Incremental MQDF Learning for Writer Adaptive Handwriting Recognition 1

Size: px
Start display at page:

Download "Incremental MQDF Learning for Writer Adaptive Handwriting Recognition 1"

Transcription

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 Chna Unversty of echnology, Guangzhou, P.R.Chna * E-Mal: lanwen.jn@gmal.com Abstract Wrter adaptaton has been proved to be an effectve approach to mprove the recognton performance of the wrter-ndependent recognzer for a partcular wrter. In ths paper, we propose a wrter adaptve handwrtng recognton approach by ncremental learnng the Modfed Quadratc Dscrmnant Functon (MQDF classfer. We derved the soluton of Incremental MQDF (IMQDF and then present a Dscrmnatve IMQDF (DIMQDF by dervng the soluton of IMQDF n the updated dscrmnatve feature space. Based on IMQDF or DIMQDF, the wrter adaptaton s fnally performed by updatng the MQDF recognzer adaptvely. he expermental results for recognzng handwrtng Chnese characters ndcate that the proposed IMQDF and DIQMDF approaches can reduce as much as 52.7% and 45.38% error rate respectvely on the wrter-dependent dataset whle only have less than 0.8% accuracy loss on the wrter-ndependent dataset. In other words, the proposed IMQDF and DIMQDF based wrter adaptaton approaches can sgnfcantly ncrease the recognton accuracy on wrter-dependent dataset whle only have lmted negatve nfluence for general wrter.. Introducton Wth the emergence of Personal Dgtal Assstants (PDA and of ablet PCs usng pen based nterfaces, the handwrtng recognton accuracy s the key factor n determnng the acceptablty of a handwrtng recognton system and the whole applcaton n whch t s mplemented. Motvated by these, many researchers devoted themselves to the feld of handwrtng character recognton and acheved great progress durng the past 40 years [-3]. However, recent researches [4-6] have shown that the recognton accuracy usng state-of-the-art technques cannot satsfy user s expectatons [7]. he tranng set of a wrter-ndependent recognzer s typcally composed of large sets of data produced by dfferent wrters to acheve a good performance for general wrter, but t s not optmal wth respect to any partcular wrter. A straghtforward dea to ncrease the recognton accuracy for a specfc user s adaptng the wrter-ndependent recognzer to the specfc wrtng style, whch s also known as wrter adaptaton [8-2]. Many the prevous works ether desgn for the small scale (small classes, small vocabulary recognton problem [8-0] or focus on feature transformaton module and gnore the classfer desgn module durng the wrter adaptaton process [-2]. In ths paper, we propose a wrter adaptaton approach for large scale onlne handwrtng Chnese character recognton by adaptvely updatng the Modfed Quadratc Dscrmnant Functon (MQDF recognzer, whch has been successful n achevng state-of-the-art performance for handwrtng character recognton [3-4]. However, the assumpton used n typcal mplementatons of the tradtonal MQDF, that a complete dataset for tranng s gven n advance, conflcts wth the prncples of the wrter adaptaton. o solve ths problem, we have derved a soluton of the MQDF model usng ncremental learnng (henceforth referred to as the IMQDF, whch enables the MQDF technque to be appled to the wrter adaptaton for the frst tme. On the other hand, n practcal, the raw feature s usually transformed to a dscrmnatve feature space to reduce the feature dmenson as well as mprove the recognton accuracy. In ths stuaton, we can drectly apply the IMQDF approach by assumng the dscrmnatve feature space s preserved durng the ncremental learnng process. hs work s supported n part by the research fundng of NSFC (no. U , , GDSP (no , 2007B , 2008A , 2009B /0 $ IEEE DOI 0.09/ICFHR

