MODEL BASED TECHNIQUE FOR VEHICLE TRACKING IN TRAFFIC VIDEO USING SPATIAL LOCAL FEATURES

Size: px
Start display at page:

Download "MODEL BASED TECHNIQUE FOR VEHICLE TRACKING IN TRAFFIC VIDEO USING SPATIAL LOCAL FEATURES"

Transcription

1 MODEL BASED TECHNIQUE FOR VEHICLE TRACKING IN TRAFFIC VIDEO USING SPATIAL LOCAL FEATURES Arun Kumar H. D. 1 and Prabhakar C. J. 2 1 Deparmen of Compuer Science, Kuvempu Universiy, Shimoga, India ABSTRACT In his paper, we proposed a novel mehod for visible vehicle racking in raffic video sequence using model based sraegy combined wih spaial local feaures. Our racking algorihm consiss of wo componens: vehicle deecion and vehicle racking. In he deecion sep, we subrac he background and obained candidae foreground objecs represened as foreground mask. Afer obaining foreground mask of candidae objecs, vehicles are deeced using Co-HOG descripor. In he racking sep, vehicle model is consruced based on shape and exure feaures exraced from vehicle regions using Co-HOG and CS- LBP mehod. Afer consrucing he vehicle model, for he curren frame, vehicle feaures are exraced from each vehicle region and hen vehicle model is updaed. Finally, vehicles are racked based on he similariy measure beween curren frame vehicles and vehicle models. The proposed algorihm is evaluaed based on precision, recall and VTA merics obained on GRAM-RTM daase and i-lids daase. The experimenal resuls demonsrae ha our mehod achieves good accuracy. KEYWORDS Traffic vehicle racking, Vehicle model, Spaial local feaures, CS-LBP, Co-HOG, Shape feaures, Texure feaures, 1. INTRODUCTION Real-ime vehicle racking in raffic video is an essenial componen of an inelligen video raffic surveillance sysem. Accurae and real-ime vehicle racking will grealy improve he performance of vehicle classificaion, road vehicle densiy esimaion, vehicle aciviy analysis and high-level abnormal evens analysis like lane crossing, sudden and long ime vehicle sop. The aim of a vehicle racker is o generae he rajecory of he vehicle over ime by locaing is posiion in every frame of raffic video. Developmen of a robus racking mehod for vehicles is challenging because of: complex vehicle appearances like pose and scale variaions, occlusion (he vehicle may be occluded by he background or oher moving vehicles), and complex vehicle moion. The feaures-based vehicle racking algorihms (Perez, P. e al., 2002; Avidan, S., 2007; Ross, D. e al., 2008; Grabner, H. e al., 2006; Wang, S. e al., 2011) are mos promising and racking is performed based on racking of feaures such as disinguishable poins or lines on he vehicle. Selecing he righ feaures plays an imporan role in order o increase he accuracy of feauresbased racking algorihms (Beymer, e al., 1997). In general, he desirable propery of a visual feaure is is uniqueness so ha he vehicle can be easily disinguished in he feaure space. For example, color, exure, inensiy, and pixel-based feaures are he spaial appearance feaures widely used o rack he vehicle. Su, X. e al. (2007) have proposed rule based muliple objecs racking sysem for raffic surveillance using a collaboraive background exracion algorihm. DOI: /mlaij

2 Jung, Y.K. e al. (2001) have proposed feaures-based vehicle racking sysem, which exracs corner feaures of he vehicle and racks he feaures using a linear Kalman filer. Babaei, P. e al. (2010) have proposed he racking sysem which is based on a combinaion of a emporal difference and correlaion maching in defined raffic zones. The sysem effecively combines simple domain knowledge abou vehicle classes wih ime domain saisical measures o recognize arge vehicles in he presence of parial occlusions. Gao, e al. (2008) have proposed paricle filering based racking mehod. A moving vehicle is deeced by redundan discree wavele ransforms mehod (RDWT), and he key poins are obained by scale-invarian feaure ransform (SIFT). The maching of key poins in he follow-up frames is obained by he SIFT mehod and are used as he firs paricles o improve he racking performance. Dahlkamp, e al. (2004) proposed Edge-Elemen Associaion (EEA) and Marginalized Conour (MCo) approaches for 3D model-based vehicle racking in raffic scenes. Based on usage of global and local feaures, feaures-based racking algorihms can be furher classified ino wo caegories: Global mehods(ha, e al., 2011), and Local mehods (Grabner, H. e al., 2006; Yu, Q. e al., 2008; Tran, S. e al., 2007; He, W. e al., 2009; Wang, S. e al., 2011). The global mehods work in many pracical applicaions, bu have several basic limiaions. Firs, i is very difficul o capure he small changes in illuminaion variaion and difficul o represen he local deails like scale and shape variaions. Second, global represenaions are no robus for parial occlusion. Once he vehicles are occluded, he whole feaure vecor of vehicle represenaion is affeced. Third, global represenaions are hard o updae. Hence, global mehods are no efficien for vehicle racking in raffic video. Recenly, local mehods have opened a promising direcion o solve hese problems by represening a vehicle as a se of local pars or sparse local feaures. Par-based rackers generally use ses of conneced or visual local properies. The pars used for vehicle represenaion are updaed during racking by removing he old pars ha exhibi signs of drifing and adding new ones for easy accommodaion of appearance changes. In order o solve he problems associaed wih he global feaures-based algorihms, researchers have developed model-based racking algorihms for vehicle racking (Liu, X. e al., 2011; Cehovin, L. e al., 2013; Kwon, J. e al., 2009). In he model-based echnique, here are wo key componens: vehicle represenaion and dynamics. Vehicle represenaion ries o model he vehicle as correcly as possible so ha he racking algorihm can correcly describe he complex vehicle appearance. The vehicle dynamics model represens how he vehicle appearance evolves over ime o be able o handle appearance variaions. These wo problems are usually coupled ogeher. The vehicle represenaion should be designed o simply updae he model based on appearance variaions, while he vehicle dynamics should be able o ake advanage of he characerisics of vehicle represenaion for model updae. In his paper, we proposed vehicle racking in raffic video using model-based sraegy combined wih spaial local feaures. We consruc a vehicle model which capures he variaion in vehicle scale, vehicle pose, and complex vehicles occlusion based on spaial local feaures such as shape and exure feaures exraced using Co-HOG and CS-LBP operaor respecively. Afer consrucing he vehicle model for he curren frame, he vehicle feaures are exraced from each foreground mask of vehicle region and hen vehicle model is updaed. Finally, he vehicles are racked based on he similariy measure beween curren frame vehicles and vehicle model. 2. RELATED WORK Tracking is used o measure vehicle pahs in video sequences. The racking generally follows wo seps: in he firs sep, feaures for he vehicle regions are generaed in every video frame, and in 2

3 he second sep, a daa associaion sep has o provide correspondences beween he regions of consecuive frames based on he feaures and dynamic model. The vehicle racking is mainly used in wo ypes of raffic videos such as highway and urban raffic scenes videos. Vehicles racking on highways are easier han in urban raffic as here are few ypes of objecs (one moorized vehicle of various sizes), lile change in he orienaion of he vehicles and few known enry and exi poins. Cameras are also usually locaed much higher han in urban scenes, which reduce he vehicle occlusions. Tracking vehicles on highways are more challenging when he raffic is slower because he iner-vehicle space is significanly reduced, increasing he occlusion beween vehicles. In urban areas, raffic includes pedesrians, moorcycles, and vehicles, and more complicaed rajecories, wih vehicles urning a inersecions, sopping and parking, and many more enry and exi poins in he scene. Differen compuer vision mehods have hus been developed for hese wo applicaions. Rad, R. (2005) has proposed real-ime racking of muliple vehicles on he highway. They used Kalman filer and background subracion echniques. They exrac he conour of he vehicle using morphological operaions, and he algorihm has hree phases, deecion of pixels on moving vehicle, deecion of a shape of ineres in frame sequences and finally deerminaion of relaion among objecs in frame sequences. Ma, C. e al. (2016) have proposed fusion based hashing mehod for visual objec racking. Nguyen, P.V. e al. (2008) have proposed Muli-modal Paricle Filer (MPF) for racking vehicles. The aim of his mehod is o build some mos basic funcions of a moorcycle surveillance sysem using MPF based on he color observaion model. Babaei, P. e al. (2013) have mehod which addresses synchronizing he cameras for racking vehicles simulaneously in overlapping fields of view. Arrospide, J. e al. (2008) have proposed muli objec feaure racking sraegy. I racks specially seleced poins of he image based on compuaion of sparse opical flow. The racking sraegy includes a cenral oulier rejecion sage, ha ensures robusness of he racker based on probabilisic echniques, and a kalman filering sage o smooh ou he rajecories. Niknejad, H. e al. (2011) have proposed an embedded real ime mehod for deecion and racking of muli objecs including vehicles, pedesrians, moorbikes and bicycles in urban environmen. The feaures of differen objecs are learned as a deformable objec model hrough he combinaion of a laen suppor vecor machine (LSVM) and hisograms of oriened gradiens(hog). Laser deph daa have been used as a priori o generae objecs hypohesis regions and HOG feaure pyramid level is used o reduce he deecion ime. Deeced objecs are racked hrough a paricle filer which fuses he observaions from laser map and sequenial images. Messelodi, S. e al. (2005) have proposed he sysem ha uses a combinaion of segmenaion and moion informaion o localize and rack moving vehicles on he urban road plane, uilizing a robus background updaing, and a feaure-based racking mehod. Srigel, E. e al. (2013) have proposed vehicle deecion and racking a inersecions by fusing muliple camera views. Using his fusion map, he pose, widh and heigh of he vehicles can be deermined. Afer ha, he deeced vehicles are racked by a Gaussian-Mixure approximaion of he Probabiliy Hypohesis Densiy filer. Zheng, Y. e al. (2012) have proposed model based vehicle localizaion and racking for urban raffic surveillance using image gradien maching. The maching beween he 3D model projecion and 2D image daa is a key echnique for model based localizaion, recogniion and racking problems. Lee, K. H. e al. (2015) have proposed a model based 3D consrained muliple kernel racking. This approach regards each pach of he 3D vehicle model as a kernel and racks he kernels under cerain consrains faciliaed by he 3D geomery of he vehicle model. A kernel densiy esimaor is designed o well fi he 3D vehicle model during racking. Kim,G. e al. (2011) proposed vehicle racking based on Kalman filer. They deec cars based in Adaboos and he vehicles are racked using Kalman filer. Barh, A. e al. (2010) have proposed real-ime muli-filer approach for vehicle racking a inersecions. Boh moion and deph informaion is combined o esimae he pose and moion parameers of an oncoming vehicle, including he yaw rae, by means of kalman filering. 3

