Local Ridge Regression for Face Recognition

Size: px
Start display at page:

Download "Local Ridge Regression for Face Recognition"

Transcription

1 Local Rge Regresson for Face Recognton Hu Xue 1,2 Yulan Zhu 1 Songcan Chen *1,2 1 Department of Computer Scence & Engneerng, Nanjng Unversty of Aeronautcs & Astronautcs, , Nanjng, P.R. Chna 2 State Key Lab. for Novel Software echnology, Nanjng Unversty, P.R. Chna Abstract: Rge regresson (RR) for classfcaton s a regularze least square metho to moel the lnear epenency between covarate varables an labels. By applyng approprate technques to encoe the multvarate labels n face recognton as the vertces of the regular smplex whch can separate ponts wth hghest egree of symmetry, RR maps the face mages nto a face subspace where the mages from each nvual wll locate near ther nvual targets. However, as a holstc metho, RR operates rectly on a whole face regon represente as a vector an thus can not effectvely recognze the faces wth llumnaton varatons an partal occlusons. In ths paper, we present a novel algorthm, terme as Local Rge Regresson (LRR). Dfferent from RR, LRR emphaszes on each local face regon matchng rather than the whole. As a result, LRR can not only enhance the robustness to the local varatons by utlzng the spatal an geometrcal nformaton of facal components, but also avo the mensonalty reucton n the holstc RR as a preprocessng. Furthermore, an effcent cross-valaton algorthm s aopte to select the regularzaton parameters n each local regon. Experments on two stanar face atabases emonstrate that the propose algorthm sgnfcantly outperforms RR an the two popular lnear face recognton technques (Egenface an Fsherface). Although we concentrate on rge regresson n ths paper, followng the propose lne of the research, many current mult-category classfers can also be apple n face recognton through combnng the characterstcs of face mages an may be obtan better recognton accuraces. * Corresponng author: el: Ext ; Fax: ; E-mal: s.chen@nuaa.eu.cn (S. Chen), xuehu@nuaa.eu.cn (H. Xue) an lany_1999@nuaa.eu.cn (Y. Zhu)

2 Keywors: Rge regresson (RR); Local nformaton; Local rge regresson (LRR); Face recognton 1. Introucton In the past few ecaes, face recognton has become a hot ssue of research n computer vson communty. Among varous evelope technques n ths fel, appearance-base metho s one of the most wely use technques whch usually represents a face mage as a hgh mensonal vector of pxels[1]. o overcome the ffculty ncurre by hgh menson, a lot of subspace methos have been propose where Egenface[2] an Fsherface[3] are two of the most popular algorthms. Egenface s an unsupervse metho whch utlzes the ea of Prncple Component Analyss (PCA) to project the orgnal hgh mensonal ata onto a low mensonal subspace that can maxmally preserve orgnal mage nformaton[4]. Fsherface s a supervse algorthm whch combnes PCA an Lnear Dscrmnant Analyss (LDA) to extract the most scrmnant features that can maxmally separate the mages of fferent classes n the resultant face subspace. Although Fsherface further ntrouces the class nformaton compare to Egenface, t s affecte heavly by the relatve postons of the labele tranng mages ue to the weakness of LDA. An et al.[1] have ncate that n the mult-category face recognton problems, whle LDA tres to maxmze the between-class stances an mnmze the wthn-class stances smultaneously, the parwse stances can be sgnfcantly unbalance an ths may result n ba performance for classes wth small parwse between-class stances n the reuce subspace. Recently, a new generalze Rge Regresson (RR) metho[1] has been propose to solve the latent problem of Fsherface. Motvate by the fact that the m vertces of a regular m-smplex s the most balance an symmetrc separate ponts n the (m-1)-mensonal space, the metho frst encoes the targets for m stnct nvuals as the m vertces an then apples the rge regresson to map the tranng face mages nto the (m-1)-mensonal subspace so as to the mages from each

3 nvual wll locate near ther nvual targets[1]. Recognton s performe by mappng the new face mages nto the subspace an comparng ts stance to all the targets. However, although RR has yele much better recognton performance than Egenface an Fsherface expermentally, lke the two methos, RR s also a holstc technque whch operates rectly on a whole face regon an neglects the local nformaton. As a result, RR s senstve to the local varatons n face mages, such as llumnaton varatons an partal occlusons. For usng as much local nformaton hen n face mages as possble to relax the nfluence of local varaton for recognton, recently varous local regon matchng technques have been evelope[4-10]. he general ea of local regon matchng technques s to frst locate several facal features (components), an then classfy the faces by comparng an combnng the corresponng local statstcs[11]. Hesele et al.[12] further ncate that comparng the component (local) system an the global systems, the former outperforms the latter n recognton rate larger than 60%[11]. Consequently, n ths paper, we also apply local regon matchng technques nto RR an present a novel algorthm, terme as Local Rge Regresson (LRR). LRR frst parttons an orgnally whole mage nto L equally sze local regons n non-overlappng or overlappng ways, an then collects all those local regons sharng the same orgnal feature components respectvely from the tranng set to compose L corresponng local regon tranng sets. Rge regresson s performe on each of such L local regon sets to rectly tran fferent L classfers by the fferent face features. he new unlabele face mage s entfe by also parttonng nto L local regons n the same way as the tranng phase an classfyng each local regon by the corresponng classfer. he fnal recognton result s obtane by assemblng the total L results from the L classfers an votng. In ths way, not only s the spatal an geometrcal nformaton n a face mage preserve n each local regon, but also the nfluence of local varatons s restrane n several local regons by the classfers votng so as to greatly mprove the recognton accuracy. here are three major contrbutons of the propose LRR. Frst, LRR s more robust to the local varatons than the holstc RR. Secon, LRR can smultaneously

