A Literature Survey in Face Recognition Techniques · Face recognition is a technique in which images and pat-terns are analyzed and recognized. Face detection is known as the identi

A Literature Survey in Face RecognitionTechniques

M.Tamilselvi1 and Dr. S.Karthikeyan2

1Research Scholar, Department of Electronics and Communication,Sathyabama university, Chennai

2Associate Professsor, Department of Electronics and Communication,Sathyabama university, Chennai

December 27, 2017

Abstract

Face recognition is a technique in which images and pat-terns are analyzed and recognized. Face detection is knownas the identification of face from a video or a image. Manyimproved techniques are implemented in face recognition inpast ten years. Some well known methods in each cate-gory are overviewed and then benefits and drawbacks arementioned and analyzed in this paper. For the purposeof recognizing the face, the most recent algorithms and theapproaching technology methods are analyzed in this paper.

Keywords: Face detection, recognition, analyse, bio-metric, pattern recognition.

1 Introduction

In recent years, the techniques based on the biological propertiesof human beings having the much significant in the identificationof individuals where the other techniques like pass codes, OTPgeneration and other types of security modes having the possibilitiesof getting stolen, misused and forged etc. Hence the biologicalproperties like identification of face finger prints, Palm, ear, Iris,

1

International Journal of Pure and Applied MathematicsVolume 118 No. 16 2018, 831-849ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version)url: http://www.ijpam.euSpecial Issue ijpam.eu

831

retina and signature can be used which are not easily accessed byanyone[3].

The major purpose of face recognition is to verify and iden-tify. Face recognition applications are playing significant role inthe following fields like security investigation, Camera surveillanceprocess, General identity verification, criminal case investigation,database management systems, application based on smart cardand other types of magnetic cards.

In addition, the underlying techniques have also been modifiedand used in relevant applications like gender classification, gesturerecognition, facial recognition and tracking. The gesture recogni-tion can be used in the field of medicine for monitoring intensivecare unit. The facial recognition can be implemented for trackingvehicle driver face where face recognition can also be hybrid withother biometrics like speech, fingerprint and gait recognition.

It is similar to object recognition. Human faces are mostly ap-pearing to be similar and differences between them are negligible.Although face is not a unique, there are several factors that makevariation in appearance of the face. It can be classified as follows;Intrinsic factors and extrinsic factors. Intrinsic represents the ob-jective of a face. It is divided into interpersonal and intrapersonal.

2 Face Recognition Techniques- Appear-

ance Based Approaches

2.1 The Eigen face Method

Firstly Kirby and Sirvoich demonstrated Eigenfaces method forrecognition. Pentland and Turk made improvements on this re-search by employing Eigenfaces method based on Principle Com-ponent Analysis for the same reason[41]. PCA is a Karhumen-Loevetransformation. PCA is a realized linear dimensionality reductionmethod used to determine a set of mutually orthogonal basis func-tions and as shown in fig 1. It uses the vanguard eigenvectors of thesample covariance matrix to characterize the lower dimensional.

It is used to reduce dimension of image matrix. Ex: If a faceimage is represented in g dimensional space, PCA aims to obtainan h dimensional sub space using linear transforms, which answers

2

International Journal of Pure and Applied Mathematics Special Issue

832

Figure 1: Feature vectors are derived using Eigen faces [5]

maximum variance in g dimensional space and g is too big accordingto h. Subtracting the normalized training images from the calcu-lated mean images thus mean centered images are calculated. If wis mean centered training image matrix Wi(i=1,2,........,L) and l isthe number of training images, matrix d is calculated from as inequation 1

D = WW T (1)

To reduce the size of covariance matrix D, we can use D =WTW instead. Eigenvectors ei and eigen values λi are obtainedfrom covariance matrix.

Zi = ETwi(i = 1, 2, ...., L) (2)

In the equation 2, Zi represents the new feature vector of lowerdimensional space. Negative aspect of this method, it tries to maxinter and intra class scattering. Inter class scattering is good forclassification where intra scattering is not. If there is variance il-lumination, increases intra class scattering very high, even classesseems stained.

