A Fast Dictionary Learning Algorithm for Image Denoising Hai-yang LI *, Chao YUAN and Heng-yuan WANG

Size: px
Start display at page:

Download "A Fast Dictionary Learning Algorithm for Image Denoising Hai-yang LI *, Chao YUAN and Heng-yuan WANG"

Transcription

1 08 Internatonal onference on Modelng, Sulaton and Optzaton (MSO 08) ISBN: A ast ctonary Learnng Algorth for Iage enosng Ha-yang LI, hao YUAN and Heng-yuan WANG School of Scence, X'an Polytechnc Unversty, X'an 70048, PR hna orrespondng author Keywords: Sparse presentaton, ctonary learnng, Graph Laplacan, lusterng Abstract he K-SV s one of the well-nown and effectve ethods to learn a unversal and overcoplete dctonary However, K-SV s very expensve because any teraton steps are needed What s ore, when t converts data patches nto vectors for tranng or learnng, K-SV breas down the nherent geoetrc structure of the data o overcoe these ltatons, eployng a subspace partton technque, we propose an effcent and fast algorth, the fast top-botto two-densonal subspace partton algorth, for learnng overcoplete dctonares Experental sulatons deonstrate that our dctonary learnng approach s effectve for age denosng Introducton Sparse and redundant representaton odelng has recently receved extensve research nterest and found successful applcatons n copressve sensng[], age processng tass (copresson, denosng, zoong)[], lnear regresson and varable selecton[3] he sparse representaton proble descrbes that a sgnal can be approxated as a lnear cobnaton of as few as possble bass functons Each bass functon s called an ato and the collecton of the s called dctonary[4] hs dctonary s overcoplete, that s, the nuber of atos s ore than the N denson of each ato More precsely, a target sgnal y R s descrbed as y x, where N M R s an overcoplete dctonary and x s a vector contanng the representaton coeffcent of y We are seeng the sparsest soluton x, the one wth the fewest nonzero entres hs can be forulated as the proble 0 n yx s t x () x where x stands for the so called l 0 0 nor that counts the nuber of nonzero eleents of x, and stands for the axu nuber of nonzero eleents Exact deternaton of sparsest representatons s nown to be an NP-hard proble hus a nuber of algorths have been proposed to provde the sparsest approxaton of a sgnal, ncludng Orthogonal Matchng Pursut (OMP)[5] and Bass N M Pursut (BP)[6] More concretely, gven a tranng data atrx Y R, contanng M sgnals N M N M y R, dctonary learnng s a procedure to fnd a dctonary R he soluton can be obtaned by solvng the followng proble M y x s t x 0 0 X, n Most dctonary learnng algorths perfor two stages[7]: sparse codng and dctonary update In the sparse codng stage, eepng dctonary fxed, X s coputed by solvng () In ctonary update stage, wth a fxed X, the dctonary s updated to reduce the approxaton error he an dfference aong ost dctonary learnng algorths, such as K-SV algorth[8], the ctonary Par Learnng on the Grassann-anfold algorth (PLG)[9] and frst dctonary learnng (L)[0], s the way of updatng the dctonary However, K-SV algorth s expensve In partcular, any teraton steps are needed snce the atos of the dctonary are updated coponent by coponent and the K-SV updates each ato along wth the coeffcents n that ultply t usng 56 ()

2 sngular value decoposton (SV) It s worth notng that the K-SV convert data patches nto vectors for tranng and learnng, and ths converson breas down the nherent structure of the data o overcoe these ltaton, other ethods for dctonary learnng cae up to replace the K-SV or exaple, Lu et al[0] and Zeng et al[] Zeng et al n[] proposed a dctonary par learnng odel (PL odel) for age denosng and desgned a correspondng algorth, called the PLG algorth hs algorth learned an ntal dctonary par by the op-botto wo-densonal Subspace Partton algorth (SP algorth) he ethods of frst dctonary learnng (L) was presented n Lu et al[0] he parttonng procedure n L s equvalent to the frst part of the SP algorth and constructon dctonary s dfferent fro the second part of the SP algorth Motvated by deas n[0] and[], we propose a fast dctonary learnng algorth for age denosng Our ethod s also a two-stage approach that ncludes dctonary learnng stage and sparse codng stage, n whch dctonary learnng stage s dfferent fro the SP algorth n[] and sparse codng stage s slar to the ethod n[] by addng the ntrnsc geoetrc structure of the data through a graph regularzed and usng the learned dctonary to provde a sparse representaton of data patches he paper s organzed as follows In Secton we descrbe the graph Laplacan and then provde a bref descrpton of the graph-based dctonary learnng he correspondng optzaton algorth, a fast top-botto two-densonal subspace partton algorth (SP algorth), s presented n Secton 3 Secton 4 presents soe experent results nally, the paper s concluded n Secton 5 Sparse odng by Learned ctonary and Graph Regularzaton Recall that sparse codng, eepng the fxed learned dctonary, tres to fnd a sparse coeffcent atrx X by solvng () However, ost of the exstng approaches to sparse codng fal to consder the geoetrcal structure of the data space In[], Zheng et al propose a graph based algorth, called graph regularzed sparse codng (GraphS), to learn the sparse representatons that explctly tae nto account the local anfold structure of the data Specfcally, the graph Laplacan s ncorporated nto the sparse codng obectve functon as a regularzer In ths way, the obtaned representatons vary soothly along the geodescs of the data anfold By preservng localty, GraphS can have ore dscrnatng power copared wth tradtonal sparse codng algorths Here we follow deas n Zheng et al[] and ntroduce GraphS n the followng or the gven set of tranng patches { Y }, we construct a weghted undrected coplete graph GV (, EW, ), where the fnte set V of vertces represents the gven patches, V Y,, Y urther, E V V s a set of weghted edges, and these weghts are collected n the weght atrx W R W W by We defne the syetrc weght atrx, Y Y exp for W h h 0 for usng the Gaussan ernel and paraeter h he degree of each vertex Y, the nuber of all edges wth weght to the vertex Y s gven by W, Introducng the dagonal degree atrx dag,,, the graph Laplacan of G s now gven by L W to acheve a sparse atrx L Hence, L s syetrc and postve se-defnte, wth non-dagonal entres beng non-postve, and the su of all entres n each colun (or row) s zero A drect coputaton shows that, r( YLY ) W Y Y W y y Y Y,,,, YY (4) 57