4 tran a set of classfers corresponng to fferent local regons an thus t s qute sutable for parallel computaton to greatly mprove the computatonal effcency of the holstc RR, especally n the large-sze mage cases. hr, the partton of local regons n LRR s nepenent on the menson of the face mage. Consequently, t can avo the mensonalty reucton n the holstc RR as a preprocessng. Furthermore, t can also escape the latent mensonal curse when the menson of the mages s qute large. he rest of the paper s organze as follows. In secton 2, we brefly revew the holstc RR metho. Secton 3 presents the propose LRR. Secton 4 proves expermental results on two face atabases to llustrate the superorty of LRR. Some conclusons are rawn n Secton Rge Regresson (RR) Suppose there are m nvuals for recognton. RR algorthm usually has three parts: labelng tranng mages, learnng classfer an recognton. Frstly, RR chooses the regular smplex vertces as the nvual targets an uses these targets as the multvarate labels of the tranng mages[1]. Let m 1 R ( 1, 2,, 1 2 = m ) are the vertces of one regular m-smplex an = [,,, ]. RR constructs as follows[1]: m 1. Let 1 = [1, 0,, 0] an 1, = 1/( m 1), for = 2,, m. 2. For 1 k m 2, k 2 k+ 1, k+ 1 1 = 1, k = k+ 1, k+ 1 k+ 1, j =, j = k+ 2,, m (1) m k 1 = k, 1 0, k + 1 < + m 1 hen, RR treats the face recognton as a rge regresson problem to locate the mages from each nvual as near ther nvual targets as possble. As a result, the task of learnng classfer n RR s to fn a matrx W that can moel the lnear

5 epenency between the mage x an the label Y, where Y = j f x belongs to the jth nvual, = 1,, n, j = 1,, m. Meanwhle, RR also penalzes the norm of W to reuce the varance of the estmate as the regularzaton term. herefore, the objectve functon (2) of RR s mnmzng n = J ( W) = Y W x + λ W (2) where λ s the regularzaton parameter to balance the bas an varance of the estmate. akng ervatve of (2) wth respect to the W an equalng t to zero, we have the matrx W as 1 ( W= XX +λi) XY (3) where X= [ x1,, xn] an Y= [ Y1,, Yn ]. Fnally, let x be a new mage. RR compares the stances between Wx an the nvual target an entfes x as that wth mnmal stance. 3. Local Rge Regresson (LRR) In ths secton, we propose a novel algorthm to solve the senstvty of local varatons n RR. Relatvely to the holstc RR, here we abuse the termnology,.e. local, to name the propose algorthm as Local Rge Regresson (LRR). Followng the lne of the research n the local regon matchng methos, LRR also nvolves three steps: local regon partton, classfer tranng an classfcaton. It s noteworthy that LRR avos the general menson reucton preprocessng n the classfer tranng phase ue to the partton, an thus t s much smpler Local Regon Partton Generally, there are two fferent technques to mplement the partton, that s, local components an local regons. Local components are areas occupe by the facal components, such as eyes, noses an mouths, an centere nepenently at the component centers; Local regons are local wnows centere at esgnate

6 coornates of a common coornate system[11]. Zou et al.[11] have verfe that comparson of corresponng local regons s better than comparng corresponng facal components. So, n ths paper, we aopt the smplest rectangular regons to partton mages, whch not only are convenently use but also can better preserve the spatal an geometrcal nformaton n the orgnal mages[4-6]. Suppose that there are n W1 W2 mages belongng to m nvuals n the tranng set, an these nvuals possess n 1, n 2,, nm face mages respectvely. Each mage s frst ve nto L equally sze local regons n a non-overlappng way whch are further concatenate nto corresponng column vectors wth mensonalty of W / 1 W L 2. hen we collect these vectors at the same poston of all face mages to form a specfc local regon tranng set, n ths way, L separate local regon sets are forme[5]. hs process s llustrate n Fg. 1. Non-overlappng partton sometmes may ve the relaton between each sub-mage an lea to totally neglect the relaton between local regons. Consequently, we also attempt the overlappng partton way whch can connect the ajacent local regons an combne the fferent nformaton n each regons. he process s llustrate n Fg. 2. In Secton 4, we wll verfy the conjecture that comparson of overlappng regons s better than comparng non-overlappng regons. Local Regon Set 1 Local Regon Set 2 Local Regon Set L Fg.1. he constructon of local regon face mage sets (Images are from the Extene Yale face atabase B[13])

7 Fg.2. he constructon of overlappng local regon face mage set (he mage s also from [13]) 3.2. Classfer ranng After parttonng L local regons, we can apply the rge regresson n each local regon set to fn the corresponng matrx W, 1,, = L. 1 W = [ X ( X ) +λi] X Y (4) where X = [ x 1, x 2,, x ] s the pxel matrx of the th local regon set of the n tranng mages, an Y= [ Y1,, Yn ] s the label matrx as n the orgnal RR algorthm. λ s the regularzaton parameter. Due to the partton nepenent on the menson of the mages, the LRR algorthm avos the menson reucton preprocessng n the orgnal RR an rectly learns the local classfers. Especally, the regularzaton parameter λ s a crucal hyper-parameter n the LRR whch controls the goo generalzaton performance of the trane classfer. Hence, here we wll scuss the opton of λ n etal. A popular way to estmate λ s cross-valaton. In k-fol cross-valaton, the tranng ataset s ranomly splt nto k sjont subsets. A classfer s trane for k tmes on stochastc k -1 subsets an a subset s left out as the valaton set to be use for estmatng the generalzaton error at the same tme[14]. Fnally, the classfer corresponng to the parameter wth the lowest average estmate rsk s chosen. However, the orgnal mplementaton of k-fol cross-valaton trans a prector for each splt of the ata an thus has much expensvely computatonal complexty f k s large[1]. An et al.[1] further evelope an effcent technque for general k-fol cross-valaton of the generalze RR wth