3


833

2.2 The Fisher face Method

Belhumeur introduced the Fisher Face method in 1997, [43] a deriva-tive of Fishers Linear Discriminant (FLD) which has linear discrim-inant analysis (LDA) to gain the vast discriminant structures. BothPCA and LDA which are used to produce a subspace projection ma-trix is similar to eigen face and Fisher face methods. LDA describesa pair of projection vectors which form the maximum between-classscatter and minimum in the class scatter matrix concurrently [44]produces lower error when compared to Eigen face method. Sixdifferent classes using LDA with large variances within classes, butlittle variance within classes are shown in Fig 2.

Kernel FLD is capable of extracting the most distinct featuresin the feature space, which is common to the nonlinear features inthe reference input space and shows better results when compareto the conventional Fisher face which is established on second or-der statistics of an image-set without considering the high orderstatistical dependencies. Few of the modern LDA-based algorithmsinclude [46]: Direct LDA [47] constructing the image scatter matrixobtained from a normal two dimensional image and it is capable ofresolving small sample size problem. Further, to resolve the sameproblem Dual-Space LDA algorithm requires full discriminative in-formation of face. Both LDA and weighted pair wise Fisher criteriaprivileges are used together by Direct-Weighted LDA. Block LDA[50] algorithm segments the entire image into several blocks andstructures each block as a row vector. Linear discrimination anal-ysis is performed on the row vectors for block which from the twodimensional matrices. The K-Nearest Neighbor approach (KNN)and the Nearest Mean approach (NM) are the two approaches fusedusing LDA and PCA [52], was done on the AT&T and Yale datasets.

Figure 2: Example of Six Classes Using LDA [17]

4


834

Fisher face or Linear Discriminant Analysis (LDA) aims to in-crease inter class differences and are not used to increase data rep-resentation.

Sw =∑R

j=1

∑Mi=1(x

ji−µj)(x

ji−µj)

T (3)

Sb =∑R

j=1(µj−µ)(µj−µ)T (4)

Above are intra class (Equation 3) and inter class (Equation4) scatter matrices respectively. The indices, i is image number, jis class. j is the mean of class j , and is mean of all classes. Mjshows the number images in class j, and R is the number of classes.Sb is maximized while Sw is minimized for the classification tobe done[3]. Intrinsic factors are independent of the observer andrepresents the objective of the face. Further it can be divided intointrapersonal and interpersonal[40].

2.3 Support Vector Machines

To improve the classification performance of the PCA and LDA sub-space features, support vector [SVM] machines came into existence[53], Supervised learning techniques are used to train SVM gen-erally. In estimating the Optimal Separating Hyper plane (OSH)[53] a set of images is used for training SVM. Bringing down therisk of misclassification among two classes of image in some featurespace. Guo et al [53] applied this technique for face recognition. Heapplied binary tree classification techniques where a face image iscontinuously grouped as belonging to one of two classes. A binarytree structure is applied until the two classes denote individual sub-jects and a final classification decision can be arrived. SVM has beopted for face recognition by some other researchers to attain goodresults.

2.3.1 Performance comparison & Experimental result

Two sets of experiments are shown to examine the performance ofindividual algorithms. For a given sample of n images in a class,a classifier is trained using (n-1) images in that class and testedon the remaining single case[63]. This test repeats n times until

5


835

each time training a classifier with leaving-one-out. This is how allimages are used for training and testing to attain good result.[63]

The ORL Database of Faces

Figure 3: Snapshot of ORL Database

AT & T Laboratories Cambridge is the first to perform exper-iments on the ORL face databases. These images are in grayscalewith a resolutions of 92 x 112 pixels containing 400 images, in-cluding 40 distinct people, each with 10 images that are differentin position, rotation, scale and expressions [63]. The images areshot under constant light exposure. Figure 3 shows a snapshot of4 individual from the ORL results