2 However, ths assumpton s not reasonable and may cause recognton accuracy decrease due to the dscrmnatve feature space s actually changed when the ncremental samples are provded. o solve ths problem, we proposed a Dscrmnatve IMQDF (DIMQDF to adaptvely learn the MQDF model n the updated dscrmnatve feature space. Based on these, the wrter adaptaton s carred out by adaptve updatng MQDF recognzer to learn the specfc wrter s wrtng styles. Expermental results on the wrter-dependent dataset verfy that both of the IMQDF and DIMQDF approaches are very effectve to mprove the recognton accuracy for a partcular wrter, and DIMQDF acheves better performance than IMQDF approach. Moreover, expermental results on the ndependent-dataset show that both of the proposed wrter adaptaton approaches have lmted accuracy loss for the general wrter. he remander of ths paper s organzed as follows: Secton2 frst revews the MQDF approach and then derves the soluton of IMQDF and DIMQDF. he WIMQDF/WDIMQDF approach s proposed n Secton3. And the proposed wrter adaptaton method s presented n Secton4. Secton5 gves expermental results and Secton6 concludes ths paper. 2. Incremental learnng of MQDF 2. MQDF classfer Based on a Bayesan decson rule that classfes the nput pattern to the class as the maxmum a posteror (MAP probablty of all the classes, the quadratc dscrmnant functon (QDF s obtaned assumng a multvarate Gaussan densty and an equal pror probablty for each class. he MQDF proposed by Kmura et al. [3] ncorporates a modfcaton to the QDF through a K-L transform and smoothng the mnor egenvalues to mprove the computaton effcency and classfcaton performance. he dscrmnant functon of a QDF classfer can be represented as [4]: g0 ( x, ω = ( x x ( x x + log Σ, =,, M ( where M s the number of classes and ω represents the th class, and x denotes the mean vector of class ω, D s the dmenson of x. By K-L transform, the covarance matrx can be dagonalzed as: Σ = ΦΛΦ (2 where Λ = dag[ λ,, λd] wth λ, j =,, D, j beng the egenvalues (ordered n decreasng order of Σ, and Φ = [ φ,, φ D ] wth φ, j j =,, D, beng the ordered egenvectors, Φ s orthonormal (untary such that Φ Φ = I. Accordng to (2, the QDF can be rewrtten n the form of egenvectors and egenvalues: g0 ( x, ω = [ Φ ( x x] Λ Φ ( x x + log Λ D D (3 [ ( ] 2 = ϕj x x + log λj λ j= j j= By replacng the mnor egenvalues wth a constantδ, the MQDF classfer can be expressed as: 2 K δ 2 g( x, ω = x x ( [ ϕj( x x] δ j= λj K + log λ + ( D Klogδ j= j (4 where K denotes the number of domnant egenvectors. Snce the tranng of the QDF classfer always underestmate the patterns egenvalues by lmted sample set, the mnor egenvalues become some knd of unstable noses and affect the classfer s robustness. By smoothng them n the MQDF classfer, not only the classfcaton performance s mproved, but also the computaton tme and storage for the parameters are saved. 2.2 Incremental learnng of MQDF (IMQDF Suppose X and Y are two sets of observatons n the feature space, where X s the gven observaton set wth N samples X= {x } (=,, N n M classes, and Y s a set of new observatons wth L ncremental samples Y = {y } j (j=,,l n P classes. It s worthwhle notng that some of classes n Y may be not avalable n X, n other words, some new classes may be ntroduced. hus, the mxed observaton set Z=X Y = {z k } (,,L+N has L+N samples n C classes, where C M and C P. Wthout loss of generalty, we assume that n of the N orgnal samples and l of the L ncremental samples belong to class C (=,, C, therefore, n the updated observaton set, the sample number belongng to the C class s s =n +l. Let x, Σ x, y, Σ represent the th class s mean vector y and covarance matrx of the gven observaton set and the new observaton set respectvely. hen we have n n k = l l k = x = x k (5 y = y k (6 560

3 Σ = x x x x = x x x x (7 x y n n ( k ( k k k n l l ( ( k k k k l Σ = y y y y = y y y y (8 Accordng to the above defntons, each class s mean vector z and covarance matrx Σ z of the mxed dataset can be updated as follows: s n l n ( x + l y = j = k + j = s n + l j= n + l s z z kzk z z s n l n ( x+ l y n ( l x+ y x k xk yk yk n+ l n+ l n+ l n+ l n n l l ( xk xk x xk ( yk yk y y n k k + l n = n+ l l = nl ( yy 2 x x x y y x ( n+ l n l nl ( y ( x y 2 x y x n+ l n+ l ( n+ l z z x y (9 Σ = = + = = Σ + Σ + (0 After updatng the mean vector and covarance matrx of each class, the covarance matrx of each class s dagonalzed by K-L transformaton to obtan the domnant egenvalues and egenvectors. Fnally the MQDF classfer can be updated accordng to Eq. ( Dscrmnatve IMQDF (DIMQDF In practcal, before usng MQDF, a dscrmnatve transform approach, such as Lnear Dscrmnatve Analyss (LDA [5], Prncpal Component Analyss (PCA[4] and General ensor Dscrmnatve Analyss (GDA[6], s usually employed to transform the raw feature to a dscrmnatve feature space to reduce the feature dmenson as well as mprove the recognton performance. After then, the MQDF classfer s performed n the dscrmnatve feature space. Due to the proposed IMQDF only can be used n the preserved feature space, therefore, when the dscrmnatve transform approach s used, the IMQDF approach can t be drectly employed wthout the assumpton that the dscrmnatve feature space s preserved durng the ncremental learnng process, but ths assumpton may cause performance loss. o solve ths problem, we proposed a Dscrmnatve IMQDF (DIMQDF to mprove performance of IMQDF by adaptvely learnng the MQDF classfer n the updated dscrmnatve feature space. Snce the LDA, also known as Fsher Dscrmnatve Analyss (FDA, s wdely used approach n the character recognton applcatons [,3,4,,2,4], n ths secton, we focus on how to ncremental learnng MQDF classfer n the LDA feature space. Suppose x, Σ x, y, Σ represents the th class s y mean vector and covarance matrx of the gven and new observaton sets respectvely n the raw feature space. And W orglda denotes the LDA transform matrx, whch s traned by the gven observaton data. When the ncremental samples are provded, the LDA transformaton matrx s updated as W nclda accordng to the ILDA approach [2]. After then, each class s mean vector y and nclda covarance matrx Σ y nclda of the ncremental data can be computed n the updated dscrmnatve feature space accordng to: y = W y ( -nclda nclda l y nclda ( Wnclda k -nclda ( Wnclda k -nclda Σ = y y y y (2 Smlar wth above steps, each class s mean vector x and covarance matrx Σ nclda x nclda of the gven data can be computed n the new dscrmnatve feature space as: x-nclda = W x (3 nclda l x nclda ( Wnclda xk Wnclda x( Wnclda xk Wnclda x l Wnclda ( xk x( xk x Wnclda Wnclda xwnclda Σ = = = Σ (4 Snce all the class s mean vector and the covarance matrx of the gven data and ncremental data are obtaned, each class s updated mean vector and covarance matrx of mxed dataset can be computed by nsertng Eq. (~(4 to Eq. (9 and Eq. (0 respectvely. After obtanng the domnant egenvalues and egenvectors of each class s covarance matrx by K-L transformaton, the MQDF classfer can be updated n the updated dscrmnatve feature space accordng to Eq. (4. 4. Wrter adaptaton Fgure depcts a flow dagram of the proposed WDIMQDF/DIMQDF based wrter adaptaton approach. he ntal MQDF modelng s traned usng the full wrter-ndependent tranng set. he preprocessng and feature extracton are performed frstly. After then, the feature space s transformed to a dscrmnatve feature space to reduce the feature dmenson and mprove the recognton accuracy by LDA. hereafter, to generate the MQDF recognzer for the wrter-ndependent dataset, the adaptaton 56

