MEAN SHIFT ALGORITHM AND ITS APPLICATION IN TRACKING OF OBJECTS
|
|
- Bathsheba Cole
- 5 years ago
- Views:
Transcription
1 Proceens o e F Inernaonal Conerence on Macne Learnn an Cbernecs Dalan 3-6 Auus 006 MEAN SHIF ALGORIHM AND IS APPLICAION IN RACKING OF OBJECS ZHI-QIANG WEN ZI-XING CAI Collee o Inormaon Scence an Enneern Cenral Sou Unvers HuNan Cansa Cna zqwen000@63.com zxca@csu.eu.cn Absrac: Mean s alorm s recenl wel use n racn clusern ec owever converence o mean s alorm as no been rorousl prove. In s paper mean s alorm w Gaussan prole s sue an apple o racn o objecs. e mprecse proos abou converence o mean s are rsl pone ou. en a converence eorem an s rorous converence proo are prove. Lasl racn approac o objecs base on mean s s moe. e resuls o expermen sow e moe approac as oo perormance o objec racn apple o occluson. e conrbuons n s paper are expece o urer su an applcaon n mean s alorm. Kewors: Mean s alorm; Converence; Kernel uncon; racn o objec; Baacara coecen. Inroucon Mean s wc was propose n 975 b Fuunaa an Hoseler[] s a nonparamerc erave proceure a ss eac aa o local maxmum o ens uncon. In spe o s oo properes as been nore unl Cen s paper[] renews our neres n. Cen n [] revse mean s evelopn a more eneral ormulaon an emonsran s poenal uses n clusern an lobal opmzaon. Snce en mean s as been wel use n objec racn[3-7] mae semenaon[89] paern reconon an clusern[0] lern[] normaon uson[3] an ec. Cen[] scusse e mean s alorm n ree was an cel sue e blurrn process. Le aa be a ne se S embee n e -mensonal Euclean space X. Le X be a ne se prouce n erave proceure. Wen S e mean s proceure s a blurrn process namel e npu aa s recursvel moe aer eac mean s sep. Cen n [] prove e proos or e converence o blurrn process. However wen S s xe rou e process an s nalze o S s no loner a blurrn process an Cen s eores o no appl. Comancu n [4] propose e mean s proceure rom ens esmaon an prove convere a neares saonar pon o e unerln ens uncon. Comancu s wor n [4] ncel s wa Cen as no eal w. L n [5] oun e msae n Comancu s converence proo n [4] an ave some counerexamples. Comancu also mae e same msae n [46]. L prove e converence o mean s n a new wa. In ac e proo n [4456] s mprecse an e converence o mean s nee o be sue. Armn a mean s alorm w Gaussan prole s paper sues above problem. e man conrbuons o s paper are: e mprecse proos abou converence o mean s n [4456] are pone ou an a rorous converence proo s prove. Moreover e racn approac o objecs base on mean s s moe. e paper s oranze as ollows: mean s alorm s nrouce n secon. Secon 3 proves e proo or e converence o mean s. Moe racn approac o objecs an s expermen are presene n secon 4. Secon 5 s e concluson.. Mean s Gven n aa pons x n n e -mensonal space R erave ormula o mean s s as ollows. n x x n x We can el λ Were λ c > /06/$ IEEE 404
2 Proceens o e F Inernaonal Conerence on Macne Learnn an Cbernecs Dalan 3-6 Auus 006 n c x x s a mulvarae ernel ens n esmaor w prole wc s Gaussan prole. c n x x were c n c s e corresponn normalzaon consan. We ene ervave n x x x 3 m x n x x From we can el m 4 were x m c sows mean s alernaes owar e raen recon an leas e ornal pons s o a local maxmum pon o srbun ens uncon. e sep sze λ canes on w e wole erave process. 3. Converence o mean s Leraure [446] prove e converence eores. However ern rom [44] Leraure [6] sue e varable-banw mean s. L n [5] oun a e proos wc were prove or converence o mean s n leraure [446] are wron. Wen leraure [4] was provn e converence o sequence {...} L n[5] consere s wron o euce {...} convere rom - converes o zero an ave some counerexamples. L n [5] use a wen parameer c n Comancu s meo. L prove e converence o mean s n a new wa. e wor L n [5] s o n above msaes an correc bu e no rascall correc. L[5] an Comancu [446] use e proper o convex uncon o prove e sequences... convere an s monooncall ncreasn bu uncon x s rewren n -x wc s no alwas a convex uncon possbl s a concave uncon or neer o em. Wen x s a convex uncon sequences are also no alwas convere o a local maxmum pon. e enons an properes abou convex se convex uncon an concave uncon ma consul leraure[7] or oer leraures. eorem. an s proo are prove as ollows or e mprecse proo n [45]. eorem. e proo o eorem