. The results suggest that the effect of increasing the number of training images has more significance on the recognition rate than changing the image ...

View PDFchevron_right

References (27)

  1. Qian Zhang, Lefei Zhang and Yiping Yang, et al. Local Patch Discriminative Metric learning for Hyperspectral Image Feature Extraction. IEEE Geoscience and Remote Sensing Letters, Vol. 11, No. 3, pp: 612-616 (2014)
  2. A.M. Martinez and A.C. Kak, "PCA versus LDA," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 23, no. 2, pp. 228-233, Feb (2001)
  3. J. Ye, R. Janardan, C. Park, and H. Park, "An Optimization Criterion for Generalized Discriminant Analysis on Undersampled Problems," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 8, pp. 982-994, Aug (2004)
  4. X. He, S. Yan, Y. Hu, P. Niyogi, and H. Zhang, "Face Recognition Using Laplacianfaces," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 27, no. 3, pp. 328-340, Mar (2005)
  5. J. Tenenbaum, V. Silva, and J. Langford, "A Global Geometric Framework for Nonlinear Dimensionality Reduction," Science, vol. 290, no. 22, pp. 2319-2323, Dec (2000)
  6. S. Roweis and L. Saul, "Nonlinear Dimensionality Reduction by Locally Linear Embedding," Science, vol. 290, no. 22, pp. 2323-2326, Dec(2000)
  7. M. Belkin and P. Niyogi, "Laplacian Eigenmaps and Spectral Techniques for Embedding and Clustering," Advances in Neural Information Processing System, vol. 14, pp. 585-591 (2001)
  8. K. Muller, S. Mika, G. Riitsch, K. Tsuda, and B. Scho¨ lkopf, "An Introduction to Kernel-Based Learning Algorithms," IEEE Trans. Neural Networks, vol. 12, pp. 181-201(2001)
  9. S. Yan, D. Xu, Q. Yang, L. Zhang, X. Tang, and H. Zhang,"Discriminant Analysis with Tensor Representation," Proc. Int'l Conf. Computer Vision and Pattern Recognition, vol. 1, pp. 526-532 (2005)
  10. J. Yang, D. Zhang, A. Frangi, and J. Yang, "Two-Dimensional PCA: A New Approach to Appearance-Based Face Representation and Recognition," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 1, pp. 131-137, Jan (2004)
  11. J. Ye, "Generalized Low Rank Approximations of Matrices," Proc. Int'l Conf. Machine Learning, pp. 895-902 (2004)
  12. J. Ye, R. Janardan, and Q. Li, "Two-Dimensional Linear Discriminant Analysis," Advances in Neural Information Processing Systems, vol. 17, pp. 1569-1576 (2005)
  13. Victor Popa, Jani Nurminen, Moncef Gabbouj. A Study of Bilinear Models in Voice Conversion. Journal of Signal and Information Processing, vol. 2.No. 2, pp: 125-139. (2011)
  14. Huachun Tan, Bin Cheng, Jianshuai Feng and et al. Tensor Recovery via Multi-linear Augmented Lagrange Multiplier Method, Sixth International Conference on Image and Graphics (ICIG), pp:141-146 (2011)
  15. Y Li, Y Gao, H Erdogan. Weighted Pairwise Scatter to Improve Linear Discriminant Analysis. In Proc. ICSLP, 608-611(2000)
  16. M Loog, R P W Duin, R Haeb-Umbach. Multiclass linear dimension reduction by weighted pairwise Fisher criteria. TPAMI, 23 (7):762-766 (2001)
  17. Hung-Shin Lee, Berlin Chen. Linear Discriminant Feature Extraction Using Weighted Classification Confusion Information. In INTERSPEECH 2008, 9th Annual Conference of the International Speech Communication Association, Brisbane, Australia, September, 2254-2257 (2008)
  18. S Yan, D Xu, B Zhang. Graph embedding and extensions: a general framework for dimensionality reduction. TPAMI, 29(1): 40-51 (2007)
  19. T Zhang, D Tao, X Li. Patch alignment for dimensionality reduction. IEEE Transactions on Knowledge and Data Engineering, 21(9): 1299-1313 (2009)
  20. Simone Bianco. Can Linear Data Projection Improve Hyperspectral Face Recognition?[J]. Lecture Notes in Computer Science, 2015,vol. 9016,pp 161-170.
  21. E. Murphy-Chutorian, M. Trivedi. Head pose estimation in computer vision: a survey [J]. IEEE Trans. Pattern Anal. Mach. Intell. 31 (4) (2009) 607-626.
  22. H. Han, C. Otto, A.K. Jain. Age estimation from face images: Human vs. machine performance[C]. International Conference on Biometrics, Madrid, Spain, Jun, 2013.
  23. AT & T Laboratories Cambridge. The ORL Database of Faces [OL]. Http://www.cam-orl. Co. uk/facedatabase.html
  24. Belhumeur P.N., Hespanha J.P., Kriengman D.J. Eigenfaces versus Fisherfaces: Recognition Using Class Specific Linear Projection [J]. TPAMI, 1997, 19(7):711-720.
  25. The CMU PIE database [OL]. http://www.ri.cmu.edu/projects/project_418.html, 2013.
  26. The FGNET Aging Database[OL]. Available at http: // www. prima.inrialpes.fr/FGnet, 2013.
  27. Liu S.F, Lin Y. Grey Information: Theory and Practical Applications [M]. London: Springer-Verlag, 2005.