Figure 4: Snapshot of cropped Yale database

6


836

Table 1: ORL ResultImage Set Eigen Fisher SVM1 92.5% 100.0% 95.0%2 85.0% 100.0% 100%3 87.5% 100.0% 100%4 90.0% 97.5% 100%5 85.0% 100.0% 100%6 87.5% 97.5% 97.5%7 82.5% 95.0% 95.0%8 92.5% 95.0% 97.5%9 90.0% 100.0% 97.5%10 85.0% 97.5% 95.0%Average 87.5% 98.3% 97.8%

Yale Face Database

Second experiment is performed on Yale face database from YaleUniversity. These images are in gray-scale and have cropped to aresolution of 116 x 136 ppi, has 165 images in group, including 15discrete people, each with 11 images that differ in both lighting andexpression. A snapshot of 4 individuals from the database is howin Figure 4. The results of FRCM performed on Yale database todistinguish the 15 people under different conditions is given in theTable 2.

3 Feature Based Approaches

3.1 Face Recognition through geometric features

In the first phase a set of fiducial points are examined in each facesand the geometric facts such as distances between these points areexplored and the image closest to the query face is nominated.This process was done by Kanade [55] who employed the Euclideandistance for correlation between 16 extracted feature vectors con-structed on image database of 20 distinct people with 2 imagerper person and attained accuracy rate of 75%. Further, Brunelliand Poggio [37] practiced the same on 35 geometric features fromimage database of 47 peculiar people with 4 images per person asdisplayed in fig 5 and attained performance rate of 95%. Most re-

7


837

Table 2: Yale ResultImage Set Eigen Fisher SVMCenterlight 53.3 93.3 86.7Glasses 80 100 86.7Happy 93.3 100 100Left light 26.7 26.7 26.7No glasses 100 100 100Normal 86.7 100 100Right light 26.7 40 13.3Sad 86.7 93.3 100Sleepy 86.7 100 100Surprised 86.7 66.7 73.3Wink 100 100 93.3

cently, Cox et al. [38] derived 35 facial features from a databasecomprised 685 images and reportedly achieved a recognition per-formance rate of 95% on a database of 685 images with one imagefor each individual.

Figure 5: Geometrical feature used by Brunelli and Poggio [24]

8


838

3.2 Hidden Markov Model (Hmm)

The HMM was first presented by Young and Samaira [55]. HMMgenerally employed on images with variations due to lighting, ori-entation and facial expression and thus it have more advancementsover than the approaches for treating images using HMM, spacesequences are considered. This procedure is named as a HiddenMarkov Model this is why because the states are invisible, only theoutput is vivid to the external use. This procedure uses pixel stripsto cover all the areas in the without finding the precise locationsof facial features. The face arrangements are identified as a con-tinuous of discrete parts. The arrangements of the system shouldbe maintained for e.g., it should start from top to bottom fromforehead, eyes, nose, mouth, and chin as in fig 6.

Figure 6: Left to Right HMM for face recognition

3.3 Active Appearance Model (AAM)-2D Mor-phable Method

Faces are highly distinct and able to be deformed. Classifying bypose, expression, lighting, and faces can have various looks in theimages. Coots, Taylor, and Edwards [56] presented Active Appear-ance Model which is strongly capable of explaining the view of facein set of model parameters. AAM is an integrated statistical model,implemented on the basis of a training set comprising labeled im-ages. The landmark points are pointed as show in fig 7. Modelparameters are found to ability matching with the image which

9


839

brings down the difference between the image and a synthesizedmodel sample projected into the image.

Figure 7: image is split into shape and shape normalized texture[19]

3.4 3D Morphable Model

To differentiate the facial variations like illumination, pose etc. 3Dmorphable model is an effective, strong and versatile representationof human faces and so it is better to represent the face by employing3D model. High quality frontal and half profile picture are takenfirst of each subject under ambient lighting conditions to make a3D model. Later these images are use as reference to the analysisby synthesis loop which results a face mode. Blanz et al. [57]proposed this method based upon a 3D morphable face model inwhich he tries to find an algorithm to reconstruct the parameterslike texture and shape from the single image of a face and encodesthem with respect to model parameters. Thus the 3D morphablemodel provides the full 3D feature information which enables forautomatic extraction of facial regions and facial components.