or example: I e ernel K as a convex an monooncall ecreasn prole e sequences { }... an { K }... convere an { K }... s monooncall ncreasn[4] n [45] s mprecse. Proo. e eals abou e proo can be oun n [45]. We suppose e proo n [45] s accurae so sequences convere an monooncall ncrease. From [45] we now convere a a pon assumn an en e expresson 0 s rue. Leraure [4] prove e proo or s eor accorn o proper o convex uncon: x x x x -x as ollows e expresson n [4] s use: c n x K j K j x j x n Here e auor consere x s a convex uncone proo n [5] s smlar o a n [4]. Accorn o e proper o convex uncon s convex an Hessan marx s a posve semene marx. Accorn o secon-orer alor seres expanson ormula aroun we ave: were η 0 <η <. Aer s replace b we el From > we ave < 0. Assumn e as connuous paral ervaves o secon orer ere exs a Neboroo Ω o ŷ sasn s a neave ene marx were Ω. Wen an. So ere a leas exs a τ τ... sasn τ Ω wc mples τ s a neave ene marx. s s paraoxcal w e proper a... s a posve semene marx. So e proos n [45] are mprecse. From e proo o eorem we suppose sequence... convere an monooncall ncrease e Hessan marx s a neave ene 405
3 Proceens o e F Inernaonal Conerence on Macne Learnn an Cbernecs Dalan 3-6 Auus 006 marx so ere s eorem. In orer o prove proo or e eorem a lemma s ven as ollows Lemma I x s Gaussan uncon e ollown expresson s rue. > Proo. From [4] we now a x s Gaussan uncon e ollown expresson s rue m m > 0 m m 6 From e 6 we can el c c > 0 So > 0 eorem. Converence eorem Assumn S R s nonemp open convex se : S R as connuous paral ervaves o secon orer n S. I S s Gaussan uncon an... e Hessan Marx o sequences convere an monooncall ncrease moreover sequences { } convere. Proo. Accorn o e properes o prole x s boune so s also boune. I wll explan e converence o monooncall ncreases. As accorn o rs-orer alor seres expanson ormula aroun an aer s replace b λ S we ave: λ λ θλ were 0 <θ <. Assumn ϕ θ λ λ θλ e ollown expresson s rue. lm θ 0 ϕ θ λ θλ λ 0 > lm ϕ θ θ λ λ λ > 0 Lemma Funcon ϕ θ s connuous an erenable mpln ϕ θλ ' θ λ θλ < 0 Easl eln θλ S. So ϕ θ monooncall ecrease an ϕ θ > 0 θ 0 an expresson > can be conclue. Consequenl sequence converes an monooncall ncreases. Assumn sequence converes o e pon S we can rve e equal lm ϕ θ 0. In aonor 0 < ϕ < ϕ θ we ave wo equales 7a an 7b lm 0 7a lm 0 7b e le erm o 7a an 7b are mulple b eac oer eln lm 0 wc mples lm 0 a s lm 0 so s converen. For e connu o uncon s easl nown o convere o pon an 0 e aapve sep szes s aope n mean s o uaranee converence wc also elmnaes e nee or aonal proceures o coose e aequae sep szes or example nn e opmum sep szes n seepes escen meo[78]. s s a major avanae over e raonal raen-base meos. e banw onl eces e number o observe pea values[0] an e sze o reon were s concave. Generall e number o peas ecreases w e ncrease o e x 0 banw. So or Hessan marx H s mean 0 s can convere o a sea pon. 4. Expermen on racn base on mean s e racn alorm base mean s n [9] s aope n s paper an more eals abou s alorm can be oun n [9]. Bu n s paper ere are some erences rom [9] as ollows. 0 x Hessan marx H s were x s e 0 w an e o objec respecvel. 406
4 Proceens o e F Inernaonal Conerence on Macne Learnn an Cbernecs Dalan 3-6 Auus 006 e prole s Gaussan uncon. x e x λ x 0 x > λ 3 Opmum x s searce sasn a s maxmum. e veo n our expermens s acqure n ouoor a srbuon o Baacara coecen Fure. Analss n process o expermen b erave mes a 0 rame b 30 rame c 70 rame 0 rame Fure. e resul o expermen wenλ Fure 3. Alerave Baacara coecen envronmen rou vson ssem nsalle n robo o our Lab[0]. A objec an s nal locaon are poecall nown. rou e mean s alorm e locaon o objec s oun aer some mes. Furea sows e srbuon o Baacara coecenρaroun e nal locaon o objec. Aer some erave mes e nex a 0 rame b 30 rame Fure 4. e resul o expermen wenλ locaon s oun. Fure b sows e erave mes n process o nn e objec rom e rs rame o 0 rame. Fure sows e resul o expermen. Aer 0 rames e objec can be oun; even ere are some occlusons. Fure 3 sows e alerave Baacara coecen n eac rame so e curve n Fure3 as wo mnmum pons A an B a sow ere are wo 407