Related papers

An efficient face recognition approach combining likelihood-based sufficient dimension reduction and LDA
Abdallah Benouareth

Multimedia Tools and Applications, 2020

In this paper, an efficient approach is proposed for face recognition (FR) under pose and illumination variations. It is based on combining likelihood-based sufficient dimension reduction (LSDR) and linear discriminant analysis (LDA) using different facial features. LDA is a well-established technique for dimensionality reduction, while LSDR is a relatively new supervised subspace learning method based on the key concept of sufficiency, and consists of estimating a central subspace for low-dimensional representation of facial images. We show, empirically, that this combination can obviously increase the separability between face classes. The facial features that were considered here are either engineered (hand-crafted) features (e.g., discrete shearlet transform coefficients) or learned features which are obtained by retraining a deep learning-based system called FaceNet. To assess the performance of the proposed approach, different classification methods were used during the evaluation phase, namely collaborative representation based classifier (CRC), KNN, linear SVM and three regression-based classifiers (LSRC, LRC and LDRC). The extensive experiments performed on four publicly available face databases-the extended Yale B, FERET, LFW and CMU Multi-PIE-have demonstrated that, on the one hand, LSDR is an effective and efficiency dimensionality reduction technique for face recognition. Particularly, it can be used to significantly increase the performance of a deep learning-based system such as FaceNet, mainly, when training samples are insufficient. On the other hand, its combination with LDA can outperform best individual face recognition algorithm based on LSDR or LDA. Furthermore, the proposed approach compares favorably to current state-of-the-art methods on face recognition, in terms of classification accuracy, under illumination and pose variations.

View PDFchevron_right
Tied Factors Analysis final
Roger Lambert III

Feature extraction from images, which are typical of high dimensionality, is crucial to the recognition performance. To explore the discriminative information while depressing the intra-class variations due to variable illumination and view conditions, we propose a factor analysis framework for separate "content" from "style", identifying a familiar face seen under unfamiliar viewing conditions, classifying familiar poses presented in an unfamiliar face, estimating age across unfamiliar faces. The framework applies efficient algorithms derived from objective factor separating functions and space mapping functions, which can produce sufficiently expressive representations of feature extraction and dimensionality reduction. We report promising results on three different tasks in the high-dimensional image perceptual domains: face identification with two benchmark face databases, facial pose classification with a benchmark facial pose database, extrapolation of age to unseen facial image. Experimental results show that our approach produced higher classification performance when compared to classical LDA, WLDA, LPP, MFA and DLA algorithms.