8 multvarate labels. So, here we also aopt ths technque to estmate λ n each local regon set. Generally, for a new mage x, we frst partton t nto L local regons as n the tranng mages. hen the corresponng precte label n each local regon s Y ( x) = ( W ) x = YX ( ) [ X( X) +λi] x 1 = Y X X +λi X x 1 [( ) ] ( ) ( A ) ( X ) x (5) where x enotes the pxel vector of the th local regon n the new mage, an 1 A = [( X ) X +λi] Y (6) In k-fol cross-valaton, we splt the ata nto k approxmately equally sze subsets { x,} nl 1, l = 1,, k. As n [1], we also splt the label matrx Y an A l = nto k sub-matrces as follows: where Y() l Yl,1, Yl,2,, Yl, nl = (1), (2),, ( k ) Y Y Y Y, A = A (1), A (2),, A ( k ) (7) =. Wthout loss of generalty, we leave the lth subset ase as the valaton set. hen the corresponng precte labels Y can be rectly compute as follows: () l cv Y = Y B A, l = 1,, k (8) () l 1 cv () l ll () l where B B B B B B [( X ) X + λi] B1k B2k Bkk k k 1 (9) n n j an B R, for, j = 1,, k. For more etals, the reaers can refer to [1]. j We entfy each local regon x cv of the valaton mages by comparng the stances from the precte labels Y () l to the nvual targets, = 1,, m. cv hen we sum up all recognton errors n the k-fol cross-valaton n the

9 corresponng local regon set an choose the optmal parameter λ wth the mnmal error for each local classfer. In summary, the proceure of classfer tranng n each local regon set can be formally state as follows: Input: he local regon set X = [ x 1, x 2,, x n ], = 1,, L Output: λ an W. 1. Label the multvarate label Y of x as the regular smplex vertces, j = 1,, m; j 2. Choose the regularzaton parameter λ : 2.1. Compute 2.2. Compute 1 X X + λi ; [( ) ] A an B ll from (6) an (9) respectvely; 2.3. Compute the precte label Y from (8), l = 1,, k ; () l cv 2.4. Ientfy the valaton mages x cv by comparng the stances from the precte labels () Y l cv to the nvual targets j, j = 1,, m; 2.5. Sum up all recognton errors n the k-fol cross-valaton an choose the optmal parameter λ wth the mnmal error; 3. Compute W from (4) Classfcaton For an unknown face mage x, we wll classfy t by classfers votng. In ths way, the nfluence of local varatons, such as llumnaton varatons an partal occlusons, wll be restrcte n the several local regons so as to greatly mprove the recognton robustness to the varatons. As escrbe n Secton 3.2, we frst partton x nto L local regons. hen n each local regon, the mage s entty s etermne by comparng the stances from the precte label prouce by the corresponng local classfer to the nvual targets s. Snce one classfcaton result for the unknown mage s generate nepenently n each local regon, there wll be total L results from L local regons. Let the probablty of the mage x belongng to the cth class s[6]

10 P c 1 L qc L = 1 = (10) where q c 1, = 0, f the th sub mage belongs to the cth class otherwse hen the fnal classfcaton result s Ientty( x ) = argmax( P ) (11) 1 c m c 4. Experments 4.1. Face Image Databases We carry out our experments on two face mage atabases: the AR face atabase[15] an the Extene Yale face atabase B[13]. he AR atabase contans 100 nvuals wth fferent facal expressons, llumnaton contons an occlusons. Each nvual has 26 face mages taken n two sessons. he frst sesson has 13 face mages name from 01 to 13, nclung neutral expresson (01), fferent facal expresson (02-04), fferent lghtng (05-07), occlusons wth sunglasses (08-10) an a scarf (11-13) uner fferent lghtng. he secon sesson exactly uplcates the frst sesson two weeks later[11]. For psychophyscal experments have ncate that eye s most mportant for recognton[16, 17], here we omt the pctures occlue by sunglasses. We use pctures n the frst sesson from each nvual as gallery. An our experments are conucte on four probe sets: AR11-13 ( 11, 12 an 13 pctures, occlusons wth a scarf n Sesson 1), AR15-17 ( 15, 16 an 17 pctures, fferent expressons n Sesson 2), AR18-20 ( 18, 19 an 20 pctures, fferent lghtng contons n Sesson 2), AR24-26 ( 24, 25 an 26 pctures, occlusons wth a scarf n Sesson 2). he 2000 mages are all croppe nto the same sze of pxels. he extene Yale face atabase B contans 38 nvuals an aroun 64 near frontal mages uner fferent llumnatons per nvual. All mage ata are manually algne, croppe an then resze to pxels just as n [1]. A ranom subset wth l (l=5, 10, 20, 30) mages per nvual s taken wth labels to form the