5 Proceens o e F Inernaonal Conerence on Macne Learnn an Cbernecs Dalan 3-6 Auus 006 occlusons n process o objec movn. For occlusons e Baacara coecen value s qucl ecreasn. Aer a s qucl ncreasn. e me o occluson can no be oo lon; oerwse e objec wll be los. λ s mporan or e perormance o objec racn apple o occluson. I λ s oo small e objec s easl los. I λ s oo lare e compue me s oo lon an real me perormance s ba. Wen λ e resuls o expermen n Fure4 sow e objec s los. Generall λ wen λ n Fure sows e oo resuls. 5. Conclusons s paper prmarl sues e converence on mean s alorm w Gaussan prole an presens a moe racn approac o objec base on mean s. A rs s paper revew e mean s alorm en e mprecse proo n [45] s pone ou. r a new converence eorem an s proo are prove; lasl a moe racn approac o objec s presene. s approac can be apple o occluson ssue. However e parameer λ as an mporan eec on perormance o racn o objec n occluson envronmen. Generall λ.5-.5 can ensure e oo resuls. Acnowleemens s wor was suppore b e Naonal Naural Scence Founaon o Cna uner Gran No e auors an e wole member n nsue o nellen robo o Cenral Sou Unvers. Beses we an e anonmous revewers or er oo avce. Reerences [] Fuunaa K an Hoseler LD e esmaon o e raen o a ens uncon w applcaons n paern reconon IEEE rans. Inormaon eor vol. pp [] Cen Y Mean s moe seen an clusern IEEE rans. on Paern Analss an Macne Inellence vol.7 no.8 pp [3] Comancu D an Rames V Mean s an opmal precon or ecen objec racn In: Mojslovc A Hu J es. Proc. o e IEEE In l Con. on Imae Processn ICIP pp [4] Comancu D Rames V an Meer P Real-me racn o non-r objecs usn mean s In: Proc. o e IEEE Con. on Compuer Vson an Paern Reconon CVPR pp [5] Collns R Mean s blob racn rou scale space In: Proc. o e Con. on Compuer Vson an Paern Reconon CVPR pp [6] San C We Y an an e al Real me Han racn b Combnn Parcle Flern an Mean S In: Proc. o e 6 IEEE Inernaonal Con. on Auomac Face an Gesure Reconon 7-9 Ma pp [7] Mao E an Cavallaro A Hbr Parcle Fler an Mean S racer w aapve ranson moel In: Proc. o IEEE Snal Proc. Soce In. Con. on Acouscs Speec an Snal Processn ICASSP Plaelpa PA USA Marc pp [8] Comancu D Imae semenaon usn clusern w sale pon eecon In: Proc. o e IEEE In l Con. on Imae Processn ICIP pp [9] Wan J esson B an Y. Xu e al Imae an Veo Semenaon b Ansoropc Kernel Mean S In: Proc. European Con. on Compuer Vson ECCV 004. [0] Comancu D Rames V an A. D. Bue Mulvarae sale pon eecon or sascal clusern In: Proc. o e European Con. Compuer Vson ECCV. Pp [] Georescu B Smson I an Meer P Mean S Base Clusern n H Dmensons: A exure Classcaon Example In: Proc. ICCV Oc. pp [] Comancu D an Meer P Mean s analss an applcaons In: Proc. o e IEEE In l Con. on Compuer Vson ICCV pp [3] Comancu D Nonparamerc normaon uson or moon esmaon In: Proc. o e IEEE Con. on Compuer Vson an Paern Reconon CVPR pp [4] Comancu D an Meer P Mean s: A robus approac owar eaure space analss IEEE rans. on Paern Analss an Macne Inellence vol.4 no.5 pp [5] L X an Wu F e al Converence o a mean s alorm Journal o Soware vol.6 no.3 pp n Cnese. [6] Comancu D Rames V an Meer P e varable banw mean s an aa-rven scale selecon In: Proc. o e IEEE In l Con. on Compuer Vson ICCV pp [7] Xe Z L J an an Z non-lnear opmzaon Cansa: Naonal Unvers o Deence ecnolo Publs House pp n Cnese. [8] S Y Globall converen alorms or unconsrane opmzaon Compuaonal Opmzaon an Applcaons vol.6 pp [9] Comancu D Rames V Meer P Kernel-Base Objec racn IEEE rans. on Paern Analss an Macne Inellence Vol. 5 no pp [0] Ca ZX Zou XB e al Desn o Dsrbue Conrol Ssem or Moble Robo J. Cen. Sou Unv. scence an ecnolo vol.36 no.5 005pp: n Cnese 408
On Convergence Rate of Concave-Convex Procedure
On Converence Rae o Concave-Conve Proceure Ian E.H. Yen Nanun Pen Po-We Wan an Shou-De Ln Naonal awan Unvers OP 202 Oulne Derence o Conve Funcons.c. Prora Applcaons n SVM leraure Concave-Conve Proceure
More informationOutline. Energy-Efficient Target Coverage in Wireless Sensor Networks. Sensor Node. Introduction. Characteristics of WSN
Ener-Effcen Tare Coverae n Wreless Sensor Newors Presened b M Trà Tá -4-4 Inroducon Bacround Relaed Wor Our Proosal Oulne Maxmum Se Covers (MSC) Problem MSC Problem s NP-Comlee MSC Heursc Concluson Sensor
More informationRelative controllability of nonlinear systems with delays in control
Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.
More informationThe Study of Target Tracking Based on ARM Embedded Platform
JOURNAL OF COMPUTERS, VOL. 7, NO. 8, AUGUST 01 015 Te Suy of Tare Trackn Base on ARM Embee Plaform Lan Pan Scool of Informaon Scence an Enneern,Wuan Unversy of Scence an Tecnoloy,Wuan Cna E-mal: plwsco@163.com
More informationPHYS 1443 Section 001 Lecture #4
PHYS 1443 Secon 001 Lecure #4 Monda, June 5, 006 Moon n Two Dmensons Moon under consan acceleraon Projecle Moon Mamum ranges and heghs Reerence Frames and relae moon Newon s Laws o Moon Force Newon s Law
More informationLecture 6: Learning for Control (Generalised Linear Regression)
Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson
More informationFTCS Solution to the Heat Equation
FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence
More informationCHAPTER 5: MULTIVARIATE METHODS
CHAPER 5: MULIVARIAE MEHODS Mulvarae Daa 3 Mulple measuremens (sensors) npus/feaures/arbues: -varae N nsances/observaons/eamples Each row s an eample Each column represens a feaure X a b correspons o he
More informationRobustness Experiments with Two Variance Components
Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference
More informationIncluding the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.
Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample
More informationCS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4
CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped
More informationVariants of Pegasos. December 11, 2009
Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on
More informationIn the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!
ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal
More informationFall 2010 Graduate Course on Dynamic Learning
Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/
More informationChapter 6: AC Circuits
Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.
More informationStability Analysis of Fuzzy Hopfield Neural Networks with Timevarying
ISSN 746-7659 England UK Journal of Informaon and Compung Scence Vol. No. 8 pp.- Sably Analyss of Fuzzy Hopfeld Neural Neworks w mevaryng Delays Qfeng Xun Cagen Zou Scool of Informaon Engneerng Yanceng
More informationMidterm Exam. Thursday, April hour, 15 minutes
Economcs of Grow, ECO560 San Francsco Sae Unvers Mcael Bar Sprng 04 Mderm Exam Tursda, prl 0 our, 5 mnues ame: Insrucons. Ts s closed boo, closed noes exam.. o calculaors of an nd are allowed. 3. Sow all
More informationTHE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS
THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he
More informationStructural Optimization Using Metamodels
Srucural Opmzaon Usng Meamodels 30 Mar. 007 Dep. o Mechancal Engneerng Dong-A Unvers Korea Kwon-Hee Lee Conens. Numercal Opmzaon. Opmzaon Usng Meamodels Impac beam desgn WB Door desgn 3. Robus Opmzaon
More informationFuzzy derivations KU-ideals on KU-algebras BY
Fu ervaons KU-eals on KU-algebras BY Sam M.Mosaa, Ahme Ab-elaem 2 sammosaa@ahoo.com ahmeabelaem88@ahoo.com,2deparmen o mahemacs -Facul o Eucaon -An Shams Unvers Ro, Caro, Egp Absrac. In hs manuscrp, we
More informationJohn Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany
Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy
More informationLecture VI Regression
Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M
More informationSensor Scheduling for Multiple Parameters Estimation Under Energy Constraint
Sensor Scheduln for Mulple Parameers Esmaon Under Enery Consran Y Wan, Mnyan Lu and Demoshens Tenekezs Deparmen of Elecrcal Enneern and Compuer Scence Unversy of Mchan, Ann Arbor, MI {yws,mnyan,eneke}@eecs.umch.edu
More informationStatistical performance analysis by loopy belief propagation in probabilistic image processing
Statstcal perormance analyss by loopy bele propaaton n probablstc mae processn Kazuyuk Tanaka raduate School o Inormaton Scences Tohoku Unversty Japan http://www.smapp.s.tohoku.ac.p/~kazu/ Collaborators
More information( 1) β function for the Higgs quartic coupling λ in the standard model (SM) h h. h h. vertex correction ( h 1PI. Σ y. counter term Λ Λ.
funon for e Hs uar oun n e sanar moe (SM verex >< sef-ener ( PI Π ( - ouner erm ( m, ( Π m s fne Π s fne verex orreon ( PI Σ (,, ouner erm, ( reen funon ({ } Σ s fne Λ Λ Bn A n ( Caan-Smanz euaon n n (
More informationLecture 12: HEMT AC Properties
Lecure : HEMT A Proeres Quas-sac oeraon Transcaacances -araeers Non-quas ac effecs Parasc ressances / caacancs f f ax ean ue for aer 6: 7-86 95-407 {407-46 sk MEFET ars} 47-44. (.e. sk an MEFET ars brefl
More informationA NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION
S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy
More informationPattern Classification (III) & Pattern Verification
Preare by Prof. Hu Jang CSE638 --4 CSE638 3. Seech & Language Processng o.5 Paern Classfcaon III & Paern Verfcaon Prof. Hu Jang Dearmen of Comuer Scence an Engneerng York Unversy Moel Parameer Esmaon Maxmum
More information12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer
d Model Cvl and Surveyng Soware Dranage Analyss Module Deenon/Reenon Basns Owen Thornon BE (Mech), d Model Programmer owen.hornon@d.com 4 January 007 Revsed: 04 Aprl 007 9 February 008 (8Cp) Ths documen
More informationKernel-Based Bayesian Filtering for Object Tracking
Kernel-Based Bayesan Flerng for Objec Trackng Bohyung Han Yng Zhu Dorn Comancu Larry Davs Dep. of Compuer Scence Real-Tme Vson and Modelng Inegraed Daa and Sysems Unversy of Maryland Semens Corporae Research
More informationAn introduction to Support Vector Machine
An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,
More informationDelay-Range-Dependent Stability Analysis for Continuous Linear System with Interval Delay
Inernaonal Journal of Emergng Engneerng esearch an echnology Volume 3, Issue 8, Augus 05, PP 70-76 ISSN 349-4395 (Prn) & ISSN 349-4409 (Onlne) Delay-ange-Depenen Sably Analyss for Connuous Lnear Sysem
More informationDisplacement, Velocity, and Acceleration. (WHERE and WHEN?)
Dsplacemen, Velocy, and Acceleraon (WHERE and WHEN?) Mah resources Append A n your book! Symbols and meanng Algebra Geomery (olumes, ec.) Trgonomery Append A Logarhms Remnder You wll do well n hs class
More informationdoi: info:doi/ /
do: nfo:do/0.063/.322393 nernaonal Conference on Power Conrol and Opmzaon, Bal, ndonesa, -3, June 2009 A COLOR FEATURES-BASED METHOD FOR OBJECT TRACKNG EMPLOYNG A PARTCLE FLTER ALGORTHM Bud Sugand, Hyoungseop
More information( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model
BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng
More informationComparison of Differences between Power Means 1
In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,
More informationDEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL
DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA
More informationMotion in Two Dimensions
Phys 1 Chaper 4 Moon n Two Dmensons adzyubenko@csub.edu hp://www.csub.edu/~adzyubenko 005, 014 A. Dzyubenko 004 Brooks/Cole 1 Dsplacemen as a Vecor The poson of an objec s descrbed by s poson ecor, r The
More informationSuyash Narayan Mishra, Piyush Kumar Tripathi & Alok Agrawal
IOSR Journal o Mahemaics IOSR-JM e-issn: 78-578 -ISSN: 39-765X. Volume Issue Ver. VI Mar - Ar. 5 PP 43-5 www.iosrjournals.org A auberian heorem or C α β- Convergence o Cesaro Means o Orer o Funcions Suash
More informationIntroduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms
Course organzaon Inroducon Wee -2) Course nroducon A bref nroducon o molecular bology A bref nroducon o sequence comparson Par I: Algorhms for Sequence Analyss Wee 3-8) Chaper -3, Models and heores» Probably
More informationEA Properties of NCGPC applied to nonlinear SISO systems with a relative degree one or two M. DABO, N. LANGLOIS & H. CHAFOUK
EA 4353 Properies o NCGPC applied o nonlinear SISO ssems wi a relaive degree one or wo. DABO, N. ANGOIS H. CHAFOU G Commande Prédicive Non inéaire ENSA, Paris 3 janvier Ouline Relaive degree o nonlinear
More informationCS 536: Machine Learning. Nonparametric Density Estimation Unsupervised Learning - Clustering
CS 536: Machne Learnng Nonparamerc Densy Esmaon Unsupervsed Learnng - Cluserng Fall 2005 Ahmed Elgammal Dep of Compuer Scence Rugers Unversy CS 536 Densy Esmaon - Cluserng - 1 Oulnes Densy esmaon Nonparamerc
More informationUS Monetary Policy and the G7 House Business Cycle: FIML Markov Switching Approach
U Monear Polc and he G7 House Busness Ccle: FML Markov wchng Approach Jae-Ho Yoon 5 h Jul. 07 Absrac n order o deermne he eec o U monear polc o he common busness ccle beween housng prce and GDP n he G7
More informationThe Analysis of the Thickness-predictive Model Based on the SVM Xiu-ming Zhao1,a,Yan Wang2,band Zhimin Bi3,c
h Naonal Conference on Elecrcal, Elecroncs and Compuer Engneerng (NCEECE The Analyss of he Thcknesspredcve Model Based on he SVM Xumng Zhao,a,Yan Wang,band Zhmn B,c School of Conrol Scence and Engneerng,
More information( ) [ ] MAP Decision Rule
Announcemens Bayes Decson Theory wh Normal Dsrbuons HW0 due oday HW o be assgned soon Proec descrpon posed Bomercs CSE 90 Lecure 4 CSE90, Sprng 04 CSE90, Sprng 04 Key Probables 4 ω class label X feaure
More informationA New Generalisation of Sam-Solai s Multivariate symmetric Arcsine Distribution of Kind-1*
IOSR Journal o Mahemacs IOSRJM ISSN: 78-578 Volume, Issue May-June 0, PP 4-48 www.osrournals.org A New Generalsaon o Sam-Sola s Mulvarae symmerc Arcsne Dsrbuon o Knd-* Dr. G.S. Davd Sam Jayaumar. Dr.A.Solarau.
More informationApproximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy
Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae
More informationCHAPTER 2: Supervised Learning
HATER 2: Supervsed Learnng Learnng a lass from Eamples lass of a famly car redcon: Is car a famly car? Knowledge eracon: Wha do people epec from a famly car? Oupu: osve (+) and negave ( ) eamples Inpu
More informationAdvanced time-series analysis (University of Lund, Economic History Department)
Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng
More informationAdvanced Machine Learning & Perception
Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel
More informationLecture 11 SVM cont
Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc
More informationOP = OO' + Ut + Vn + Wb. Material We Will Cover Today. Computer Vision Lecture 3. Multi-view Geometry I. Amnon Shashua
Comuer Vson 27 Lecure 3 Mul-vew Geomer I Amnon Shashua Maeral We Wll Cover oa he srucure of 3D->2D rojecon mar omograh Marces A rmer on rojecve geomer of he lane Eolar Geomer an Funamenal Mar ebrew Unvers
More informationNormal Random Variable and its discriminant functions
Noral Rando Varable and s dscrnan funcons Oulne Noral Rando Varable Properes Dscrnan funcons Why Noral Rando Varables? Analycally racable Works well when observaon coes for a corruped snle prooype 3 The
More information( ) () we define the interaction representation by the unitary transformation () = ()
Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger
More informationON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS
ON THE WEA LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS FENGBO HANG Absrac. We denfy all he weak sequenal lms of smooh maps n W (M N). In parcular, hs mples a necessary su cen opologcal
More informationDepartment of Economics University of Toronto
Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of
More informationMAXIMIN POWER DESIGNS IN TESTING LACK OF FIT Douglas P. Wiens 1
MAXIMIN POWER DEIGN IN TETING LACK OF FIT Douglas P. Wens Absrac We nd desgns wc maxmze e mnmum power, over a broad class of alernaves, of e es for Lack of F, n dscree desgn spaces. Ts complemens prevous
More informationOn One Analytic Method of. Constructing Program Controls
Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna
More informationOnline Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading
Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng
More informationWiH Wei He
Sysem Idenfcaon of onlnear Sae-Space Space Baery odels WH We He wehe@calce.umd.edu Advsor: Dr. Chaochao Chen Deparmen of echancal Engneerng Unversy of aryland, College Par 1 Unversy of aryland Bacground
More informationOutline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model
Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon
More informationTwo-Handed Gesture Tracking Incorporating Template Warping With Static Segmentation
Two-Handed Gesure Trackng Incorporang Templae Warpng Wh Sac Segmenaon Yu Huang *, Thomas S. Huang *, Henrch Nemann + * Beckman Insue, U. o Illnos a Urbana-Champagn, Urbana, IL61801, US + Dep. Compuer Scence,
More informationBayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance
INF 43 3.. Repeon Anne Solberg (anne@f.uo.no Bayes rule for a classfcaon problem Suppose we have J, =,...J classes. s he class label for a pxel, and x s he observed feaure vecor. We can use Bayes rule
More informationCHAPTER 10: LINEAR DISCRIMINATION
CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g
More informationChapters 2 Kinematics. Position, Distance, Displacement
Chapers Knemacs Poson, Dsance, Dsplacemen Mechancs: Knemacs and Dynamcs. Knemacs deals wh moon, bu s no concerned wh he cause o moon. Dynamcs deals wh he relaonshp beween orce and moon. The word dsplacemen
More informationDynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005
Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc
More informationA Principled Approach to MILP Modeling
A Prncpled Approach o MILP Modelng John Hooer Carnege Mellon Unvers Augus 008 Slde Proposal MILP modelng s an ar, bu need no be unprncpled. Slde Proposal MILP modelng s an ar, bu need no be unprncpled.