4 Niknejad, H. T. e al. (2012) have proposed muli-vehicle deecion and racking using vehicle mouned monocular camera. The feaures of vehicles are learned by Laen Suppor Vecor Machine(LSVM) and Hisograms of Oriened Gradiens (HOG). The deecion algorihm combines boh global and local feaures of he vehicle as a deformable objec model. Deeced vehicles are racked hrough a paricle filer. Sivaraman, S. e al., (2011) have proposed sereomonocular fusion approach o on-road localizaion and racking of vehicles. Uilizing a calibraed sereo-vision rig, he proposed approach combines monocular deecion wih sereo-vision for on road vehicle localizaion and racking for driver assisance. The sysem iniially acquires synchronized monocular frames and calculaes deph maps from he sereo rig. The sysem hen deecs vehicles in he image plane using an acive learning-based monocular vision approach. Using he image coordinaes of deeced vehicles, he sysem hen localizes he vehicles in realworld coordinaes using he calculaed deph map. The vehicles are racked boh in he image plane, and in real-world coordinaes. 3. PROPOSED WORK Our approach for racking of vehicles in raffic video based on model-based sraegy involves wo seps. In he firs sep, background is subraced and vehicles are deeced in frame. We subrac he background and obained candidae foreground objecs represened as foreground mask using our previous work (Arun Kumar, H.D. e al., 2015). The background subracion reduces compuaion ime and removes complex background. Afer obaining foreground mask of candidae objecs in frame, vehicles are deeced using Co-occurrence Hisograms of Oriened Gradien (Co-HOG) descripor. In he second sep, we consruc a vehicle model for each vehicle in frame based on shape and exure feaures exraced from he foreground mask of vehicle regions using Co-HOG and CS-LBP operaor. The vehicle model capures he variaion in vehicle scale, vehicle pose, and complex vehicles occlusion. Afer consrucing he vehicle model for he curren frame, he vehicle feaures are exraced from each deeced vehicle image and hen vehicle model is updaed. Finally, he vehicles are racked based on he similariy measure beween curren frame vehicles and vehicle model. Vehicle posiion is locaed by inegraing all maching in he vehicle model. Since, feaures may appear and disappear due o viewpoin changes and occlusions, our dynamic model is designed o be able o add a new feaure model and remove expired ones adapively and dynamically. The Hungarian algorihm is used o link deecions for racking. The flow diagram of proposed vehicle racking approach is shown in Figure 1. 4

5 3.1 Formulaion of proposed approach Figure 1. The flow diagram of our approach Traffic vehicle racking can generally be formulaed as a muli-variable esimaion problem. i Given he video sequence { V1, V2,..., V m } as inpu, we use S o indicae he sae of he h h i vehicle in he frame. We use S = ( S1,, S2,,..., SM ) o indicae he saes of all he h M vehicles in he frame, O { O }, i 1, 2,..., S = { S, S,..., S } o indicae he sequenial saes of he i,1: i,1 i,2 i, = = as all he deecions, h i vehicle from he firs frame o h he frame, and S1: = { S1, S2,..., S} o indicae all he sequenial saes of all he vehicles from h he firs frame o he frame. S = arg max P( S O ), (1) 1: 1: 1: The sae of he vehicle in he curren frame only depends on he sae of he vehicle in previous frame 1. When we process frame, only he racking in frame 1 and he image in he h frame I are involved in he calculaion. P( S O ) = P( S O, S ) = P( S O, I, S ) P( O I, S ) do, (2) 1: where P( O I, S 1) indicaes how realisic he deecion in frame is, P( S O, I, S 1) describes how well he deecions in frame maches racking s in frame 1. 5

6 3.2 Background Subracion and Vehicles Deecion In his secion, we inroduce he mehod for background subracion and vehicles deecion. The formulaion for background subracion (foreground deecion) and vehicles deecion process is P( O I, S ) defined in equaion (2), which is described as: 1 Pfg Pde if he caegory of objec is moving vehicle, P( O I, S 1) = (3) P oherwise. bg We subrac he background and obained he foreground mask of candidae moving objecs using our approach proposed in he previous work (Arun Kumar, H.D. e al. 2015), which reduces he compuaion ime and removes he complex background. Our previous approach has verified o be an efficien and effecive background subracion mehod. Our approach for background subracion uses modified SXCS-LBP exure descripor for finding foreground mask of candidae objecs, 1 if O is in foreground regions by background subracion, Pfg ( O I, S 1) = 0 oherwise. (4) Afer background subracion, each pixel is labelled as foreground or background using 0 or 1, 0 indicaes background and 1 indicaes foreground. Once he background subracion process is compleed, we obain foreground mask for each candidae objec. Once foreground mask of each objec is obained, i is necessary o deec vehicles among candidae objecs in he frame. The deecion of vehicles is described as P de. In order o deec vehicles, we uilize he approach proposed by Tomoki Waanabe, e al. (2009), in which Cooccurrence Hisograms of Oriened Gradien (Co-HOG) feaures are exraced o represen moving vehicles. P 1 if O is obained by he deecion, ( O I, S ) = (5) 0 oherwise. de Vehicle Tracking Afer deecing he vehicles in frame, we rack he moving vehicles in raffic video. There are wo subsecions. Firs, we iniialize he racking, in he second sep, we calculae he similariy beween racking and observaion, and hen observaion could be linked o being a racking Iniializaion of racking In order o iniialize racking, vehicle model is consruced for each deeced vehicle a ime. The appearance of he vehicle under racking may change over ime due o changes of vehicle scale, vehicle pose, complex vehicle occlusion, and he appearance variaion would lead o losing rack. Hence, vehicle model capures hese changes occurred in vehicle while i is moving. In he saring sage of he vehicle racking, he arge vehicle model in he firs frame is only iniialized and all he remaining vehicle models are empy. Generally, he vehicle model is updaed incremenally. Whenever a larger appearance variaion is deeced, he updaed arge vehicle models are sored in he empy models. 6

7 For M vehicles a ime, we consruc M vehicle models represened by a shape and exure feaures obained using Co-HOG and CS-LBP mehod respecively. In he following subsecions, we presen procedure involved in exracing he Co-HOG and CS-LBP feaures from each deeced vehicle image. Co-occurrence Hisogram of Oriened Gradiens (Co-HOG) Co-occurrence Hisogram of Oriened Gradiens (Co-HOG) (Tomoki Waanabe, e al., 2009) descripor is an exension of he original HOG shape descripor ha capures he spaial informaion of neighboring pixels. Insead of couning he occurrence of he gradien orienaion of a single pixel, gradien orienaions of wo or more neighboring pixels are considered. For each pixel in an image block, he gradien orienaions of he pixel pair formed by is neighbor and iself are examined. The Co-HOG has wo imporan meris. One is he robusness agains illuminaion variaion because gradien orienaions are compued from he local inensiy difference. The oher meri is he robusness agains deformaions because sligh shifs deformaions make small hisogram value changes. The co-occurrence marix expresses he disribuion of gradien orienaions a a given offse over an image. The combinaions of neighbor gradien orienaions can express shapes in deail. Mahemaically, a co-occurrence marix K is defined a an each block N Χ M of an image I, parameerized by an offse (x, y) as: K x, y ( i, j) 1 if I( p, q) = i and I( p + x, q + y) = j, (6) 0 oherwise. N M = p= 1 q= 1 We describe he process of Co-HOG calculaion as follows. Iniially, we compue gradien orienaions from an image by I y θ = arcan, I x (7) where arcan( ) reurns he inverse angen of he elemens in degrees. I y and I x are verical and horizonal gradien respecively calculaed by Gaussian filer. We label each pixel wih one of eigh discree orienaions. In our approach, all 0 0 o orienaions are spli up ino eigh orienaions per Then, we compue co-occurrence marices using Eq. (6). We used 31 offses, including a zero offse. In mos of oher applicaions, he auhors proceed by dividing an image ino a number of blocks and from each block exrac co-occurrence marices. We divide he vehicle image pach ino non overlapping blocks of size N Χ M, he co-occurrence marices are compued for each block. Finally, he componens of all he co-occurrence marices are concaenaed ino a vecor. We divide he vehicle image ino 2 Χ 4 (he accuracy of our approach is considerably beer han for oher number of blocks. Hence, in all he experimens, we divide he vehicle image ino 2 Χ 4 blocks) blocks and Co-HOGs a each block are compued. Figure 2 gives an illusraion of Co- HOG feaure descripor exracion process from a given vehicle image. 7

8 Figure 2. Illusraion of Co-HOG feaure descripor exracion process The dimension of Co-HOG feaure vecor for he given vehicle image is 15,424 when we divide he vehicle image ino 2 Χ 4 blocks. From one small region or block, Co-HOG obains 31 cooccurrence marices. A co-occurrence marix has 64 componens. Thus, Co-HOG obains ( ) (2 4) = 15,424 componens for each vehicle image. Cener Symmeric Local Binary Paern (CS-LBP) Heikkila, e al. (2002) have proposed a novel ineres region descripor called as Cener Symmeric Local Binary Paern (CS-LBP) descripor which is an exension of LBP exure operaor. The CS-LBP descripor has several advanages such as olerance o illuminaion changes, robusness on fla image areas, and compuaional efficiency. The CS-LBP compares he gray values of pairs of pixels in he cener-symmeric direcion. For 8 neighbors, LBP produces 256 differen binary paerns, whereas for CS-LBP produces only 16 binary paerns, Figure 3 gives an illusraion of CS-LBP feaure descripor compuaion process, and CS-LBP is mahemaically defined as follows: where g p and p p+ 2 p 1 2 p CS LBPR, P = S p 2, p g g p= 0 p+ (8) 2 g correspond o gray values of he cener-symmery pair of pixels and he funcion S( x) is hreshold funcion defined as follows: 1 x 0 S( x) = (9) 0 oherwise. Figure 3. LBP and CS-LBP feaures for neighborhood of 8 pixels 8