4 Wrter-ndependent ranng set Preprocessng & Feature Extracton LDA ransform Estmate each class s mean and covarance MQDF recognzer LDA {W orglda} ILDA {W nclda} IMQDF modelng Re-estmate each class s mean and covarance n updated dscrmnatve space Wrter-ndependent ranng set LDA ransform Estmate each class s mean and covarance Estmate each class s mean and covarance Smoothng of mnor egenvalues Updated MQDF recognzer Fgure An overvew of wrter adaptaton process (dashed lnes show the flow of events for wrter adaptaton approach tres to fnd a statstcal model and estmates the model s parameters to ft the feature dstrbuton of each class. When addtonal samples for a partcular wrter are provded, the LDA feature space s updated frst, and then the parameters of each class s feature he evaluaton of the proposed wrter adaptaton approach was conducted usng two subsets of SCU- COUCH handwrtng dataset [8], GB subset (henceforth referred to as CouchGB dataset and Word8888 subset. he GB subset contans 68 dstrbuton are re-estmated n the updated wrters samples of 3755 categores of Chnese dscrmnatve feature space to learn the wrtng styles of the partcular wrter adaptvely. Fnally, the updated MQDF recognzer s generated to mprove the recognton accuracy for the partcular wrter. characters, whereas the Word8888 subset conssts of 30 wrters samples of 8888 categores of most frequently used handwrtten words. Fgure2 shows some samples of the character 我 from the GB In the wrter-ndependent recognzer tranng subset and the words contans character 我 n process, the feature of the ndependent tranng dataset Word8888 subset. It s worthwhle notcng that many s frst extracted usng an 8-drectonal feature dfferent words contan same character. For example, extracton method [7]. And then, the parameters of all of the words n fgure2 (b contan the character each class s dstrbuton are estmated to generate the 我. hs ndcates that data collected n ths way, n wrter-ndependent MQDF recognzer accordng to the whch the same character s collected for many tmes methods descrbed n Secton 2.. n accordance wth the context of the word corpus, he accuracy of a recognton system for a provdes us wth partcularly realstc wrterndependent ncremental handwrtten samples. partcular wrter can be mproved by adjustng the recognzer s model to match the dstrbuton of the partcular wrter s handwrtng more closely, rather than usng the more generalzed models traned for a large number of wrters. herefore, our wrter adaptaton approach begns by adaptve updatng the new LDA feature space usng the ncremental samples. After transformng the gven samples dstrbuton parameters nto updated dscrmnatve feature space, we also compute the dstrbuton models of the wrterdependent dataset n the update dscrmnatve feature space. In ths way, the dstrbuton models of both the (a Character 我 (b Words contan 我 wrter-ndependent and wrter-dependent datasets are Fgure2: Samples from the SCU-COUCH database obtaned n the updated dscrmnatve feature space. o buld a general purpose classfer, we randomly hereafter, the dstrbuton models of the mxed selected 34 sets of data from the CouchGB to buld dataset can be generated accordng to Eq. (9 and (0. a wrter-ndependent baselne classfer, wth the Fnally, the updated MQDF recognzer s developed remanng 34 sets used to test the performance of the accordng to Eq. (4 usng the dstrbuton models of baselne classfer, as well as to evaluate the nfluence the mxed dataset. after the adaptaton. In order to generate partcularly 5. Experments and Analyss 5. Data preparaton and expermental setup realstc wrter-dependent ncremental handwrtten samples, we frst manually segmented all of the handwrtten words samples nto separate characters, whch results n 2078 categores of 9595 solated Chnese characters, to form a new dataset, the 562