More informationSingle-loop System Reliability-Based Design & Topology Optimization (SRBDO/SRBTO): A Matrix-based System Reliability (MSR) Method
10 h US Naonal Congress on Compuaonal Mechancs Columbus, Oho 16-19, 2009 Sngle-loop Sysem Relably-Based Desgn & Topology Opmzaon (SRBDO/SRBTO): A Marx-based Sysem Relably (MSR) Mehod Tam Nguyen, Junho
More informationSPE Copyright 2007, Society of Petroleum Engineers
SPE 463 Smulaneous Esmaon of Aqufer Parameers and Ornal Hydrocarbons n Place From Producon aa Usn umercal Inverson of Laplace Transform oaman El-Khab, SPE, Sudan U. for Scence & Technoloy Copyrh 7, Socey
More informationOn Construction of Odd-fractional Factorial Designs
J. Sa. Appl. Pro., o., -7 SP Journal of Sascs Applcaons & Proaly --- An Inernaonal Journal @ SP aural Scences Pulsn Cor. On Consrucon of Odd-fraconal Facoral Desns Ike Basl Onukou Deparmen of Sascs, Unversy
More informationV.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS
R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon
More informationOptimization. Nuno Vasconcelos ECE Department, UCSD
Optmzaton Nuno Vasconcelos ECE Department, UCSD Optmzaton many engneerng problems bol on to optmzaton goal: n mamum or mnmum o a uncton Denton: gven unctons, g,,...,k an h,,...m ene on some oman Ω R n
More informationStructure Restricted (PID) Controller Performance Assessment for Multi-stage Batch Processes with Tracking and Regulatory Requirements
Proceengs o he 4h Inernaonal Symposum on Avance Conrol o Inusral Processes housan Islans Lake Hangzhou P Chna May 3-6 ub3 Srucure esrce (PID) Conroller Perormance Assessmenor Mul-sage Bach Processes wh
More informationReview of Numerical Schemes for Two Point Second Order Non-Linear Boundary Value Problems
Proceedngs of e Pasan Academ of Scences 5 (: 5-58 (5 Coprg Pasan Academ of Scences ISS: 377-969 (prn, 36-448 (onlne Pasan Academ of Scences Researc Arcle Revew of umercal Scemes for Two Pon Second Order
More informationCH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC
CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal
More informationRobust and Accurate Cancer Classification with Gene Expression Profiling
Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem
More informationTheoretical Analysis of Biogeography Based Optimization Aijun ZHU1,2,3 a, Cong HU1,3, Chuanpei XU1,3, Zhi Li1,3
6h Inernaonal Conference on Machnery, Maerals, Envronmen, Boechnology and Compuer (MMEBC 6) Theorecal Analyss of Bogeography Based Opmzaon Aun ZU,,3 a, Cong U,3, Chuanpe XU,3, Zh L,3 School of Elecronc
More informationCS537. Numerical Analysis
CS57 Numerical Analsis Lecure Numerical Soluion o Ordinar Dierenial Equaions Proessor Jun Zang Deparmen o Compuer Science Universi o enuck Leingon, Y 4006 0046 April 5, 00 Wa is ODE An Ordinar Dierenial
More informationApproximating the Powers with Large Exponents and Bases Close to Unit, and the Associated Sequence of Nested Limits
In. J. Conemp. Ma. Sciences Vol. 6 211 no. 43 2135-2145 Approximaing e Powers wi Large Exponens and Bases Close o Uni and e Associaed Sequence of Nesed Limis Vio Lampre Universiy of Ljubljana Slovenia
More informationFX-IR Hybrids Modeling
FX-IR Hybr Moeln Yauum Oajma Mubh UFJ Secure Dervave Reearch Dep. Reearch & Developmen Dvon Senor Manaer oajma-yauum@c.mu.jp Oaka Unvery Workhop December 5 h preenaon repreen he vew o he auhor an oe no
More informationTrack Properities of Normal Chain
In. J. Conemp. Mah. Scences, Vol. 8, 213, no. 4, 163-171 HIKARI Ld, www.m-har.com rac Propes of Normal Chan L Chen School of Mahemacs and Sascs, Zhengzhou Normal Unversy Zhengzhou Cy, Hennan Provnce, 4544,
More informationImprovement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling
Rajesh ngh Deparmen of ascs, Banaras Hndu Unvers(U.P.), Inda Pankaj Chauhan, Nrmala awan chool of ascs, DAVV, Indore (M.P.), Inda Florenn marandache Deparmen of Mahemacs, Unvers of New Meco, Gallup, UA
More informationMachine Learning Linear Regression
Machne Learnng Lnear Regresson Lesson 3 Lnear Regresson Bascs of Regresson Leas Squares esmaon Polynomal Regresson Bass funcons Regresson model Regularzed Regresson Sascal Regresson Mamum Lkelhood (ML)
More informationCS 4495 Computer Vision Tracking 1- Kalman,Gaussian