9 The concaenaed Co-HOG and CS-LBP feaure vecor of each vehicle represens vehicle model for each vehicle and concaenaed feaure vecor (fv) of each vehicle (O i ) is defined as N { hv j} j= 1 where N is he oal number of feaure vecor bins and fv( O ) =, (10) i hv j is he value of h j bin. For each rajecory, he vehicle model is updaed frame by frame. The appearance in he curren frame is he mos imporan one, so we updae as follows: ( α ) fv = k 1 1 fv + + k α fvc, (11) h where fvk is he feaure vecor afer updaing in k frame,α is a consan which is se o 0.8 and fv c is he feaure vecor calculaed in he curren frame. Each vehicle is racked based on minimum disance beween feaure vecor of he vehicle in he curren frame and is associaed vehicle model. The Hellinger disance is used o compare he feaure vecors N 1 d fv, fv = 1 hv hv, (12) ( ) 1 2 q,1 q,2 2 fv 1 1, fv q 2N = where fv1 and fv 2 are wo feaure vecors, and 1 fv = hv k = k N q, k, 1,2 (13) N q = Similariy beween racking and observaion We simplify P( S O, I, S 1) is probabiliy for a racking and a deecion, where ( ) P( S O, I, S 1) = Pa, (14) ( ( ), ) 2 1 d fv D fv Pa = exp, (15) 2 2πσ 2σ a a fv D is he feaure vecor of deeced vehicle in frame, fv is he feaure vecor of racking afer updaing in frame, d (.) is he funcion o calculae Hellinger disance, and σ a is a given hreshold. 4. EXPERIMENTAL RESULTS The performance evaluaion of he proposed mehod for vehicle-racking is a frame-by-frame evaluaion process. We carry ou experimens on challenging wo raffic surveillance video daases such as GRAM-Road Traffic Monioring (GRAM-RTM) (Guerrero-Gómez-Olmedo, e al., 2013) and i-lids. In order o measure he performance of our approach for vehicle-racking, we used evaluaion merics such as precision, recall, and Vehicle-Tracking Accuracy (VTA ) 9

10 (Smih, K. e al., 2005). The precision measures how much of he esimaes (ε is racker oupus are referred o as esimaes) cover he ground ruh (GT ) vehicle and can ake values beween 0 (no overlap) and 1 (full overlap). I is possible o have high precision wih poor qualiy racking as depiced in Figure 4(a). The recall measures how much of hegt is covered by hen ε and can ake values beween 0 (no overlap) and 1 (full overlap). I is possible o have a high recall ye have poor qualiy racking (Figure 4(b)). Vehicle-Tracking Accuracy (VTA ) is oal posiion error for maching vehicle hypohesis pair over all frames, averaged by he oal number of maches. The precision ( v i, j ), recall ( ρ i, j ), and VTA are defined as follows: v ε GT = (16) i j i, j, ε i ρ ε GT = (17) i j i, j, GTj ( M + FP + MM ) ( GT ) ( ) VTA = 1 = 1 M + FP + MM, (18) where GT j is he ground ruh for racking arge vehicles and indexed by j, ε i is he racker oupu are referred o as esimaes and indexed by i, M is he number of missed vehicles a ime. FP is he number of false posiives ha correspond o deec vehicles and ha do no overlap any real vehicles in he scene. MM is he number of mismaches a ime. number of vehicles a ime. M, FP and MM represen he corresponding raio. GTj is he oal Figure 4. a) precision ( v ) b) recall ( ρ ) c) boh precision and recall should have high values There are four basic ypes of errors ha our sysem can make. The firs ype of error may occur when a vehicle exiss, bu he sysem does no recognize i (False Negaive: A ground ruh objec exiss ha is no associaed wih an esimae). The second ype of error occurs when he sysem may indicae he presence of a vehicle which does no exis (False Posiive: An esimae exiss ha is no associaed wih a ground ruh objec). The hird ype of error occurs when one vehicle is racked by muliple esimaes (Muliple Trackers: Two are more esimaes are associaed wih he same ground ruh). The las ype of error occurs when muliple vehicles are racked by one esimae (Muliple Objecs: Two or more ground ruh objecs are associaed wih he same esimae). These errors are depiced in Figure 5. 10

11 Figure 5. Types of configuraion errors ε s (1, 2,3, 4) aemp o rack four GT s ( a, b, c, d ) 4.1. Experimens on GRAM-RTM daase In he firs se of experimens, we evaluaed he performance of our approach quaniaively using GRAM Road-Traffic Monioring (GRAM-RTM) daase. The GRAM Road-Traffic Monioring daase consiss of hree raffic video sequences and hese video sequences were recorded under differen condiions and wih differen plaforms. The firs video sequence is M-30 video (7529 frames), was recorded on a sunny day and resoluion for each frame is 800 Χ 480. The second video sequence is M-30-HD (9390 frames), was recorded in he same locaion, bu cloudy day and resoluion for each frame is 1200 Χ 720. The hird video sequence Urban1 (2345 frames), was recorded a a busy urban inersecion and he resoluion of each frame is 600 Χ 360. The ground ruhs of hese hree video sequences were manually obained. We compared he performance of our approach wih CS-LBP alone (Texure descripor) and Co-HOG alone (Shape Descripor). Table 1 shows he comparaive sudy of our approach (CS-LBP + Co-HOG) wih CS-LBP alone and CO-HOG alone. All of he values represen he average of he considered crierion obained for he whole hree M-30, M-30-HD, and Urban1 video sequences. I is observed ha our approach achieves he highes precision and highes recall compared o CS-LBP alone and CO- HOG alone for M-30, M-30-HD, and Urban1 video sequences. The VTA of our approach is 89%, 88% and 81% for he hree challenging video sequences, which means ha he racking of almos all of he vehicles is deeced. The low FP level of our approach (03%, 02%, and 06%) shows ha almos all of he deeced racking s correspond o a real vehicle in he scene. 03% of he vehicles are oally missing from he M-30 video sequence, while 04% are missing from he M-30-HD video sequence, and 06% are missing from he Urban1 video sequence. The increase in FP for Urban1 video sequence is mainly due o some pedesrians locaed a he scene of he ROI. Our approach decreases he mismach rae (03%, 02% and 06% for respecive video sequences) compared o shape and a exure descripor alone. The combinaion of CS-LBP and Co-HOG improves he VTA, precision, recall and reduce false deecion, mismach and missed vehicles. Figure 6 shows he qualiaive performance of our approach for all hree GRAM Road-Traffic Monioring daase video sequences. The red color windows in he sample video frames describe racked vehicles. 11

12 Table 1. The resuls comparison of our approach wih CS-LBP alone and Co-HOG alone based on precision, recall and VTA for M-30, M-30-HD, and Urban1 video sequences Experimens on i-lids daase In he second se of experimens, we evaluaed he performance of our approach quaniaively using i-lids daase. The i-lids daase consiss of seven raffic video sequences, among seven video sequences, we seleced AVSS PV Easy video sequence which includes scenes of vehicle urning, illuminaion changes, and vehicles moving from far o near. The resoluion for each frame is 720 Χ 576. The ground ruhs of his video sequence were manually obained. Table 2 shows he precision, recall, and VTA based quaniaive comparison of our approach resul wih CS-LBP alone and Co-HOG alone for he AVSS PV easy video sequence. The proposed approach (CS-LBP + Co-HOG) has achieved highes precision, recall and VTA compared o CS- LBP alone and Co-HOG alone. The FP rae of our approach is very high because AVSS PV easy video sequence conains some pedesrians, moorcycles, and vehicles moving from far o near. Figure 7 shows he racking resuls of our approach for sample frames of AVSS PV easy i-lids daase. The firs row of Figure 7 shows sample original video frames where vehicles pose are changing because of curved lane. The second row shows racking resul represened using red color window. I is observed ha our approach accuraely rack he vehicles even hough vehicle pose changes. This is because, in each frame, he vehicle model is updaed efficienly in order o capure he variaion in pose of he vehicles. 12

13 Figure 6. Vehicle racking resul of our approach for all hree video sequences of GRAM Road-Traffic Monioring daase a) M-30 b) M-30-HD, and c) Urban1(firs row is exclusion area shown using red color). Table 2. The resuls comparison of our approach wih CS-LBP alone and Co-HOG alone based on precision, recall and VTA for AVSS PV Easy video sequence. 13

14 Figure 7. Vehicle racking resul of our approach for AVSS PV easy video sequence 4.3. Visual comparison wih exising mehod We compared racking resuls of our approach obained on i-lids daase visually wih he racking resuls of SIFT-based Mean Shif algorihm proposed by Liang e al. (2014). The firs row of Figure 8 shows sample original video frames, such as Frame #26, Frame #55, Frame #75, and Frame #85. The second and hird row shows he resuls of Mean Shif mehod and our proposed approach resuls for he sample frames. The racking window of Mean shif mehod deviaes a he Frame #55, Frame #75 and Frame #85. This is because he color hisogram of he candidae emplae is changed when he moving vehicle is urning lef, and he illuminaion is affeced by he shadow of he building. For Mean Shif algorihm, when he illuminaion and shape of he vehicle changes, he number of mached poins grealy increases and he performance of he mehod decreases, which records false racking rae. Our approach resuls presened in he hird row demonsrae ha he moving vehicles are racked more accuraely for he Frame #26, Frame #55, Frame #75, and Frame #85. The increase in racking rae of our approach is due o he fac ha adapaion of CS-LBP descripor, which is illuminaion invarian and exracs accurae vehicles exure feaures, and he Co-HOG descripor gives accurae shape feaures even hough he vehicle changes is pose. Hence, he combinaion of shape and exure descripors increases he vehicle racking resul. 14

