High-level Hierarchical Semantic Processing Framework for Smart Sensor Networks
|
|
- Juniper Floyd
- 6 years ago
- Views:
Transcription
1 HSI 2008 Kakow, Poland, May 25-27, 2008 Hgh-level Heachcal Semanc Pocessng Famewok fo Sma Senso Newoks Dema Buckne, Membe, IEEE, Jamal Kasb Rosemae Velk, Membe, IEEE, and Wolfgang Hezne, Membe, IEEE Venna Unvesy of Technology, Venna, Ausa {buckne, Ausan Reseach Cenes, Venna, Ausa {jamal.kasb Absac Ths pape pesens he famewok of a novel appoach o combne mul-modal senso nfomaon fom audo and vdeo modales o gan valuable supplemenay nfomaon compaed o adonal vdeo-based obsevaon sysems o even jus CCTV sysems. A heachcal, mulmodal senso pocessng achecue fo obsevaon and suvellance sysems s poposed. I ecognzes a se of pedefned behavo and leans abou usual behavo. Devaons fom nomaly ae epoed n a way undesandable even fo saff whou specal anng. The pocessng achecue ncludng he physcal senso nodes s called SENSE (sma embedded newok of sensng enes [1, 4]. Keywods senso newoks, senso fuson, semanc symbols, daa mnng, heachcal model I I. INTRODUCTION N cuen mes, obsevaon sysems fo publc spaces become moe wdespead. They ae a vsble eacon fo he publc on heas lke eosm and cme. Ths pape descbes he conceps of he semanc pocessng layes n a newok of SENSE nodes [1], [4]. The goal of hese layes s o lean he "nomaly" n he envonmen of a SENSE newok, n ode o deec unusual behavo, suaons, o evens and o nfom he cusome n such cases [5]. SENSE consss of a newok of communcang senso nodes, each equpped wh a camea and a mcophone aay. Those senso modales obseve he envonmen and delve a seam of socalled low-level symbols (LLS, e.g. movng objecs, o sounds. In he easonng un he LLSs ae pocessed n ode o nfom he peson n chage n case of above menoned evens occu. The fs applcaon aea of SENSE wll be an apo, heefoe all alams and ohe consdeaons ae akng no accoun he needs of he apo saff. The goal wll be acheved n seveal seps. Fs, he nfomaon eceved fom he senso laye,.e. he unmodal (audo and vdeo symbols, s pe-pocessed, whch ncludes deleon of spuous objecs and yng o ack symbols by coelang he eceved senso messages (snapshos a dsnc pons n me ove me. Second, he pe-pocessed un-modal symbols ae fused, whch esuls n negaed hgh-level symbols (HLS chaacezed by he combned un-modal popees. These symbols ae npu o he semanc symbol leanng pocess, whch deves he models fo ypcal behavo and popees of he dffeen objecs caegoes pesons, luggage, ec. These models descbe nomal behavo wh dsbuons of popees lke speed and decon wh espec of locaon. As he pahs whch symbols ake hough he (vsual sensng aea of a SENSE node ae an mpoan aspec of behavo, ajecoes ae deved fom he semanc symbol models. All of he mehods used n he pacula layes ae wdely used n e.g. obsevaon sysems and many ohe applcaons, bu o ou knowledge no ohe sysem uses a combnaon heeof n ode o le he messages of he sysem eally look sma and meanngful o he use. Ths pape s sucued as follows: he nex chape oulnes he sysem achecue, whle chape 3 descbes he ndvdual layes n moe deal. Chape 4 dscusses ackng of low-level symbols, and chape 5 conans concluson and oulook. II. ARCHITECTURE OVERVIEW An 8-e daa pocessng achecue s used, n whch he lowe levels wll be esponsble fo a sable and compehensve wold epesenaon o be evaluaed n he hghe levels (Fg. 1. A fs, he modaly symbols wll be checked egadng he plausbly he audo symbols wh espec o poson, he vdeo symbols wh espec o poson, and sze. In he second pocessng laye (n hs pape, we wll use e and laye as synonyms, symbol ackng wll ake place. Hee, symbols whch pass he fs (spaal check wll be checked egadng he empoal behavo usng a Makov Chan Mone Calo based pacle fleng appoach. The oupu of hs e s a sable and compehensve wold epesenaon ncludng un-modal symbols. Te 3 s he senso fuson e, n whch he un-modal symbols ae fused o fom mul-modal symbols. Laye 4 s he paamee nfeence machne, n whch pobablsc model(s fo symbols paamees and evens ae opmzed. The esuls of hs e ae models of hghlevel symbols and feaues ha descbe behavo. In e 5, he sysem leans abou ajecoes of symbols. Typcal /08/$ IEEE
2 pahs ough he vew of he senso node ae soed. The 6 h laye has he ask of managng he communcaon o ohe nodes and esablshng a global wold vew. The ajecoes ae also used n hs laye o fnd ou f a ajecoy of one node can be polonged ove a neghbong node. In e 7, he ecognon of unusual behavo and evens occus wh wo appoaches. One pa compaes cuen symbols wh he leaned models and ajecoes. Theefoe, hs sub-laye calculaes pobables fo he exsence of symbols wh espec o he poson, velocy, decon, and pobables fo ajecoes of symbols. I also calculaes pobables fo he duaon of say of symbols n aeas, pobables fo he movemen along ajecoes, also acoss nodes (global map, scenao ecognon. Symbols wh such pobables below defned hesholds ase "unusual behavo" alams. The second pa of e 7 s concened wh he ecognon of pe-defned scenaos and he ceaon of alams n case pe-defned condons ae me. Fnally, laye 8 s esponsble fo communcaon o he use. I geneaes alam o saus messages and fles hem f pacula condons would be announced oo ofen o he same even s ecognzed by boh mehods n laye 7. Neghbou node Hgh-level layes semanc pocessng Low-level laye Fg. 1. Semanc Pocessng Laye Sofwae Achecue The vsual feaue exacon (laye 0 pocesses fame by fame fom he camea n 2D camea coodnaes. Fs esuls fom he es vdeos show ha vsual deecon can delve sgnfcanly changng daa fom one fame o he nex. In case of unfounae condons fo he camea (many pesons, hey exchange posons n he cowd, bad lgh condons, ec, deeced symbols can change he label fom peson o objec o peson-goup and back (fo he same physcal objec. The sze of deeced symbols can change fom small elemens lke bags o lage goups of pesons coveng ens of squae mees and ncludng pevously deeced sngle pesons and ohe objecs. Consequenly, he hghe levels have o be pepaed o wok wh mpefecly deeced symbols. III. DESCRIPTION OF TIERS A. Low-level feaue exacon Descpon Ths laye s he pocessng un povdng he semanc pocessng laye wh un-modal daa seams descbng wha s obseved by he sensos of a SENSE node a me (.e. n he envonmen whee he node s embedded. The audo and vdeo low-level symbols (LLS epesen defned pmves (Vdeo: Peson, Peson Flow, Luggage, ec. Audo: Seps, Gun shoong, ec.. Funconaly The pacula funconaly of hs laye s no whn he scope of hs pape. We jus wan o pon ou ha he audo and vdeo daa mus be povded synchonzed,.e. me-samped wh efeence o he same clock. Inefaces Conssng of wo componens, one fo he audo and he ohe fo he vdeo modaly, hs laye povdes wo nefaces: Though he Audo Ineface, he audo symbols ae sen o laye 1. An audo LLS consss of a label, descbng he ype of he deeced audo symbols, he decon of aval, he loudness, ec. Though he Vdeo Ineface, he vdeo symbols ae popagaed. A vdeo LLS consss of a label, descbng he ype of he deeced vdeo symbols, he poson n pxel, he sze, he velocy, ec. B. Pe-pocessng ncludng plausbly checks Descpon Dung vsual feaue exacon, a se of emplaes s mached wh he cuen fame. The emplaes ae scaled n ode o fnd objecs of vaous szes. In ode o fle unealsc pmves fom he daa seam, a fs we nend o lean abou aveage sze of pmve symbols dependng on he ype and poson n camea coodnaes. The second plausbly check s done on boundng boxes. The boundng boxes of symbols ae aken o deemne whehe some peson o objec s blend no a lage objec. In hs case he coun of smalle and lage symbols decdes whch knd of symbol s mos pobable and wll be used fo fuhe pocessng. Smla o he sze of symbols, also he aveage speed wll be leaned by he senso. Ths nfomaon wll be used fo symbol ackng, oo. We assume he exsence of pons n mes, whee no pesons ae n he sensve aea of he