3 easurng the slarty of neghborhood patches n the graph, where we have used the notaton Y Y or each, the vector x s assued to be a good approxaton y Snce the transfor atrx nduces a lnear appng, we can suppose that the vectors x,,, possess a slar topologcal structure as y,,,, and partcularly that, f y and y are -neghbors wth a sall dstance sall herefore, we ncorporate the ter y y, we also have that x x s r( XLX ) W x x = x x,, YY and obtan the new nzaton proble (5) n Y X r( XLX ) X (6) MM NM XR, R where the Laplacan atrx L only depends on the tranng data Y that generates the graph he dctonary learnng algorth by eployng GraphS, called dctonary learnng based on graph regularzaton, s outlned n Algorth Algorth ctonary learnng based on graph regularzaton Input: ranng data Y [ Y ]; Paraeters and Procedures: : opute the graph Laplacan L for the gven tranng set Y : eterne the dctonary by a dctonary learnng algorth based on Y 3: Solve the nzaton proble n Y X r( XLX ) X MM XR 4: Reconstruct the data Y X Loop steps untl the gven nuber of teratons s acheved Output ata Y In the followng, we ntroduce the algorth for solvng the thrd step n Algorth brefly hat s to say, we wll solve the nzaton proble n Y X r( XLX ) X MM XR for gven (nosy) tranng data Y and the dctonary =[ d,, d ], where d col (the vectorzed ) are the dctonary eleents constructed n above We suggest here to solve the proble usng the splt Bregan teraton see eg Goldsten and Osher[3]; Plona and Ma[4] whch s n the consdered case equvalent to the Alternatng recton Method of Multplers(AMM), see Yanelevsy and Elad[5] We outlne the algorth n the followng Algorth Graph regularzed sparse codng Input: ranng data Y [ Y ] ; Laplacan atrx L ; Learned dctonary ; 0 0 X Z B 0 =0 ; Paraeters, u, 0; Nuber of teratons Algorth Iterate untl the gven nuber of teratons s acheved: : opute X as the soluton to ( uix ) XL YuZ ( B ) : opute Z coponentwsely by eployng shrnage z, \ (x, B, ) 3: Update B B Z X Output X 58

4 A ast ctonary Learnng Method Here we wll propose a fast dctonary learnng ethod whch s based on a specal partton tree structure Our ethod s also a two-stage approach that ncludes dctonary learnng stage and sparse codng stage In dctonary learnng stage, we construct the dctonary n two steps We frst obtan a tree structure to partton the set of our tranng patches and then construct the dctonary based on the coputed subset parttons n the tree, usng a fast top-botto two-densonal subspace partton algorth (SP algorth) he frst step, the tree constructon, s dfferent fro the frst part of the SP algorth, whle the second step, constructng the dctonary, s equvalent to the second part of ethod SL In sparse codng stage, we add the ntrnsc geoetrc structure of the data through a graph regularzed and use the learned dctonary to provde a sparse representaton of data patches he SP algorth s outlned n Algorth 3 Step: onstructon of the partton tree or the gven tranng set,, n n Y Y R of age patches We copute the ean nn : Y R (7) and the two non-syetrc (n n) covarance atrces : ( Y )( Y ) L : ( Y ) ( Y ) (8) R Now, the noralzed egenvectors u, u andv, and R s coputed By the frst two egenvalues we refer to the two largest egenvalues v correspondng to the frst two egenvalues of L u: argaxx Lx, x v: argaxx Rx, (9) x representng the an structures of the tranng patches beng not captured by the ean patch We copute the nubers s uyv, s uyv,,, (0) hese nubers { s,, } and { s,, } gve us a easure, how strong each sngle patch s correlated to the found structure and wll be used to obtan a partton of the set of all patches { Y } nto four partal sets rst, n the frst level, we partton the one-densonal real nuber { s,, } nto two clusters { s } and { s } by K-eans, n whch {,, } and hen, n the second level, we partton the one-densonal real nuber sets { s } and { s } nto two clusters { s }, { s } and { s }, { s } by K-eans respectvely, n whch, and, herefore, { Y }= { Y} { Y} and { Y}={ Y} { Y}, { Y}={ Y} { Y} In ths way, we can dvde two level tree structure wth every calculaton Rears Havng found ths frst partton, we can proceed to partton the obtaned subsets further, usng the sae schee hs procedure yelds a bnary tree, where we stop the further partton of a subset, f t contans two tranng data that autoatcally separate the two classes Snce two level tree structures are obtaned n every calculaton, the algorth of our artcle speeds up SP algorth 3 If we yeld a bnary tree wth the frst three egenvalues or ore egenvalues n above procedure, then the algorth has a better convergence rate but a weaer perforances such as structural slarty ndex (SSIM), pea sgnal to nose rato (PSNR) and root ean square error (RMSE) n general 59

5 Step : eterne the dctonary fro the partton tree Each not n the tree s now assocated wth a subset of tranng patches { Y }, where {,, } denotes the subset of ndces of these patches We assue that the root of the tree has the not nuber (e = {,, } ) and we proceed nuberng by gong through each level fro left to rght or each not, we copute the ean (center) Y () and the noralzed sngular vectors to the axal sngular value of and, e u : argax x x v : argax x (3) x If denotes the axal sngular value of then uv s the best ran- approxaton of, snceu and v are the frst vectors n the sngular value decoposton of he dctonary s now deterned as follows We fx the frst dctonary eleent uv (4) capturng the an structure of the ean urther, for each par of chldren nots and to the sae parent wth eans and +, we set u v u v (5) :, : thereby capturng the dfference of an structures of,,, ), we construct the dctonary =[ d,, d ] and + Let d col (the vectorzed Algorth 3 (SP algorth) ast top-botto two-densonal subspace partton algorth Input: ranng age patches, the axu depth of the bnary tree Procedures: : he frst node s the root node ncludng all age patches : or all age patches n the current leaf node, run the followng )-4) steps: ) opute egenvectors u, v and u, v correspondng to the frst two egenvalues of the two covarance atrxes ) opute the two-densonal proecton representatons of all age patches fro ths node, that s, s u Yv and s u Yv,,, 3) Partton the one-densonal real nuber set s nto two clusters { s } and { s } by K-eans hen partton the age patches correspondng to these two clusters nto the left chld { Y } and the rght chld { Y } Sultaneously the depth of the node s added one 4) Partton the one-densonal real nuber sets { s } nto two clusters { s } and { s } by K-eans hen partton the age patches { Y } correspondng to these two clusters nto the left chld { Y} and the rght chld{ Y } Partton the age patches { Y } n an analogous anner Sultaneously the depth of the node s added one 3: I the depth of the node s larger than the axu depth or the nuber of age patches n ths leaf node s saller than the row nuber or colun nuber of the age patches, HEN stop the partton ELSE repeat Step recursvely for the left chld node and the rght chld node 60