15 Figure 8. Vehicle racking resul for AVSS PV Easy video sequences for Mean Shif (red), and proposed mehod (yellow) a) Frame #26, b) Frame #55, c) Frame #75, and d) Frame # CONCLUSION In his paper, we proposed a model-based vehicle racking echnique using spaial local feaures such as shape and exure feaures. The shape descripor such as Co-HOG is used for he represenaion of vehicle shape and CS-LBP exure descripor is used for represenaion of vehicle exure. The vehicle model is consruced which capures he variaion in illuminaion, vehicle scale, vehicle pose and complex vehicles occlusion. The evaluaion process conduced on wo popular daases such as GRAM-RTM and i-lids demonsrae ha our approach achieves highes vehicle racking accuracy. The visual comparison wih exising mehod shows ha our approach yields accurae racking even vehicle pose and illuminaion changes. The drawback of our approach is ha when he pedesrians and moorcycles are presen in he video sequence, our approach racks hese objecs as vehicles and i reduces he racking accuracy. REFERENCES [1] Pérez, P., Hue, C., Vermaak, J., & Gangne, M. (2002), Color-based probabilisic racking, In Compuer vision ECCV 2002 (pp ). Springer Berlin Heidelberg. [2] Avidan, S. (2007), Ensemble racking, Paern Analysis and Machine Inelligence, IEEE Transacions on, 29(2), [3] Ross, D. A., Lim, J., Lin, R. S., & Yang, M. H. (2008), Incremenal learning for robus visual racking, Inernaional Journal of Compuer Vision, 77(1-3), [4] Grabner, H., & Bischof, H. (2006, June), On-line boosing and vision, In Compuer Vision and Paern Recogniion, 2006 IEEE Compuer Sociey Conference on (Vol. 1, pp ). IEEE. [5] Grabner, H., Grabner, M., & Bischof, H. (2006, Sepember), Real-Time Tracking via On-line Boosing, In BMVC (Vol. 1, No. 5, p. 6). [6] Wang, S., Lu, H., Yang, F., & Yang, M. H. (2011, November), Superpixel racking, In Compuer Vision (ICCV), 2011 IEEE Inernaional Conference on(pp ). IEEE. 15

16 [7] Beymer, D., McLauchlan, P., Coifman, B., & Malik, J. (1997, June), A real-ime compuer vision sysem for measuring raffic parameers, In Compuer Vision and Paern Recogniion, Proceedings., 1997 IEEE Compuer Sociey Conference on (pp ). [8] Su, X., Khoshgofaar, T. M., Zhu, X., & Folleco, A. (2007), Rule-based muliple objec racking for raffic surveillance using collaboraive background exracion, In Advances in Visual Compuing (pp ). Springer Berlin Heidelberg. [9] Jung, Y. K., & Ho, Y. S. (2001), A feaure-based vehicle racking sysem in congesed raffic video sequences, In Advances in Mulimedia Informaion Processing PCM 2001 (pp ). Springer Berlin Heidelberg. [10] Babaei, P. (2010, December), Vehicles racking and classificaion using raffic zones in a hybrid scheme for inersecion raffic managemen by smar cameras, In Signal and Image Processing (ICSIP), 2010 Inernaional Conference on (pp ). IEEE. [11] Gao, T., Liu, Z. G., Gao, W. C., & Zhang, J. (2008), Moving vehicle racking based on SIFT acive paricle choosing, In Advances in Neuro-Informaion Processing (pp ). Springer Berlin Heidelberg. [12] Dahlkamp, H., Pece, A. E., Olik, A., & Nagel, H. H. (2004), Differenial analysis of wo modelbased vehicle racking 0020 approaches, In Paern Recogniion (pp ). Springer Berlin Heidelberg. [13] Ha, S. W., & Moon, Y. H. (2011), Muliple objec racking using SIFT feaures and locaion maching, Inernaional Journal of Smar Home, 5(4), [14] Yu, Q., Dinh, T. B., & Medioni, G. (2008), Online racking and reacquisiion using co-rained generaive and discriminaive rackers, In Compuer Vision ECCV 2008 (pp ). Springer Berlin Heidelberg. [15] Tran, S., & Davis, L. (2007, Ocober), Robus Objec Tracking wih Regional Affine Invarian Feaures, In Compuer Vision, ICCV IEEE 11h Inernaional Conference on (pp. 1-8). IEEE. [16] He, W., Yamashia, T., Lu, H., & Lao, S. (2009, Sepember), Surf racking, In Compuer Vision, 2009 IEEE 12h Inernaional Conference on (pp ). IEEE. [17] Liu, X., Lin, L., Yan, S., Jin, H., & Jiang, W. (2011), Adapive objec racking by learning hybrid emplae online, Circuis and Sysems for Video Technology, IEEE Transacions on, 21(11), [18] Cehovin, L., Krisan, M., & Leonardis, A. (2013), Robus visual racking using an adapive coupled-layer visual model, Paern Analysis and Machine Inelligence, IEEE Transacions on, 35(4), [19] Kwon, J., & Lee, K. M. (2009, June), Tracking of a non-rigid objec via pach-based dynamic appearance modeling and adapive basin hopping Mone Carlo sampling, In Compuer Vision and Paern Recogniion, CVPR IEEE Conference on (pp ). IEEE. [20] Rad, R., & Jamzad, M. (2005), Real ime classificaion and racking of muliple vehicles in highways, Paern Recogniion Leers, 26(10), [21] Ma, C., Liu, C., Peng, F., & Liu, J. (2016), Muli-feaure Hashing Tracking, Paern Recogniion Leers, 69, [22] Nguyen, P. V., & Le, H. B. (2008), A muli-modal paricle filer based moorcycle racking sysem, In PRICAI 2008: Trends in Arificial Inelligence (pp ). Springer Berlin Heidelberg. [23] Babaei, P., Fahy, M., & Berangi, R. (2013), Consisen Vehicles Tracking By Using A Cooperaive Disribued Video Surveillance Sysem, Inernaional Research Journal of Applied and Basic Sciences, 4(10), [24] Arróspide, J., Salgado, L., Nieo, M., & Jaureguizar, F. (2008, Ocober), On-board robus vehicle deecion and racking using adapive qualiy evaluaion, In h IEEE Inernaional Conference on Image Processing. IEEE. [25] Niknejad, H. T., Takahashi, K., Mia, S., & McAlleser, D. (2011, June), Embedded muli-sensors objecs deecion and racking for urban auonomous driving, In Inelligen Vehicles Symposium (IV), 2011 IEEE (pp ). IEEE. [26] Messelodi, S., Modena, C. M., & Zanin, M. (2005), A compuer vision sysem for he deecion and classificaion of vehicles a urban road inersecions, Paern analysis and applicaions, 8(1-2),

17 [27] Srigel, E., Meissner, D., & Diemayer, K. (2013, June), Vehicle deecion and racking a inersecions by fusing muliple camera views, In Inelligen Vehicles Symposium (IV), 2013 IEEE (pp ). IEEE [28] Zheng, Y., & Peng, S. (2012, Sepember), Model based vehicle localizaion for urban raffic surveillance using image gradien based maching, In Inelligen Transporaion Sysems (ITSC), h Inernaional IEEE Conference on (pp ). IEEE. [29] Lee, K. H., Hwang, J. N., & Chen, S. I. (2015), Model-Based Vehicle Localizaion Based on 3-D Consrained Muliple-Kernel Tracking, Circuis and Sysems for Video Technology, IEEE Transacions on, 25(1), [30] Kim, G., Kim, H., Park, J., & Yu, Y. (2011), Vehicle racking based on kalman filer in unnel, In Informaion Securiy and Assurance (pp ). Springer Berlin Heidelberg. [31] Barh, A., & Franke, U. (2010, Sepember), Tracking oncoming and urning vehicles a inersecions, In Inelligen Transporaion Sysems (ITSC), h Inernaional IEEE Conference on (pp ). IEEE. [32] Niknejad, H. T., Takeuchi, A., Mia, S., & McAlleser, D. (2012), On-road mulivehicle racking using deformable objec model and paricle filer wih improved likelihood esimaion, Inelligen Transporaion Sysems, IEEE Transacions on, 13(2), [33] Sivaraman, S., & Trivedi, M. M. (2011, Ocober), Combining monocular and sereo-vision for real-ime vehicle ranging and racking on mulilane highways In h Inernaional IEEE Conference on Inelligen Transporaion Sysems (ITSC) (pp ). IEEE. [34] Arun Kumar, H. D., Prabhakar, C. J. (2015), Moving Vehicles Deecion in Traffic Video Using Modified SXCS-LBP Texure Descripor, Inernaional Journal of Compuer Vision and Image Processing (IJCVIP), Vol 5, No. 2, pp [35] Waanabe, T., Io, S., & Yokoi, K. (2009), Co-occurrence hisograms of oriened gradiens for pedesrian deecion, In Advances in Image and Video Technology (pp ). Springer Berlin Heidelberg. [36] Heikkilä, M., & Pieikäinen, M. (2002), A exure-based mehod for modeling he background and deecing moving objecs, Paern Analysis and Machine Inelligence, IEEE Transacions on, 28(4), [37] Guerrero-Gómez-Olmedo, R., López-Sasre, R. J., Maldonado-Bascón, S., & Fernández-Caballero, A. (2013, June), Vehicle racking by simulaneous deecion and viewpoin esimaion, In Inernaional Work-Conference on he Inerplay Beween Naural and Arificial Compuaion (pp ). Springer Berlin Heidelberg. [38] Smih, K., Gaica-Perez, D., Odobez, J. M., & Ba, S. (2005, June), Evaluaing muli-objec racking, In 2005 IEEE Compuer Sociey Conference on Compuer Vision and Paern Recogniion (CVPR'05)-Workshops (pp ). IEEE. [39] i-lids: hp:// Auhors Arun Kumar H. D. Received M.Sc. degree in compuer Science from Kuvempu Universiy, Karnaaka, India in 2009, He is pursuing Ph.D. degree in Kuvempu Universiy, Karnaaka, India. His research ineress are image and video processing, Compuer Vision and Machine Vision Prabhakar C.J. received Ph.D. degree in Compuer Science in he year 2009 from Gulbarga Universiy, Gulbarga, Karnaaka, India. He is currenly working as Assisan Professor in he deparmen of Compuer Science, Kuvempu Universiy, Karnaaka, India. His research ineress are compuer vision, Image and video processing. 17