11 tranng set, an the rest of the atabase s the testng set[1]. For each gven l, the experments s repeate over 50 ranom splts by usng the matlab ata fles n [13], an the average results are reporte Evaluaton of Classfcaton Performance We compare the propose LRR wth the most popular face recognton methos: Egenface, Fsherface an the orgnally holstc RR on the two face atabases. In LRR, we attempt the fferent partton szes accorng to the fferent szes of the mages n the atabases by cross-valaton an the selecte expermental results are reporte. In the AR atabase, the szes of local regons are 11 8, an 33 24; An n the extene Yale atabase B, the szes are esgnate to 4 4, 8 8 an respectvely. We also attempt the non-overlappng an overlappng partton ways. In the overlappng way, the ajacent local regons overlap each other almost 50%. he corresponng classfcaton results are lste n able 1 an 2 respectvely, where Egenface, Fsherface an the holstc RR all nvolve the menson reucton an the optmal results are reporte. As shown n able 1, LRR s sgnfcantly superor to the other three holstc methos n all the probe sets n the AR atabase, bascally wthn all combnatons of the sze of local regons an non-overlappng or overlappng way. Especally n the AR11-13, AR18-20 an AR24-26 corresponng to fferent lght contons an occlusons wth a scarf, LRR shows the surprsngly hgh robustness to these local varatons an the optmal recognton error rates are less than 50% of those n the other algorthms. he smlar concluson can also be rawn n the extene Yale atabase B n able 2. he face mages n the atabase mostly have pose an llumnaton varatons. LRR also shows the best classfcaton performance corresponng to all the fferent tranng sets. Especally, when the numbers of the tranng mages are smaller, such as n 5ran an 10ran, LRR acheves much better recognton accuraces. Furthermore, t s noteworthy that, the optons of the approprate sze of local regons an overlappng way are stll open problems. However, here we can capture

12 some emprcal observatons about the optons n LRR. Obvously, from ables 1 an 2, n the two atabases, the optmal classfcaton performances of LRR are accomplshe both n the mle partton sze an n overlappng way, whch even excee the other combnatons n LRR over 50%. hese exactly accors wth our conjecture n Secton 3.1, that s, the mle-sze an overlappng partton way can connect more spatal an geometrcal nformaton n ajacent local regons. able 1. Classfcaton performance (error rate %) comparson on the AR face atabase AR11-13 AR15-17 AR18-20 AR24-26 Egenface Fsherface RR LRR 11 8, non-overlappng LRR 11 8, overlappng LRR 22 16, non-overlappng LRR 22 16, overlappng LRR 33 24, non-overlappng LRR 33 24, overlappng able 2. Classfcaton performance (error rate %) comparson on the extene Yale face atabase B 5 ran 10 ran 20 ran 30 ran Egenface Fsherface RR LRR 4 4, non-overlappng LRR 4 4, overlappng LRR 8 8, non-overlappng LRR 8 8, overlappng LRR 16 16, non-overlappng LRR 16 16, overlappng Conclusons In ths paper, we have propose a new face recognton technque LRR base on the nsghts of the orgnally holstc RR. o overcome the senstvty of RR to local varatons, LRR aopts the popular local regon matchng technques. As a result, LRR not only enhances the robustness to the varatons, but also effectvely avos the

13 latent mensonal curse when the menson of the mages s very large. Expermental results emonstrate the surprsngly goo classfcaton performance of LRR. It s worth to note that, although we concentrate on the mprovement of RR n the whole paper, the propose lne of the research about LRR s general. hrough combnng the spatal an geometrcal nformaton of facal components n local way, many current mult-category classfers can also be apple n face recognton an may obtan better recognton performance, whch eserves our future researches. Furthermore, tensor subspace moels are one of the moern research rectons n the face recognton. Many researches have showe that representng the mages as tensors of arbtrary orer can further mprove the performance of algorthms n most cases[18-24]. Consequently, how to generalze LRR to tensor learnng s another nterestng topc for future stuy. Acknowlegment hs work was supporte by Natonal Natural Scence Founatons of Chna ( an ) an Natural Scence Founatons of Jangsu Provnce (BK an BK2008xxx). We also thank Prof. Xaoyang an for benefcal scusson. References [1] S. An, W. Lu, an S. Venkatesh. Face recognton usng kernel rge regresson. CVPR, [2] M. urk an A. Pentlan. Egenfaces for recognton. J. Cogntve Neurosc,3(1): 71-86, [3] P. Belhumeur, J. Hespanha, an D. Kregman. Egenfances vs. fsherfaces: recognton usng class specfc lnear projecton. IEEE rans. Pattern Analyss an Machne Intellgence,19(7): , [4] S. Chen an Y. Zhu. Subpattern-base prncple component analyss. Patter Recognton,37(5): , [5] K. an an S. Chen. Aaptvely weghte sub-pattern PCA for face recognton. Neurocomputng,64: , [6] Q. Hong, S. Chen, an X. N. Sub-pattern canoncal correlaton analyss wth applcaton n face recognton. ACA AUOMAICA SINICA,34(1): 21-30, 2008 (In Chnese). [7] A. Pentlan, B. Moghaam, an. Starner. Vew-base an moular egenspaces for face recognton. CVPR: 84-91, [8] A.M. Martnez. Recognzng mprecsely localze, partcally occlue an expresson varant faces