5 Fgure 4 Performance of dfferent wrter adaptaton approaches for each partcular wrter IncCouchDB dataset. For each partcular wrter s handwrtten samples from the IncCouchDB dataset, we randomly selected 50% of each category s data for adaptve learnng the IMQDF model and then usng the remanng 50% to test the wrter adapton performance. It s worth notng that the two datasets don t share any common wrters. 5.2 Baselne Performance on CouchGB and IncCouchDB before wrter adaptaton After the classfer was traned by 34 sets of CouchGB data, ts performance was evaluated on both the testng sets of CouchGB dataset and IncCouchDB dataset respectvely. able shows the average recognton rate of the baselne MQDF classfer on these two datasets. he results demonstrate that, snce many of the wrtng styles of IncCouchDB are unseen n the tranng dataset, the recognton accuracy on IncCouchDB testng datasets s much lower than whch on CouchGB testng dataset. In other words, the recognton accuracy for the specfc wrter should be further mproved. able Baselne performance on two datasets op op5 op0 CouchGB 96.03% 99.30% 99.49% IncCouchDB 88.56% 95.93% 96.82% 5.3 Performance comparson wth dfferent wrter adaptaton approaches In ths experment, we examned the performance of the varous wrter adaptaton approaches based on ILDA [2] and the proposed IMQDF and DMQDF approaches on the wrter-dependent IncCouchDB dataset. It worth notng that snce the MQDF classfer s not used n ths experment n prevous method [2], we mplement an ILDA plus MQDF based wrter adaptaton approach (we refer t as ILDAMQDF, where the LDA transform matrx s frst updated usng the new ncremental data accordng to ILDA, and then the MQDF classfer model s transformed to the updated LDA feature space wthout employng the IMQDF approach. From the results gven n table 2, t can be seen that the ILDA and ILDAMQDF based wrter adaptaton can only reduce less than 26% error rate and acheve less than 9.0% recognton rate for the partcular wrter, whle the proposed IMQDF and DIMQDF based wrter adaptaton approaches can reduce above 45% error rate and acheve above 93.75% recognton rate. In other words, they can mprove the recognton accuracy from 88.56% to 93.75% and 94.59% respectvely. On the other hand, t can be also found that the DIMQDF sgnfcantly outperform IMQDF, ths ndcates that by ncrementally learnng the MQDF model n the updated dscrmnatve feature space, the performance of the wrter adaptaton can be further mproved. Generally, the expermental results demonstrate that the proposed IMQDF based wrter adaptaton approach s very effectve to mprove the recognton accuracy for the specfc wrter, and by adaptve learnng the MQDF recognzer n the updated dscrmnatve feature space, the performance of the wrter adaptaton can be further mproved. able2 Recognton performance of varous wrter adaptaton approaches Approach Intal Adapted Error rate reducton ILDA[2] 82.64% 87.08% 25.62% ILDAMQDF 88.56% 9.0% 2.39% IMQDF 88.56% 93.75% 45.38% DIMQDF 88.56% 94.59% 52.7% o examne the nfluence of wrter adaptaton approaches for total 30 dfferent wrters nvolved, fgure 4 shows the error rate reducton of these four adaptaton approaches for each partcular wrter. From the result gven n fgure 4, we can see that the proposed wrter adaptaton approaches sgnfcantly outperforms ILDA and ILDAMQDF approach for every partcular wrter. In addton to evaluatng the performance mprovement of the IMQDF and DIMQDF based wrter adaptaton approaches on the wrter-dependent dataset; we also examned the mpact of the proposed wrter adaptaton approaches for the general purpose 563