Improved TLD Algorithm for Face Tracking

Improved TLD Algorithm for Face Tracking Absrac Improved TLD Algorihm for Face Tracking Huimin Li a, Chaojing Yu b and Jing Chen c Chongqing Universiy of Poss and Telecommunicaions, Chongqing 400065, China a li.huimin666@163.com, b 15023299065@163.com,

More information

CAMERA CALIBRATION BY REGISTRATION STEREO RECONSTRUCTION TO 3D MODEL

CAMERA CALIBRATION BY REGISTRATION STEREO RECONSTRUCTION TO 3D MODEL CAMERA CALIBRATION BY REGISTRATION STEREO RECONSTRUCTION TO 3D MODEL Klečka Jan Docoral Degree Programme (1), FEEC BUT E-mail: xkleck01@sud.feec.vubr.cz Supervised by: Horák Karel E-mail: horak@feec.vubr.cz

More information

A Fast Stereo-Based Multi-Person Tracking using an Approximated Likelihood Map for Overlapping Silhouette Templates

A Fast Stereo-Based Multi-Person Tracking using an Approximated Likelihood Map for Overlapping Silhouette Templates A Fas Sereo-Based Muli-Person Tracking using an Approximaed Likelihood Map for Overlapping Silhouee Templaes Junji Saake Jun Miura Deparmen of Compuer Science and Engineering Toyohashi Universiy of Technology

More information

Wheelchair-user Detection Combined with Parts-based Tracking

Wheelchair-user Detection Combined with Parts-based Tracking Wheelchair-user Deecion Combined wih Pars-based Tracking Ukyo Tanikawa 1, Yasuomo Kawanishi 1, Daisuke Deguchi 2,IchiroIde 1, Hiroshi Murase 1 and Ryo Kawai 3 1 Graduae School of Informaion Science, Nagoya

More information

Visual Perception as Bayesian Inference. David J Fleet. University of Toronto

Visual Perception as Bayesian Inference. David J Fleet. University of Toronto Visual Percepion as Bayesian Inference David J Flee Universiy of Torono Basic rules of probabiliy sum rule (for muually exclusive a ): produc rule (condiioning): independence (def n ): Bayes rule: marginalizaion:

More information

Image segmentation. Motivation. Objective. Definitions. A classification of segmentation techniques. Assumptions for thresholding

Image segmentation. Motivation. Objective. Definitions. A classification of segmentation techniques. Assumptions for thresholding Moivaion Image segmenaion Which pixels belong o he same objec in an image/video sequence? (spaial segmenaion) Which frames belong o he same video sho? (emporal segmenaion) Which frames belong o he same

More information

STEREO PLANE MATCHING TECHNIQUE

STEREO PLANE MATCHING TECHNIQUE STEREO PLANE MATCHING TECHNIQUE Commission III KEY WORDS: Sereo Maching, Surface Modeling, Projecive Transformaion, Homography ABSTRACT: This paper presens a new ype of sereo maching algorihm called Sereo

More information

A Hierarchical Object Recognition System Based on Multi-scale Principal Curvature Regions

A Hierarchical Object Recognition System Based on Multi-scale Principal Curvature Regions A Hierarchical Objec Recogniion Sysem Based on Muli-scale Principal Curvaure Regions Wei Zhang, Hongli Deng, Thomas G Dieerich and Eric N Morensen School of Elecrical Engineering and Compuer Science Oregon

More information

Robust Multi-view Face Detection Using Error Correcting Output Codes

Robust Multi-view Face Detection Using Error Correcting Output Codes Robus Muli-view Face Deecion Using Error Correcing Oupu Codes Hongming Zhang,2, Wen GaoP P, Xilin Chen 2, Shiguang Shan 2, and Debin Zhao Deparmen of Compuer Science and Engineering, Harbin Insiue of Technolog

More information

Multiple View Discriminative Appearance Modeling with IMCMC for Distributed Tracking

Multiple View Discriminative Appearance Modeling with IMCMC for Distributed Tracking Muliple View Discriminaive ing wih IMCMC for Disribued Tracking Sanhoshkumar Sunderrajan, B.S. Manjunah Deparmen of Elecrical and Compuer Engineering Universiy of California, Sana Barbara {sanhosh,manj}@ece.ucsb.edu

More information

A Matching Algorithm for Content-Based Image Retrieval

A Matching Algorithm for Content-Based Image Retrieval A Maching Algorihm for Conen-Based Image Rerieval Sue J. Cho Deparmen of Compuer Science Seoul Naional Universiy Seoul, Korea Absrac Conen-based image rerieval sysem rerieves an image from a daabase using

More information

Occlusion-Free Hand Motion Tracking by Multiple Cameras and Particle Filtering with Prediction

Occlusion-Free Hand Motion Tracking by Multiple Cameras and Particle Filtering with Prediction 58 IJCSNS Inernaional Journal of Compuer Science and Nework Securiy, VOL.6 No.10, Ocober 006 Occlusion-Free Hand Moion Tracking by Muliple Cameras and Paricle Filering wih Predicion Makoo Kao, and Gang

More information

A Face Detection Method Based on Skin Color Model

A Face Detection Method Based on Skin Color Model A Face Deecion Mehod Based on Skin Color Model Dazhi Zhang Boying Wu Jiebao Sun Qinglei Liao Deparmen of Mahemaics Harbin Insiue of Technology Harbin China 150000 Zhang_dz@163.com mahwby@hi.edu.cn sunjiebao@om.com

More information

Joint Feature Learning With Robust Local Ternary Pattern for Face Recognition

Joint Feature Learning With Robust Local Ternary Pattern for Face Recognition Join Feaure Learning Wih Robus Local Ternary Paern for Face Recogniion Yuvaraju.M 1, Shalini.S 1 Assisan Professor, Deparmen of Elecrical and Elecronics Engineering, Anna Universiy Regional Campus, Coimbaore,

More information

LAMP: 3D Layered, Adaptive-resolution and Multiperspective Panorama - a New Scene Representation

LAMP: 3D Layered, Adaptive-resolution and Multiperspective Panorama - a New Scene Representation Submission o Special Issue of CVIU on Model-based and Image-based 3D Scene Represenaion for Ineracive Visualizaion LAMP: 3D Layered, Adapive-resoluion and Muliperspecive Panorama - a New Scene Represenaion

More information

We are IntechOpen, the world s leading publisher of Open Access books Built by scientists, for scientists. International authors and editors

We are IntechOpen, the world s leading publisher of Open Access books Built by scientists, for scientists. International authors and editors We are InechOpen, he world s leading publisher of Open Access books Buil by scieniss, for scieniss 4,000 116,000 120M Open access books available Inernaional auhors and ediors Downloads Our auhors are

More information

Upper Body Tracking for Human-Machine Interaction with a Moving Camera

Upper Body Tracking for Human-Machine Interaction with a Moving Camera The 2009 IEEE/RSJ Inernaional Conference on Inelligen Robos and Sysems Ocober -5, 2009 S. Louis, USA Upper Body Tracking for Human-Machine Ineracion wih a Moving Camera Yi-Ru Chen, Cheng-Ming Huang, and

More information

Image Based Computer-Aided Manufacturing Technology

Image Based Computer-Aided Manufacturing Technology Sensors & Transducers 03 by IFSA hp://www.sensorsporal.com Image Based Compuer-Aided Manufacuring Technology Zhanqi HU Xiaoqin ZHANG Jinze LI Wei LI College of Mechanical Engineering Yanshan Universiy

More information

Multi-Target Detection and Tracking from a Single Camera in Unmanned Aerial Vehicles (UAVs)

Multi-Target Detection and Tracking from a Single Camera in Unmanned Aerial Vehicles (UAVs) 2016 IEEE/RSJ Inernaional Conference on Inelligen Robos and Sysems (IROS) Daejeon Convenion Cener Ocober 9-14, 2016, Daejeon, Korea Muli-Targe Deecion and Tracking from a Single Camera in Unmanned Aerial

More information

Robust Visual Tracking for Multiple Targets

Robust Visual Tracking for Multiple Targets Robus Visual Tracking for Muliple Targes Yizheng Cai, Nando de Freias, and James J. Lile Universiy of Briish Columbia, Vancouver, B.C., Canada, V6T 1Z4 {yizhengc, nando, lile}@cs.ubc.ca Absrac. We address

More information

FACIAL ACTION TRACKING USING PARTICLE FILTERS AND ACTIVE APPEARANCE MODELS. Soumya Hamlaoui & Franck Davoine

FACIAL ACTION TRACKING USING PARTICLE FILTERS AND ACTIVE APPEARANCE MODELS. Soumya Hamlaoui & Franck Davoine FACIAL ACTION TRACKING USING PARTICLE FILTERS AND ACTIVE APPEARANCE MODELS Soumya Hamlaoui & Franck Davoine HEUDIASYC Mixed Research Uni, CNRS / Compiègne Universiy of Technology BP 20529, 60205 Compiègne

More information

A new algorithm for small object tracking based on super-resolution technique

A new algorithm for small object tracking based on super-resolution technique A new algorihm for small objec racking based on super-resoluion echnique Yabunayya Habibi, Dwi Rana Sulisyaningrum, and Budi Seiyono Ciaion: AIP Conference Proceedings 1867, 020024 (2017); doi: 10.1063/1.4994427

More information

Visual Indoor Localization with a Floor-Plan Map

Visual Indoor Localization with a Floor-Plan Map Visual Indoor Localizaion wih a Floor-Plan Map Hang Chu Dep. of ECE Cornell Universiy Ihaca, NY 14850 hc772@cornell.edu Absrac In his repor, a indoor localizaion mehod is presened. The mehod akes firsperson

More information

Moving Object Detection Using MRF Model and Entropy based Adaptive Thresholding

Moving Object Detection Using MRF Model and Entropy based Adaptive Thresholding Moving Objec Deecion Using MRF Model and Enropy based Adapive Thresholding Badri Narayan Subudhi, Pradipa Kumar Nanda and Ashish Ghosh Machine Inelligence Uni, Indian Saisical Insiue, Kolkaa, 700108, India,