3 node. All nenal scenaos and symbols wll be ese a hese momens. When new LLSs appea, hey wll be acked ove me wh espec o he poson, sze, and speed. Funconaly Fo he aveage sze of symbols, a Gaussan o mxue of Gaussans model wll be ulzed. One aveage sze model wll be used fo pxel cluses, so ha he 640x480 camea pxels anslae o 40x30 pxel cluses, each 16x16 pxels lage. Each pxel cluse has models fo each ype of symbol. The models may need dffeen paamees dependen on he numbe of pesons,. e. may un ou ha people behave dffeenly n goups han hey do alone o n pas. Due o he fac ha symbols can change he ype fom fame o fame would no be a good soluon o delee all mpobable symbols, heefoe hey ae jus maked. Nex, we wll use fuzzy logc o fnd ou symbols ha do no change much fom fame o fame and so hem ou as beng able o ack. Thd, all ohe symbols ha may appea, dsappea, change he sze, ec.: o fnd ou f hey ovelap n he fame and ae no ecognzed as befoe because of he ovelap we ake he boundng boxes and es f a smalle objec blended no a lage o f a lage objec spl no smalle ones. If so, hese symbols may be coec deecons, bu we canno assgn speed, and we do no know f he deecon as small o bg symbol s domnan ove me. So, wh he me, he mely symbol coun fo lage o small symbol wll deemne how hs symbol s handed ove o he nex laye. Fnally, we wll compue he speed of symbols. In hs pocess, s necessay o evaluae moe han he mmedae pas fame. Rescons on he compuaonal powe wll show he possbles n hs espec. Afe all plausbly checks, a voe wll decde f a symbol s handed ove o he nex laye. C. Tackng Descpon Ths laye uses pacle fle echnques o ack he pepocessed symbols. The basc assumpons fo he algohm ae pesened n deal n he nex chape. Tadonally, mulple objecs (n he aea of pacle fles, objecs ae acked, no symbols, heefoe hs em s used hee ae acked by mulple sngle-objecackng fles. Whle usng ndependen fles s compuaonally acable, he esul s pone o fequen falues. Each pacle fle samples n a small space, and he esulng jon fle s complexy s lnea n he numbe of ages n. Howeve, n cases whee ages do neac, as n many of ou scenaos, sngle pacle fles ae suscepble o falues exacly when neacons occu. In a ypcal falue mode, seveal ackes wll sa ackng he sngle age wh he hghes lkelhood scoe. Funconaly A pacle fle specfcally desgned fo ackng neacng objecs [2] s used o ack he pe-pocessed symbols. The appoach fo addessng acke falues esulng fom neacons s o noduce a moon model based on Makov andom felds (MRFs [11]. D. Senso fuson Descpon Ths e ges as npu he sable un-modal symbols. Is ask s o fuse audo and vdeo symbols. One possbly s o use faco analyss [10, 12] o deemne he coelaon beween audo and vdeo LLSs. The oupu of hs e wll be a symbolc epesenaon of he eal wold n fom of a collecon of mul-modal symbols [13]. Funconaly Fuson of he audo and he vdeo daa s a ask ha can be done usng he coelaon beween he povded daa seams. Based on he me coelaon of LLS, feaues ha can be aken no accoun fo hs pupose ae: loudness, decon of avals, powe specum, sze of he vdeo symbol n pxels, and poson. E. Paamee Infeence Descpon Ths laye wll pocess he ncomng symbols of fused LLSs and adap he paamees of he used pobablsc model(s o f he daa. The daa ae defned as he se of all he nsances of semanc symbols. Funconaly Geng any symbol fom he senso fuson laye, he ask of hs laye s o nfe he paamees of he undelyng pobablsc model (Mxue of Gaussans, hsogams, o mxue of faco analyzes [3]. Usng an onlne veson of he Expecaon Maxmzaon (EM algohm o fnd he paamees of he model(s, we can focus on he vaan whee he sysem leans he behavo only of he ecen pas (me wndow, by usng onlne aveage movng. We also can assume ha he daa fs a sac model and heefoe we can use he "gaden descen vaan". The use of non-paamec mehods lke hsogams (n pacula, k neaes neghbos can be aken no accoun. Geneally, we ae usng onlne cluseng mehods lke descbed n [8], [9], [10], [11], [12]. F. Tajecoes Descpon Tajecoes n a node wll be deved hough he use of a leaned anson max conssng of ansons beween model cluses. Ths could be done by usng a local seach fo he mos lkely sequence. Each HLS (model mus heefoe keep a ls of all local ajecoes o whch he symbol s belongng and a he me, n whch an nsance s beng obseved and belongng o ha symbol, he suable ajecoy (whch should be acve fo he nsance mus be seleced. The node mus acvae all he local ajecoes whch ae possble a me fo he obseved nsance. Addonally, all he neghbong nodes, o whch he node has some coelaons, mus also acvae he suable ajecoes. Due o he fac ha one HLS could belong o moe han one ajecoy (afe leanng, he possble ajecoy wll no be necessaly unque. Bu addng he eal me nfomaon,.e. he nsance a me, whch s belongng o jus one HLS, and whch s n un belongng o one o moe ajecoes, he