14 from a sngle sample per class. IEEE rans. Pattern Analyss an Machne Intellgence,24(6): , [9] R. Sngh, M. Vatsa, an A. Noore. Face recognton wth sguse an sngle gallery mages. Image an Vson Computng, [10] X. Geng an Z. Zhou. Image regon selecton an ensemble for face recognton. JCS, [11] J. Zou, Q. J, an G. Nagy. A comparatve stuy of local matchng approach for face recognton. IEEE rans. on Image Processng,16(10): , [12] B. Hesele, P. Ho, J. Wu, an. Poggo. Face recognton: component-base versus global approaches. Comput. Vs. Image Unerstan.,91(1): 6-12, [13] he Extene Yale Face Database B. [14] R.O. Dua, P.E. Hart, an D.G. Stork. Pattern Classfcaton. Wley, [15] A.M. Martnez an R. Benavente. he AR Face Database. CVC echncal Report #24, June, [16] D.I. Perrett, P.A.J. Smth, D.D. Potter, A.J. Mstln, A.S. Hea, A.D. Mlner, an M.A. Jeeves. Vsual cells n the temporal cortex senstve to face vew an gaze recton. Proc. of the Royal Socety of Lonon. Seres B: Bologcal Scences,223: , [17] E. Bart an S. Ullman. Class-base feature matchng across unrestrcte transformatons. IEEE rans. on Pattern Analyss an Machne Intellgence,30(9): , [18] H. Zhou, Y. Yuan, an A.H. Saka. Applcaton of semantc features n face recognton. Patter Recognton,41(10): , [19] D. Xu, S. Yan, D. ao, S. Ln, an H. Zhang. Margnal Fsher analyss an ts varants for human gat recognton an content-base mage retreval. IEEE rans. on Image Processng,16(11): , [20] D. ao, X. L, X. Wu, an S.J. Maybank. General tensor scrmnant analyss an Gabor features for gat recognton. IEEE rans. on Pattern Analyss an Machne Intellgence,29(10): , [21] D. ao, X. L, X. Wu, W. Hu, an S.J. Maybank. Supervse tensor learnng. Knowlege an Informaton Systems,13(1): 1-42, [22]. Zhang, D. ao, an J. Yang. Dscrmnatve localty algnment. he 10th European Conference on Computer Vson (ECCV), [23] D. ao, J. Sun, J. Shen, X. Wu, X. L, S.J. Maybank, an C. Faloutsos. Bayesan tensor analyss. IEEE Internatonal Jont Conference on Neural Networks (IJCNN), [24]. Zhang, K. Huang, X. L, J. Yang, an D. ao. Dscrmnatve orthogonal neghborhoo preservng projectons for classfcaton. IEEE rans. on Systems, Man, an Cybernetcs, Part: B, 2008.

Recognizing Faces. Outline

Recognizing 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 information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning 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 information

Feature Reduction and Selection

Feature 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 information

Faces Recognition with Image Feature Weights and Least Mean Square Learning Approach

Faces Recognition with Image Feature Weights and Least Mean Square Learning Approach Faces Recognton wth Image Feature Weghts an Least Mean Square Learnng Approach We-L Fang, Yng-Kue Yang an Jung-Kue Pan Dept. of Electrcal Engneerng, Natonal Tawan Un. of Sc. & Technology, Tape, Tawan Emal:

More information

Computer Aided Drafting, Design and Manufacturing Volume 25, Number 2, June 2015, Page 14

Computer 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 information

Cluster Analysis of Electrical Behavior

Cluster 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 information

The Research of Support Vector Machine in Agricultural Data Classification

The 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 information

Human Face Recognition Using Generalized. Kernel Fisher Discriminant

Human 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 information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL 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 information

Two-Dimensional Supervised Discriminant Projection Method For Feature Extraction

Two-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 information

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

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 information

Competitive Sparse Representation Classification for Face Recognition

Competitive Sparse Representation Classification for Face Recognition Vol. 6, No. 8, 05 Compettve Sparse Representaton Classfcaton for Face Recognton Yng Lu Chongqng Key Laboratory of Computatonal Intellgence Chongqng Unversty of Posts and elecommuncatons Chongqng, Chna

More information

Face Recognition by Fusing Binary Edge Feature and Second-order Mutual Information

Face Recognition by Fusing Binary Edge Feature and Second-order Mutual Information Face Recognton by Fusng Bnary Edge Feature and Second-order Mutual Informaton Jatao Song, Bejng Chen, We Wang, Xaobo Ren School of Electronc and Informaton Engneerng, Nngbo Unversty of Technology Nngbo,

More information

Combination of Local Multiple Patterns and Exponential Discriminant Analysis for Facial Recognition

Combination of Local Multiple Patterns and Exponential Discriminant Analysis for Facial Recognition Sensors & ransducers 203 by IFSA http://.sensorsportal.com Combnaton of Local Multple Patterns and Exponental Dscrmnant Analyss for Facal Recognton, 2 Lfang Zhou, 2 Bn Fang, 3 Wesheng L, 3 Ldou Wang College

More information

Tone-Aware Sparse Representation for Face Recognition

Tone-Aware Sparse Representation for Face Recognition Tone-Aware Sparse Representaton for Face Recognton Lngfeng Wang, Huayu Wu and Chunhong Pan Abstract It s stll a very challengng task to recognze a face n a real world scenaro, snce the face may be corrupted

More information

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS P.G. Demdov Yaroslavl State Unversty Anatoly Ntn, Vladmr Khryashchev, Olga Stepanova, Igor Kostern EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS Yaroslavl, 2015 Eye

More information

Detection of an Object by using Principal Component Analysis

Detection 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 information

Collaboratively Regularized Nearest Points for Set Based Recognition

Collaboratively Regularized Nearest Points for Set Based Recognition Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

Face Recognition Based on SVM and 2DPCA

Face 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 information

K-means Clustering Algorithm in Projected Spaces

K-means Clustering Algorithm in Projected Spaces K-means Clusterng Algorthm n Projecte paces Alssar NAER, Dens HAMAD.A.. -U..C.O 50 rue F. Busson, BP 699, 68 Calas, France Emal: nasser@lasl.unv-lttoral.fr Chaban NAR ebanese Unversty E.F Rue Al-Arz, rpol