More information

Weighted Voting in 3D Random Forest Segmentation

Weighted Voting in 3D Random Forest Segmentation Weighed Voing in 3D Random Fores Segmenaion M. Yaqub,, P. Mahon 3, M. K. Javaid, C. Cooper, J. A. Noble NDORMS, Universiy of Oxford, IBME, Deparmen of Engineering Science, Universiy of Oxford, 3 MRC Epidemiology

More information

Implementing Ray Casting in Tetrahedral Meshes with Programmable Graphics Hardware (Technical Report)

Implementing Ray Casting in Tetrahedral Meshes with Programmable Graphics Hardware (Technical Report) Implemening Ray Casing in Terahedral Meshes wih Programmable Graphics Hardware (Technical Repor) Marin Kraus, Thomas Erl March 28, 2002 1 Inroducion Alhough cell-projecion, e.g., [3, 2], and resampling,

More information

Robot localization under perceptual aliasing conditions based on laser reflectivity using particle filter

Robot localization under perceptual aliasing conditions based on laser reflectivity using particle filter Robo localizaion under percepual aliasing condiions based on laser refleciviy using paricle filer DongXiang Zhang, Ryo Kurazume, Yumi Iwashia, Tsuomu Hasegawa Absrac Global localizaion, which deermines

More information

Real-time 2D Video/3D LiDAR Registration

Real-time 2D Video/3D LiDAR Registration Real-ime 2D Video/3D LiDAR Regisraion C. Bodenseiner Fraunhofer IOSB chrisoph.bodenseiner@iosb.fraunhofer.de M. Arens Fraunhofer IOSB michael.arens@iosb.fraunhofer.de Absrac Progress in LiDAR scanning

More information

Gauss-Jordan Algorithm

Gauss-Jordan Algorithm Gauss-Jordan Algorihm The Gauss-Jordan algorihm is a sep by sep procedure for solving a sysem of linear equaions which may conain any number of variables and any number of equaions. The algorihm is carried

More information

J. Vis. Commun. Image R.

J. Vis. Commun. Image R. J. Vis. Commun. Image R. 20 (2009) 9 27 Conens liss available a ScienceDirec J. Vis. Commun. Image R. journal homepage: www.elsevier.com/locae/jvci Face deecion and racking using a Boosed Adapive Paricle

More information

Video Content Description Using Fuzzy Spatio-Temporal Relations

Video Content Description Using Fuzzy Spatio-Temporal Relations Proceedings of he 4s Hawaii Inernaional Conference on Sysem Sciences - 008 Video Conen Descripion Using Fuzzy Spaio-Temporal Relaions rchana M. Rajurkar *, R.C. Joshi and Sananu Chaudhary 3 Dep of Compuer

More information

Modeling and Tracking of Dynamic Obstacles for Logistic Plants using Omnidirectional Stereo Vision

Modeling and Tracking of Dynamic Obstacles for Logistic Plants using Omnidirectional Stereo Vision Modeling and Tracing of Dynamic Obsacles for Logisic Plans using Omnidirecional Sereo Vision Andrei Vaavu, Arhur D. Cosea, and Sergiu Nedevschi, Members, IEEE Absrac In his wor we presen an obsacle deecion

More information

Simultaneous Localization and Mapping with Stereo Vision

Simultaneous Localization and Mapping with Stereo Vision Simulaneous Localizaion and Mapping wih Sereo Vision Mahew N. Dailey Compuer Science and Informaion Managemen Asian Insiue of Technology Pahumhani, Thailand Email: mdailey@ai.ac.h Manukid Parnichkun Mecharonics

More information

Evaluation and Improvement of Region-based Motion Segmentation

Evaluation and Improvement of Region-based Motion Segmentation Evaluaion and Improvemen of Region-based Moion Segmenaion Mark Ross Universiy Koblenz-Landau, Insiue of Compuaional Visualisics, Universiässraße 1, 56070 Koblenz, Germany Email: ross@uni-koblenz.de Absrac

More information

Real time 3D face and facial feature tracking

Real time 3D face and facial feature tracking J Real-Time Image Proc (2007) 2:35 44 DOI 10.1007/s11554-007-0032-2 ORIGINAL RESEARCH PAPER Real ime 3D face and facial feaure racking Fadi Dornaika Æ Javier Orozco Received: 23 November 2006 / Acceped:

More information

A Novel Approach for Monocular 3D Object Tracking in Cluttered Environment

A Novel Approach for Monocular 3D Object Tracking in Cluttered Environment Inernaional Journal of Compuaional Inelligence Research ISSN 0973-1873 Volume 13, Number 5 (2017), pp. 851-864 Research India Publicaions hp://www.ripublicaion.com A Novel Approach for Monocular 3D Objec

More information

IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART A: SYSTEMS AND HUMANS 1

IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART A: SYSTEMS AND HUMANS 1 TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART A: SYSTEMS AND HUMANS 1 Adapive Appearance Model and Condensaion Algorihm for Robus Face Tracking Yui Man Lui, Suden Member,, J. Ross Beveridge, Member,,

More information

Learning in Games via Opponent Strategy Estimation and Policy Search

Learning in Games via Opponent Strategy Estimation and Policy Search Learning in Games via Opponen Sraegy Esimaion and Policy Search Yavar Naddaf Deparmen of Compuer Science Universiy of Briish Columbia Vancouver, BC yavar@naddaf.name Nando de Freias (Supervisor) Deparmen

More information

Detection and segmentation of moving objects in highly dynamic scenes

Detection and segmentation of moving objects in highly dynamic scenes Deecion and segmenaion of moving objecs in highly dynamic scenes Aurélie Bugeau Parick Pérez INRIA, Cenre Rennes - Breagne Alanique Universié de Rennes, Campus de Beaulieu, 35 042 Rennes Cedex, France

More information

MORPHOLOGICAL SEGMENTATION OF IMAGE SEQUENCES

MORPHOLOGICAL SEGMENTATION OF IMAGE SEQUENCES MORPHOLOGICAL SEGMENTATION OF IMAGE SEQUENCES B. MARCOTEGUI and F. MEYER Ecole des Mines de Paris, Cenre de Morphologie Mahémaique, 35, rue Sain-Honoré, F 77305 Fonainebleau Cedex, France Absrac. In image

More information

TrackNet: Simultaneous Detection and Tracking of Multiple Objects

TrackNet: Simultaneous Detection and Tracking of Multiple Objects TrackNe: Simulaneous Deecion and Tracking of Muliple Objecs Chenge Li New York Universiy cl2840@nyu.edu Gregory Dobler New York Universiy greg.dobler@nyu.edu Yilin Song New York Universiy ys1297@nyu.edu

More information

Research Article Auto Coloring with Enhanced Character Registration

Research Article Auto Coloring with Enhanced Character Registration Compuer Games Technology Volume 2008, Aricle ID 35398, 7 pages doi:0.55/2008/35398 Research Aricle Auo Coloring wih Enhanced Characer Regisraion Jie Qiu, Hock Soon Seah, Feng Tian, Quan Chen, Zhongke Wu,

More information

Multi-camera multi-object voxel-based Monte Carlo 3D tracking strategies

Multi-camera multi-object voxel-based Monte Carlo 3D tracking strategies RESEARCH Open Access Muli-camera muli-objec voxel-based Mone Carlo 3D racking sraegies Crisian Canon-Ferrer *, Josep R Casas, Monse Pardàs and Enric Mone Absrac This aricle presens a new approach o he

More information

An Adaptive Spatial Depth Filter for 3D Rendering IP

An Adaptive Spatial Depth Filter for 3D Rendering IP JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE, VOL.3, NO. 4, DECEMBER, 23 175 An Adapive Spaial Deph Filer for 3D Rendering IP Chang-Hyo Yu and Lee-Sup Kim Absrac In his paper, we presen a new mehod

More information

Probabilistic Detection and Tracking of Motion Discontinuities

Probabilistic Detection and Tracking of Motion Discontinuities Probabilisic Deecion and Tracking of Moion Disconinuiies Michael J. Black David J. Flee Xerox Palo Alo Research Cener 3333 Coyoe Hill Road Palo Alo, CA 94304 fblack,fleeg@parc.xerox.com hp://www.parc.xerox.com/fblack,fleeg/

More information

Vision-Based Traffic Measurement System

Vision-Based Traffic Measurement System *J. M. Wang, **S. L. Chang, **Y. C. Chung, and **S. W. Chen Deparmen of Informaion and Compuer Educaion *Naional Taiwan Universiy **Naional Taiwan Normal Universiy Taipei, Taiwan schen@csie.nnu.edu.w Absrac

More information

Robust Segmentation and Tracking of Colored Objects in Video

Robust Segmentation and Tracking of Colored Objects in Video IEEE TRANSACTIONS ON CSVT, VOL. 4, NO. 6, 2004 Robus Segmenaion and Tracking of Colored Objecs in Video Theo Gevers, member, IEEE Absrac Segmening and racking of objecs in video is of grea imporance for

More information

Sequential Monte Carlo Tracking for Marginal Artery Segmentation on CT Angiography by Multiple Cue Fusion

Sequential Monte Carlo Tracking for Marginal Artery Segmentation on CT Angiography by Multiple Cue Fusion Sequenial Mone Carlo Tracking for Marginal Arery Segmenaion on CT Angiography by Muliple Cue Fusion Shijun Wang, Brandon Peplinski, Le Lu, Weidong Zhang, Jianfei Liu, Zhuoshi Wei, and Ronald M. Summers

More information

A Bayesian Approach to Video Object Segmentation via Merging 3D Watershed Volumes

A Bayesian Approach to Video Object Segmentation via Merging 3D Watershed Volumes A Bayesian Approach o Video Objec Segmenaion via Merging 3D Waershed Volumes Yu-Pao Tsai 1,3, Chih-Chuan Lai 1,2, Yi-Ping Hung 1,2, and Zen-Chung Shih 3 1 Insiue of Informaion Science, Academia Sinica,