4 local and global ajecoy s unambguously denfed. Funconaly The dynamcs of he deeced objecs n he envonmen of he node ae descbed by buldng he local ajecoes map of a node. The map buldng pocess makes use of he abues of he HLS; especally, uses he velocy abues mean and vaance n he coesponden poson. Each of he nvolved nodes has o lean he global possble ajecoes (ncludng he pobably of havng he ajecoy acve, he densy - hgh, mddle, low - he decon, and he velocy. Then each node keeps a lookup able o a max whee all he possble ajecoes ae egseed. A smple swch fom one local o global pah o anohe one mus cause an alam (because s unusual ha people ake hs pah. Ths max of all he possble ajecoes wll be bul usng he local anson maces and he wegh maces (coelaons beween he HLS n he dffeen nodes. G. Ine-Node Communcaon Descpon Ine-node communcaon s based on he Loopy-Belef Popagaon (LBP algohm [6, 7]. LBP wll be used o fom collecons of neghbong nodes and o map he posons whn one node s vew no anohe one s vew. Ths nfomaon can fnally be used o e.g. soe he ajecoy of pesons ove mulple nodes, o o fnd ou f somebody es o escape obsevaon. Funconaly Ths module wll be a he hea of he nfomaon exchange whn he SENSE sysem. Ths laye sends and eceves messages fom and o ohe nodes n he SENSE sysem. Those nodes mus be eachable by he nodes (.e. "neghbos" wh sensng aeas whch ovelap wh ha of he especve node. The nfomaon of neghbohood s soed n a max and s heefoe a dynamc paamee ha can change ove me. Anohe ask of hs laye wll be o updae he compables and heefoe he weghs beween HLSs. Inefaces Laye 6 (Ine-Node communcaon Laye 7 (Alam Geneao: We assume ha he sub laye nesenso communcaon povdes he sub laye alam geneao wh he global coelaons ha concens he map (ajecoes, he hgh-level symbols, he global ackng of pesons, and he coesponden abues. Ths affecs he node-o-node neface. Laye 6 (Ine-Node communcaon Neghbos: Fo leanng he global sucue hough he hgh-level symbols, only wo vaables wll be exchanged: he belef of he node gven he evdence a me and also he belef of he node whou he evdence a me. (The evdence s he senso eadng. The exchange of he HLS ncludng s paamees (abues, deeced nsances a me, local ajecoes, and global ajecoes should also be done. I wll be nvesgaed whehe ohe nfomaon can also be exchanged. H. Alam Geneao Descpon Ths e seves o deec pedefned alams and fuhe unusual behavo. I geneaes alams n case of vey unlkely symbols (as a esul of he above menoned ackng and pobably esmang asks. A second paly ndependen pa wll be he deecon of pedefned scenaos: The nfomaon of pesons movng, meeng, havng luggage, ec. wll be compaed wh emplae scenaos. If possble, hese scenaos (e.g. one peson comng wh luggage, doppng luggage, leavng wll be used o mono he behavo of pesons and fo alamng. The majo dffeence beween pedefned alams and ecognzed unusual behavo s ha he fs can be assocaed wh a pedefned, human-eadable ex, lke "sceam nose" o "peson dopped luggage". In conas, wh unusual behavo only he nvolved symbols(s can be denfed (n he use neface, whou fuhe explanaoy ex assocaed (besdes poenal namng of he feaues deeced as beng ou of nomal. Funconaly 1: (Pedefned scenao ecognon Ths mehod akes he symbolc epesenaon on he level of he modaly elaed symbols as npu. I uses a ule-base o combne hese symbols n a heachcal way o ceae symbols wh hghe (moe absac semanc meanng ou of symbols wh lowe semanc level. Pedefned alams ae no flexble and hus canno adap o new suaons. Whaeve s ecognzed has o be bul no he sysem befoe becomes opeave as opposed o ecognzng unusual behavou, whch leans dung opeaon. The advanage of ecognzng pedefned alams s he ably o povde a human use wh a semanc descpon of he ype of alam ha he sysem has deeced (e.g. Unaended Luggage nsead of Unusual Behavo, also fo complex scenaos. The pedefned alam scenaos ha he sysem a he apo shall deec ae: Unaended luggage Loeng peson Ca n pakng aea has exceeded maxmum pakng me Sceamng peson Gunfe Beakng glass Whle he pedefned alams Sceamng peson, Gunfe and Beakng glass meely ely on nfomaon avalable fom low-level symbols of he audo modaly and wll be handed hough o he Use Nofcaon Fle, he alams Unaended luggage and Loeng peson eque a symbolc pocessng of dffeen nfomaon souces as descbed n [10]. Unaended luggage To deec unaended luggage, he sysem has o deec peson objecs and luggage objecs. When hese wo objecs can be elably deeced, he sysem can eason abou he assocaons beween peson and luggage symbols. The fs scenao s a luggage symbol, whch canno be assgned o a peson elably. If he sysem fals
5 o assgn he luggage fo a cean peod, assumes he luggage o be unaended. Ths scenao s a successful connecon beween a peson and a pece of luggage. In he second scenao, he sysem has o ack boh peson and luggage and deec f he peson has moved away fom he luggage fo a longe peod. I s expeced ha he fs scenao yelds moe false alams, snce eles only on he successful vsual deecon of luggage. The second scenao, howeve, s expeced o be moe elable and fuhemoe possbly allows denfcaon of he peson who has lef he pece of luggage. Loeng peson Peson symbols ha can be successfully acked fo a longe peod can ase an alam fo loeng f hey eman n he same locaon fo an exended peod of me. The challenge hee s o pove ha loeng can be deeced, even f he peson shfs s poson beween he vewpo of dffeen cameas. Thus he sysem shall be able o deec a peson ha has been movng aound he aea fo e.g. a whole day. Whou ne-node communcaon, he sysem can be fooled by changng he poson beween dffeen camea posons. By handng ove denfed pesons beween nodes, hs should no be possble. Funconaly 2: Unusual behavo ecognon Pobably of exsence: Ths mehod calculaes he pobably of a symbol wh espec o s poson whn he senso s vew, s decon of movemen, and s speed. The esul wll be Pobably Mounans ove he menoned paamees. Pobably of duaon: In addon o he pobably of he smple exsence of symbols, hs ask measues he duaon of exsence of symbols whn some ange. The ange can be composed of pas of he senso s vew up o he vew acoss seveal sensos. Pobably of movemen: The pobably of he movemen of a HLS whn he node s vew (e.g. s usual ha a peson comng fom one ajecoy moves o anohe one? and acoss nodes: Gven an nsance ha s obseved a me, he pobably ha hs nsance wll follow he same pah (o akes he same ajecoy as ohe nsances should be calculaed. Also a swch fom one ajecoy o anohe should be handled as an alam. I. Use Nofcaon Fle Descpon Ths e bulds he neface of he hgh-level senso pocessng. I delves alams o he use neface and can be asked abou he saus of a node o seveal nodes. I fles dencal alams, e.g. when he same lukng peson s epoed seveal mes fom he pedefned scenao ecognon, o a epoed unusual behavo can be mached wh a pedefned alam, only he lae wll be delveed. Addonally, he use can apply fleng ules o om o polong alams va he GUI. Funconaly A basc fleng mechansm fo avodng sendng he same alam seveal mes o sendng a pe-defned alam and an unusual behavo fo he same hng s appled. An addonal ule-base wh use pefeences s also consdeed. Noe: alhough by means of LBP a global vew wll be esablshed among he nodes and despe he fleng, canno be excluded ha he GUI wll eceve dencal alams fom dffeen nodes. IV. DETAILED DESCRIPTION OF TRACKING OF LLS We ae concened wh he poblem of ackng mulple neacng ages. Ou objecve s o oban a ecod of he ajecoes of ages ove me and o manan coec, unque denfcaon of each age. Tackng mulple dencal ages becomes challengng when he ages pass close o one anohe o mege as pesons do n a cowd. In ecen mes, an appoach ha eles on he use of a moon model ha s able o adequaely descbe age behavo houghou an neacon even was developed [2]. Ths appoach has a moon model ha eflecs he addonal complexy of he age behavo. The eason fo usng hs appoach les n he fac ha he numbe of symbols n he obsevaon model can change fom senso obsevaon o senso obsevaon. E.g. f seveal pesons ae gong hough a codo, he vsual feaue exacon algohms mgh deec a sasfacoy numbe of pesons n one mage and jus a goup of pesons n he consecuve one. In case of unlucky condons, he deecon can change ofen whn sho peods of me fo he same physcal objec. The mulple age ackng poblem can be expessed as a Bayesan fle. We ecusvely updae he poseo dsbuon P ( Z ove he jon sae of all n ages { 1.. n} gven all obsevaons Z = Z1.. Z up o and ncludng me, accodng o: P ( Z = kp ( Z 1 ( Z 1 1 Z 1 The lkelhood P expesses he measuemen model, he pobably we obseved he measuemen Z gven he sae a me, whch s a model fo he modaly-elaed feaue exacon algohms. The moon model P ( 1 pedcs he sae a me gven he pevous sae 1. In all ha follows we wll assume ha he lkelhood P Z facos acoss ages as Z = n = 1 ( Z and ha he appeaances of ages ae condonally ndependen, whch may no be compleely ue n case of people, who belong ogehe, bu wll hold mos of he me beween all pesons n he cowd. If we assume he ages as beng ndependen, o nonneacng, hey can be acked wh sngle-age pacle
6 fles. In ohe wods, he moon model s facoed n a poduc of moon models fo ndvdual ages P ( =, 1 1. The ask s o appoxmae he poseo P ( Z ove each age s sae. In ohe wods, he pobably ha he cuen obsevaons ae made, because he obseved objecs behave n a pacula way. One vew on pacle fles s o see hem as mpoance fle fo hs poseo. Theefoe, we assume he poseo of he pevous me beng appoxmaed by a se of weghed pacles, 1 ( ( { } N Z 1, π 1. = 1 Then, fo he cuen me sep, we daw N samples (s fom a poposal dsbuon ~ q( ( = π ( s 1 ( 1 ( whch s a mxue of moon models, 1. Fnally, we wegh each sample by s lkelhood. The ( s ( s ( s N esulng se {, π = Z } s= 1 s a weghed appoxmaon fo he poseo ove he age s sae a me. The MRF-based appoach fo he moon model uses pa wse MRFs, whee he ψ (, j ae pa wse neacon poenals (along edges and E he space of edges beween objecs: ( 1 ( 1 ψ (, j P Each wo objecs shae a pacula poenal. Ths em can be ncopoaed no he Bayesan fle easly, bu now we appoxmae he jon sae of all ages, whch would esul n he necessy of dawng an ncedble hgh numbe of pacles o be able o fnd a good appoxmaon. Applyng he Mone Calo appoxmaon o he ognal Bayesan poseo, we ge ( ( π, P ( Z k Z 1 1 and ncopoang he MRF moon model, we oban ( ( P ( Z k Z ψ (, π j 1 1 Founaely, we see ha he neacon poenal s ndependen on eale saes, so can be eaed as addonal faco. Unfounaely, appoxmang hs em appoxmaes he jon poson of all ages, whch s no ou scope. Theefoe, we apply MCMC (Makov chan Mone Calo samplng, so ha he saonay dsbuon of he chan s exacly he age dsbuon, and we change only he sae of one age a a me by samplng decly fom he moon model of ha age / 1 / 1 /' ( / Q ( = Q(, = 1 δ ( j = N N j The accepance ao of hs samplng mehod s only / / / Z ψ (, j as = mn 1,. Z ψ (, j Ths mehod mnmzes he compuaonal effo n compason o jon pacle fles fo ackng of mulple objecs and also mnmzes he faul deecon ae compaed o a se of sngle-objec-ackng pacle fles. V. CONCLUSION AND OUTLOOK Ths pape pesens a heachcal pocessng achecue fo sma senso newoks. The nnovave aspec les n he sep-by-sep pocessng, hough whch fom he low-level symbols hee emeges nfomaon beng moe and moe meanngful o a human peson n chage. Expeced esuls ae descbed fo he deploymen n an apo envonmen. All layes of he heachcal pocessng famewok ae descbed n ode o undesand he dea behnd and he algohm fo ackng of mulple objecs s descbed n moe deal. Ieave ess and nevews wh he apo saff ae planned o gan measues fo evaluang he esuls. REFERENCES [1] (SENSE pojec webse [2] Z. Khan, T. Balch, and F. Dellae: An MCMC-based Pacle Fle fo Tackng Mulple Ineacng Tages, IEEE Tansacons on Paen Analyss and Machne Inellgence, 2006 [3] Bshop, C. M.: Neual Newoks fo Paen Recognon. New Yok NY.: Oxfod Unvesy Pess Inc., 1995 [4] G. Zucke (ne Pal, and L. Fangu: Sma Nodes fo Semanc Analyss of Vsual and Aual Daa, Poceedngs of he IEEE INDIN, p , 2007 [5] B. Sallans, D. Buckne, and G. Russ: Sascal Deecon of Alam Condons n Buldng Auomaon Sysems. In: Poceedngs of 2006 IEEE INDIN, S. 6, 2006 [6] C. Cck and A. Pfeffe: Loopy belef popagaon as a bass fo communcaon n senso newoks, In Poceedngs of Unceany n Afcal Inellgence (UAI, 2003 [7] J.S. Yedda, W.T. Feeman, and Y. Wess: Undesandng Belef Popagaon and Is Genealzaons, IJCAI 2001 [8] D.-S. Lee: Onlne adapve gaussan mxue leanng fo vdeo. applcaons, n ECCV 2004 Wokshop on Sascal fo Vdeo Pocessng [9] N. Vlasss, and A. Lkas: A Geedy EM Algohm fo Gaussan Mxue Leanng, Neual Pocessng Lees, Volume 15, Numbe 1, 2002 [10] Z. Ghahaman and M. J. Beal: Vaaonal Infeence fo Bayesan Mxues of Faco Analyses, Advances n Neual Infomaon Pocessng Sysems. 2000, vol. 12, MIT Pess [11] N. Ueda, R. Nakano, Z. Ghahaman and G. E. Hnon: SMEM Algohm fo Mxue Models, Neual Compuaon achve Volume 12 [12] Z. Ghahaman and G. E. Hnon: The EM Algohm fo Mxue of Faco Analyzes, Techncal Repo CRG-TR-96-1, Depamen of Compue Scence, Unvesy of Toono, [13] P. Lombad: A sudy on daa fuson echnques fo vsual modules, Techncal. epo, Unvesy of Pava, 2002 [14] W. Bugsalle: Inepeaon of Suaons n Buldngs, Dsseaon hess, Venna Unvesy of Technology, 2007 [15] R. Kndemann, and J. L. Snell: Makov Random Felds and The Applcaons, AMS Books Onlne, ISBN: ,
CHAPTER 10: LINEAR DISCRIMINATION
HAPER : LINEAR DISRIMINAION Dscmnan-based lassfcaon 3 In classfcaon h K classes ( k ) We defned dsmnan funcon g () = K hen gven an es eample e chose (pedced) s class label as f g () as he mamum among g
More information5-1. We apply Newton s second law (specifically, Eq. 5-2). F = ma = ma sin 20.0 = 1.0 kg 2.00 m/s sin 20.0 = 0.684N. ( ) ( )
5-1. We apply Newon s second law (specfcally, Eq. 5-). (a) We fnd he componen of he foce s ( ) ( ) F = ma = ma cos 0.0 = 1.00kg.00m/s cos 0.0 = 1.88N. (b) The y componen of he foce s ( ) ( ) F = ma = ma
More informationName of the Student:
Engneeng Mahemacs 05 SUBJEC NAME : Pobably & Random Pocess SUBJEC CODE : MA645 MAERIAL NAME : Fomula Maeal MAERIAL CODE : JM08AM007 REGULAION : R03 UPDAED ON : Febuay 05 (Scan he above QR code fo he dec
More information1 Constant Real Rate C 1
Consan Real Rae. Real Rae of Inees Suppose you ae equally happy wh uns of he consumpon good oday o 5 uns of he consumpon good n peod s me. C 5 Tha means you ll be pepaed o gve up uns oday n eun fo 5 uns
More informationModern Energy Functional for Nuclei and Nuclear Matter. By: Alberto Hinojosa, Texas A&M University REU Cyclotron 2008 Mentor: Dr.
Moden Enegy Funconal fo Nucle and Nuclea Mae By: lbeo noosa Teas &M Unvesy REU Cycloon 008 Meno: D. Shalom Shlomo Oulne. Inoducon.. The many-body poblem and he aee-fock mehod. 3. Skyme neacon. 4. aee-fock
More information( ) ( )) ' j, k. These restrictions in turn imply a corresponding set of sample moment conditions:
esng he Random Walk Hypohess If changes n a sees P ae uncoelaed, hen he followng escons hold: va + va ( cov, 0 k 0 whee P P. k hese escons n un mply a coespondng se of sample momen condons: g µ + µ (,,
More informationChapter 3: Vectors and Two-Dimensional Motion
Chape 3: Vecos and Two-Dmensonal Moon Vecos: magnude and decon Negae o a eco: eese s decon Mulplng o ddng a eco b a scala Vecos n he same decon (eaed lke numbes) Geneal Veco Addon: Tangle mehod o addon
More informationOutline. GW approximation. Electrons in solids. The Green Function. Total energy---well solved Single particle excitation---under developing
Peenaon fo Theoecal Condened Mae Phyc n TU Beln Geen-Funcon and GW appoxmaon Xnzheng L Theoy Depamen FHI May.8h 2005 Elecon n old Oulne Toal enegy---well olved Sngle pacle excaon---unde developng The Geen
More informationLecture 5. Plane Wave Reflection and Transmission
Lecue 5 Plane Wave Reflecon and Tansmsson Incden wave: 1z E ( z) xˆ E (0) e 1 H ( z) yˆ E (0) e 1 Nomal Incdence (Revew) z 1 (,, ) E H S y (,, ) 1 1 1 Refleced wave: 1z E ( z) xˆ E E (0) e S H 1 1z H (
More informationHandling Fuzzy Constraints in Flow Shop Problem
Handlng Fuzzy Consans n Flow Shop Poblem Xueyan Song and Sanja Peovc School of Compue Scence & IT, Unvesy of Nongham, UK E-mal: {s sp}@cs.no.ac.uk Absac In hs pape, we pesen an appoach o deal wh fuzzy
More informationCHAPTER 3 DETECTION TECHNIQUES FOR MIMO SYSTEMS
4 CAPTER 3 DETECTION TECNIQUES FOR MIMO SYSTEMS 3. INTRODUCTION The man challenge n he paccal ealzaon of MIMO weless sysems les n he effcen mplemenaon of he deeco whch needs o sepaae he spaally mulplexed
More informationCptS 570 Machine Learning School of EECS Washington State University. CptS Machine Learning 1
ps 57 Machne Leann School of EES Washnon Sae Unves ps 57 - Machne Leann Assume nsances of classes ae lneal sepaable Esmae paamees of lnea dscmnan If ( - -) > hen + Else - ps 57 - Machne Leann lassfcaon
More informationA Methodology for Detecting the Change of Customer Behavior based on Association Rule Mining
A Mehodology fo Deecng he Change of Cusome Behavo based on Assocaon Rule Mnng Hee Seo Song, Soung He Km KAIST Gaduae School of Managemen Jae Kyeong Km KyungHee Unvesy Absac Undesandng and adapng o changes
More informationFIRMS IN THE TWO-PERIOD FRAMEWORK (CONTINUED)
FIRMS IN THE TWO-ERIO FRAMEWORK (CONTINUE) OCTOBER 26, 2 Model Sucue BASICS Tmelne of evens Sa of economc plannng hozon End of economc plannng hozon Noaon : capal used fo poducon n peod (decded upon n
More informationChapter Finite Difference Method for Ordinary Differential Equations
Chape 8.7 Fne Dffeence Mehod fo Odnay Dffeenal Eqaons Afe eadng hs chape, yo shold be able o. Undesand wha he fne dffeence mehod s and how o se o solve poblems. Wha s he fne dffeence mehod? The fne dffeence
More informationModeling Background from Compressed Video
Modelng acgound fom Compessed Vdeo Weqang Wang Daong Chen Wen Gao Je Yang School of Compue Scence Canege Mellon Unvesy Psbugh 53 USA Insue of Compung echnology &Gaduae School Chnese Academy of Scences
More informationSimulation of Non-normal Autocorrelated Variables
Jounal of Moden Appled Sascal Mehods Volume 5 Issue Acle 5 --005 Smulaon of Non-nomal Auocoelaed Vaables HT Holgesson Jönöpng Inenaonal Busness School Sweden homasholgesson@bshse Follow hs and addonal
More informations = rθ Chapter 10: Rotation 10.1: What is physics?
Chape : oaon Angula poson, velocy, acceleaon Consan angula acceleaon Angula and lnea quanes oaonal knec enegy oaonal nea Toque Newon s nd law o oaon Wok and oaonal knec enegy.: Wha s physcs? In pevous
More informationI-POLYA PROCESS AND APPLICATIONS Leda D. Minkova
The XIII Inenaonal Confeence Appled Sochasc Models and Daa Analyss (ASMDA-009) Jne 30-Jly 3, 009, Vlns, LITHUANIA ISBN 978-9955-8-463-5 L Sakalaskas, C Skadas and E K Zavadskas (Eds): ASMDA-009 Seleced
More informationA Monte Carlo Sequential Estimation of Point Process Optimum Filtering for Brain Machine Interfaces
A Mone Calo Sequenal Esmaon of Pon Pocess Opmum Fleng fo Ban Machne Inefaces Ywen Wang, Suden Membe, Anóno R. C. Pava, Suden Membe, José C. Píncpe, Fellow, Jusn C. Sanchez, Membe Absac The pevous decodng
More informationajanuary't I11 F or,'.