4 Hybrid Methods

These methods show better results using both the holistic andfeature-based methods to recognize the face. Eigen modules pro-posed by Pentland et al. [58], which applied both local Eigen fea-

10


840

tures and global Eigen faces and shows much better results than theholistic eigen faces. Penev and Atiek [59], used a method called Hy-brid LFA (Local Feature Analysis). Shape-normalized Flexible ap-pearance technique by lantis et al. [61] which combines component-based Face region and components by Huang et al.[61] which com-bines component based recognition and 3D morphable model forface recognition.

The important phase is to generate 3D face models using 3Dmorphable model form the three reference images of each person.These images are furnished under variable illumination conditionsand pose to populate a large set of synthetic images, are used totrain a component-based face recognition system [62].A SupportVector Machine (SVM) based recognition system is used to decom-pose the face into a set of components that are interconnected by aflexible geometrical model so that it can keep track for the changesin the head pose leading to changes in the position of the facialcomponents.

5 Conclusions

Face recognition is a highly challenging task in the domain of imageanalysis and computer vision that has received an immense deal ofattention over the last few decades because of its many applica-tions in vast domains. Few classical face recognition techniquesare cited in this paper. In some face database, the methods ofSVM and HMM can produce better face recognition results, butthey use more complex algorithms. Research has been conductedexorbitantly in this area and immense progress has been attained,notable results have been obtained and present face recognition sys-tems have elevated to a certain measure of maturity when imposedunder constrained conditions; however, these methods are far fromachieving the ideal of being able to perform adequately in all thevarious situations that are commonly faced by the applications em-ploying these procedures in practical life. The fundamental goal ofresearchers in this domain is to enable computers to emulate the hu-man vision system and, as has aptly pointed out by Torres, Strongand coordinated effort between the computer vision, psychophysicsand signal processing and neurosciences communities is needed to

11


841

derive this objective.

References

[1] Ahmet zdil Metin Mete zbilen A Survey on Comparison ofFace Recognition Algorithms IEEE 2015

[2] G. Ramkumar, M. Manikandan - Face Recognition SurveyInternational Journal of Advances in Science and Technology(IJAST) 2014 ISSN 2348-5426

[3] Sharkas, M. Abou Elenien, Eigenfaces vs. Fisherfaces vs. ICAfor Face Recognition; A Comparative Study, 9th InternationalConference on Signal Processing, 2008, ICSP 2008., 2008, pp.914919

[4] K. Kim, Intelligent Immigration Control System by UsingPassport Recognition and Face Verification, in InternationalSymposium on Neural Networks Chongqing, China, 2005,pp.147-156.

[5] J. N. K. Liu, M. Wang, and B. Feng, iBotGuard: an Internet-based intelligent robot security system using invariant facerecognition against intruder, IEEE Transactions on SystemsMan And Cybernetics Part C-Applications And Reviews,Vol.35, pp.97-105, 2005.

[6] H. Moon, Biometrics Person Authentication Using Projection-Based Face Recognition System in Verification Scenario, in In-ternational Conference on Bioinformatics and its Applications.Hong Kong, China, 2004, pp.207-213.

[7] D. McCullagh, Call It Super Bowl Face Scan 1, in Wired Mag-azine, 2001.

[8] CNN, Education School face scanner to search for sex offend-ers. Phoenix, Arizona: The Associated Press, 2003.

[9] P. J. Phillips, H. Moon, P. J. Rauss, and S. A. Rizvi, TheFERET Evaluation Methodology for Face Recognition Algo-rithms, IEEE Transactions on Pattern Analysis and MachineIntelligence, Vol.22, pp.1090-1104, 2000.

12


842

[10] T. Choudhry, B. Clarkson, T. Jebara, and A. Pentland, Multi-modal person recognition using unconstrained audio and video,in Proceedings, International Conference on Audio and Video-Based Person Authentication, 1999, pp.176-181.

[11] S. L. Wijaya, M. Savvides, and B. V. K. V. Ku-mar, Illumination-tolerant face verification of low-bit-rateJPEG2000 wavelet images with advanced correlation filters forhandheld devices, Applied Optics, Vol.44, pp.655-665, 2005.

[12] E. Acosta, L. Torres, A. Albiol, and E. J. Delp, An automaticface detection and recognition system for video indexing appli-cations, in Proceedings of the IEEE International Conferenceon Acoustics, Speech and Signal Processing, Vol.4. Orlando,Florida, 2002, pp.3644-3647.

[13] J.-H. Lee and W.-Y. Kim, Video Summarization and RetrievalSystem Using Face Recognition and MPEG-7 Descriptors, inImage and Video Retrieval, Vol.3115.

[14] Lecture Notes in Computer Science : Springer Berlin / Heidel-berg, 2004, pp.179-188.

[15] C. G. Tredoux, Y. Rosenthal, L. d. Costa, and D. Nunez,Face reconstruction using a configural, eigenface-based compos-ite system, in 3rd Biennial Meeting of the Society for AppliedResearch in Memory and Cognition (SARMAC). Boulder, Col-orado, USA, 1999.

[16] K. Balci and V. Atalay, PCA for Gender Estimation: WhichEigenvectors Contribute? in Proceedings of Sixteenth Interna-tional Conference on Pattern Recognition, Vol.3. Quebec City,Canada, 2002, pp. 363-366.

[17] B. Moghaddam and M. H. Yang, Learning Gender with SupportFaces, IEEE Transactions on Pattern Analysis and MachineIntelligence, Vol.24, pp.707-711, 2002.

[18] R. Brunelli and T. Poggio, HyperBF Networks for GenderClassification, Proceedings of DARPA Image UnderstandingWorkshop, pp.311-314, 1992.

13


843

[19] A. Colmenarez, B. J. Frey, and T. S. Huang, A probabilisticframework for embedded face and facial expression recognition,in Proceedings of the IEEE Conference on Computer Visionand Pattern Recognition, Vol.1. Ft. Collins, CO, USA, 1999,pp. 1592-1597.

[20] Y. Shinohara and N. Otsu, Facial Expression Recognition Us-ing Fisher Weight Maps, in Sixth IEEE International Con-ference on Automatic Face and Gesture Recognition, Vol.100,2004, pp.499-504.

[21] F. Bourel, C. C. Chibelushi, and A. A. Low, Robust Facial Fea-ture Tracking, in British Machine Vision Conference. Bristol,2000, pp.232-241.

[22] K. Morik, P. Brockhausen, and T. Joachims, Combining statis-tical learning with a knowledge-based approach – A case studyin intensive care monitoring, in 16th International Conferenceon Machine Learning (ICML-99). San Francisco, CA, USA:Morgan Kaufmann, 1999, pp.268-277.

[23] S. Singh and N. Papanikolopoulos, Vision-based detection ofdriver fatigue, Department of Computer Science, University ofMinnesota, Technical report 1997.

[24] D. N. Metaxas, S. Venkataraman, and C. Vogler, Image-BasedStress Recognition Using a Model-Based Dynamic Face Track-ing System, International Conference on Computational Sci-ence , pp.813-821, 2004.

[25] M. M. Rahman, R. Hartley, and S. Ishikawa, A Passive AndMultimodal Biometric System for Personal Identification, inInternational Conference on Visualization, Imaging and ImageProcessing Spain, 2005, pp.89-92.

[26] R. Brunelli and D. Falavigna, Person identification using mul-tiple cues, IEEE Transactions on Pattern Analysis and Ma-chine Intelligence, Vol.17, pp.955-966, 1995.

[27] M. Viswanathan, H. S. M. Beigi, A. Tritschler, and F. Maali,Information access using speech, speaker and face recognition,

14


844

in IEEE International Conference on Multimedia and Expo,Vol.1, 2000, pp. 493–496.