More information

Detection Tracking and Recognition of Human Poses for a Real Time Spatial Game

Detection Tracking and Recognition of Human Poses for a Real Time Spatial Game Deecion Tracking and Recogniion of Human Poses for a Real Time Spaial Game Feifei Huo, Emile A. Hendriks, A.H.J. Oomes Delf Universiy of Technology The Neherlands f.huo@udelf.nl Pascal van Beek, Remco

More information

IntentSearch:Capturing User Intention for One-Click Internet Image Search

IntentSearch:Capturing User Intention for One-Click Internet Image Search JOURNAL OF L A T E X CLASS FILES, VOL. 6, NO. 1, JANUARY 2010 1 InenSearch:Capuring User Inenion for One-Click Inerne Image Search Xiaoou Tang, Fellow, IEEE, Ke Liu, Jingyu Cui, Suden Member, IEEE, Fang

More information

An Iterative Scheme for Motion-Based Scene Segmentation

An Iterative Scheme for Motion-Based Scene Segmentation An Ieraive Scheme for Moion-Based Scene Segmenaion Alexander Bachmann and Hildegard Kuehne Deparmen for Measuremen and Conrol Insiue for Anhropomaics Universiy of Karlsruhe (H), 76 131 Karlsruhe, Germany

More information

4.1 3D GEOMETRIC TRANSFORMATIONS

4.1 3D GEOMETRIC TRANSFORMATIONS MODULE IV MCA - 3 COMPUTER GRAPHICS ADMN 29- Dep. of Compuer Science And Applicaions, SJCET, Palai 94 4. 3D GEOMETRIC TRANSFORMATIONS Mehods for geomeric ransformaions and objec modeling in hree dimensions

More information

DAGM 2011 Tutorial on Convex Optimization for Computer Vision

DAGM 2011 Tutorial on Convex Optimization for Computer Vision DAGM 2011 Tuorial on Convex Opimizaion for Compuer Vision Par 3: Convex Soluions for Sereo and Opical Flow Daniel Cremers Compuer Vision Group Technical Universiy of Munich Graz Universiy of Technology

More information

Analysis of Various Types of Bugs in the Object Oriented Java Script Language Coding

Analysis of Various Types of Bugs in the Object Oriented Java Script Language Coding Indian Journal of Science and Technology, Vol 8(21), DOI: 10.17485/ijs/2015/v8i21/69958, Sepember 2015 ISSN (Prin) : 0974-6846 ISSN (Online) : 0974-5645 Analysis of Various Types of Bugs in he Objec Oriened

More information

Rao-Blackwellized Particle Filtering for Probing-Based 6-DOF Localization in Robotic Assembly

Rao-Blackwellized Particle Filtering for Probing-Based 6-DOF Localization in Robotic Assembly MITSUBISHI ELECTRIC RESEARCH LABORATORIES hp://www.merl.com Rao-Blackwellized Paricle Filering for Probing-Based 6-DOF Localizaion in Roboic Assembly Yuichi Taguchi, Tim Marks, Haruhisa Okuda TR1-8 June

More information

Detection of salient objects with focused attention based on spatial and temporal coherence

Detection of salient objects with focused attention based on spatial and temporal coherence ricle Informaion Processing Technology pril 2011 Vol.56 No.10: 1055 1062 doi: 10.1007/s11434-010-4387-1 SPECIL TOPICS: Deecion of salien objecs wih focused aenion based on spaial and emporal coherence

More information

A High-Speed Adaptive Multi-Module Structured Light Scanner

A High-Speed Adaptive Multi-Module Structured Light Scanner A High-Speed Adapive Muli-Module Srucured Ligh Scanner Andreas Griesser 1 Luc Van Gool 1,2 1 Swiss Fed.Ins.of Techn.(ETH) 2 Kaholieke Univ. Leuven D-ITET/Compuer Vision Lab ESAT/VISICS Zürich, Swizerland

More information

Robust LSTM-Autoencoders for Face De-Occlusion in the Wild

Robust LSTM-Autoencoders for Face De-Occlusion in the Wild IEEE TRANSACTIONS ON IMAGE PROCESSING, DRAFT 1 Robus LSTM-Auoencoders for Face De-Occlusion in he Wild Fang Zhao, Jiashi Feng, Jian Zhao, Wenhan Yang, Shuicheng Yan arxiv:1612.08534v1 [cs.cv] 27 Dec 2016

More information

Improving Occupancy Grid FastSLAM by Integrating Navigation Sensors

Improving Occupancy Grid FastSLAM by Integrating Navigation Sensors Improving Occupancy Grid FasSLAM by Inegraing Navigaion Sensors Chrisopher Weyers Sensors Direcorae Air Force Research Laboraory Wrigh-Paerson AFB, OH 45433 Gilber Peerson Deparmen of Elecrical and Compuer

More information

Real Time Integral-Based Structural Health Monitoring

Real Time Integral-Based Structural Health Monitoring Real Time Inegral-Based Srucural Healh Monioring The nd Inernaional Conference on Sensing Technology ICST 7 J. G. Chase, I. Singh-Leve, C. E. Hann, X. Chen Deparmen of Mechanical Engineering, Universiy

More information

Real-Time Non-Rigid Multi-Frame Depth Video Super-Resolution

Real-Time Non-Rigid Multi-Frame Depth Video Super-Resolution Real-Time Non-Rigid Muli-Frame Deph Video Super-Resoluion Kassem Al Ismaeil 1, Djamila Aouada 1, Thomas Solignac 2, Bruno Mirbach 2, Björn Oersen 1 1 Inerdisciplinary Cenre for Securiy, Reliabiliy, and

More information

In Proceedings of CVPR '96. Structure and Motion of Curved 3D Objects from. using these methods [12].

In Proceedings of CVPR '96. Structure and Motion of Curved 3D Objects from. using these methods [12]. In Proceedings of CVPR '96 Srucure and Moion of Curved 3D Objecs from Monocular Silhouees B Vijayakumar David J Kriegman Dep of Elecrical Engineering Yale Universiy New Haven, CT 652-8267 Jean Ponce Compuer

More information

Large-scale 3D Outdoor Mapping and On-line Localization using 3D-2D Matching

Large-scale 3D Outdoor Mapping and On-line Localization using 3D-2D Matching Large-scale 3D Oudoor Mapping and On-line Localizaion using 3D-D Maching Takahiro Sakai, Kenji Koide, Jun Miura, and Shuji Oishi Absrac Map-based oudoor navigaion is an acive research area in mobile robos

More information

MATH Differential Equations September 15, 2008 Project 1, Fall 2008 Due: September 24, 2008

MATH Differential Equations September 15, 2008 Project 1, Fall 2008 Due: September 24, 2008 MATH 5 - Differenial Equaions Sepember 15, 8 Projec 1, Fall 8 Due: Sepember 4, 8 Lab 1.3 - Logisics Populaion Models wih Harvesing For his projec we consider lab 1.3 of Differenial Equaions pages 146 o

More information

MOTION TRACKING is a fundamental capability that

MOTION TRACKING is a fundamental capability that TECHNICAL REPORT CRES-05-008, CENTER FOR ROBOTICS AND EMBEDDED SYSTEMS, UNIVERSITY OF SOUTHERN CALIFORNIA 1 Real-ime Moion Tracking from a Mobile Robo Boyoon Jung, Suden Member, IEEE, Gaurav S. Sukhame,

More information

Graffiti Detection Using Two Views

Graffiti Detection Using Two Views Graffii Deecion Using wo Views Luigi Di Sefano Federico ombari Alessandro Lanza luigi.disefano@unibo.i federico.ombari@unibo.i alanza@arces.unibo.i Sefano Maoccia sefano.maoccia@unibo.i Sefano Moni sefano.moni3@sudio.unibo.i

More information

Video-Based Face Recognition Using Probabilistic Appearance Manifolds

Video-Based Face Recognition Using Probabilistic Appearance Manifolds Video-Based Face Recogniion Using Probabilisic Appearance Manifolds Kuang-Chih Lee Jeffrey Ho Ming-Hsuan Yang David Kriegman klee10@uiuc.edu jho@cs.ucsd.edu myang@honda-ri.com kriegman@cs.ucsd.edu Compuer

More information

Gender Classification of Faces Using Adaboost*

Gender Classification of Faces Using Adaboost* Gender Classificaion of Faces Using Adaboos* Rodrigo Verschae 1,2,3, Javier Ruiz-del-Solar 1,2, and Mauricio Correa 1,2 1 Deparmen of Elecrical Engineering, Universidad de Chile 2 Cener for Web Research,

More information

Track and Cut: simultaneous tracking and segmentation of multiple objects with graph cuts

Track and Cut: simultaneous tracking and segmentation of multiple objects with graph cuts INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE Track and Cu: simulaneous racking and segmenaion of muliple objecs wih graph cus Aurélie Bugeau Parick Pérez N 6337 Ocober 2007 Thèmes COM

More information

A METHOD OF MODELING DEFORMATION OF AN OBJECT EMPLOYING SURROUNDING VIDEO CAMERAS

A METHOD OF MODELING DEFORMATION OF AN OBJECT EMPLOYING SURROUNDING VIDEO CAMERAS A METHOD OF MODELING DEFORMATION OF AN OBJECT EMLOYING SURROUNDING IDEO CAMERAS Joo Kooi TAN, Seiji ISHIKAWA Deparmen of Mechanical and Conrol Engineering Kushu Insiue of Technolog, Japan ehelan@is.cnl.kuech.ac.jp,

More information

Reinforcement Learning by Policy Improvement. Making Use of Experiences of The Other Tasks. Hajime Kimura and Shigenobu Kobayashi

Reinforcement Learning by Policy Improvement. Making Use of Experiences of The Other Tasks. Hajime Kimura and Shigenobu Kobayashi Reinforcemen Learning by Policy Improvemen Making Use of Experiences of The Oher Tasks Hajime Kimura and Shigenobu Kobayashi Tokyo Insiue of Technology, JAPAN genfe.dis.iech.ac.jp, kobayasidis.iech.ac.jp