CS 4495 Compuer Vision A. Bobick CS 4495 Compuer Vision - KalmanGaussian Aaron Bobick School of Ineracive Compuing CS 4495 Compuer Vision A. Bobick Adminisrivia S5 will be ou his Thurs Due Sun Nov h :55pm
More informationComb Filters. Comb Filters
The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of
More informationComprehensive Integrated Simulation and Optimization of LPP for EUV Lithography Devices
Comprehense Inegraed Smulaon and Opmaon of LPP for EUV Lhograph Deces A. Hassanen V. Su V. Moroo T. Su B. Rce (Inel) Fourh Inernaonal EUVL Smposum San Dego CA Noember 7-9 2005 Argonne Naonal Laboraor Offce
More informationWebAssign HW Due 11:59PM Tuesday Clicker Information
WebAssgn HW Due 11:59PM Tuesday Clcker Inormaon Remnder: 90% aemp, 10% correc answer Clcker answers wll be a end o class sldes (onlne). Some days we wll do a lo o quesons, and ew ohers Each day o clcker
More informationPhysics Notes - Ch. 2 Motion in One Dimension
Physics Noes - Ch. Moion in One Dimension I. The naure o physical quaniies: scalars and ecors A. Scalar quaniy ha describes only magniude (how much), NOT including direcion; e. mass, emperaure, ime, olume,
More informationDensity estimation III. Linear regression.
Lecure 6 Mlos Hauskrec mlos@cs.p.eu 539 Seo Square Des esmao III. Lear regresso. Daa: Des esmao D { D D.. D} D a vecor of arbue values Obecve: r o esmae e uerlg rue probabl srbuo over varables X px usg
More informationIntroduction. Voice Coil Motors. Introduction - Voice Coil Velocimeter Electromechanical Systems. F = Bli
UNIVERSITY O TECHNOLOGY, SYDNEY ACULTY O ENGINEERING 4853 Elecroechncl Syses Voce Col Moors Topcs o cover:.. Mnec Crcus 3. EM n Voce Col 4. orce n Torque 5. Mhecl Moel 6. Perornce Voce cols re wely use
More informationSupporting Information: The integrated Global Temperature change Potential (igtp) and relationships between emission metrics
2 3 4 5 6 7 8 9 Supporng Informaon: Te negraed Global Temperaure cange Poenal (GTP) and relaonsps beween emsson mercs Glen P. Peers *, Borgar Aamaas, Tere Bernsen,2, Jan S. Fuglesved Cener for Inernaonal
More informationDelay Dependent Robust Stability of T-S Fuzzy. Systems with Additive Time Varying Delays
Appled Maemacal Scences, Vol. 6,, no., - Delay Dependen Robus Sably of -S Fuzzy Sysems w Addve me Varyng Delays Idrss Sad LESSI. Deparmen of Pyscs, Faculy of Scences B.P. 796 Fès-Alas Sad_drss9@yaoo.fr
More informationLinear Response Theory: The connection between QFT and experiments
Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure
More informationSHOALING OF NONLINEAR INTERNAL WAVES ON A UNIFORMLY SLOPING BEACH
SHOALING OF NONLINEAR INTERNAL WAVES ON A UNIFORMLY SLOPING BEACH Ke Yamasa Taro Kaknuma and Kesuke Nakayama 3 Te nernal waves n e wo-layer sysems ave been numercally smulaed by solvng e se o nonlnear
More informationImprovement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling
Improvemen n Esmang Populaon Mean usng Two Auxlar Varables n Two-Phase amplng Rajesh ngh Deparmen of ascs, Banaras Hndu Unvers(U.P.), Inda (rsnghsa@ahoo.com) Pankaj Chauhan and Nrmala awan chool of ascs,
More informationHEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD
Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,
More informationHomework 2 Solutions
Mah 308 Differenial Equaions Fall 2002 & 2. See he las page. Hoework 2 Soluions 3a). Newon s secon law of oion says ha a = F, an we know a =, so we have = F. One par of he force is graviy, g. However,
More informationMechanics Physics 151
Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu
More informationSemi-Lagrangian Method for Advection Equation on GPU in Unstructured R Mesh for Fluid Dynamics Application
Worl Acaemy o Scence, Engneerng an Technology Inernaonal Journal o Physcal an Mahemacal Scences Sem-Lagrangan Meho or Avecon Equaon on GPU n Unsrucure R Mesh or Flu Dynamcs Applcaon Irakl V. Gugushvl,
More informationTesting a new idea to solve the P = NP problem with mathematical induction
Tesng a new dea o solve he P = NP problem wh mahemacal nducon Bacground P and NP are wo classes (ses) of languages n Compuer Scence An open problem s wheher P = NP Ths paper ess a new dea o compare he
More information