6 wrter, snce we don t expect the adaptaton to the partcular wrter s handwrtng style s at the cost of losng too much generalty for other wrter styles. he expermental result demonstrates that the accuracy loss on wrter-ndependent CouchGB dataset are only 0.02%, 0.8%, for IMQDF and DIMQDF based wrter adaptaton approaches respectvely. It ndcates that whle the IMQDF and DIMQDF approaches can sgnfcantly mprove the accuracy for dfferent specfc wrters, they have very lmted negatve nfluence to general wrter. 6. Concluson In ths paper, we proposed a new wrter adapton method to mprove the recognton of handwrtten Chnese character by ncremental learnng of MQDF model. We proposed the soluton of IMQDF for the frst tme and then presented the DIMQDF approach. Based on IMQDF or DIMQDF, the wrter adaptaton was fnally carred out by updatng the MQDF classfer model adaptvely. Expermental results demonstrated that both of the proposed IMQDF and DIMQDF based wrter adaptaton approaches can sgnfcantly ncrease the recognton accuracy for a specfc wrter, whle at the same tme, havng lmted negatve mpact for the general wrter. References [] C. Lu, S. Jaeger, M. Nakagawa, Onlne Recognton of Chnese Characters: he State-of-the-Art, IEEE rans. PAMI, 26(2: 98-23, [2] R. Plamondon and S.N. Srhar, On-lne and off-lne handwrtng recognton: A comprehensve survey, IEEE rans. PAMI, 22(: , [3]. Long, L. Jn, A Novel Orentaton Free Method for Onlne Unconstraned Cursve Handwrtten Chnese Word Recognton, ICPR 2008: [4] K. Dng, G. Deng, L. Jn, An Investgaton of Imagnary Stroke echnque for Cursve Onlne Handwrtng Chnese Character Recognton, ICDAR 2009:53-535, [5] D. Wang, C. Lu, et al., CASIA-OLHWDB: A Database of Onlne Handwrtten Chnese Characters, ICDAR2009: [6] C.Lu, X.Zhou, Onlne Japanese character recognton usng trajectory-based normalzaton and drecton feature extracton, IWFHR2006: , [7] R.A. Cole, J. Maran et al., Survey of the state of the art n human language technology, Cteseer, 997. [8] S.D.Connel, and A.K. Jan, Wrter Adaptaton of Onlne Handwrtng Models, IEEE rans. PAMI., 24(3: pp , [9] J.J. LaVola, and R.C.Zeleznk, A Practcal Approach for Wrter-Dependent Symbol Recognton Usng a Wrter- Independent Symbol Recognzer, IEEE rans. PAMI., 29(: , [0] Vuor, V. and eknllnen Korkeakoulu, Adaptve methods for on-lne recognton of solated handwrtten characters, Helsnk Unversty of echnology, 2002 [] Z. Huang, K. Dng, L. Jn, Wrter Adaptve Onlne Handwrtng Recognton Usng Incremental Lnear Dscrmnant Analyss, ICDAR2009: [2] L. Jn, K. Dng, Z. Huang, Incremental Learnng of LDA model for Chnese Wrter Adaptaton, accepted by NeuroComputng 200. [3] F.Kmura, K.akashna et al, Modfed quadratc dscrmnant functons and the applcaton to Chnese character recognton, IEEE rans. PAMI., 9: 49-53, 987. [4]. Long and L. Jn, Buldng Compact MQDF Classfer for Large Character Set Recognton by Subspace Dstrbuton Sharng, Pattern Recognton, 4(9:pp , [5]R.O. Duda, P.E. Hart and D.G. Stork, Pattern Classfcaton, 200 [6] D. ao, X. L, et al., General ensor Dscrmnant Analyss and Gabor Features for Gat Recognton, IEEE rans. on PAMI, 9(0 ( [7] Z. Ba and Q. Huo, A Study On the Use of 8- Drectonal Features For Onlne Handwrtten Chnese Character Recognton, ICDAR 2005: [8] SCU-COUCH dataset, Webste: 564

ARTICLE IN PRESS. Neurocomputing

ARTICLE IN PRESS. Neurocomputing Neurocomputng 73 (2010) 1614 1623 Contents lsts avalable at ScenceDrect Neurocomputng journal homepage: www.elsever.com/locate/neucom Incremental learnng of LDA model for Chnese wrter adaptaton Lanwen

More information

An Empirical Comparative Study of Online Handwriting Chinese Character Recognition:Simplified v.s.traditional

An Empirical Comparative Study of Online Handwriting Chinese Character Recognition:Simplified v.s.traditional 2013 12th Internatonal Conference on Document Analyss and Recognton An Emprcal Comparatve Study of Onlne Handwrtng Chnese Recognton:Smplfed v.s.tradtonal Yan Gao, Lanwen Jn +, Wexn Yang School of Electronc

More information

Accumulated-Recognition-Rate Normalization for Combining Multiple On/Off-Line Japanese Character Classifiers Tested on a Large Database

Accumulated-Recognition-Rate Normalization for Combining Multiple On/Off-Line Japanese Character Classifiers Tested on a Large Database 4 th Internatonal Workshop on Multple Classfer Systems (MCS23) Guldford, UK Accumulated-Recognton-Rate Normalzaton for Combnng Multple On/Off-Lne Japanese Character Classfers Tested on a Large Database

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

An Approach to Real-Time Recognition of Chinese Handwritten Sentences

An Approach to Real-Time Recognition of Chinese Handwritten Sentences An Approach to Real-Tme Recognton of Chnese Handwrtten Sentences Da-Han Wang, Cheng-Ln Lu Natonal Laboratory of Pattern Recognton, Insttute of Automaton of Chnese Academy of Scences, Bejng 100190, P.R.