View PDFchevron_right
Holistic Face Recognition by Dimension Reduction
Aydin Alatan

2003

, 104 pages Face recognition is a popular research area where there are different approaches studied in the literature. In this thesis, a holistic Principal Component Analysis (PCA) based method, namely Eigenface method is studied in detail and three of the methods based on the Eigenface method are compared. These are the Bayesian PCA where Bayesian classifier is applied after dimension reduction with PCA, the Subspace Linear Discriminant Analysis (LDA) where LDA is applied after PCA and Eigenface where Nearest Mean Classifier applied after PCA. All the three methods are implemented on the Olivetti Research Laboratory (ORL) face database, the Face Recognition Technology (FERET) database and the CNN-TURK Speakers face database. The results are compared with respect to the effects of changes in illumination, pose and aging. Simulation results show that Subspace LDA and Bayesian PCA perform slightly well with respect to PCA under changes in pose; however, even Subspace LDA and Bayesian PCA do not perform well under changes in illumination and aging although they perform better than PCA.

View PDFchevron_right
Face Recognition by Exploring Information Jointly in Space, Scale and Orientation
Stan Li

IEEE Transactions on Image Processing, 2011

Information jointly contained in image space, scale and orientation domains can provide rich important clues not seen in either individual of these domains. The position, spatial frequency and orientation selectivity properties are believed to have an important role in visual perception. This paper proposes a novel face representation and recognition approach by exploring information jointly in image space, scale and orientation domains. Specifically, the face image is first decomposed into different scale and orientation responses by convolving multiscale and multiorientation Gabor filters. Second, local binary pattern analysis is used to describe the neighboring relationship not only in image space, but also in different scale and orientation responses. This way, information from different domains is explored to give a good face representation for recognition. Discriminant classification is then performed based upon weighted histogram intersection or conditional mutual information with linear discriminant analysis techniques. Extensive experimental results on FERET, AR, and FRGC ver 2.0 databases show the significant advantages of the proposed method over the existing ones. Index Terms-Conditional mutual information (CMI), face recognition, Gabor volume based local binary pattern (GV-LBP), Gabor volume representation, local binary pattern (LBP). I. INTRODUCTION F ACE recognition has attracted much attention due to its potential value for applications and its theoretical challenges. In real world, the face images are usually affected by different expressions, poses, occlusions and illuminations, and the difference of face images from the same person could be larger than those from different ones. Therefore, how to extract robust and discriminant features which make the intraperson faces compact and enlarge the margin among different persons becomes a critical and difficult problem in face recognition. Up to now, many face representation approaches have been introduced, including subspace based holistic features and local appearance features [1], [2]. Typical holistic features include the well known principal component analysis (PCA) [3], linear dis-Manuscript

View PDFchevron_right
Face recognition across large pose variations via Boosted Tied Factor Analysis
Salman Khaleghian

Applications of Computer …, 2011

In this paper, we propose an ensemble-based approach to boost performance of Tied Factor Analysis(TFA) to overcome some of the challenges in face recognition across large pose variations. We use Adaboost.m1 to boost TFA which has shown to possess state-of-the-art face recognition performance under large pose variations. To this end, we have employed boosting as a discriminative training in the TFA as a generative model. In this model, TFA is used as a base classiēr for the boosting algorithm and a weighted likelihood model for TFA is proposed to adjust the importance of each training data. Moreover, a modiēd weighting and a diversity criterion are used to generate more diverse classiērs in the boosting process. Experimental results on the FERET data set demonstrated the improved performance of the Boosted Tied Factor Analysis(BTFA) in comparison with TFA for lower dimensions when a holistic approach is being used. 978-1-4244-9497-2/10/$26.00 ©2010 IEEE

View PDFchevron_right
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved