Discriminative Projection Selection Based Face Image Hashing
|
|
- Theresa Fowler
- 5 years ago
- Views:
Transcription
1 IEICE TRANS. INF. & SYST., VOL.Exx D, NO.xx XXXX 00x LETTER Dscrmnatve Projecton Selecton Based Face Image Hashng Cagatay KARABAT a), Student Member and Hakan ERDOGAN, Nonmember SUMMARY Face mage hashng s an emergng method used n bometrc verfcaton systems. In ths paper, we propose a novel face mage hashng method based on a new technque called dscrmnatve projecton selecton. We apply the Fsher crteron for selectng the rows of a random projecton matrx n a user-dependent fashon. Moreover, another contrbuton of ths paper s to employ a bmodal Gaussan mxture model at the quantzaton step. Our smulaton results on three dfferent databases demonstrate that the proposed method has superor performance n comparson to prevously proposed random projecton based methods. key words: face mage hashng, bometrc securty, prvacy Fg. Basc steps of the bometrc hashng methods. Introducton In recent years, bometrcs has acheved wde-spread usage n varous applcatons. In most of these applcatons, bometrc templates are stored n databases and/or smart cards, thus rasng questons such as data securty and prvacy. Varous bometrc hashng methods, whch mostly depend on random projectons, are proposed to protect the bometrcs data []-[5] n the lterature. See Fgure for an llustraton of the bascs steps of such bometrc hashng methods. Ngo et al.[], [] employ feature extracton methods.e. Prncple Component Analyss PCA), Fsher Lnear Dscrmnant FLD), Wavelet transformaton, Wavelet Transform wth PCA, Wavelet Transform wth Fourer Melln Transform etc.) to the face mages and then make and use a random projecton RP) matrx for reducng the dmenson of the feature vectors. Fnally, they employ bnary quantzaton to obtan face mage hash vectors. We mprove upon ths method n our work. In ths work, we develop a new face mage hashng method based on a proposed technque that we call dscrmnatve projecton selecton to reduce verfcaton errors. Ths technque selects the rows of an RP matrx, whch s a user dependent dmenson reducton matrx, by usng the Fsher crteron [6]. Moreover, we employ Gaussan mxture model at the quantzaton step to obtan more dstnct face mage hash vectors for each user. Manuscrpt receved June 7, 0. The author s wth TUBITAK BILGEM Center of Research for Advanced Technologes on Informatcs and Informaton Securty) The author s wth the Faculty of Engneerng and Natural Scences, Sabanc Unversty The author s wth the Faculty of Engneerng and Natural Scences, Sabanc Unversty a) E-mal: cagatay@uekae.tubtak.gov.tr DOI: 0.587/transnf.E0.D.. The Proposed Bometrc Verfcaton Method In ths secton, we ntroduce a bometrc verfcaton system whch employs the proposed face mage hashng method based on the dscrmnatve projecton selecton technque.. Enrollment Stage There are three man steps at the enrollment stage: ) Feature extracton, ) Dmenson reducton, 3) Quantzaton... Feature Extracton At the feature extracton phase, we use two sets of data: tranng set and others set. The tranng set has tranng face mages of regstered users, I, j R m n, =,..., K where K denotes number of users and j=,..., L where L denotes number of tranng mages per user. We lexcographcally re-order them and obtan tranng face vectors, x, j R mn). The others set contans randomly selected face mages whch do not belong to any regstered users Ĩ s R m n, s =,..., M where M denotes the number of face mages belongng to the others set. We agan lexcographcally re-order them and obtan face vectors, x s R mn) of the others set. We apply PCA to the face mages n the tranng set for feature extracton. y, j = Ax, j µ), ) where A R q mn) s the PCA matrx traned usng the face mages n the tranng set, y, j R q s the PCA coeffcent vector belongng to the j th tranng mage of the th user and µ s the mean face vector. We project the face mages n the others set onto the PCA subspace as follows: ỹ s = A x s µ), ) where ỹ s R q s the PCA coeffcent vector belongng to Copyrght c 00x The Insttute of Electroncs, Informaton and Communcaton Engneers
2 IEICE TRANS. INF. & SYST., VOL.Exx D, NO.xx XXXX 00x the s th mage of the others set. We use these PCA coeffcent vectors n the dscrmnatve projecton selecton technque to fnd the most valuable features, whch maxmze the dstance between the face mages of a user n the tranng set and the face mages n the others set, n the lowerdmensonal subspace... Dmenson Reducton At the dmenson reducton phase, we generate an RP matrx, T R l q, for each user to reduce the dmenson of her feature vector. The RP matrx elements are dentcally and ndependently dstrbuted..d) and generated from a Gaussan dstrbuton wth zero mean and unt varance by usng a random number generator RNG) wth a seed derved from the user s secret key. We apply the Gram-Schmdt GS) procedure to obtan an orthonormal projecton matrx R R l q to have more dstnct projectons. Then, we project the PCA coeffcent vectors of the th user onto a lowerl-dmensonal subspace. z, j = R y, j, 3) where z, j R l s the ntermedate face mage hash vector belongng to the j th tranng mage of the th user. To determne the competng hash vectors, we also project the others set usng the RP matrx as follows: z s = R ỹ s, 4) where z s R l s the ntermedate face mage hash vector of the s th face mage of the others set and R s the orthonormal random projecton matrx of the th user. The proposed dscrmnatve projecton selecton technque selects the rows of the matrx R usng the Fsher crteron [6] and creates the dscrmnatve random projectons. Thus, we am to ncrease dscrmnablty due to mappng the PCA coeffcent vectors nto a more dscrmnant subspace. Fsher crteron s a feature selecton method and n ths case the features are the features obtaned after the random projecton of a PCA coeffcent vectory, namely: zk)=r T,ky, 5) for k=,...,l, where zk) s a scalar value and r T,k R q denotes the k th row of R. The k th feature for the th user s j th mage s z, j k) and the same feature for the s th element of the others set s z s k) whch are the k th elements of the correspondng vectors defned n Equatons 3) and 4). The features are already uncorrelated due to the GS procedure whch ensures r T,k r,m= 0 for k m. We defne e k [ z, k),...,z,l k) ], 6) whch s a collecton of the k th dmenson or bt poston) coeffcents of the ntermedate hash vectors belongng to the th user for each k=,...,l. Frst, we compute the sample mean value, ˆµ,k, of each e k vector for each bt poston k=,...,l as follows: ˆµ,k = L L j= e k j), 7) where e k j) s the jth element of the vector e k whch s defned n Equaton 6). Then, we compute the sample standard devaton, ˆσ,k, of each e k vector for each bt poston k=,...,l as follows: ˆσ,k = L L j= e k j) ˆµ,k ). 8) Smlarly, we collect together the k th dmenson values of the ntermedate hash vectors of the others data set as q k [ z k),..., z M k)] 9) for all k=,...,l. Frst, we compute the sample mean value, ˆµ,k, of each q k for each bt poston k=,...,l as follows: ˆµ,k = M M q k s), 0) where q k s) s the s th element of the vector q k whch s defned n Equaton 9). Next, we compute the sample standard devaton, ˆσ,k, of each q k for each bt poston k=,...,l as follows: ˆσ,k = M s= M q k s) ˆµ,k) s=. ) By applyng the Fsher crteron, we try to select the rows that have hgher contrast between genune user s data and the others set. In other words, we am to reduce the dstance between the genune user s dfferent face mage hash vectors whle at the same tme we am to maxmze the dstance between the th user s data and the others set. We compute the Fsher score for each row of R as follows: η k)= ˆµ,k ˆσ,k ˆµ,k ) + ˆσ,k ). ) We use these Fsher scores obtaned for each dmenson k to rank the rows of the random projecton matrx R. We defne r T,k to be the kth row of R. That s R = r T,.. r T,l. 3) We choose top rankng rows from the random projecton matrx. Let c be the ndex vector whch ranks the rows of the matrx n a descendng manner from towwherew
3 LETTER 3 s the number of desred rows. That s, c ) s the row ndex of R that has the hghest Fsher score, c ) s the ndex wth the second hghest score and so on. Thus, we obtan the dscrmnatve random projecton matrx ˆR R w q and the ndex vector, c, whch contans the ndces of topwrows for each user. We defne ˆr T, ˆR =.., 4) ˆr T,w where ˆr,p = r,c p) for p=,...,w. We only store the ndex vector, c, for the th user n the database for verfcaton at the test stage. Next, we project the PCA coeffcents, whch belong to the tranng face mages of the th user, onto a lowerwdmensonal subspace by usng the calculated ˆR as follows: f, j = ˆR y, j, 5) where f, j R w s the raw face mage hash vector belongng to the j th tranng mage of the th user. Ngo et.al [], [] uses FLD as a feature extracton method whch s appled before the random projecton step n the algorthm. Ther FLD transform s not user specfc and ams to dscrmnate face mages belongng to dfferent users n the database. In our case, we employ the Fsher crteron for projecton selecton for bometrc verfcaton. Therefore, n our case, the projecton selecton s user-specfc and ams to dscrmnate the clamed user s bometrc hash vector from all other possble ones that may come from other face mages. In our case, other face mages may even be from outsde the database whch s more realstc n a real scenaro. Our others set s chosen from another database for ths purpose. The projecton selecton s done after the random projecton step. FLD can reduce dmenson at most to K dmensons, where K s the number of users n the database, due to the maxmal rank of the between class covarance matrx. However, n our method, we do not have such a lmtaton snce the selecton s done by rankng the Fsher crtera obtaned from each projecton. In summary, there are fundamental dfferences between usng FLD as a dmenson reducng feature extracton method and usng the Fsher crtera for selecton of random projectons that best separate the clamed dentty from all others.. Quantzaton In ths subsecton, we dscuss the quantzaton methods used n ths work. We employ two dfferent quantzaton methods: ) Bnary quantzaton BQ) [], [] and ) The proposed Gaussan mxture model GMM) based quantzaton method. In our smulatons, we employ these quantzaton methods separately to show the performance of the system... Bnary Quantzaton Method wth a Fxed Threshold Ths technque s employed n Ngo et al. s method [], []. The raw face mage hash vector f, j elements are bnarzed wth respect to a pre-determned fxed threshold as follows: { f f λ, j k)=, j k) µ, 6) 0 Otherwse. where the threshold µ s chosen as the sample mean value of elements of the vector f, j and k=,...,w. The computed reference face mage hash vectorsλ, j R w are stored n the database... The Proposed GMM Based Quantzaton Method To the best of our knowledge, there s no face mage hashng method employng GMM n the quantzaton step. GMM s one of the most wdely used data clusterng methods n the lterature [7]. Let us assume that we have a set of numbers whch are obtaned by collectng the k th elements of raw face mage hash vectors and we want to bnarze the element of ths set. Snce our am s to make bnarzaton, we ft two Gaussan dstrbutons to the hstogram of the k th elements of the raw face mage hash vectors by usng the GMM. Then, we choose the average of the mean values of these two Gaussan dstrbutons as an optmum threshold for partton of these two dstrbutons. We repeat t for each bt locaton k =,..., w separately. In other words, we employ bmodal GMM to fnd an optmum threshold for each bt poston for bnarzaton. Let f, j k) denote the k th bt of f, j R w, we defne the vector d k as the collecton of all k th dmenson values of the raw mage hashes n the database. d k [ f, j k) : =,..., K, j=,..., L ] R r, 7) where k=,...,w, r=k L, K s the number of users and L s the number of tranng mages per user. Assume that the elements of the vector d k are observatons of a sngle random varable d. p d Ψ k) = S α k s p d θs) k, 8) s= whereα k s s a mxture weght, S s= αk s = where S = due to bnarzaton,ψ k = { α k,αk,θk,θk } and p d Ψ k ) s a one-dmensonal Gaussan densty wth ts own parameters θ k s= { µ k s,σ k s} as follows: p ) d θ k s = e d µk s) / σ k s) ), 9) π σ k s whereµ k s andσ k s denote mean varance for the s th component of the GMM respectvely for s={, }. We fnd an optmum threshold for each bt poston as follows: T k = µk +µk, 0) where T k denotes the optmum threshold for the k th bt poston of the raw face mage hash vector f, j,, j. Note that,
4 4 IEICE TRANS. INF. & SYST., VOL.Exx D, NO.xx XXXX 00x the GMM s traned usng the whole tranng set for each bt poston. Thus, the GMM parameters are not user dependent. Fnally, the elements of f, j are bnarzed wth respect to the optmum system-level thresholds as follows: { f f λ, j k)=, j k) T k, ) 0 Otherwse. where k =,...,w. The computed reference face mage hash vectorsλ, j R wx are stored n the database..3 Test Stage At the test stage, a clamer clams that she s the th user and sends her face mage and her secret key to the system. The system computes her test face mage hash vector by usng her face mage, her secret key to generate a RP matrx) and the ndex vector, c, whch belongs to the th user. Recall that ndex vectors and the reference face mage hash vectors of the regstered users are stored n the database; however, the secret keys are not stored n the database. Then, the Hammng dstance [8] s computed between the test face mage hash vector and the reference face mage hash vectors whch belong to the th user and were generated at the enrollment stage. If t s below the pre-determned dstance threshold, the clamer s accepted; otherwse, the clamer s rejected. We smulate two scenaros n our experments. These scenarous are descrbed n detal below.. Key-Unknown Scenaro: In ths scenaro, an unauthorzed mpostor has nether the secret key nor the face mage template belongng to the genune user. Note that the ndex vectors of the users are stored n the database. Therefore, whenever a clamer clams that she s the th user and sends her face mage and a secret key to the system, the system computes a test face mage hash vector by usng the data sent by the clamer and the ndex vector, c, whch belongs to the th user.. Key Stolen Scenaro: In ths scenaro, an unauthorzed mpostor acqures the secret key of the th genune user but does not have the clamed person s face mage. When an mpostor sends her face mage and the secret key of the th user to the system, the system computes a test face mage hash vector by usng the data sent by the mpostor and the ndex vector c that belongs to the th user whch s stored n the database. 3. Smulaton Results In ths secton, we dscuss our expermental results. We test the performance of the proposed method on AT&T [9], AR [0] and the Sheffeld prevously UMIST) face databases []. AT&T database has 400 dfferent face mages correspondng to 40 dstnct people. AR database has 30 face mages belongng to 0 dfferent people s faces. The Sheffeld database has 564 dfferent face mages belongng to 0 dfferent people. Besdes, we randomly select 04 face mages from Carnege Mellon Unversty database [] and create the others set. We compare the performance of the proposed method to the Ngo et al. s PCA+RP and FLD+RP methods that were ntroduced n [], [] as shown n Table. We automatcally select face mages for tranng and test sets and evaluate the performance of the proposed method and Ngo et al. s methods [], []. We use 04-length PCA coeffcent vectors for the face mages belongng to the tranng, test and others sets n the smulatons. In our experments, pre-processng technques such as eye algnment, head regon maskng, lghtng adjustment are not appled to the face mages. In our smulatons for both scenaros; for mpostor tests, each face mage n the test set s compared aganst each face mage of all other users n the tranng set. The faled mposter test results n False Acceptance error. For the genune tests, each face mage n the test set s compared aganst all face mages of the same user n the tranng set. The faled genune test results n False Rejecton error. The detaled nformaton on the data sets used n the experments are gven n Table. The proposed Table Data sets and Expermental Set-up Database Number of Face Images Tran set AR 30 mages from The frst 7 mages 0 people AT&T 400 mages from The frst 5 mages 40 people Sheffeld 564 mages from The frst 8 mages 0 people Test set The last mages The rest 5 mages The followng 8 mages method has better performance n terms of equal error rate EER) n comparson to the Ngo et al. s methods [], [] whereas Ngo et al. s PCA+RP and FLD+RP methods have comparable performances wth each other as shown n Table. As the length of face mage hash vector decreases, the proposed method shows better mprovement snce the proposed dmenson reducton matrx better preserves the Mss probablty n %) DET plots for the AT&T Database Ngo PCA+RP Ngo FLD+RP The Proposed Method wth BQ The Proposed Method wth GMM False Alarm probablty n %) Fg. DET plots for the methods wth 56 bt face mage hash vector length for key stolen scenaro - AT&T database
5 LETTER 5 Length of Face Hash Vector Table [], [] EER %) of Ngo et al. s Method [], [] PCA+RP) EER Performances of the Proposed Face Image Hashng Method and Ngo et al. s Methods EER %) of Ngo et al. s Method [], [] FLD+RP) EER %) of The Proposed Method wth Bnary Quantzaton Method wth Fxed Threshold EER %) of The Proposed Method wth GMM Based Quantzaton Method Scenaro 64 bt %.9 % 0.46 % 3.83 %.73 Key Unknown AT&T 8 bt % 7.36 % 8.5 %.3 %.57 Key Unknown AT&T 56 bt % 5.8 % 5.50 %.80 %.5 Key Unknown AT&T 5 bt % 3.79 % 4.7 %.48 %.0 Key Unknown AT&T 64 bt % 6.93 % 8.3 % 4.40 % 3.58 Key Stolen AT&T 8 bt % 3.97 % 6.73 %.0 %.4 Key Stolen AT&T 56 bt %.76 % 4.50 % 0.80 % 0.3 Key Stolen AT&T 5 bt %.34 % 3.55 % 0.5 % 9.73 Key Stolen AT&T 64 bt % 3.63 % 3.34 % 9.08 % 8.96 Key Unknown AR 8 bt % 8.4 % 8.05 % 8.67 % 8.7 Key Unknown AR 56 bt % 5.8 % 3.93 % 7.8 % 8. Key Unknown AR 5 bt %.38 %.9 % 8.33 % 8.57 Key Unknown AR 64 bt % 8.7 % 8.5 % 8.07 % 8.46 Key Stolen AR 8 bt % 7.7 % 7.56 % 8.06 % 8.05 Key Stolen AR 56 bt % 5.50 % 6.44 % 9.0 % 8.83 Key Stolen AR 5 bt % 4.89 % 5.04 % 0.95 % 0.4 Key Stolen AR Database 64 bt % 7.09 %.00 % 5.75 % 6.3 Key Unknown Sheffeld 8 bt % 6.38 % 9.0 % 3.33 % 4.03 Key Unknown Sheffeld 56 bt % 5.05 % 4.93 %.45 %.05 Key Unknown Sheffeld 5 bt % 4.97 % 4. % 0.44 %.0 Key Unknown Sheffeld 64 bt %.40 % 4.50 % 9.38 % 0.68 Key Stolen Sheffeld 8 bt %.9 % 4.30 % 7.5 % 9.7 Key Stolen Sheffeld 56 bt %.53 %.0 % 6.96 % 7.80 Key Stolen Sheffeld 5 bt % 3.47 %.55 % 9.7 % 8. Key Stolen Sheffeld par-wse dstances between feature vectors n the reduced dmenson subspace n comparson wth the tradtonal random projecton matrx. The best results are usually obtaned wth 8 or 56 bts. Besdes, we plot the detecton error trade-off DET) curves [3] for key stolen scenaro of the 56 bt face mage hash length wth the AT&T database n Fgure. Ngo et al. [], [] employs bnary quantzaton method wth a fxed threshold for all bts. Ths method may be suboptmal n some cases. The proposed GMM-based quantzaton method reduces EER most of the tme n comparson to the bnary quantzaton snce t fnds approxmately optmum threshold values for each bt poston. 4. Concluson In ths paper, we propose a novel face mage hashng method for a bometrc verfcaton system. The proposed method s based on dscrmnatve projecton selecton dependng on Fsher crtera. Another novelty of the proposed method s to employ bmodal GMM n the quantzaton step. The smulatons demonstrate that the proposed method has better performance n comparson wth the random projecton based face mage hashng methods proposed n the lterature. References Vdeo Tech., vol.6, no.6, June 006. [] D.C.L.Ngo, A.B.J. Teoh, A. Goh, Egenspace-based face hashng, Proc. of the Internatonal Conference on Bometrc Authentcaton ICBA) n LNCS, vol. 307, pp , Hong Kong, July 004. [3] C. Karabat, H.Erdogan, A Cancelable Bometrc Hashng for Secure Bometrc Verfcaton, Proc. IEEE IIH-MSP 009, pp , Kyoto, Japan, 009. [4] A. Lumn, L. Nann, An Improved BoHashng for Human Authentcaton, Proc. of the Pattern Recognton, vol.40, no.3, pp , March 007. [5] A.Kong, K.H. Cheung, D. Zhang, M. Kamel, J. You, An Analyss of BoHashng and Its Varants, Proc. of the Pattern Recognton, vol.39, no.7, pp , 007. [6] G. J. McLachlan ed., Dscrmnant Analyss and Statstcal Pattern Recognton, John Wley&Sons, Inc., 99. [7] R.O. Duda and P.E. Hart ed., Pattern Classfcaton and Scene Analyss, John Wley&Sons, Inc., 973. [8] R.W. Hammng, Error Detectng and Error Correctng Codes, Proc. Bell Sys. Techn. Journal, vol.9, no., pp.47-60, 950. [9] Cambrdge Unversy AT&T face database, [0] A.M.Martnez and R.Benavente, The AR Face Database, CVC Techncal report #4, June 998. [] Danel B. Graham and Ngerl M. Allnson, Face Recognton: From Theory to Applcatons, Proc. NATO ASI Seres F, Computer and Systems Scences, Vol. 63. H. Wechler, P.J.Phlps, V.Bruce, F. Fogelman-Soule and T.S. Huang ed.), pp , 998. [] Carnege Mellon Unversty Face Database, [3] A. Martn, G. Doddngton, T. Kamm, M. Ordowsk, and M.A. Przybock, The det curve n assessment of detecton task performance, n Proc. of Eurospeech Rhodes, Greece, Sept., 997, pp [] D.C.L.Ngo, A.B.J. Teoh, A.Goh, Bometrc Hash: Hgh Confdence Face Recognton, Proc. IEEE Trans. on Crc. and Sys. for
Face Recognition University at Buffalo CSE666 Lecture Slides Resources:
Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural
More informationContent Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers
IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth
More informationMULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION
MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and
More informationUser Authentication Based On Behavioral Mouse Dynamics Biometrics
User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA
More informationLearning the Kernel Parameters in Kernel Minimum Distance Classifier
Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department
More informationA Binarization Algorithm specialized on Document Images and Photos
A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a
More informationAn Image Fusion Approach Based on Segmentation Region
Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua
More informationParallelism for Nested Loops with Non-uniform and Flow Dependences
Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr
More informationA Fast Content-Based Multimedia Retrieval Technique Using Compressed Data
A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,
More informationFeature Reduction and Selection
Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components
More informationTerm Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task
Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto
More informationClassifier Selection Based on Data Complexity Measures *
Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.
More informationCluster Analysis of Electrical Behavior
Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School
More informationFace Recognition Based on SVM and 2DPCA
Vol. 4, o. 3, September, 2011 Face Recognton Based on SVM and 2DPCA Tha Hoang Le, Len Bu Faculty of Informaton Technology, HCMC Unversty of Scence Faculty of Informaton Scences and Engneerng, Unversty
More informationThe Research of Support Vector Machine in Agricultural Data Classification
The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou
More informationFeature Extraction and Dimensionality Reduction in SVM Speaker Recognition
Feature Extracton and Dmensonalty Reducton n SVM Speaker Recognton hembsle Mazbuko, themb@crg.ee.uct.ac.za Danel J. Mashao, danel@ebe.uct.ac.za Department of Electrcal Engneerng, Unversty of Cape own Abstract
More informationA Fast Visual Tracking Algorithm Based on Circle Pixels Matching
A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng
More informationSLAM Summer School 2006 Practical 2: SLAM using Monocular Vision
SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,
More informationModular PCA Face Recognition Based on Weighted Average
odern Appled Scence odular PCA Face Recognton Based on Weghted Average Chengmao Han (Correspondng author) Department of athematcs, Lny Normal Unversty Lny 76005, Chna E-mal: hanchengmao@163.com Abstract
More informationRecognizing Faces. Outline
Recognzng Faces Drk Colbry Outlne Introducton and Motvaton Defnng a feature vector Prncpal Component Analyss Lnear Dscrmnate Analyss !"" #$""% http://www.nfotech.oulu.f/annual/2004 + &'()*) '+)* 2 ! &
More informationNovel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition
Mathematcal Methods for Informaton Scence and Economcs Novel Pattern-based Fngerprnt Recognton Technque Usng D Wavelet Decomposton TUDOR BARBU Insttute of Computer Scence of the Romanan Academy T. Codrescu,,
More informationLarge-scale Web Video Event Classification by use of Fisher Vectors
Large-scale Web Vdeo Event Classfcaton by use of Fsher Vectors Chen Sun and Ram Nevata Unversty of Southern Calforna, Insttute for Robotcs and Intellgent Systems Los Angeles, CA 90089, USA {chensun nevata}@usc.org
More informationHuman Face Recognition Using Generalized. Kernel Fisher Discriminant
Human Face Recognton Usng Generalzed Kernel Fsher Dscrmnant ng-yu Sun,2 De-Shuang Huang Ln Guo. Insttute of Intellgent Machnes, Chnese Academy of Scences, P.O.ox 30, Hefe, Anhu, Chna. 2. Department of
More informationProblem Definitions and Evaluation Criteria for Computational Expensive Optimization
Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty
More informationX- Chart Using ANOM Approach
ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are
More informationFeature Extractions for Iris Recognition
Feature Extractons for Irs Recognton Jnwook Go, Jan Jang, Yllbyung Lee, and Chulhee Lee Department of Electrcal and Electronc Engneerng, Yonse Unversty 134 Shnchon-Dong, Seodaemoon-Gu, Seoul, KOREA Emal:
More informationSubspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;
Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features
More informationUsing Fuzzy Logic to Enhance the Large Size Remote Sensing Images
Internatonal Journal of Informaton and Electroncs Engneerng Vol. 5 No. 6 November 015 Usng Fuzzy Logc to Enhance the Large Sze Remote Sensng Images Trung Nguyen Tu Huy Ngo Hoang and Thoa Vu Van Abstract
More informationModeling Inter-cluster and Intra-cluster Discrimination Among Triphones
Modelng Inter-cluster and Intra-cluster Dscrmnaton Among Trphones Tom Ko, Bran Mak and Dongpeng Chen Department of Computer Scence and Engneerng The Hong Kong Unversty of Scence and Technology Clear Water
More informationDetection of an Object by using Principal Component Analysis
Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,
More informationISSN: International Journal of Engineering and Innovative Technology (IJEIT) Volume 1, Issue 4, April 2012
Performance Evoluton of Dfferent Codng Methods wth β - densty Decodng Usng Error Correctng Output Code Based on Multclass Classfcaton Devangn Dave, M. Samvatsar, P. K. Bhanoda Abstract A common way to
More informationLocal Quaternary Patterns and Feature Local Quaternary Patterns
Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents
More informationAn efficient method to build panoramic image mosaics
An effcent method to buld panoramc mage mosacs Pattern Recognton Letters vol. 4 003 Dae-Hyun Km Yong-In Yoon Jong-Soo Cho School of Electrcal Engneerng and Computer Scence Kyungpook Natonal Unv. Abstract
More informationS1 Note. Basis functions.
S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type
More informationTsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance
Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for
More informationPCA Based Gait Segmentation
Honggu L, Cupng Sh & Xngguo L PCA Based Gat Segmentaton PCA Based Gat Segmentaton Honggu L, Cupng Sh, and Xngguo L 2 Electronc Department, Physcs College, Yangzhou Unversty, 225002 Yangzhou, Chna 2 Department
More informationFeature Extraction and Test Algorithm for Speaker Verification
Feature Extracton and Test Algorthm for Speaker Verfcaton Wu Guo, Renhua Wang and Lrong Da Unversty of Scence and Technology of Chna, Hefe guowu@mal.ustc.edu.cn,{rhw, lrda}@ustc,edu.cn Abstract. In ths
More informationTN348: Openlab Module - Colocalization
TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages
More informationDetermining the Optimal Bandwidth Based on Multi-criterion Fusion
Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn
More informationPositive Semi-definite Programming Localization in Wireless Sensor Networks
Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer
More informationFEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur
FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents
More informationLoad Balancing for Hex-Cell Interconnection Network
Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,
More informationOutline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:
Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A
More informationEnhanced Watermarking Technique for Color Images using Visual Cryptography
Informaton Assurance and Securty Letters 1 (2010) 024-028 Enhanced Watermarkng Technque for Color Images usng Vsual Cryptography Enas F. Al rawashdeh 1, Rawan I.Zaghloul 2 1 Balqa Appled Unversty, MIS
More informationClassification of Face Images Based on Gender using Dimensionality Reduction Techniques and SVM
Classfcaton of Face Images Based on Gender usng Dmensonalty Reducton Technques and SVM Fahm Mannan 260 266 294 School of Computer Scence McGll Unversty Abstract Ths report presents gender classfcaton based
More informationA fast algorithm for color image segmentation
Unersty of Wollongong Research Onlne Faculty of Informatcs - Papers (Arche) Faculty of Engneerng and Informaton Scences 006 A fast algorthm for color mage segmentaton L. Dong Unersty of Wollongong, lju@uow.edu.au
More informationRelated-Mode Attacks on CTR Encryption Mode
Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory
More informationResearch and Application of Fingerprint Recognition Based on MATLAB
Send Orders for Reprnts to reprnts@benthamscence.ae The Open Automaton and Control Systems Journal, 205, 7, 07-07 Open Access Research and Applcaton of Fngerprnt Recognton Based on MATLAB Nng Lu* Department
More informationCS 534: Computer Vision Model Fitting
CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust
More informationNUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS
ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana
More informationMachine Learning 9. week
Machne Learnng 9. week Mappng Concept Radal Bass Functons (RBF) RBF Networks 1 Mappng It s probably the best scenaro for the classfcaton of two dataset s to separate them lnearly. As you see n the below
More informationCorner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity
Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent
More informationOnline Detection and Classification of Moving Objects Using Progressively Improving Detectors
Onlne Detecton and Classfcaton of Movng Objects Usng Progressvely Improvng Detectors Omar Javed Saad Al Mubarak Shah Computer Vson Lab School of Computer Scence Unversty of Central Florda Orlando, FL 32816
More informationSmoothing Spline ANOVA for variable screening
Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory
More informationSecurity Enhanced Dynamic ID based Remote User Authentication Scheme for Multi-Server Environments
Internatonal Journal of u- and e- ervce, cence and Technology Vol8, o 7 0), pp7-6 http://dxdoorg/07/unesst087 ecurty Enhanced Dynamc ID based Remote ser Authentcaton cheme for ult-erver Envronments Jun-ub
More informationType-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data
Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES
More informationUnsupervised Learning
Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and
More informationFacial Expression Recognition Based on Local Binary Patterns and Local Fisher Discriminant Analysis
WSEAS RANSACIONS on SIGNAL PROCESSING Shqng Zhang, Xaomng Zhao, Bcheng Le Facal Expresson Recognton Based on Local Bnary Patterns and Local Fsher Dscrmnant Analyss SHIQING ZHANG, XIAOMING ZHAO, BICHENG
More informationy and the total sum of
Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton
More informationSixth Indian Conference on Computer Vision, Graphics & Image Processing
Sxth Indan Conference on Computer Vson, Graphcs & Image Processng Incorporatng Cohort Informaton for Relable Palmprnt Authentcaton Ajay Kumar Bometrcs Research Laboratory, Department of Electrcal Engneerng
More informationPerformance Evaluation of Information Retrieval Systems
Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence
More informationEmpirical Distributions of Parameter Estimates. in Binary Logistic Regression Using Bootstrap
Int. Journal of Math. Analyss, Vol. 8, 4, no. 5, 7-7 HIKARI Ltd, www.m-hkar.com http://dx.do.org/.988/jma.4.494 Emprcal Dstrbutons of Parameter Estmates n Bnary Logstc Regresson Usng Bootstrap Anwar Ftranto*
More informationA PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION
1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute
More informationFeature Selection for Target Detection in SAR Images
Feature Selecton for Detecton n SAR Images Br Bhanu, Yngqang Ln and Shqn Wang Center for Research n Intellgent Systems Unversty of Calforna, Rversde, CA 95, USA Abstract A genetc algorthm (GA) approach
More information2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements
Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.
More informationAn Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices
Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal
More informationReal-time Joint Tracking of a Hand Manipulating an Object from RGB-D Input
Real-tme Jont Tracng of a Hand Manpulatng an Object from RGB-D Input Srnath Srdhar 1 Franzsa Mueller 1 Mchael Zollhöfer 1 Dan Casas 1 Antt Oulasvrta 2 Chrstan Theobalt 1 1 Max Planc Insttute for Informatcs
More informationA Novel Video Retrieval Method Based on Web Community Extraction Using Features of Video Materials
IEICE TRANS. FUNDAMENTALS, VOL.E92 A, NO.8 AUGUST 2009 1961 PAPER Specal Secton on Sgnal Processng A Novel Vdeo Retreval Method Based on Web Communty Extracton Usng Features of Vdeo Materals Yasutaka HATAKEYAMA
More informationCoding Artifact Reduction Using Edge Map Guided Adaptive and Fuzzy Filter
MEL A MITSUBISHI ELECTIC ESEACH LABOATOY http://www.merl.com Codng Artfact educton Usng Edge Map Guded Adaptve and Fuzzy Flter Hao-Song Kong Yao Ne Anthony Vetro Hufang Sun Kenneth E. Barner T-2004-056
More informationA Background Subtraction for a Vision-based User Interface *
A Background Subtracton for a Vson-based User Interface * Dongpyo Hong and Woontack Woo KJIST U-VR Lab. {dhon wwoo}@kjst.ac.kr Abstract In ths paper, we propose a robust and effcent background subtracton
More informationRobust visual tracking based on Informative random fern
5th Internatonal Conference on Computer Scences and Automaton Engneerng (ICCSAE 205) Robust vsual trackng based on Informatve random fern Hao Dong, a, Ru Wang, b School of Instrumentaton Scence and Opto-electroncs
More informationOutline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1
4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:
More informationImprovement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration
Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,
More informationCompiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz
Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster
More informationA Computer Vision System for Automated Container Code Recognition
A Computer Vson System for Automated Contaner Code Recognton Hsn-Chen Chen, Chh-Ka Chen, Fu-Yu Hsu, Yu-San Ln, Yu-Te Wu, Yung-Nen Sun * Abstract Contaner code examnaton s an essental step n the contaner
More informationAnalysis of Continuous Beams in General
Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,
More informationDiscriminative Dictionary Learning with Pairwise Constraints
Dscrmnatve Dctonary Learnng wth Parwse Constrants Humn Guo Zhuoln Jang LARRY S. DAVIS UNIVERSITY OF MARYLAND Nov. 6 th, Outlne Introducton/motvaton Dctonary Learnng Dscrmnatve Dctonary Learnng wth Parwse
More informationHigh Dimensional Data Clustering
Hgh Dmensonal Data Clusterng Charles Bouveyron 1,2, Stéphane Grard 1, and Cordela Schmd 2 1 LMC-IMAG, BP 53, Unversté Grenoble 1, 38041 Grenoble Cede 9, France charles.bouveyron@mag.fr, stephane.grard@mag.fr
More informationHigh-Boost Mesh Filtering for 3-D Shape Enhancement
Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,
More informationOutline. Type of Machine Learning. Examples of Application. Unsupervised Learning
Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton
More informationEnhanced AMBTC for Image Compression using Block Classification and Interpolation
Internatonal Journal of Computer Applcatons (0975 8887) Volume 5 No.0, August 0 Enhanced AMBTC for Image Compresson usng Block Classfcaton and Interpolaton S. Vmala Dept. of Comp. Scence Mother Teresa
More informationUnsupervised Learning and Clustering
Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned
More informationAppearance-based Statistical Methods for Face Recognition
47th Internatonal Symposum ELMAR-2005, 08-10 June 2005, Zadar, Croata Appearance-based Statstcal Methods for Face Recognton Kresmr Delac 1, Mslav Grgc 2, Panos Latss 3 1 Croatan elecom, Savsa 32, Zagreb,
More informationA Robust Method for Estimating the Fundamental Matrix
Proc. VIIth Dgtal Image Computng: Technques and Applcatons, Sun C., Talbot H., Ourseln S. and Adraansen T. (Eds.), 0- Dec. 003, Sydney A Robust Method for Estmatng the Fundamental Matrx C.L. Feng and Y.S.
More informationComputer Aided Drafting, Design and Manufacturing Volume 25, Number 2, June 2015, Page 14
Computer Aded Draftng, Desgn and Manufacturng Volume 5, Number, June 015, Page 14 CADDM Face Recognton Algorthm Fusng Monogenc Bnary Codng and Collaboratve Representaton FU Yu-xan, PENG Lang-yu College
More informationWireless Sensor Network Localization Research
Sensors & Transducers 014 by IFSA Publshng, S L http://wwwsensorsportalcom Wreless Sensor Network Localzaton Research Lang Xn School of Informaton Scence and Engneerng, Hunan Internatonal Economcs Unversty,
More informationR s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes
SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges
More informationUB at GeoCLEF Department of Geography Abstract
UB at GeoCLEF 2006 Mguel E. Ruz (1), Stuart Shapro (2), June Abbas (1), Slva B. Southwck (1) and Davd Mark (3) State Unversty of New York at Buffalo (1) Department of Lbrary and Informaton Studes (2) Department
More informationFrom Comparing Clusterings to Combining Clusterings
Proceedngs of the Twenty-Thrd AAAI Conference on Artfcal Intellgence (008 From Comparng Clusterngs to Combnng Clusterngs Zhwu Lu and Yuxn Peng and Janguo Xao Insttute of Computer Scence and Technology,
More informationFACIAL FEATURE EXTRACTION TECHNIQUES FOR FACE RECOGNITION
Journal of omputer Scence 10 (12): 2360-2365, 2014 ISSN: 1549-3636 2014 Rahb H. Abyev, hs open access artcle s dstrbuted under a reatve ommons Attrbuton (-BY) 3.0 lcense do:10.3844/jcssp.2014.2360.2365
More informationProper Choice of Data Used for the Estimation of Datum Transformation Parameters
Proper Choce of Data Used for the Estmaton of Datum Transformaton Parameters Hakan S. KUTOGLU, Turkey Key words: Coordnate systems; transformaton; estmaton, relablty. SUMMARY Advances n technologes and
More informationA New Approach For the Ranking of Fuzzy Sets With Different Heights
New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays
More informationAn Improved Image Segmentation Algorithm Based on the Otsu Method
3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,
More informationIncremental MQDF Learning for Writer Adaptive Handwriting Recognition 1
200 2th Internatonal Conference on Fronters n Handwrtng Recognton Incremental MQDF Learnng for Wrter Adaptve Handwrtng Recognton Ka Dng, Lanwen Jn * School of Electronc and Informaton Engneerng, South
More informationTwo-Dimensional Supervised Discriminant Projection Method For Feature Extraction
Appl. Math. Inf. c. 6 No. pp. 8-85 (0) Appled Mathematcs & Informaton cences An Internatonal Journal @ 0 NP Natural cences Publshng Cor. wo-dmensonal upervsed Dscrmnant Proecton Method For Feature Extracton
More informationREFRACTION. a. To study the refraction of light from plane surfaces. b. To determine the index of refraction for Acrylic and Water.
Purpose Theory REFRACTION a. To study the refracton of lght from plane surfaces. b. To determne the ndex of refracton for Acrylc and Water. When a ray of lght passes from one medum nto another one of dfferent
More informationRECOGNIZING GENDER THROUGH FACIAL IMAGE USING SUPPORT VECTOR MACHINE
Journal of Theoretcal and Appled Informaton Technology 30 th June 06. Vol.88. No.3 005-06 JATIT & LLS. All rghts reserved. ISSN: 99-8645 www.jatt.org E-ISSN: 87-395 RECOGNIZING GENDER THROUGH FACIAL IMAGE
More informationIdentification of a Person using Multimodal Biometric System
Internatonal Journal of Computer Applcatons (0975 8887) Identfcaton of a Person usng Multmodal Bometrc System.C. Sabareeswar M.E-Appled Electroncs Anna Unversty, runelvel. S. Lenty Stuwart Assstant Professor,
More informationComparison Study of Textural Descriptors for Training Neural Network Classifiers
Comparson Study of Textural Descrptors for Tranng Neural Network Classfers G.D. MAGOULAS (1) S.A. KARKANIS (1) D.A. KARRAS () and M.N. VRAHATIS (3) (1) Department of Informatcs Unversty of Athens GR-157.84
More informationK-means and Hierarchical Clustering
Note to other teachers and users of these sldes. Andrew would be delghted f you found ths source materal useful n gvng your own lectures. Feel free to use these sldes verbatm, or to modfy them to ft your
More information