6 4: opute the dctonary for each node by the followng )-3) steps: ) opute the center and the noralzed sngular vectors u and v to the axal sngular value of and n the root node onstruct the frst ato of dctonary uv ) or each par of chldren nots and to the sae parent wth eans and +, copute the noralzed sngular vectors u, and + sngular value 3) opute the ato :, where v and u +, v +to the axal : uv u v 5: ollect the ato of all leaf nodes nto the dctonary d d vectorzed ) Output the dctonary,,, where d col (the Experents In ths secton, we present experents to evaluate the dctonary perforance of our proposed algorth copared wth other algorths ntroduced n the paper In the part, we present experental result, wth the a of tranng a dctonary for sparsely representng natural age patches We then turn to test the age denosng perforance of the dctonary learned by our approach Our sulatons were perfored n MALAB R00b envronent on a syste wth 38 GHz PU and 4 GB RAM, under Mcrosoft Wndows 7 operatng syste As a rough easure of coplexty, we wll enton the runnng tes of the algorths We show the proveent acheved by applyng the above ethods to the age denosng proble In ths set of experents, we eploy the ethodology[8] gven by Elad We choose three well-nown ages of sze as test ages, ncludng Barbara, Boat, House, Each age s contanated by artfcally addng zeros-ean whte Gaussan nose at fve dfferent varances An obectve age qualty etrc plays an portant role n age denosng applcatons urrently, three classcal age qualty assessent etrcs are typcally used: the Pea Sgnal-to-Nose Rato (PSNR) and the easure of Structural Slarty (SSIM) We cut the nosed age nto sall patches of sze 8 8 he regularzaton paraeters have been eprcally chosen to be 6; = 30 ; u 005 It s also noted that the denosed age s obtaned wth an average constant 6

7 able PSNR values of the denosed results Iage Barbara =5 =0 =5 =0 =5 PSNR IME PSNR IME PSNR IME PSNR IME PSNR IME Algorth K-SV L Ours Iage Boat =5 =0 =5 =0 =5 PSNR IME PSNR IME PSNR IME PSNR IME PSNR IME Algorth K-SV L Ours Iage House =5 =0 =5 =0 =5 PSNR IME PSNR IME PSNR IME PSNR IME PSNR IME Algorth K-SV L Ours able SSIM values of the denosed result Iage K-SV PS Ours Barbara Boat House

8 gure splayng denosng results for, (a) orgnal age (b) Nosy age, (c) Iage denosed usng K-SVdctonary (d) Iage denosed usng L dctonary(e) Iage denosed usng Our dctonary able presents the fnal denosng PSNR results obtaned fro K-SV, L algorths wth addtonally the fxed ours ethods he SSIM results of the three test ethods are reported n able gure dsplays the orgnal, nosy and denosed Barbara, Boat, House, ages for nose level =0 Based on these results, we can observe that our proposed algorth, SP algorth, and L algorth n general not only cost less te but also provde hgher PSNR result and SSIM values n age denosng copared wth the K-SV algorth Although SP algorth and L algorth have the slar results n general, t s notceable that SP algorth needs less te copared wth L Suary In the paper, we present a fast top-botto two-densonal subspace partton algorth (SP algorth) for learnng overcoplete dctonary, whch s based on a specal partton tree structure In constructon of the partton tree step, our algorth can obtan two level tree structures n every calculaton, and hence t costs less te than the frst part of the SP algorth and L It s equvalent to the second part of ethod SL n constructon dctonary step Experental results on synthetc data and age patches llustrate that SP and L not only have hgher qualty but also cost less te than K-SV, and that at the sae te, SP needs less te than L In the future, we wll consder provng SP algorth and applyng t to other applcatons Acnowledgeent he authors would le to than edtoral and referees for ther coents whch help us to enrch the content and prove the presentaton of the results n ths paper he wor was supported by the Natonal Natural Scence oundatons of hna (797) and the Scence oundatons of Shaanx Provnce of hna (05JM0) 63

9 References [] L onoho, opressed sensng, IEEE rans on Inforaton heory 5(4) (006) [] M Elad, MA gueredo and Y Ma, On the role of sparse and redundant representatons n age processng, Proceedngs of the IEEE, 98 (6) (00), [3] R bshran, Regresson shrnage and selecton va the lasso, J Royal Statst Soc B 58() (006) [4] S Mallat and Z Zhang, Matchng pursuts wth te-frequency dctonares, IEEE rans on Sgnal Proc 4() (993), [5] Y Pat, R Rezafar, P Krshnaprasad, Orthogonal atchng pursut: Recursve functon approxaton wth applcatons to wavelet decoposton, n Proceedngs of the 7th Asloar onference on Sgnals, Systes & oputers (993) [6] S hen, onoho, M Saunders, Atoc decoposton by bass pursut, SIAM Revew 43() (00), 9-59 [7] W a, Xu, W Wang, Sultaneous codeword optzaton (SO) for dctonary update and learnng, IEEE rans on Sgnal Process 60() (0) [8] M Elad, M Aharon, Iage denosng va sparse and redundant representatons over learned dctonares, IEEE rans on Iage Process 5 (006) [9] X Zeng, W Ban, W Lu, J Shen and ao, ctonary par learnng on Grassan anfolds for age denosng, IEEE rans on Iage Processng 4 () (05) [0] L Lu, J Ma, G Plona, Graph regularzed sesc dctonary learnng, [] X Zeng, W Ban, W Lu, J Shen and ao, ctonary par learnng on Grassan anfolds for age denosng, IEEE rans on Iage Processng 4 (05) [] M Zheng, JJ Bu, hen and Wang, Graph regularzed sparse codng for age representaton, IEEE rans on Iage Processng 0(5) (0), [3] Goldsten and S Osher, he splt Bregan ethod for L-regularzed probles, SIAM J Iagng Scences () (009) [4] G Plona and J Ma, urvelet-wavelet regularzed splt Bregan teraton for copressed sensng, Int J Wavelets Multresolut Inf Process 9() (0) 79-0 [5] Y Yanelevsy and M Elad, ual graph regularzed dctonary learnng, IEEE ransactons on Sgnal and Inforaton Processng over Networs (4) (06)