',f,". ; q - c. ^. L.+T,..LJ.\ ; - ~,.,.,.,,,E k }."...,'s Y l.+ : '. " = /.. :4.,Y., _.,,. "-.. - '// ' 7< s k," ;< - " fn 07 265.-.-,... - ma/ \/ e 3 p~~f v-acecu ean d a e.eng nee ng sn ~yoo y namcs
More informationWhen to Treat Prostate Cancer Patients Based on their PSA Dynamics
When o Tea Posae Cance Paens Based on he PSA Dynamcs CLARA day on opeaons eseach n cance eamen & opeaons managemen Novembe 7 00 Mael S. Lave PhD Man L. Pueman PhD Sco Tyldesley M.D. Wllam J. Mos M.D CIHR
More informationComputer Propagation Analysis Tools
Compue Popagaion Analysis Tools. Compue Popagaion Analysis Tools Inoducion By now you ae pobably geing he idea ha pedicing eceived signal sengh is a eally impoan as in he design of a wieless communicaion
More informationField due to a collection of N discrete point charges: r is in the direction from
Physcs 46 Fomula Shee Exam Coulomb s Law qq Felec = k ˆ (Fo example, f F s he elecc foce ha q exes on q, hen ˆ s a un veco n he decon fom q o q.) Elecc Feld elaed o he elecc foce by: Felec = qe (elecc
More informationFast Calibration for Robot Welding System with Laser Vision
Fas Calbaon fo Robo Weldng Ssem h Lase Vson Lu Su Mechancal & Eleccal Engneeng Nanchang Unves Nanchang, Chna Wang Guoong Mechancal Engneeng Souh Chna Unves of echnolog Guanghou, Chna Absac Camea calbaon
More informationRepresenting Knowledge. CS 188: Artificial Intelligence Fall Properties of BNs. Independence? Reachability (the Bayes Ball) Example
C 188: Aificial Inelligence Fall 2007 epesening Knowledge ecue 17: ayes Nes III 10/25/2007 an Klein UC ekeley Popeies of Ns Independence? ayes nes: pecify complex join disibuions using simple local condiional
More informationN 1. Time points are determined by the
upplemena Mehods Geneaon of scan sgnals In hs secon we descbe n deal how scan sgnals fo 3D scannng wee geneaed. can geneaon was done n hee seps: Fs, he dve sgnal fo he peo-focusng elemen was geneaed o
More informationReflection and Refraction
Chape 1 Reflecon and Refacon We ae now neesed n eplong wha happens when a plane wave avelng n one medum encounes an neface (bounday) wh anohe medum. Undesandng hs phenomenon allows us o undesand hngs lke:
More informationMonetary policy and models
Moneay polcy and odels Kes Næss and Kes Haae Moka Noges Bank Moneay Polcy Unvesy of Copenhagen, 8 May 8 Consue pces and oney supply Annual pecenage gowh. -yea ovng aveage Gowh n oney supply Inflaon - 9
More informationMCTDH Approach to Strong Field Dynamics
MCTDH ppoach o Song Feld Dynamcs Suen Sukasyan Thomas Babec and Msha Ivanov Unvesy o Oawa Canada Impeal College ondon UK KITP Sana Babaa. May 8 009 Movaon Song eld dynamcs Role o elecon coelaon Tunnel
More informationMaximum Likelihood Estimation
Mau Lkelhood aon Beln Chen Depaen of Copue Scence & Infoaon ngneeng aonal Tawan oal Unvey Refeence:. he Alpaydn, Inoducon o Machne Leanng, Chape 4, MIT Pe, 4 Saple Sac and Populaon Paaee A Scheac Depcon
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 informationOptimal Sensor Placement for Cooperative Distributed Vision
Opmal Senso Placemen fo Coopeave Dsbued Vson Lus E. Navao-Semen, John M. Dolan, and Padeep K. Khosla, Depamen of Eleccal and Compue Engneeng he Robocs Insue Canege Mellon Unves Psbugh, PA, USA {lenscmu,
More informationA hybrid method to find cumulative distribution function of completion time of GERT networks
Jounal of Indusal Engneeng Inenaonal Sepembe 2005, Vol., No., - 9 Islamc Azad Uvesy, Tehan Souh Banch A hybd mehod o fnd cumulave dsbuon funcon of compleon me of GERT newos S. S. Hashemn * Depamen of Indusal
More informationInteger Programming Models for Decision Making of. Order Entry Stage in Make to Order Companies 1. INTRODUCTION
Inege Pogammng Models fo Decson Makng of Ode Eny Sage n Make o Ode Companes Mahendawah ER, Rully Soelaman and Rzal Safan Depamen of Infomaon Sysems Depamen of Infomacs Engneeng Insu eknolog Sepuluh Nopembe,
More informationCS 188: Artificial Intelligence Fall Probabilistic Models
CS 188: Aificial Inelligence Fall 2007 Lecue 15: Bayes Nes 10/18/2007 Dan Klein UC Bekeley Pobabilisic Models A pobabilisic model is a join disibuion ove a se of vaiables Given a join disibuion, we can
More informationp E p E d ( ) , we have: [ ] [ ] [ ] Using the law of iterated expectations, we have:
Poblem Se #3 Soluons Couse 4.454 Maco IV TA: Todd Gomley, gomley@m.edu sbued: Novembe 23, 2004 Ths poblem se does no need o be uned n Queson #: Sock Pces, vdends and Bubbles Assume you ae n an economy
More informationClustering (Bishop ch 9)
Cluserng (Bshop ch 9) Reference: Daa Mnng by Margare Dunham (a slde source) 1 Cluserng Cluserng s unsupervsed learnng, here are no class labels Wan o fnd groups of smlar nsances Ofen use a dsance measure
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 informationHierarchical Production Planning in Make to Order System Based on Work Load Control Method
Unvesal Jounal of Indusal and Busness Managemen 3(): -20, 205 DOI: 0.389/ujbm.205.0300 hp://www.hpub.og Heachcal Poducon Plannng n Make o Ode Sysem Based on Wok Load Conol Mehod Ehsan Faah,*, Maha Khodadad
More informationBasic molecular dynamics
1.1, 3.1, 1.333,. Inoducon o Modelng and Smulaon Spng 11 Pa I Connuum and pacle mehods Basc molecula dynamcs Lecue Makus J. Buehle Laboaoy fo Aomsc and Molecula Mechancs Depamen of Cvl and Envonmenal Engneeng
More informationSolution of Non-homogeneous bulk arrival Two-node Tandem Queuing Model using Intervention Poisson distribution
Volume-03 Issue-09 Sepembe-08 ISSN: 455-3085 (Onlne) RESEARCH REVIEW Inenaonal Jounal of Muldscplnay www.jounals.com [UGC Lsed Jounal] Soluon of Non-homogeneous bulk aval Two-node Tandem Queung Model usng
More informationUniversity of California, Davis Date: June xx, PRELIMINARY EXAMINATION FOR THE Ph.D. DEGREE ANSWER KEY
Unvesy of Calfona, Davs Dae: June xx, 009 Depamen of Economcs Tme: 5 hous Mcoeconomcs Readng Tme: 0 mnues PRELIMIARY EXAMIATIO FOR THE Ph.D. DEGREE Pa I ASWER KEY Ia) Thee ae goods. Good s lesue, measued
More informationSolution in semi infinite diffusion couples (error function analysis)
Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of
More informationEfficient Bayesian Network Learning for System Optimization in Reliability Engineering
Qualy Technology & Quanave Managemen Vol. 9, No., pp. 97-, 202 QTQM ICAQM 202 Effcen Bayesan Newok Leanng fo Sysem Opmzaon n Relably Engneeng A. Gube and I. Ben-Gal Depamen of Indusal Engneeng, Faculy
More informationTecnologia e Inovação, Lisboa, Portugal. ABB Corporate Research Center, Wallstadter Str. 59, Ladenburg, Germany,
A New Connuous-Tme Schedulng Fomulaon fo Connuous Plans unde Vaable Eleccy Cos Pedo M. Caso * Io Hajunkosk and Ignaco E. Gossmann Depaameno de Modelação e Smulação de Pocessos Insuo Naconal de Engenhaa
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 information) from i = 0, instead of i = 1, we have =
Chape 3: Adjusmen Coss n he abou Make I Movaonal Quesons and Execses: Execse 3 (p 6): Illusae he devaon of equaon (35) of he exbook Soluon: The neempoal magnal poduc of labou s epesened by (3) = = E λ
More informationL4:4. motion from the accelerometer. to recover the simple flutter. Later, we will work out how. readings L4:3
elave moon L4:1 To appl Newon's laws we need measuemens made fom a 'fed,' neal efeence fame (unacceleaed, non-oang) n man applcaons, measuemens ae made moe smpl fom movng efeence fames We hen need a wa
More informationAn Automatic Door Sensor Using Image Processing
An Auomaic Doo Senso Using Image Pocessing Depamen o Elecical and Eleconic Engineeing Faculy o Engineeing Tooi Univesiy MENDEL 2004 -Insiue o Auomaion and Compue Science- in BRNO CZECH REPUBLIC 1. Inoducion
More informationMolecular dynamics modeling of thermal and mechanical properties
Molecula dynamcs modelng of hemal and mechancal popees Alejando Sachan School of Maeals Engneeng Pudue Unvesy sachan@pudue.edu Maeals a molecula scales Molecula maeals Ceamcs Meals Maeals popees chas Maeals
More informationCubic Bezier Homotopy Function for Solving Exponential Equations
Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.