More information

Learning a Locality Preserving Subspace for Visual Recognition

Learning a Locality Preserving Subspace for Visual Recognition Learnng a Localty Preservng Subspace for Vsual Recognton Xaofe He *, Shucheng Yan #, Yuxao Hu, and Hong-Jang Zhang Mcrosoft Research Asa, Bejng 100080, Chna * Department of Computer Scence, Unversty of

More information

Discriminative Dictionary Learning with Pairwise Constraints

Discriminative 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 information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A 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 information

On Modeling Variations For Face Authentication

On Modeling Variations For Face Authentication On Modelng Varatons For Face Authentcaton Xaomng Lu Tsuhan Chen B.V.K. Vjaya Kumar Department of Electrcal and Computer Engneerng, Carnege Mellon Unversty Abstract In ths paper, we present a scheme for

More information

Object Recognition Based on Photometric Alignment Using Random Sample Consensus

Object Recognition Based on Photometric Alignment Using Random Sample Consensus Vol. 44 No. SIG 9(CVIM 7) July 2003 3 attached shadow photometrc algnment RANSAC RANdom SAmple Consensus Yale Face Database B RANSAC Object Recognton Based on Photometrc Algnment Usng Random Sample Consensus

More information

Level set segmentation using image second order statistics

Level set segmentation using image second order statistics Level set segmentaton usng mage secon orer statstcs Bo Ma, Yuwe Wu, Pe L Bejng Laboratory of Intellgent Informaton Technology, School of omputer Scence, Bejng Insttute of Technology (BIT), Bejng, P.R.

More information

Facial Expression Recognition Based on Local Binary Patterns and Local Fisher Discriminant Analysis

Facial 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 information

Lecture 4: Principal components

Lecture 4: Principal components /3/6 Lecture 4: Prncpal components 3..6 Multvarate lnear regresson MLR s optmal for the estmaton data...but poor for handlng collnear data Covarance matrx s not nvertble (large condton number) Robustness

More information

Face Detection with Deep Learning

Face Detection with Deep Learning Face Detecton wth Deep Learnng Yu Shen Yus122@ucsd.edu A13227146 Kuan-We Chen kuc010@ucsd.edu A99045121 Yzhou Hao y3hao@ucsd.edu A98017773 Mn Hsuan Wu mhwu@ucsd.edu A92424998 Abstract The project here

More information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. 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 information

Learning Depth from Single Still Images: Approximate Inference 1

Learning Depth from Single Still Images: Approximate Inference 1 Learnng Depth from Sngle Stll Images: Approxmate Inference 1 MS&E 211 course project Ashutosh Saxena, Ilya O. Ryzhov Channng Wong, Janln Wang June 7th, 2006 1 In ths report, Saxena, et. al. [1] somethng

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism 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 information

An Efficient Illumination Normalization Method with Fuzzy LDA Feature Extractor for Face Recognition

An Efficient Illumination Normalization Method with Fuzzy LDA Feature Extractor for Face Recognition www.mer.com Vol.2, Issue.1, pp-060-065 ISS: 2249-6645 An Effcent Illumnaton ormalzaton Meod w Fuzzy LDA Feature Extractor for Face Recognton Behzad Bozorgtabar 1, Hamed Azam 2 (Department of Electrcal

More information

Kernel Collaborative Representation Classification Based on Adaptive Dictionary Learning

Kernel Collaborative Representation Classification Based on Adaptive Dictionary Learning Internatonal Journal of Intellgent Informaton Systems 2018; 7(2): 15-22 http://www.scencepublshnggroup.com/j/js do: 10.11648/j.js.20180702.11 ISSN: 2328-7675 (Prnt); ISSN: 2328-7683 (Onlne) Kernel Collaboratve

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content 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 information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement 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 information

Identifying Efficient Kernel Function in Multiclass Support Vector Machines

Identifying Efficient Kernel Function in Multiclass Support Vector Machines Internatonal Journal of Computer Applcatons (0975 8887) Volume 8 No.8, August 0 Ientfng Effcent Kernel Functon n Multclass Support Vector Machnes R.Sangeetha Ph.D Research Scholar Department of Computer

More information

Classifier Selection Based on Data Complexity Measures *

Classifier 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 information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge Detection in Noisy Images Using the Support Vector Machines Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona

More information

PCA Based Gait Segmentation

PCA 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 information

Local Quaternary Patterns and Feature Local Quaternary Patterns

Local 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 information

Neurocomputing 101 (2013) Contents lists available at SciVerse ScienceDirect. Neurocomputing

Neurocomputing 101 (2013) Contents lists available at SciVerse ScienceDirect. Neurocomputing Neurocomputng (23) 4 5 Contents lsts avalable at ScVerse ScenceDrect Neurocomputng journal homepage: www.elsever.com/locate/neucom Localty constraned representaton based classfcaton wth spatal pyramd patches

More information

Learning a Class-Specific Dictionary for Facial Expression Recognition

Learning a Class-Specific Dictionary for Facial Expression Recognition BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 4 Sofa 016 Prnt ISSN: 1311-970; Onlne ISSN: 1314-4081 DOI: 10.1515/cat-016-0067 Learnng a Class-Specfc Dctonary for

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. 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 information

The Objective Function Value Optimization of Cloud Computing Resources Security

The Objective Function Value Optimization of Cloud Computing Resources Security Open Journal of Optmzaton, 2015, 4, 40-46 Publshe Onlne June 2015 n ScRes. http://www.scrp.org/journal/ojop http://x.o.org/10.4236/ojop.2015.42005 The Objectve Functon Value Optmzaton of Clou Computng