Comparative Study between different Eigenspace-based Approaches for Face Recognition

Comparative Study between different Eigenspace-based Approaches for Face Recognition Coparatve Study between dfferent Egenspace-based Approaches for Face Recognton Pablo Navarrete and Javer Ruz-del-Solar Departent of Electrcal Engneerng, Unversdad de Chle, CHILE Eal: {pnavarre, jruzd}@cec.uchle.cl

More information

Generating Fuzzy Term Sets for Software Project Attributes using and Real Coded Genetic Algorithms

Generating Fuzzy Term Sets for Software Project Attributes using and Real Coded Genetic Algorithms Generatng Fuzzy Ter Sets for Software Proect Attrbutes usng Fuzzy C-Means C and Real Coded Genetc Algorths Al Idr, Ph.D., ENSIAS, Rabat Alan Abran, Ph.D., ETS, Montreal Azeddne Zah, FST, Fes Internatonal

More information

Local Subspace Classifiers: Linear and Nonlinear Approaches

Local Subspace Classifiers: Linear and Nonlinear Approaches Local Subspace Classfers: Lnear and Nonlnear Approaches Hakan Cevkalp, Meber, IEEE, Dane Larlus, Matths Douze, and Frederc Jure, Meber, IEEE Abstract he -local hyperplane dstance nearest neghbor (HNN algorth

More information

Solutions to Programming Assignment Five Interpolation and Numerical Differentiation

Solutions to Programming Assignment Five Interpolation and Numerical Differentiation College of Engneerng and Coputer Scence Mechancal Engneerng Departent Mechancal Engneerng 309 Nuercal Analyss of Engneerng Systes Sprng 04 Nuber: 537 Instructor: Larry Caretto Solutons to Prograng Assgnent

More information

Optimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming

Optimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming Optzaton Methods: Integer Prograng Integer Lnear Prograng Module Lecture Notes Integer Lnear Prograng Introducton In all the prevous lectures n lnear prograng dscussed so far, the desgn varables consdered

More information

Merging Results by Using Predicted Retrieval Effectiveness

Merging Results by Using Predicted Retrieval Effectiveness Mergng Results by Usng Predcted Retreval Effectveness Introducton Wen-Cheng Ln and Hsn-Hs Chen Departent of Coputer Scence and Inforaton Engneerng Natonal Tawan Unversty Tape, TAIWAN densln@nlg.cse.ntu.edu.tw;

More information

Handwritten English Character Recognition Using Logistic Regression and Neural Network

Handwritten English Character Recognition Using Logistic Regression and Neural Network Handwrtten Englsh Character Recognton Usng Logstc Regresson and Neural Network Tapan Kuar Hazra 1, Rajdeep Sarkar 2, Ankt Kuar 3 1 Departent of Inforaton Technology, Insttute of Engneerng and Manageent,

More information

A Semantic Model for Video Based Face Recognition

A Semantic Model for Video Based Face Recognition Proceedng of the IEEE Internatonal Conference on Inforaton and Autoaton Ynchuan, Chna, August 2013 A Seantc Model for Vdeo Based Face Recognton Dhong Gong, Ka Zhu, Zhfeng L, and Yu Qao Shenzhen Key Lab

More information

Generalized Spatial Kernel based Fuzzy C-Means Clustering Algorithm for Image Segmentation

Generalized Spatial Kernel based Fuzzy C-Means Clustering Algorithm for Image Segmentation Internatonal Journal of Scence and Research (IJSR, Inda Onlne ISSN: 39-7064 Generalzed Spatal Kernel based Fuzzy -Means lusterng Algorth for Iage Segentaton Pallav Thakur, helpa Lnga Departent of Inforaton

More information

IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. *, NO. *, Dictionary Pair Learning on Grassmann Manifolds for Image Denoising

IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. *, NO. *, Dictionary Pair Learning on Grassmann Manifolds for Image Denoising IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. *, NO. *, 2015 1 Dctonary Par Learnng on Grassmann Manfolds for Image Denosng Xanhua Zeng, We Ban, We Lu, Jale Shen, Dacheng Tao, Fellow, IEEE Abstract Image

More information

Color Image Segmentation Based on Adaptive Local Thresholds

Color Image Segmentation Based on Adaptive Local Thresholds Color Iage Segentaton Based on Adaptve Local Thresholds ETY NAVON, OFE MILLE *, AMI AVEBUCH School of Coputer Scence Tel-Avv Unversty, Tel-Avv, 69978, Israel E-Mal * : llero@post.tau.ac.l Fax nuber: 97-3-916084

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

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

On-line Scheduling Algorithm with Precedence Constraint in Embeded Real-time System

On-line Scheduling Algorithm with Precedence Constraint in Embeded Real-time System 00 rd Internatonal Conference on Coputer and Electrcal Engneerng (ICCEE 00 IPCSIT vol (0 (0 IACSIT Press, Sngapore DOI: 077/IPCSIT0VNo80 On-lne Schedulng Algorth wth Precedence Constrant n Ebeded Real-te

More information

Pattern Classification of Back-Propagation Algorithm Using Exclusive Connecting Network

Pattern Classification of Back-Propagation Algorithm Using Exclusive Connecting Network World Acade of Scence, Engneerng and Technolog 36 7 Pattern Classfcaton of Bac-Propagaton Algorth Usng Eclusve Connectng Networ Insung Jung, and G-Na Wang Abstract The obectve of ths paper s to a desgn

More information

A new Fuzzy Noise-rejection Data Partitioning Algorithm with Revised Mahalanobis Distance

A new Fuzzy Noise-rejection Data Partitioning Algorithm with Revised Mahalanobis Distance A new Fuzzy ose-reecton Data Parttonng Algorth wth Revsed Mahalanobs Dstance M.H. Fazel Zarand, Mlad Avazbeg I.B. Tursen Departent of Industral Engneerng, Arabr Unversty of Technology Tehran, Iran Departent

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

Large Margin Nearest Neighbor Classifiers

Large Margin Nearest Neighbor Classifiers Large Margn earest eghbor Classfers Sergo Bereo and Joan Cabestany Departent of Electronc Engneerng, Unverstat Poltècnca de Catalunya (UPC, Gran Captà s/n, C4 buldng, 08034 Barcelona, Span e-al: sbereo@eel.upc.es