More information

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying Acoustic Transient Signals Using Artificial Intelligence Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)

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

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

th International Conference on Document Analysis and Recognition

th International Conference on Document Analysis and Recognition 2013 12th Internatonal Conference on Document Analyss and Recognton Onlne Handwrtten Cursve Word Recognton Usng Segmentaton-free n Combnaton wth P2DBM-MQDF Blan Zhu 1, Art Shvram 2, Srrangaraj Setlur 2,

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

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

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

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

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

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

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

Modeling Inter-cluster and Intra-cluster Discrimination Among Triphones

Modeling 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 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

Modular PCA Face Recognition Based on Weighted Average

Modular 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 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

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

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

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. 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

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

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

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term 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 information

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

Deep Classification in Large-scale Text Hierarchies

Deep Classification in Large-scale Text Hierarchies Deep Classfcaton n Large-scale Text Herarches Gu-Rong Xue Dkan Xng Qang Yang 2 Yong Yu Dept. of Computer Scence and Engneerng Shangha Jao-Tong Unversty {grxue, dkxng, yyu}@apex.sjtu.edu.cn 2 Hong Kong

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

PRÉSENTATIONS DE PROJETS

PRÉSENTATIONS DE PROJETS PRÉSENTATIONS DE PROJETS Rex Onlne (V. Atanasu) What s Rex? Rex s an onlne browser for collectons of wrtten documents [1]. Asde ths core functon t has however many other applcatons that make t nterestng

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

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

Implementation Naïve Bayes Algorithm for Student Classification Based on Graduation Status

Implementation Naïve Bayes Algorithm for Student Classification Based on Graduation Status Internatonal Journal of Appled Busness and Informaton Systems ISSN: 2597-8993 Vol 1, No 2, September 2017, pp. 6-12 6 Implementaton Naïve Bayes Algorthm for Student Classfcaton Based on Graduaton Status

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

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 Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

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

RECOGNIZING GENDER THROUGH FACIAL IMAGE USING SUPPORT VECTOR MACHINE

RECOGNIZING 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 information

X- Chart Using ANOM Approach

X- 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 information

FACE RECOGNITION USING MAP DISCRIMINANT ON YCBCR COLOR SPACE

FACE RECOGNITION USING MAP DISCRIMINANT ON YCBCR COLOR SPACE FAC RCOGNIION USING MAP DISCRIMINAN ON YCBCR COLOR SPAC I Gede Pasek Suta Wjaya lectrcal ngneerng Department, ngneerng Faculty, Mataram Unversty. Jl. Majapaht 62 Mataram, West Nusa enggara, Indonesa. mal:

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

Fast Sparse Gaussian Processes Learning for Man-Made Structure Classification

Fast Sparse Gaussian Processes Learning for Man-Made Structure Classification Fast Sparse Gaussan Processes Learnng for Man-Made Structure Classfcaton Hang Zhou Insttute for Vson Systems Engneerng, Dept Elec. & Comp. Syst. Eng. PO Box 35, Monash Unversty, Clayton, VIC 3800, Australa

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

A Robust Method for Estimating the Fundamental Matrix

A 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 information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

High-Boost Mesh Filtering for 3-D Shape Enhancement

High-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 information

UB at GeoCLEF Department of Geography Abstract

UB 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 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

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

Learning Non-Linearly Separable Boolean Functions With Linear Threshold Unit Trees and Madaline-Style Networks

Learning Non-Linearly Separable Boolean Functions With Linear Threshold Unit Trees and Madaline-Style Networks In AAAI-93: Proceedngs of the 11th Natonal Conference on Artfcal Intellgence, 33-1. Menlo Park, CA: AAAI Press. Learnng Non-Lnearly Separable Boolean Functons Wth Lnear Threshold Unt Trees and Madalne-Style

More information

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation Intellgent Informaton Management, 013, 5, 191-195 Publshed Onlne November 013 (http://www.scrp.org/journal/m) http://dx.do.org/10.36/m.013.5601 Qualty Improvement Algorthm for Tetrahedral Mesh Based on

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

Learning-Based Top-N Selection Query Evaluation over Relational Databases

Learning-Based Top-N Selection Query Evaluation over Relational Databases Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **

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

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

Online Handwritten Cursive Word Recognition Using Segmentation-free and Segmentation-based Methods

Online Handwritten Cursive Word Recognition Using Segmentation-free and Segmentation-based Methods 3th Asan Conference on Pattern Recognton (ACPR2015), November 3-6, 2015, Kuala Lumpur, Malasa. Onlne Handwrtten Cursve Word Recognton Usng Segmentaton-free and Segmentaton-based Methods Blan Zhua, Art

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

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems A Unfed Framework for Semantcs and Feature Based Relevance Feedback n Image Retreval Systems Ye Lu *, Chunhu Hu 2, Xngquan Zhu 3*, HongJang Zhang 2, Qang Yang * School of Computng Scence Smon Fraser Unversty

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

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler 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 information

Investigating the Performance of Naïve- Bayes Classifiers and K- Nearest Neighbor Classifiers

Investigating the Performance of Naïve- Bayes Classifiers and K- Nearest Neighbor Classifiers Journal of Convergence Informaton Technology Volume 5, Number 2, Aprl 2010 Investgatng the Performance of Naïve- Bayes Classfers and K- Nearest Neghbor Classfers Mohammed J. Islam *, Q. M. Jonathan Wu,

More information

Three supervised learning methods on pen digits character recognition dataset

Three supervised learning methods on pen digits character recognition dataset Three supervsed learnng methods on pen dgts character recognton dataset Chrs Flezach Department of Computer Scence and Engneerng Unversty of Calforna, San Dego San Dego, CA 92093 cflezac@cs.ucsd.edu Satoru

More information

INTELLECT SENSING OF NEURAL NETWORK THAT TRAINED TO CLASSIFY COMPLEX SIGNALS. Reznik A. Galinskaya A.