More information

Smoothing Spline ANOVA for variable screening

Smoothing 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 information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua 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 information

Gender Classification using Interlaced Derivative Patterns

Gender Classification using Interlaced Derivative Patterns Gender Classfcaton usng Interlaced Dervatve Patterns Author Shobernejad, Ameneh, Gao, Yongsheng Publshed 2 Conference Ttle Proceedngs of the 2th Internatonal Conference on Pattern Recognton (ICPR 2) DOI

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A 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 information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE 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 information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM 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 information

Announcements. Supervised Learning

Announcements. Supervised Learning Announcements See Chapter 5 of Duda, Hart, and Stork. Tutoral by Burge lnked to on web page. Supervsed Learnng Classfcaton wth labeled eamples. Images vectors n hgh-d space. Supervsed Learnng Labeled eamples

More information

BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION

BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION BAYESIAN MULTI-SOURCE DOMAIN ADAPTATION SHI-LIANG SUN, HONG-LEI SHI Department of Computer Scence and Technology, East Chna Normal Unversty 500 Dongchuan Road, Shangha 200241, P. R. Chna E-MAIL: slsun@cs.ecnu.edu.cn,

More information

Classification of Face Images Based on Gender using Dimensionality Reduction Techniques and SVM

Classification 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 information

Incremental MQDF Learning for Writer Adaptive Handwriting Recognition 1

Incremental 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 information

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010 Smulaton: Solvng Dynamc Models ABE 5646 Week Chapter 2, Sprng 200 Week Descrpton Readng Materal Mar 5- Mar 9 Evaluatng [Crop] Models Comparng a model wth data - Graphcal, errors - Measures of agreement

More information

Appearance-based Statistical Methods for Face Recognition

Appearance-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 information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace 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 information

Palmprint Recognition Using Directional Representation and Compresses Sensing

Palmprint Recognition Using Directional Representation and Compresses Sensing Research Journal of Appled Scences, Engneerng and echnology 4(22): 4724-4728, 2012 ISSN: 2040-7467 Maxwell Scentfc Organzaton, 2012 Submtted: March 31, 2012 Accepted: Aprl 30, 2012 Publshed: November 15,

More information

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples

More information

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity

Corner-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 information

WIRELESS CAPSULE ENDOSCOPY IMAGE CLASSIFICATION BASED ON VECTOR SPARSE CODING.

WIRELESS CAPSULE ENDOSCOPY IMAGE CLASSIFICATION BASED ON VECTOR SPARSE CODING. WIRELESS CAPSULE ENDOSCOPY IMAGE CLASSIFICATION BASED ON VECTOR SPARSE CODING Tao Ma 1, Yuexan Zou 1 *, Zhqang Xang 1, Le L 1 and Y L 1 ADSPLAB/ELIP, School of ECE, Pekng Unversty, Shenzhen 518055, Chna

More information

Efficient Load-Balanced IP Routing Scheme Based on Shortest Paths in Hose Model. Eiji Oki May 28, 2009 The University of Electro-Communications

Efficient Load-Balanced IP Routing Scheme Based on Shortest Paths in Hose Model. Eiji Oki May 28, 2009 The University of Electro-Communications Effcent Loa-Balance IP Routng Scheme Base on Shortest Paths n Hose Moel E Ok May 28, 2009 The Unversty of Electro-Communcatons Ok Lab. Semnar, May 28, 2009 1 Outlne Backgroun on IP routng IP routng strategy

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An 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 information

An Image Fusion Approach Based on Segmentation Region

An 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 information

Machine Learning 9. week

Machine 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 information

A Binarization Algorithm specialized on Document Images and Photos

A 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 information

Available online at Available online at Advanced in Control Engineering and Information Science

Available online at   Available online at   Advanced in Control Engineering and Information Science Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced

More information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

An efficient method to build panoramic image mosaics

An 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 information

Journal of Chemical and Pharmaceutical Research, 2014, 6(6): Research Article. A selective ensemble classification method on microarray data

Journal of Chemical and Pharmaceutical Research, 2014, 6(6): Research Article. A selective ensemble classification method on microarray data Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(6):2860-2866 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A selectve ensemble classfcaton method on mcroarray

More information

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET 1 BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET TZU-CHENG CHUANG School of Electrcal and Computer Engneerng, Purdue Unversty, West Lafayette, Indana 47907 SAUL B. GELFAND School

More information

Robust Kernel Representation with Statistical Local Features. for Face Recognition

Robust Kernel Representation with Statistical Local Features. for Face Recognition Robust Kernel Representaton wth Statstcal Local Features for Face Recognton Meng Yang, Student Member, IEEE, Le Zhang 1, Member, IEEE Smon C. K. Shu, Member, IEEE, and Davd Zhang, Fellow, IEEE Dept. of

More information

Recognizing Partially Occluded, Expression Variant Faces from Single Training Image per Person with SOM and soft knn Ensemble

Recognizing Partially Occluded, Expression Variant Faces from Single Training Image per Person with SOM and soft knn Ensemble IEEE Transactons on Neural Networs 1 Recognzng Partally Occluded, Expresson Varant Faces from Sngle Tranng Image per Person wth SOM and soft NN Ensemble Xaoyang Tan, Songcan Chen, Zh-Hua Zhou, Member,

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

More information

The Discriminate Analysis and Dimension Reduction Methods of High Dimension

The Discriminate Analysis and Dimension Reduction Methods of High Dimension Open Journal of Socal Scences, 015, 3, 7-13 Publshed Onlne March 015 n ScRes. http://www.scrp.org/journal/jss http://dx.do.org/10.436/jss.015.3300 The Dscrmnate Analyss and Dmenson Reducton Methods of