[28] A. K. Jain, K. Nandakumar, X. Lu, and U. Park, IntegratingFaces, Fingerprints, and Soft Biometric Traits for User Recog-nition, Proceedings of Biometric Authentication Workshop, inconjunction with ECCV2004, LNCS 3087, pp.259-269, 2004.

[29] P. Melin and O. Castillo, Human Recognition using Face, Fin-gerprint and Voice, in Hybrid Intelligent Systems for PatternRecognition Using Soft Computing, Vol.172, Studies in Fuzzi-ness and Soft Computing: Springer Berlin / Heidelberg, 2005,pp.241-256.

[30] K. Chang, K. W. Bowyer, S. Sarkar, and B. Victor, Compari-son and Combination of Ear and Face Images in Appearance-Based Biometrics, IEEE Transactions on Pattern Analysis andMachine Intelligence, Vol.25, pp.1160-1165, 2003.

[31] R. Chellappa, A. Roy-Chowdhury, and S. Zhou, Human Iden-tification Using Gait and Face, in The Electrical EngineeringHandbook, 3rd ed: CRC Press, 2004.

[32] S. Ben-Yacoub, J. Luttin, K.Jonsson, J. Matas, and J. Kit-tler, Audio-visual person verification, in IEEE Computer Soci-ety Conference on Computer Vision and Pattern Recognition,Vol.1. Fort Collins, CO, USA, 1999, pp.580-585.

[33] X. Zhou and B. Bhanu, Feature fusion of side face and gait forvideo-based human identification, Pattern Recognition, Vol.41,pp.778-795, 2008.

[34] D. Bouchaffra and A. Amira, Structural hidden Markov modelsfor biometrics: Fusion of face and fingerprint, Pattern Recog-nition, Vol.41, pp.852-867, 2008.

[35] H. Vajaria, T. Islam, P. Mohanty, S. Sarkar, R. Sankar, and R.Kasturi, Evaluation and analysis of a face and voice outdoormulti-biometric system, Pattern Recognition Letters , Vol.28,pp.1572-1580, 2007.

15


845

[36] Y.-F. Yao, X.-Y. Jing, and H.-S. Wong, Face and palmprintfeature level fusion for single sample biometrics recognition,Neurocomputing, Vol.70, pp. 1582-1586, 2007.

[37] J. Zhou, G. Su, C. Jiang, Y. Deng, and C. Li, A face andfingerprint identity authentication system based on multi-routedetection, Neurocomputing, Vol.70, pp.922-931, 2007.

[38] C. Nastar and M. Mitschke, Real time face recognition usingfeature combination, in Third IEEE International Conferenceon Automatic Face and Gesture Recognition. Nara, Japan,1998, pp. 312-317.

[39] S. Gong, S. J. McKenna, and A. Psarrou., Dynamic Vision:From Images to Face Recognition: Imperial College Press(World Scientific Publishing Company), 2000.

[40] T. Jebara, 3D Pose Estimation and Normalization for FaceRecognition, Center for Intelligent Machines, McGill Univer-sity, Undergraduate Thesis May, 1996.

[41] M. Turk and A. Pentland. Face recognition using eigenfaces.In Proceedings of the IEEE Conference on Computer Visionand Pattern Recognition, 1991

[42] Face Recognition: National Science and Technology Council(NSTC), Committee on Technology, Committee on Homelandand National Security, Subcommittee on biometrics.

[43] Peter N.Belhumeur,JoaoP.Hespanha and David Kreigman,Eigenfaces vs. Fisherfaces Recognition using class specific Lin-ear Projection, IEEE Trans. PAMI, 1997.

[44] Chen, H.Liao, M.Ko, J.Lin and G.Yu,A New LDA-Based FaceRecognition System which can Solve the Small Sample SizeProblem, Pattern Recognition, 2000.

[45] S.Mika, G.Ratsch, J Weston, Scholkopf, Smola and K RMuller, Invariant Feature extraction and classification in ker-nel spaces. In NIPS-12, 2000.