More informationNumerical Study of Large-area Anti-Resonant Reflecting Optical Waveguide (ARROW) Vertical-Cavity Semiconductor Optical Amplifiers (VCSOAs)
USOD 005 uecal Sudy of Lage-aea An-Reonan Reflecng Opcal Wavegude (ARROW Vecal-Cavy Seconduco Opcal Aplfe (VCSOA anhu Chen Su Fung Yu School of Eleccal and Eleconc Engneeng Conen Inoducon Vecal Cavy Seconduco
More informationMultistage Median Ranked Set Sampling for Estimating the Population Median
Jounal of Mathematcs and Statstcs 3 (: 58-64 007 ISSN 549-3644 007 Scence Publcatons Multstage Medan Ranked Set Samplng fo Estmatng the Populaton Medan Abdul Azz Jeman Ame Al-Oma and Kamaulzaman Ibahm
More informationSuppose we have observed values t 1, t 2, t n of a random variable T.
Sppose we have obseved vales, 2, of a adom vaable T. The dsbo of T s ow o belog o a cea ype (e.g., expoeal, omal, ec.) b he veco θ ( θ, θ2, θp ) of ow paamees assocaed wh s ow (whee p s he mbe of ow paamees).
More informationLecture 17: Kinetics of Phase Growth in a Two-component System:
Lecue 17: Kineics of Phase Gowh in a Two-componen Sysem: descipion of diffusion flux acoss he α/ ineface Today s opics Majo asks of oday s Lecue: how o deive he diffusion flux of aoms. Once an incipien
More informationSCIENCE CHINA Technological Sciences
SIENE HINA Technologcal Scences Acle Apl 4 Vol.57 No.4: 84 8 do:.7/s43-3-5448- The andom walkng mehod fo he seady lnea convecondffuson equaon wh axsymmec dsc bounday HEN Ka, SONG MengXuan & ZHANG Xng *
More informationMechanics Physics 151
Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ
More informationAPPROXIMATIONS FOR AND CONVEXITY OF PROBABILISTICALLY CONSTRAINED PROBLEMS WITH RANDOM RIGHT-HAND SIDES
R U C O R R E S E A R C H R E P O R APPROXIMAIONS FOR AND CONVEXIY OF PROBABILISICALLY CONSRAINED PROBLEMS WIH RANDOM RIGH-HAND SIDES M.A. Lejeune a A. PREKOPA b RRR 7-005, JUNE 005 RUCOR Ruges Cene fo
More informationCourse Outline. 1. MATLAB tutorial 2. Motion of systems that can be idealized as particles
Couse Oulne. MATLAB uoal. Moon of syses ha can be dealzed as pacles Descpon of oon, coodnae syses; Newon s laws; Calculang foces equed o nduce pescbed oon; Deng and solng equaons of oon 3. Conseaon laws
More informationGarrettsville, Ohio Public Water System Lead (Pb) Components
Gaesvlle, Oho Publc Wae Sysem Lead (Pb) Componens PWSID# OH6701412 Pepaed Febuay 2017 To comply wh Secon 6109.21 of he Oho Revsed Code, enaced n Sepembe 2016, he Vllage of Gaesvlle n Poage Couny, Oho has
More informationDensity Matrix Description of NMR BCMB/CHEM 8190
Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary
More informationBackcalculation Analysis of Pavement-layer Moduli Using Pattern Search Algorithms
Bakallaon Analyss of Pavemen-laye Modl Usng Paen Seah Algohms Poje Repo fo ENCE 74 Feqan Lo May 7 005 Bakallaon Analyss of Pavemen-laye Modl Usng Paen Seah Algohms. Inodon. Ovevew of he Poje 3. Objeve
More information, on the power of the transmitter P t fed to it, and on the distance R between the antenna and the observation point as. r r t
Lecue 6: Fiis Tansmission Equaion and Rada Range Equaion (Fiis equaion. Maximum ange of a wieless link. Rada coss secion. Rada equaion. Maximum ange of a ada. 1. Fiis ansmission equaion Fiis ansmission
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 informationPerformance-Driven Resource Management In Layered Sensing
h Inenaonal Confeence on Infomaon Fuson Seale, WA, USA, July 6-9, 009 Pefomance-Dven Resouce Managemen In Layeed Sensng Chun Yang Sgem echnology, Inc. San Maeo, CA 9440 chunynag@sgem.com Ivan Kada Inelnk
More informationReal-coded Quantum Evolutionary Algorithm for Global Numerical Optimization with Continuous Variables
Chnese Jounal of Eleconcs Vol.20, No.3, July 2011 Real-coded Quanum Evoluonay Algohm fo Global Numecal Opmzaon wh Connuous Vaables GAO Hu 1 and ZHANG Ru 2 (1.School of Taffc and Tanspoaon, Souhwes Jaoong
More informationTSS = SST + SSE An orthogonal partition of the total SS
ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally
More informationReinforcement learning
Lecue 3 Reinfocemen leaning Milos Hauskech milos@cs.pi.edu 539 Senno Squae Reinfocemen leaning We wan o lean he conol policy: : X A We see examples of x (bu oupus a ae no given) Insead of a we ge a feedback
More informationThe Unique Solution of Stochastic Differential Equations. Dietrich Ryter. Midartweg 3 CH-4500 Solothurn Switzerland
The Unque Soluon of Sochasc Dffeenal Equaons Dech Rye RyeDM@gawne.ch Mdaweg 3 CH-4500 Solohun Swzeland Phone +4132 621 13 07 Tme evesal n sysems whou an exenal df sngles ou he an-iô negal. Key wods: Sochasc
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 informationcalculating electromagnetic
Theoeal mehods fo alulang eleomagne felds fom lghnng dshage ajeev Thoapplll oyal Insue of Tehnology KTH Sweden ajeev.thoapplll@ee.kh.se Oulne Despon of he poblem Thee dffeen mehods fo feld alulaons - Dpole
More informationAccelerated Sequen.al Probability Ra.o Test (SPRT) for Ongoing Reliability Tes.ng (ORT)
cceleaed Sequen.al Pobably Ra.o Tes (SPRT) fo Ongong Relably Tes.ng (ORT) Mlena Kasch Rayheon, IDS Copygh 25 Rayheon Company. ll ghs eseved. Cusome Success Is Ou Msson s a egseed adema of Rayheon Company
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 information. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.
Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons
More informationPhysics 2A Chapter 11 - Universal Gravitation Fall 2017
Physcs A Chapte - Unvesal Gavtaton Fall 07 hese notes ae ve pages. A quck summay: he text boxes n the notes contan the esults that wll compse the toolbox o Chapte. hee ae thee sectons: the law o gavtaton,
More informationNanoparticles. Educts. Nucleus formation. Nucleus. Growth. Primary particle. Agglomeration Deagglomeration. Agglomerate
ucs Nucleus Nucleus omaon cal supesauaon Mng o eucs, empeaue, ec. Pmay pacle Gowh Inegaon o uson-lme pacle gowh Nanopacles Agglomeaon eagglomeaon Agglomeae Sablsaon o he nanopacles agans agglomeaon! anspo
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 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 informationCombinatorial Approach to M/M/1 Queues. Using Hypergeometric Functions
Inenaional Mahemaical Foum, Vol 8, 03, no 0, 463-47 HIKARI Ld, wwwm-hikaicom Combinaoial Appoach o M/M/ Queues Using Hypegeomeic Funcions Jagdish Saan and Kamal Nain Depamen of Saisics, Univesiy of Delhi,
More informationProbabilistic Models. CS 188: Artificial Intelligence Fall Independence. Example: Independence. Example: Independence? Conditional Independence
C 188: Aificial Inelligence Fall 2007 obabilisic Models A pobabilisic model is a join disibuion ove a se of vaiables Lecue 15: Bayes Nes 10/18/2007 Given a join disibuion, we can eason abou unobseved vaiables
More informationOn The Estimation of Two Missing Values in Randomized Complete Block Designs
Mahemaical Theoy and Modeling ISSN 45804 (Pape ISSN 505 (Online Vol.6, No.7, 06 www.iise.og On The Esimaion of Two Missing Values in Randomized Complee Bloc Designs EFFANGA, EFFANGA OKON AND BASSE, E.
More informationLecture 5. Chapter 3. Electromagnetic Theory, Photons, and Light
Lecue 5 Chape 3 lecomagneic Theo, Phoons, and Ligh Gauss s Gauss s Faada s Ampèe- Mawell s + Loen foce: S C ds ds S C F dl dl q Mawell equaions d d qv A q A J ds ds In mae fields ae defined hough ineacion
More information3. A Review of Some Existing AW (BT, CT) Algorithms
3. A Revew of Some Exstng AW (BT, CT) Algothms In ths secton, some typcal ant-wndp algothms wll be descbed. As the soltons fo bmpless and condtoned tansfe ae smla to those fo ant-wndp, the pesented algothms
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 informationScienceDirect. Behavior of Integral Curves of the Quasilinear Second Order Differential Equations. Alma Omerspahic *
Avalable onlne a wwwscencedeccom ScenceDec oceda Engneeng 69 4 85 86 4h DAAAM Inenaonal Smposum on Inellgen Manufacung and Auomaon Behavo of Inegal Cuves of he uaslnea Second Ode Dffeenal Equaons Alma
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 informationto Assess Climate Change Mitigation International Energy Workshop, Paris, June 2013
Decomposng he Global TIAM-Maco Maco Model o Assess Clmae Change Mgaon Inenaonal Enegy Wokshop Pas June 2013 Socaes Kypeos (PSI) & An Lehla (VTT) 2 Pesenaon Oulne The global ETSAP-TIAM PE model and he Maco
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 informationA. Thicknesses and Densities
10 Lab0 The Eath s Shells A. Thcknesses and Denstes Any theoy of the nteo of the Eath must be consstent wth the fact that ts aggegate densty s 5.5 g/cm (ecall we calculated ths densty last tme). In othe
More informationLecture 18: Kinetics of Phase Growth in a Two-component System: general kinetics analysis based on the dilute-solution approximation
Lecue 8: Kineics of Phase Gowh in a Two-componen Sysem: geneal kineics analysis based on he dilue-soluion appoximaion Today s opics: In he las Lecues, we leaned hee diffeen ways o descibe he diffusion
More information. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.
Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are
More informationFI 3103 Quantum Physics
/9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon
More informationToday - Lecture 13. Today s lecture continue with rotations, torque, Note that chapters 11, 12, 13 all involve rotations
Today - Lecue 13 Today s lecue coninue wih oaions, oque, Noe ha chapes 11, 1, 13 all inole oaions slide 1 eiew Roaions Chapes 11 & 1 Viewed fom aboe (+z) Roaional, o angula elociy, gies angenial elociy
More informationDelay-Dependent Control for Time-Delayed T-S Fuzzy Systems Using Descriptor Representation
82 Inenaonal Jounal of Conol Auomaon and Sysems Vol 2 No 2 June 2004 Delay-Dependen Conol fo me-delayed -S Fuzzy Sysems Usng Descpo Repesenaon Eun ae Jeung Do Chang Oh and Hong Bae ak Absac: hs pape pesens
More informationDensity Matrix Description of NMR BCMB/CHEM 8190
Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon If we say wh one bass se, properes vary only because of changes n he coeffcens weghng each bass se funcon x = h< Ix > - hs s how we calculae
More informationGraduate Macroeconomics 2 Problem set 5. - Solutions
Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K
More informationA Novel Fast Otsu Digital Image Segmentation Method
The Inenaonal Aab Jounal of Infomaon Technology, Vol. 3, No. 4, July 06 47 A Novel Fas Osu Dgal Image Segmenaon Mehod Duaa AlSaeed,, Ahmed oudane,, and Al El-Zaa 3 Depamen of Compue Scence and Dgal Technologes,
More informationUnsupervised Cross-Domain Transfer in Policy Gradient Reinforcement Learning via Manifold Alignment
Unsupevsed Coss-Doman ansfe n Polcy Gaden Renfocemen Leanng va Manfold Algnmen Haham Bou Amma Unv. of Pennsylvana hahamb@seas.upenn.edu Ec Eaon Unv. of Pennsylvana eeaon@cs.upenn.edu Paul Ruvolo Oln College
More informationThe Application of Fuzzy Comprehensive Evaluations in The College Education Informationization Level
IOSR Jounal of Reseach & Mehod n Educaon IOSR-JRME) e- ISSN: 3 7388,p-ISSN: 3 737X Volume 8, Issue 3 Ve IV May June 8), PP -7 wwwosjounalsog The Applcaon of Fuzzy Compehensve Evaluaons n The College Educaon
More information