More information

Human Face Recognition Using Radial Basis Function Neural Network

Human Face Recognition Using Radial Basis Function Neural Network Huan Face Recognton Usng Radal Bass Functon eural etwor Javad Haddadna Ph.D Student Departent of Electrcal and Engneerng Arabr Unversty of Technology Hafez Avenue, Tehran, Iran, 594 E-al: H743970@cc.au.ac.r

More information

A Bayesian Mixture Model for Multi-view Face Alignment

A Bayesian Mixture Model for Multi-view Face Alignment A Bayesan Mxture Model for Mult-vew Face Algnent Y Zhou, We Zhang, Xaoou Tang, and Harry Shu Mcrosoft Research Asa Bejng, P. R. Chna {t-yzhou, xtang, hshu}@crosoft.co DCST, Tsnghua Unversty Bejng, P. R.

More information

Determination of Body Sway Area by Fourier Analysis of its Contour

Determination of Body Sway Area by Fourier Analysis of its Contour PhUSE 213 Paper SP8 Deternaton of Body Sway Area by Fourer Analyss of ts Contour Abstract Thoas Wollsefen, InVentv Health Clncal, Eltvlle, Gerany Posturography s used to assess the steadness of the huan

More information

Performance Analysis of Coiflet Wavelet and Moment Invariant Feature Extraction for CT Image Classification using SVM

Performance Analysis of Coiflet Wavelet and Moment Invariant Feature Extraction for CT Image Classification using SVM Perforance Analyss of Coflet Wavelet and Moent Invarant Feature Extracton for CT Iage Classfcaton usng SVM N. T. Renukadev, Assstant Professor, Dept. of CT-UG, Kongu Engneerng College, Perundura Dr. P.

More information

Key-Words: - Under sear Hydrothermal vent image; grey; blue chroma; OTSU; FCM

Key-Words: - Under sear Hydrothermal vent image; grey; blue chroma; OTSU; FCM A Fast and Effectve Segentaton Algorth for Undersea Hydrotheral Vent Iage FUYUAN PENG 1 QIAN XIA 1 GUOHUA XU 2 XI YU 1 LIN LUO 1 Electronc Inforaton Engneerng Departent of Huazhong Unversty of Scence and

More information

An Efficient Fault-Tolerant Multi-Bus Data Scheduling Algorithm Based on Replication and Deallocation

An Efficient Fault-Tolerant Multi-Bus Data Scheduling Algorithm Based on Replication and Deallocation BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volue 16, No Sofa 016 Prnt ISSN: 1311-970; Onlne ISSN: 1314-4081 DOI: 10.1515/cat-016-001 An Effcent Fault-Tolerant Mult-Bus Data

More information

Introduction. Leslie Lamports Time, Clocks & the Ordering of Events in a Distributed System. Overview. Introduction Concepts: Time

Introduction. Leslie Lamports Time, Clocks & the Ordering of Events in a Distributed System. Overview. Introduction Concepts: Time Lesle Laports e, locks & the Orderng of Events n a Dstrbuted Syste Joseph Sprng Departent of oputer Scence Dstrbuted Systes and Securty Overvew Introducton he artal Orderng Logcal locks Orderng the Events

More information

Pose Invariant Face Recognition using Hybrid DWT-DCT Frequency Features with Support Vector Machines

Pose Invariant Face Recognition using Hybrid DWT-DCT Frequency Features with Support Vector Machines Proceedngs of the 4 th Internatonal Conference on 7 th 9 th Noveber 008 Inforaton Technology and Multeda at UNITEN (ICIMU 008), Malaysa Pose Invarant Face Recognton usng Hybrd DWT-DCT Frequency Features

More information

A Theory of Non-Deterministic Networks

A Theory of Non-Deterministic Networks A Theory of Non-Deternstc Networs Alan Mshcheno and Robert K rayton Departent of EECS, Unversty of Calforna at ereley {alan, brayton}@eecsbereleyedu Abstract oth non-deterns and ult-level networs copactly

More information

Robust Dictionary Learning with Capped l 1 -Norm

Robust Dictionary Learning with Capped l 1 -Norm Proceedngs of the Twenty-Fourth Internatonal Jont Conference on Artfcal Intellgence (IJCAI 205) Robust Dctonary Learnng wth Capped l -Norm Wenhao Jang, Fepng Ne, Heng Huang Unversty of Texas at Arlngton

More information

Optimally Combining Positive and Negative Features for Text Categorization

Optimally Combining Positive and Negative Features for Text Categorization Optally Cobnng Postve and Negatve Features for Text Categorzaton Zhaohu Zheng ZZHENG3@CEDAR.BUFFALO.EDU Rohn Srhar ROHINI@CEDAR.BUFFALO.EDU CEDAR, Dept. of Coputer Scence and Engneerng, State Unversty

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

What is Object Detection? Face Detection using AdaBoost. Detection as Classification. Principle of Boosting (Schapire 90)

What is Object Detection? Face Detection using AdaBoost. Detection as Classification. Principle of Boosting (Schapire 90) CIS 5543 Coputer Vson Object Detecton What s Object Detecton? Locate an object n an nput age Habn Lng Extensons Vola & Jones, 2004 Dalal & Trggs, 2005 one or ultple objects Object segentaton Object detecton

More information

Low training strength high capacity classifiers for accurate ensembles using Walsh Coefficients

Low training strength high capacity classifiers for accurate ensembles using Walsh Coefficients Low tranng strength hgh capacty classfers for accurate ensebles usng Walsh Coeffcents Terry Wndeatt, Cere Zor Unv Surrey, Guldford, Surrey, Gu2 7H t.wndeatt surrey.ac.uk Abstract. If a bnary decson s taken

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications 14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of

More information

Research on action recognition method under mobile phone visual sensor Wang Wenbin 1, Chen Ketang 2, Chen Liangliang 3

Research on action recognition method under mobile phone visual sensor Wang Wenbin 1, Chen Ketang 2, Chen Liangliang 3 Internatonal Conference on Autoaton, Mechancal Control and Coputatonal Engneerng (AMCCE 05) Research on acton recognton ethod under oble phone vsual sensor Wang Wenbn, Chen Ketang, Chen Langlang 3 Qongzhou

More information

Multimodal Biometric System Using Face-Iris Fusion Feature

Multimodal Biometric System Using Face-Iris Fusion Feature JOURNAL OF COMPUERS, VOL. 6, NO. 5, MAY 2011 931 Multodal Boetrc Syste Usng Face-Irs Fuson Feature Zhfang Wang, Erfu Wang, Shuangshuang Wang and Qun Dng Key Laboratory of Electroncs Engneerng, College

More information

Outline. Third Programming Project Two-Dimensional Arrays. Files You Can Download. Exercise 8 Linear Regression. General Regression

Outline. Third Programming Project Two-Dimensional Arrays. Files You Can Download. Exercise 8 Linear Regression. General Regression Project 3 Two-densonal arras Ma 9, 6 Thrd Prograng Project Two-Densonal Arras Larr Caretto Coputer Scence 6 Coputng n Engneerng and Scence Ma 9, 6 Outlne Quz three on Thursda for full lab perod See saple

More information

Multiple Instance Learning via Multiple Kernel Learning *

Multiple Instance Learning via Multiple Kernel Learning * The Nnth nternatonal Syposu on Operatons Research and ts Applcatons (SORA 10) Chengdu-Juzhagou, Chna, August 19 23, 2010 Copyrght 2010 ORSC & APORC, pp. 160 167 ultple nstance Learnng va ultple Kernel

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

User Behavior Recognition based on Clustering for the Smart Home

User Behavior Recognition based on Clustering for the Smart Home 3rd WSEAS Internatonal Conference on REMOTE SENSING, Vence, Italy, Noveber 2-23, 2007 52 User Behavor Recognton based on Clusterng for the Sart Hoe WOOYONG CHUNG, JAEHUN LEE, SUKHYUN YUN, SOOHAN KIM* AND

More information

A system based on a modified version of the FCM algorithm for profiling Web users from access log

A system based on a modified version of the FCM algorithm for profiling Web users from access log A syste based on a odfed verson of the FCM algorth for proflng Web users fro access log Paolo Corsn, Laura De Dosso, Beatrce Lazzern, Francesco Marcellon Dpartento d Ingegnera dell Inforazone va Dotsalv,

More information

Hermite Splines in Lie Groups as Products of Geodesics

Hermite Splines in Lie Groups as Products of Geodesics Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the

More information

IMAGE REPRESENTATION USING EPANECHNIKOV DENSITY FEATURE POINTS ESTIMATOR

IMAGE REPRESENTATION USING EPANECHNIKOV DENSITY FEATURE POINTS ESTIMATOR Sgnal & Iage Processng : An Internatonal Journal (SIPIJ) Vol.4, No., February 03 IMAGE REPRESENTATION USING EPANECHNIKOV DENSITY FEATURE POINTS ESTIMATOR Tranos Zuva, Kenelwe Zuva 3, Sunday O. Ojo, Selean

More information

ENSEMBLE learning has been widely used in data and

ENSEMBLE learning has been widely used in data and IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 9, NO. 5, SEPTEMBER 2012 943 Sparse Kernel-Based Hyperspectral Anoaly Detecton Prudhv Gurra, Meber, IEEE, Heesung Kwon, Senor Meber, IEEE, andtothyhan Abstract

More information

Relevance Feedback in Content-based 3D Object Retrieval A Comparative Study

Relevance Feedback in Content-based 3D Object Retrieval A Comparative Study 753 Coputer-Aded Desgn and Applcatons 008 CAD Solutons, LLC http://www.cadanda.co Relevance Feedback n Content-based 3D Object Retreval A Coparatve Study Panagots Papadaks,, Ioanns Pratkaks, Theodore Trafals

More information

Multicast Tree Rearrangement to Recover Node Failures. in Overlay Multicast Networks

Multicast Tree Rearrangement to Recover Node Failures. in Overlay Multicast Networks Multcast Tree Rearrangeent to Recover Node Falures n Overlay Multcast Networks Hee K. Cho and Chae Y. Lee Dept. of Industral Engneerng, KAIST, 373-1 Kusung Dong, Taejon, Korea Abstract Overlay ultcast

More information

A Study on Clustering for Clustering Based Image De-Noising

A Study on Clustering for Clustering Based Image De-Noising Journal of Informaton Systems and Telecommuncaton, Vol. 2, No. 4, October-December 2014 196 A Study on Clusterng for Clusterng Based Image De-Nosng Hossen Bakhsh Golestan* Department of Electrcal Engneerng,

More information

Algorithm To Convert A Decimal To A Fraction

Algorithm To Convert A Decimal To A Fraction Algorthm To Convert A ecmal To A Fracton by John Kennedy Mathematcs epartment Santa Monca College 1900 Pco Blvd. Santa Monca, CA 90405 jrkennedy6@gmal.com Except for ths comment explanng that t s blank

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

AN ADAPTIVE APPROACH TO THE SEGMENTATION OF DCE-MR IMAGES OF THE BREAST: COMPARISON WITH CLASSICAL THRESHOLDING ALGORITHMS

AN ADAPTIVE APPROACH TO THE SEGMENTATION OF DCE-MR IMAGES OF THE BREAST: COMPARISON WITH CLASSICAL THRESHOLDING ALGORITHMS A ADAPTIVE APPROACH TO THE SEGMETATIO OF DCE-MR IMAGES OF THE BREAST: COMPARISO WITH CLASSICAL THRESHOLDIG ALGORITHMS Fath Kalel a zaettn Aydn a Gohan Ertas H.Ozcan Gulcur a Bahcesehr Unversty Engneerng

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

SUV Color Space & Filtering. Computer Vision I. CSE252A Lecture 9. Announcement. HW2 posted If microphone goes out, let me know

SUV Color Space & Filtering. Computer Vision I. CSE252A Lecture 9. Announcement. HW2 posted If microphone goes out, let me know SUV Color Space & Flterng CSE5A Lecture 9 Announceent HW posted f cropone goes out let e now Uncalbrated Potoetrc Stereo Taeaways For calbrated potoetrc stereo we estated te n by 3 atrx B of surface norals

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

Nighttime Motion Vehicle Detection Based on MILBoost

Nighttime Motion Vehicle Detection Based on MILBoost Sensors & Transducers 204 by IFSA Publshng, S L http://wwwsensorsportalco Nghtte Moton Vehcle Detecton Based on MILBoost Zhu Shao-Png,, 2 Fan Xao-Png Departent of Inforaton Manageent, Hunan Unversty of