INTELLECT SENSING OF NEURAL NETWORK THAT TRAINED TO CLASSIFY COMPLEX SIGNALS. Reznik A. Galinskaya A. Internatonal Journal "Informaton heores & Applcatons" Vol.10 173 INELLEC SENSING OF NEURAL NEWORK HA RAINED O CLASSIFY COMPLEX SIGNALS Reznk A. Galnskaya A. Abstract: An expermental comparson of nformaton

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

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

Face Recognition Method Based on Within-class Clustering SVM

Face Recognition Method Based on Within-class Clustering SVM Face Recognton Method Based on Wthn-class Clusterng SVM Yan Wu, Xao Yao and Yng Xa Department of Computer Scence and Engneerng Tong Unversty Shangha, Chna Abstract - A face recognton method based on Wthn-class

More information

Feature Extraction and Dimensionality Reduction in SVM Speaker Recognition

Feature 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 information

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram Shape Representaton Robust to the Sketchng Order Usng Dstance Map and Drecton Hstogram Department of Computer Scence Yonse Unversty Kwon Yun CONTENTS Revew Topc Proposed Method System Overvew Sketch Normalzaton

More information

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines A Modfed Medan Flter for the Removal of Impulse Nose Based on the Support Vector Machnes H. GOMEZ-MORENO, S. MALDONADO-BASCON, F. LOPEZ-FERRERAS, M. UTRILLA- MANSO AND P. GIL-JIMENEZ Departamento de Teoría

More information

A Multivariate Analysis of Static Code Attributes for Defect Prediction

A Multivariate Analysis of Static Code Attributes for Defect Prediction Research Paper) A Multvarate Analyss of Statc Code Attrbutes for Defect Predcton Burak Turhan, Ayşe Bener Department of Computer Engneerng, Bogazc Unversty 3434, Bebek, Istanbul, Turkey {turhanb, bener}@boun.edu.tr

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

MULTISPECTRAL REMOTE SENSING IMAGE CLASSIFICATION WITH MULTIPLE FEATURES

MULTISPECTRAL REMOTE SENSING IMAGE CLASSIFICATION WITH MULTIPLE FEATURES MULISPECRAL REMOE SESIG IMAGE CLASSIFICAIO WIH MULIPLE FEAURES QIA YI, PIG GUO, Image Processng and Pattern Recognton Laboratory, Bejng ormal Unversty, Bejng 00875, Chna School of Computer Scence and echnology,

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

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

Benchmarking of Update Learning Strategies on Digit Classifier Systems

Benchmarking of Update Learning Strategies on Digit Classifier Systems 2012 Internatonal Conference on Fronters n Handwrtng Recognton Benchmarkng of Update Learnng Strateges on Dgt Classfer Systems D. Barbuzz, D. Impedovo, G. Prlo Dpartmento d Informatca Unverstà degl Stud

More information

Non-Negative Matrix Factorization and Support Vector Data Description Based One Class Classification

Non-Negative Matrix Factorization and Support Vector Data Description Based One Class Classification IJCSI Internatonal Journal of Computer Scence Issues, Vol. 9, Issue 5, No, September 01 ISSN (Onlne): 1694-0814 www.ijcsi.org 36 Non-Negatve Matrx Factorzaton and Support Vector Data Descrpton Based One

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

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc.

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 97-735 Volume Issue 9 BoTechnology An Indan Journal FULL PAPER BTAIJ, (9), [333-3] Matlab mult-dmensonal model-based - 3 Chnese football assocaton super league

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Keywords - Wep page classification; bag of words model; topic model; hierarchical classification; Support Vector Machines

Keywords - Wep page classification; bag of words model; topic model; hierarchical classification; Support Vector Machines (IJCSIS) Internatonal Journal of Computer Scence and Informaton Securty, Herarchcal Web Page Classfcaton Based on a Topc Model and Neghborng Pages Integraton Wongkot Srura Phayung Meesad Choochart Haruechayasak

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

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:

Outline. 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 information

Feature Selection By KDDA For SVM-Based MultiView Face Recognition

Feature Selection By KDDA For SVM-Based MultiView Face Recognition SEI 007 4 rth Internatonal Conference: Scences of Electronc, echnologes of Informaton and elecommuncatons March 5-9, 007 UNISIA Feature Selecton By KDDA For SVM-Based MultVew Face ecognton Seyyed Majd

More information

Performance Assessment and Fault Diagnosis for Hydraulic Pump Based on WPT and SOM

Performance Assessment and Fault Diagnosis for Hydraulic Pump Based on WPT and SOM Performance Assessment and Fault Dagnoss for Hydraulc Pump Based on WPT and SOM Be Jkun, Lu Chen and Wang Zl PERFORMANCE ASSESSMENT AND FAULT DIAGNOSIS FOR HYDRAULIC PUMP BASED ON WPT AND SOM. Be Jkun,

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

Fitting & Matching. Lecture 4 Prof. Bregler. Slides from: S. Lazebnik, S. Seitz, M. Pollefeys, A. Effros.