16


846

[46] SeemaAsht, RajeshwarDass, Pattern Recognition Techniques:A Review; International journal of Computer Science Telecom-munications, vol 3, Issue 8, August 2012.

[47] H. Yu and J. Yang, A Direct LDA Algorithm for High-dimensional Data with Application to Face Recognition, Pat-tern Recognition, vol.34, pp.2067- 2070, 2001.

[48] X. Wang and X. Tang, Dual-space Linear Discriminant Anal-ysis for Face Recognition, in Proceedings of IEEE Interna-tional Conference on Computer Vision and Pattern Recogni-tion, 2004, pp.564569.

[49] D. Zhou and X. Yang, Face Recognition Using Direct-WeightedLDA, in 8th Pacific Rim International Conference on ArtificialIntelligence. Auckland, New Zealand, pp.760-768, 2004.

[50] V. D. M. Nhat and S. Lee, Block LDA for Face Recognition, inComputational Intelligence and BioinspiredSystems, vol.3512,Lecture Notes in Computer Science: Springer Berlin / Heidel-berg, pp.899-905, 2005.

[51] Gian Luca Marcialis and Fabio Roli, Fusion of LDA and PCAfor Face Recognition, Department of Electrical and ElectronicEngineering - University of Cagliari Piazza dArmi - 09123Cagliari (Italy).

[52] B. Heisele, P. Ho, and T. Poggio, Face recognition with supportvector machines: Global versus component-based approach,Proceeding of ICCV, 2001.

[53] G. Guo, S.Z. Li, K. Chan, Face Recognition by Support Vec-tor Machines. Proc. of the IEEE International Conference onAutomatic Face and Gesture Recognition, 2000.

[54] F.Samaira and S. Young, HMM based architecture for FaceIdentification, Image and Computer Vision,vol. 12, pp. 537-583,October 1994.

[55] T. Kanade, Computer Recognition of Human Faces. Basel andStuttgart Birkhauser, 1977

17


847

[56] T.F. Cootes, G.J. Edwards, and C.J. Taylor, Active appear-ance models, IEEE Trans. Pattern Analysis and Machine In-telligence, vol. 23, no. 6, pp. 681685, Jun. 2001

[57] V. Blanz and T. Vetter, A morphable model for the synthesis of3D faces, in Proc. ACM SIGGRAPH, Mar. 1999, pp. 187194.

[58] A. Pentland, B. Moghaddam, T. Starner, View-Based andModular Eigenspaces for Face Recognition , Proceedings of theIEEE Conference on Computer Vision and Pattern Recogni-tion, Seattle, Washington, USA, pp. 84-91, 21-23 June 1994.

[59] P. Penev and J. Atick, Local Feature Analysis: A General Sta-tistical Theory for Object Representation, Network: Computa-tion Neural, 1996.

[60] A. Lanitis, C. J. Taylor and T. F. Cootes, Automatic FaceIdentification System Using Flexible Appearance Models, ImageVis. Comput., 1995.

[61] Huang, J., Heisele, B., and Blanz, V., Component-based facerecognition with 3D morphable models. In Proceedings, Inter-national Conference on Audio- and Video-Based Person Au-thentication, 2003

[62] Brunelli and Poggio, Face Recognition: Features Vs Templates,IEEE Transactions on Pattern Ananlysis and Machine Intelli-gence, vol 15, no. 10, October 1993.

[63] Ho-Man Tang, Michael R. Lyu and Irwin KingFACE RECOGNITION COMMITTEE MACHINEhttp://citeseerx.ist.psu.edu/

[64] Li Xianwei Zhang Haiyang A Survey of Face Recognition Meth-ods Published by Atlantis Press, Paris, France

[65] Rabia Jafri* and Hamid R. Arabnia A Survey of Face Recog-nition Techniques Journal of Information Processing Systems,Vol.5, No.2, June 2009

18


848

849

850

Documents

A Literature Survey in Face Recognition Techniques · Face recognition is a technique in which images and pat-terns are analyzed and recognized. Face detection is known as the identi