More information

A Novel System for Document Classification Using Genetic Programming

A Novel System for Document Classification Using Genetic Programming Journal of Advances n Inforaton Technology Vol. 6, No. 4, Noveber 2015 A Novel Syste for Docuent Classfcaton Usng Genetc Prograng Saad M. Darwsh, Adel A. EL-Zoghab, and Doaa B. Ebad Insttute of Graduate

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

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

Using Gini-Index for Feature Selection in Text Categorization

Using Gini-Index for Feature Selection in Text Categorization 3rd Internatonal Conference on Inforaton, Busness and Educaton Technology (ICIBET 014) Usng Gn-Index for Feature Selecton n Text Categorzaton Zhu Wedong 1, Feng Jngyu 1 and Ln Yongn 1 School of Coputer

More information

Heuristic Methods for Locating Emergency Facilities

Heuristic Methods for Locating Emergency Facilities Heurstc Methods for Locatng Eergency Facltes L. Caccetta and M. Dzator Western Australan Centre of Excellence n Industral Optsaton, Curtn Unversty of Technology, Kent Street, Bentley WA 602, Australa E-Mal:

More information

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005 Exercses (Part 4) Introducton to R UCLA/CCPR John Fox, February 2005 1. A challengng problem: Iterated weghted least squares (IWLS) s a standard method of fttng generalzed lnear models to data. As descrbed

More information

Aircraft Engine Gas Path Fault Diagnosis Based on Fuzzy Inference

Aircraft Engine Gas Path Fault Diagnosis Based on Fuzzy Inference 202 Internatonal Conference on Industral and Intellgent Inforaton (ICIII 202) IPCSIT vol.3 (202) (202) IACSIT Press, Sngapore Arcraft Engne Gas Path Fault Dagnoss Based on Fuzzy Inference Changzheng L,

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

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

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

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

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

RESEARCH ON CLOSE-RANGE PHOTOGRAMMETRY WITH BIG ROTATION ANGLE

RESEARCH ON CLOSE-RANGE PHOTOGRAMMETRY WITH BIG ROTATION ANGLE RESEARCH ON CLOSE-RANGE PHOOGRAMMERY WIH BIG ROAION ANGLE Lu Jue a a he Departent of Surveyng and Geo-nforatcs Engneerng, ongj Unversty, Shangha, 9. - lujue985@6.co KEY WORDS: Bg Rotaton Angle; Colnearty

More information

Joint Registration and Active Contour Segmentation for Object Tracking

Joint Registration and Active Contour Segmentation for Object Tracking Jont Regstraton and Actve Contour Segentaton for Object Trackng Jfeng Nng a,b, Le Zhang b,1, Meber, IEEE, Davd Zhang b, Fellow, IEEE and We Yu a a College of Inforaton Engneerng, Northwest A&F Unversty,

More information

Laplacian Eigenmap for Image Retrieval

Laplacian Eigenmap for Image Retrieval Laplacan Egenmap for Image Retreval Xaofe He Partha Nyog Department of Computer Scence The Unversty of Chcago, 1100 E 58 th Street, Chcago, IL 60637 ABSTRACT Dmensonalty reducton has been receved much

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

AUTOMATIC IMAGE SEQUENCE REGISTRATION BASED ON A LINEAR SOLUTION AND SCALE INVARIANT KEYPOINT MATCHING

AUTOMATIC IMAGE SEQUENCE REGISTRATION BASED ON A LINEAR SOLUTION AND SCALE INVARIANT KEYPOINT MATCHING AUOMAIC IMAGE SEQUENCE REGISRAION BASED ON A LINEAR SOLUION AND SCALE INVARIAN KEPOIN MACHING Z. Shraga, S. Barnea, S. Fln, G. Zalanson,. Doytsher Departent of ransportaton and Geo-Inforaton, echnon Israel

More information

RECONSTRUCTION OF 3D FACES BY SHAPE ESTIMATION AND TEXTURE INTERPOLATION

RECONSTRUCTION OF 3D FACES BY SHAPE ESTIMATION AND TEXTURE INTERPOLATION http://www.ts.uk.y/apjt Asa-Pacc Journal o Inoraton echnology and Multeda Jurnal eknolog Makluat dan Multeda Asa-Pask Vol. 3 No. 1, June 2014 : 15-21 e-issn: 2289-2192 RECONSRUCION OF 3D FACES BY SHAPE

More information

Structure from motion (SfM) Simultaneous Localization and Mapping (SLAM)

Structure from motion (SfM) Simultaneous Localization and Mapping (SLAM) Structure fro oton (SfM) Sultaneous Localzaton and Mappng (SLAM) Schedule Feb 20 Feb 27 Mar 5 Mar 2 Mar 9 Mar 26 Apr 2 Apr 9 Apr 6 Apr 23 Apr 30 May 7 May 4 May 2 May 28 ue., May 29 Introducton Lecture:

More information

Modelling Spatial Substructure in Wildlife Populations using an Approximation to the Shortest Path Voronoi Diagram

Modelling Spatial Substructure in Wildlife Populations using an Approximation to the Shortest Path Voronoi Diagram 18 th World IMACS / MODSIM Congress, Carns, Australa 13-17 July 2009 http://ssanz.org.au/ods09 Modellng Spatal Substructure n Wldlfe Populatons usng an Approxaton to the Shortest Path Vorono Dagra Stewart,

More information

IMPROVED INITIAL VALUE PREDICTION FOR GLOBAL MOTION ESTIMATION

IMPROVED INITIAL VALUE PREDICTION FOR GLOBAL MOTION ESTIMATION 18th European Sgnal Processng Conference (EUSIPCO-010) Aalborg, Denark, August 3-7, 010 IMPROVED INITIAL VALUE PREDICTION FOR GLOBAL MOTION ESTIMATION Adel Ahad 1, Hojjat Salehnejad 1, Saak Taleb 1,, and

More information

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal

More information

A Cluster Tree Method For Text Categorization

A Cluster Tree Method For Text Categorization Avalable onlne at www.scencedrect.co Proceda Engneerng 5 (20) 3785 3790 Advanced n Control Engneerngand Inforaton Scence A Cluster Tree Meod For Text Categorzaton Zhaoca Sun *, Yunng Ye, Weru Deng, Zhexue

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

Parallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016)

Parallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016) Technsche Unverstät München WSe 6/7 Insttut für Informatk Prof. Dr. Thomas Huckle Dpl.-Math. Benjamn Uekermann Parallel Numercs Exercse : Prevous Exam Questons Precondtonng & Iteratve Solvers (From 6)

More information

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1) Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A

More information

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges

More information

Robust Low-Rank Regularized Regression for Face Recognition with Occlusion

Robust Low-Rank Regularized Regression for Face Recognition with Occlusion Robust Low-Rank Regularzed Regresson for ace Recognton wth Occluson Janjun Qan, Jan Yang, anlong Zhang and Zhouchen Ln School of Computer Scence and ngneerng, Nanjng Unversty of Scence and echnology Key

More information

2016 International Conference on Sustainable Energy, Environment and Information Engineering (SEEIE 2016) ISBN:

2016 International Conference on Sustainable Energy, Environment and Information Engineering (SEEIE 2016) ISBN: 06 Internatonal Conference on Sustanable Energy, Envronent and Inforaton Engneerng (SEEIE 06) ISBN: 978--60595-337-3 A Study on IEEE 80. MAC Layer Msbehavor under Dfferent Back-off Algorths Trong Mnh HOANG,,

More information

AN ALGORITHM FOR ODD GRACEFULNESS OF THE TENSOR PRODUCT OF TWO LINE GRAPHS

AN ALGORITHM FOR ODD GRACEFULNESS OF THE TENSOR PRODUCT OF TWO LINE GRAPHS Internatonal ournal on applcatons of graph theory n wreless ad hoc networks and sensor networks (GRAPH-HOC) Vol.3, No., March 0 AN ALGORITHM FOR ODD GRACEFULNESS OF THE TENSOR PRODUCT OF TWO LINE GRAPHS

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

Reading. 14. Subdivision curves. Recommended:

Reading. 14. Subdivision curves. Recommended: eadng ecommended: Stollntz, Deose, and Salesn. Wavelets for Computer Graphcs: heory and Applcatons, 996, secton 6.-6., A.5. 4. Subdvson curves Note: there s an error n Stollntz, et al., secton A.5. Equaton

More information

Face Detection and Tracking in Video Sequence using Fuzzy Geometric Face Model and Mean Shift

Face Detection and Tracking in Video Sequence using Fuzzy Geometric Face Model and Mean Shift Internatonal Journal of Advanced Trends n Coputer Scence and Engneerng, Vol., No.1, Pages : 41-46 (013) Specal Issue of ICACSE 013 - Held on 7-8 January, 013 n Lords Insttute of Engneerng and Technology,

More information

Realistic 3D Face Modeling by Fusing Multiple 2D Images

Realistic 3D Face Modeling by Fusing Multiple 2D Images Realstc 3D Face Modelng by Fusng Multple D ages Changhu Wang EES Departent, Unversty of Scence and echnology of Chna, wch@ustc.edu Shucheng Yan, Hongjang Zhang, Weyng Ma Mcrosoft Research Asa, Bejng,.R.

More information

FUZZY C-MEANS ALGORITHMS IN REMOTE SENSING

FUZZY C-MEANS ALGORITHMS IN REMOTE SENSING FUZZY C-MEAS ALGORITHMS I REMOTE SESIG Andrej Turčan, Eva Ocelíková, Ladslav Madarász Dept. of Cybernetcs and Artfcal Intellgence Faculty of Electrcal Engneerng and Inforatcs Techncal Unversty of Košce

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

Wavefront Reconstructor

Wavefront Reconstructor A Dstrbuted Smplex B-Splne Based Wavefront Reconstructor Coen de Vsser and Mchel Verhaegen 14-12-201212 2012 Delft Unversty of Technology Contents Introducton Wavefront reconstructon usng Smplex B-Splnes

More information

A Partial Decision Scheme for Local Search Algorithms for Distributed Constraint Optimization Problems

A Partial Decision Scheme for Local Search Algorithms for Distributed Constraint Optimization Problems A Partal Decson Schee for Local Search Algorths for Dstrbuted Constrant Optzaton Probles Zhepeng Yu, Zyu Chen*, Jngyuan He, Yancheng Deng College of Coputer Scence, Chongqng Unversty, Chongqng, Chna {204402053,

More information

Non-Split Restrained Dominating Set of an Interval Graph Using an Algorithm

Non-Split Restrained Dominating Set of an Interval Graph Using an Algorithm Internatonal Journal of Advancements n Research & Technology, Volume, Issue, July- ISS - on-splt Restraned Domnatng Set of an Interval Graph Usng an Algorthm ABSTRACT Dr.A.Sudhakaraah *, E. Gnana Deepka,

More information

A Novel Fuzzy Classifier Using Fuzzy LVQ to Recognize Online Persian Handwriting

A Novel Fuzzy Classifier Using Fuzzy LVQ to Recognize Online Persian Handwriting A Novel Fuzzy Classfer Usng Fuzzy LVQ to Recognze Onlne Persan Handwrtng M. Soleyan Baghshah S. Bagher Shourak S. Kasae Departent of Coputer Engneerng, Sharf Unversty of Technology, Tehran, Iran soleyan@ce.sharf.edu

More information

Review of approximation techniques

Review of approximation techniques CHAPTER 2 Revew of appromaton technques 2. Introducton Optmzaton problems n engneerng desgn are characterzed by the followng assocated features: the objectve functon and constrants are mplct functons evaluated

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

STATIC MAPPING FOR OPENCL WORKLOADS IN HETEROGENEOUS COMPUTER SYSTEMS

STATIC MAPPING FOR OPENCL WORKLOADS IN HETEROGENEOUS COMPUTER SYSTEMS STATIC MAPPING FOR OPENCL WORKLOADS IN HETEROGENEOUS COMPUTER SYSTEMS 1 HENDRA RAHMAWAN, 2 KUSPRIYANTO, 3 YUDI SATRIA GONDOKARYONO School of Electrcal Engneerng and Inforatcs, Insttut Teknolog Bandung,

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

Measuring Cohesion of Packages in Ada95

Measuring Cohesion of Packages in Ada95 Measurng Coheson of Packages n Ada95 Baowen Xu Zhenqang Chen Departent of Coputer Scence & Departent of Coputer Scence & Engneerng, Southeast Unversty Engneerng, Southeast Unversty Nanjng, Chna, 20096

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