Fitting & Matching. Lecture 4 Prof. Bregler. Slides from: S. Lazebnik, S. Seitz, M. Pollefeys, A. Effros. Fttng & Matchng Lecture 4 Prof. Bregler Sldes from: S. Lazebnk, S. Setz, M. Pollefeys, A. Effros. How do we buld panorama? We need to match (algn) mages Matchng wth Features Detect feature ponts n both

More information

PERFORMANCE COMPARISON OF ONLINE HANDWRITING RECOGNITION SYSTEM FOR ASSAMESE LANGUAGE BASED ON HMM AND SVM MODELLING

PERFORMANCE COMPARISON OF ONLINE HANDWRITING RECOGNITION SYSTEM FOR ASSAMESE LANGUAGE BASED ON HMM AND SVM MODELLING PERFORMANCE COMPARISON OF ONLINE HANDWRIING RECOGNIION SYSEM FOR ASSAMESE LANGUAGE BASED ON HMM AND SVM MODELLING Deepoy Das, Rtuparna Dev, SRM Prasanna, Subhankar Ghosh, Krshna Nak Department of EEE,

More information

D. Barbuzzi* and G. Pirlo

D. Barbuzzi* and G. Pirlo Int. J. Sgnal and Imagng Systems Engneerng, Vol. 7, No. 4, 2014 245 bout retranng rule n mult-expert ntellgent system for sem-supervsed learnng usng SVM classfers D. Barbuzz* and G. Prlo Department of

More information

Japanese Dependency Analysis Based on Improved SVM and KNN

Japanese Dependency Analysis Based on Improved SVM and KNN Proceedngs of the 7th WSEAS Internatonal Conference on Smulaton, Modellng and Optmzaton, Bejng, Chna, September 15-17, 2007 140 Japanese Dependency Analyss Based on Improved SVM and KNN ZHOU HUIWEI and

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

Robust Face Recognition through Local Graph Matching

Robust Face Recognition through Local Graph Matching Robust Face Recognton through Local Graph Matchng Ehsan Fazl-Ers, John S. Zele and John K. Tsotsos, Department of Computer Scence and Engneerng, Yor Unversty, Toronto, Canada E-Mal: [efazl, tsotsos]@cse.yoru.ca

More information

Writer Identification using a Deep Neural Network

Writer Identification using a Deep Neural Network Wrter Identfcaton usng a Deep Neural Network Jun Chu and Sargur Srhar Department of Computer Scence and Engneerng Unversty at Buffalo, The State Unversty of New York Buffalo, NY 1469, USA {jchu6, srhar}@buffalo.edu

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

An Improved Neural Network Algorithm for Classifying the Transmission Line Faults

An Improved Neural Network Algorithm for Classifying the Transmission Line Faults 1 An Improved Neural Network Algorthm for Classfyng the Transmsson Lne Faults S. Vaslc, Student Member, IEEE, M. Kezunovc, Fellow, IEEE Abstract--Ths study ntroduces a new concept of artfcal ntellgence

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

Vol. 5, No. 3 March 2014 ISSN Journal of Emerging Trends in Computing and Information Sciences CIS Journal. All rights reserved.

Vol. 5, No. 3 March 2014 ISSN Journal of Emerging Trends in Computing and Information Sciences CIS Journal. All rights reserved. Journal of Emergng Trends n Computng and Informaton Scences 009-03 CIS Journal. All rghts reserved. http://www.csjournal.org Unhealthy Detecton n Lvestock Texture Images usng Subsampled Contourlet Transform

More information

General Vector Machine. Hong Zhao Department of Physics, Xiamen University

General Vector Machine. Hong Zhao Department of Physics, Xiamen University General Vector Machne Hong Zhao (zhaoh@xmu.edu.cn) Department of Physcs, Xamen Unversty The support vector machne (SVM) s an mportant class of learnng machnes for functon approach, pattern recognton, and

More information

Manifold-Ranking Based Keyword Propagation for Image Retrieval *

Manifold-Ranking Based Keyword Propagation for Image Retrieval * Manfold-Rankng Based Keyword Propagaton for Image Retreval * Hanghang Tong,, Jngru He,, Mngjng L 2, We-Yng Ma 2, Hong-Jang Zhang 2 and Changshu Zhang 3,3 Department of Automaton, Tsnghua Unversty, Bejng

More information