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining 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 information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1

A New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1 A New Feature of Unformty of Image Texture Drectons Concdng wth the Human Eyes Percepton Xng-Jan He, De-Shuang Huang, Yue Zhang, Tat-Mng Lo 2, and Mchael R. Lyu 3 Intellgent Computng Lab, Insttute of Intellgent

More information

Hyperspectral Image Classification Based on Local Binary Patterns and PCANet

Hyperspectral Image Classification Based on Local Binary Patterns and PCANet Hyperspectral Image Classfcaton Based on Local Bnary Patterns and PCANet Huzhen Yang a, Feng Gao a, Junyu Dong a, Yang Yang b a Ocean Unversty of Chna, Department of Computer Scence and Technology b Ocean

More information

General Regression and Representation Model for Face Recognition

General Regression and Representation Model for Face Recognition 013 IEEE Conference on Computer Vson and Pattern Recognton Workshops General Regresson and Representaton Model for Face Recognton Janjun Qan, Jan Yang School of Computer Scence and Engneerng Nanjng Unversty

More information

Orthogonal Complement Component Analysis for Positive Samples in SVM Based Relevance Feedback Image Retrieval

Orthogonal Complement Component Analysis for Positive Samples in SVM Based Relevance Feedback Image Retrieval Orthogonal Complement Component Analyss for ostve Samples n SVM Based Relevance Feedback Image Retreval Dacheng Tao and Xaoou Tang Department of Informaton Engneerng The Chnese Unversty of Hong Kong {dctao2,

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

Unsupervised Learning

Unsupervised 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 information

An Evaluation of Divide-and-Combine Strategies for Image Categorization by Multi-Class Support Vector Machines

An Evaluation of Divide-and-Combine Strategies for Image Categorization by Multi-Class Support Vector Machines An Evaluaton of Dvde-and-Combne Strateges for Image Categorzaton by Mult-Class Support Vector Machnes C. Demrkesen¹ and H. Cherf¹, ² 1: Insttue of Scence and Engneerng 2: Faculté des Scences Mrande Galatasaray

More information

Face Recognition using 3D Directional Corner Points

Face Recognition using 3D Directional Corner Points 2014 22nd Internatonal Conference on Pattern Recognton Face Recognton usng 3D Drectonal Corner Ponts Xun Yu, Yongsheng Gao School of Engneerng Grffth Unversty Nathan, QLD, Australa xun.yu@grffthun.edu.au,

More information

Computer Science Technical Report

Computer Science Technical Report Computer Scence echncal Report NLYSIS OF PCSED ND FISHER DISCRIMINNSED IMGE RECOGNIION LGORIHMS Wendy S. Yambor July echncal Report CS3 Computer Scence Department Colorado State Unversty Fort Collns, CO

More information

A Deflected Grid-based Algorithm for Clustering Analysis

A Deflected Grid-based Algorithm for Clustering Analysis A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan

More information

Semi-Supervised Discriminant Analysis Based On Data Structure

Semi-Supervised Discriminant Analysis Based On Data Structure IOSR Journal of Computer Engneerng (IOSR-JCE) e-issn: 2278-0661,p-ISSN: 2278-8727, Volume 17, Issue 3, Ver. VII (May Jun. 2015), PP 39-46 www.osrournals.org Sem-Supervsed Dscrmnant Analyss Based On Data

More information

Histogram-Enhanced Principal Component Analysis for Face Recognition

Histogram-Enhanced Principal Component Analysis for Face Recognition Hstogram-Enhanced Prncpal Component Analyss for Face ecognton Ana-ara Sevcenco and Wu-Sheng Lu Dept. of Electrcal and Computer Engneerng Unversty of Vctora sevcenco@engr.uvc.ca, wslu@ece.uvc.ca Abstract

More information

RECOGNITION AND AGE PREDICTION WITH DIGITAL IMAGES OF MISSING CHILDREN

RECOGNITION AND AGE PREDICTION WITH DIGITAL IMAGES OF MISSING CHILDREN RECOGNIION AND AGE PREDICION WIH DIGIAL IMAGES OF MISSING CHILDREN A Wrtng Project Presented to he Faculty of the Department of Computer Scence San Jose State Unversty In Partal Fulfllment of the Requrements

More information

Integrated Expression-Invariant Face Recognition with Constrained Optical Flow

Integrated Expression-Invariant Face Recognition with Constrained Optical Flow Integrated Expresson-Invarant Face Recognton wth Constraned Optcal Flow Chao-Kue Hseh, Shang-Hong La 2, and Yung-Chang Chen Department of Electrcal Engneerng, Natonal Tsng Hua Unversty, Tawan 2 Department

More information

Feature Extraction Based on Maximum Nearest Subspace Margin Criterion

Feature Extraction Based on Maximum Nearest Subspace Margin Criterion Neural Process Lett DOI 10.7/s11063-012-9252-y Feature Extracton Based on Maxmum Nearest Subspace Margn Crteron Y Chen Zhenzhen L Zhong Jn Sprnger Scence+Busness Meda New York 2012 Abstract Based on the

More information

IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 24, NO. 1, JANUARY

IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 24, NO. 1, JANUARY IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 24, NO. 1, JANUARY 2015 189 Dscrmnatve Shared Gaussan Processes for Multvew and Vew-Invarant Facal Expresson Recognton Stefanos Eleftherads, Student Member,

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem 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 information

Online Detection and Classification of Moving Objects Using Progressively Improving Detectors

Online 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 information

CS 534: Computer Vision Model Fitting

CS 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 information