More information

Dynamic Route Planning and Obstacle Avoidance Model for Unmanned Aerial Vehicles

Dynamic Route Planning and Obstacle Avoidance Model for Unmanned Aerial Vehicles Volume 116 No. 24 2017, 315-329 ISSN: 1311-8080 (prined version); ISSN: 1314-3395 (on-line version) url: hp://www.ijpam.eu ijpam.eu Dynamic Roue Planning and Obsacle Avoidance Model for Unmanned Aerial

More information

Track-based and object-based occlusion for people tracking refinement in indoor surveillance

Track-based and object-based occlusion for people tracking refinement in indoor surveillance Trac-based and objec-based occlusion for people racing refinemen in indoor surveillance R. Cucchiara, C. Grana, G. Tardini Diparimeno di Ingegneria Informaica - Universiy of Modena and Reggio Emilia Via

More information

LOW-VELOCITY IMPACT LOCALIZATION OF THE COMPOSITE TUBE USING A NORMALIZED CROSS-CORRELATION METHOD

LOW-VELOCITY IMPACT LOCALIZATION OF THE COMPOSITE TUBE USING A NORMALIZED CROSS-CORRELATION METHOD 21 s Inernaional Conference on Composie Maerials Xi an, 20-25 h Augus 2017 LOW-VELOCITY IMPACT LOCALIZATION OF THE COMPOSITE TUBE USING A NORMALIZED CROSS-CORRELATION METHOD Hyunseok Kwon 1, Yurim Park

More information

Viewpoint Invariant 3D Landmark Model Inference from Monocular 2D Images Using Higher-Order Priors

Viewpoint Invariant 3D Landmark Model Inference from Monocular 2D Images Using Higher-Order Priors Viewpoin Invarian 3D Landmark Model Inference from Monocular 2D Images Using Higher-Order Priors Chaohui Wang 1,2, Yun Zeng 3, Loic Simon 1, Ioannis Kakadiaris 4, Dimiris Samaras 3, Nikos Paragios 1,2

More information

Tracking a Large Number of Objects from Multiple Views

Tracking a Large Number of Objects from Multiple Views Tracking a Large Number of Objecs from Muliple Views Zheng Wu 1, Nickolay I. Hrisov 2, Tyson L. Hedrick 3, Thomas H. Kun 2, Margri Beke 1 1 Deparmen of Compuer Science, Boson Universiy 2 Deparmen of Biology,

More information

Robust 3D Visual Tracking Using Particle Filtering on the SE(3) Group

Robust 3D Visual Tracking Using Particle Filtering on the SE(3) Group Robus 3D Visual Tracking Using Paricle Filering on he SE(3) Group Changhyun Choi and Henrik I. Chrisensen Roboics & Inelligen Machines, College of Compuing Georgia Insiue of Technology Alana, GA 3332,

More information

Tracking a Large Number of Objects from Multiple Views

Tracking a Large Number of Objects from Multiple Views Boson Universiy Compuer Science Deparmen Technical Repor BUCS-TR 2009-005 Tracking a Large Number of Objecs from Muliple Views Zheng Wu 1, Nickolay I. Hrisov 2, Tyson L. Hedrick 3, Thomas H. Kun 2, Margri

More information

Landmarks: A New Model for Similarity-Based Pattern Querying in Time Series Databases

Landmarks: A New Model for Similarity-Based Pattern Querying in Time Series Databases Lmarks: A New Model for Similariy-Based Paern Querying in Time Series Daabases Chang-Shing Perng Haixun Wang Sylvia R. Zhang D. So Parker perng@cs.ucla.edu hxwang@cs.ucla.edu Sylvia Zhang@cle.com so@cs.ucla.edu

More information

Scale Recovery for Monocular Visual Odometry Using Depth Estimated with Deep Convolutional Neural Fields

Scale Recovery for Monocular Visual Odometry Using Depth Estimated with Deep Convolutional Neural Fields Scale Recovery for Monocular Visual Odomery Using Deph Esimaed wih Deep Convoluional Neural Fields Xiaochuan Yin, Xiangwei Wang, Xiaoguo Du, Qijun Chen Tongji Universiy yinxiaochuan@homail.com,wangxiangwei.cpp@gmail.com,

More information

Hidden Markov Model and Chapman Kolmogrov for Protein Structures Prediction from Images

Hidden Markov Model and Chapman Kolmogrov for Protein Structures Prediction from Images Hidden Markov Model and Chapman Kolmogrov for Proein Srucures Predicion from Images Md.Sarwar Kamal 1, Linkon Chowdhury 2, Mohammad Ibrahim Khan 2, Amira S. Ashour 3, João Manuel R.S. Tavares 4, Nilanjan

More information

Design Alternatives for a Thin Lens Spatial Integrator Array

Design Alternatives for a Thin Lens Spatial Integrator Array Egyp. J. Solids, Vol. (7), No. (), (004) 75 Design Alernaives for a Thin Lens Spaial Inegraor Array Hala Kamal *, Daniel V azquez and Javier Alda and E. Bernabeu Opics Deparmen. Universiy Compluense of

More information

Image Content Representation

Image Content Representation Image Conen Represenaion Represenaion for curves and shapes regions relaionships beween regions E.G.M. Perakis Image Represenaion & Recogniion 1 Reliable Represenaion Uniqueness: mus uniquely specify an

More information

RGB-D Object Tracking: A Particle Filter Approach on GPU

RGB-D Object Tracking: A Particle Filter Approach on GPU RGB-D Objec Tracking: A Paricle Filer Approach on GPU Changhyun Choi and Henrik I. Chrisensen Cener for Roboics & Inelligen Machines College of Compuing Georgia Insiue of Technology Alana, GA 3332, USA

More information

Fusion of Multiple Cues from Color and Depth Domains using Occlusion Aware Bayesian Tracker

Fusion of Multiple Cues from Color and Depth Domains using Occlusion Aware Bayesian Tracker Fusion of Muliple Cues from Color and Deph Domains using Occlusion Aware Bayesian Tracker Kourosh MESHGI Shin-ichi MAEDA Shigeyuki OBA and Shin ISHII Graduae School of Informaics, Kyoo Universiy, Gokasho,

More information

IROS 2015 Workshop on On-line decision-making in multi-robot coordination (DEMUR 15)

IROS 2015 Workshop on On-line decision-making in multi-robot coordination (DEMUR 15) IROS 2015 Workshop on On-line decision-making in muli-robo coordinaion () OPTIMIZATION-BASED COOPERATIVE MULTI-ROBOT TARGET TRACKING WITH REASONING ABOUT OCCLUSIONS KAROL HAUSMAN a,, GREGORY KAHN b, SACHIN

More information

COSC 3213: Computer Networks I Chapter 6 Handout # 7

COSC 3213: Computer Networks I Chapter 6 Handout # 7 COSC 3213: Compuer Neworks I Chaper 6 Handou # 7 Insrucor: Dr. Marvin Mandelbaum Deparmen of Compuer Science York Universiy F05 Secion A Medium Access Conrol (MAC) Topics: 1. Muliple Access Communicaions:

More information

AUTOMATIC 3D FACE REGISTRATION WITHOUT INITIALIZATION

AUTOMATIC 3D FACE REGISTRATION WITHOUT INITIALIZATION Chaper 3 AUTOMATIC 3D FACE REGISTRATION WITHOUT INITIALIZATION A. Koschan, V. R. Ayyagari, F. Boughorbel, and M. A. Abidi Imaging, Roboics, and Inelligen Sysems Laboraory, The Universiy of Tennessee, 334

More information

A Framework for Applying Point Clouds Grabbed by Multi-Beam LIDAR in Perceiving the Driving Environment

A Framework for Applying Point Clouds Grabbed by Multi-Beam LIDAR in Perceiving the Driving Environment Sensors 215, 15, 21931-21956; doi:1.339/s15921931 Aricle OPEN ACCESS sensors ISSN 1424-822 www.mdpi.com/journal/sensors A Framewor for Applying Poin Clouds Grabbed by Muli-Beam LIDAR in Perceiving he Driving

More information

Deformable Parts Correlation Filters for Robust Visual Tracking

Deformable Parts Correlation Filters for Robust Visual Tracking PAPER UNDER REVISION Deformable Pars Correlaion Filers for Robus Visual Tracking Alan Lukežič, Luka Čehovin, Member, IEEE, and Maej Krisan, Member, IEEE ha par-based models should be considered in a layered

More information

BEHAVIOR recognition for human subjects moving in an

BEHAVIOR recognition for human subjects moving in an 294 IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE, VOL. 1, NO. 4, AUGUST 2017 Behavior Recogniion Using Muliple Deph Cameras Based on a Time-Varian Skeleon Vecor Projecion Chien-Hao

More information

STRING DESCRIPTIONS OF DATA FOR DISPLAY*

STRING DESCRIPTIONS OF DATA FOR DISPLAY* SLAC-PUB-383 January 1968 STRING DESCRIPTIONS OF DATA FOR DISPLAY* J. E. George and W. F. Miller Compuer Science Deparmen and Sanford Linear Acceleraor Cener Sanford Universiy Sanford, California Absrac

More information

Object Trajectory Proposal via Hierarchical Volume Grouping

Object Trajectory Proposal via Hierarchical Volume Grouping Objec Trajecory Proposal via Hierarchical Volume Grouping Xu Sun 1, Yuanian Wang 1, Tongwei Ren 1,, Zhi Liu 2, Zheng-Jun Zha 3, and Gangshan Wu 1 1 Sae Key Laboraory for Novel Sofware Technology, Nanjing

More information

Improving Ranking of Search Engines Results Based on Power Links

Improving Ranking of Search Engines Results Based on Power Links IPASJ Inernaional Journal of Informaion Technology (IIJIT) Web Sie: hp://www.ipasj.org/iijit/iijit.hm A Publisher for Research Moivaion... Email: edioriiji@ipasj.org Volume 2, Issue 9, Sepember 2014 ISSN

More information

Tracking Appearances with Occlusions

Tracking Appearances with Occlusions Tracking ppearances wih Occlusions Ying Wu, Ting Yu, Gang Hua Deparmen of Elecrical & Compuer Engineering Norhwesern Universiy 2145 Sheridan oad, Evanson, IL 60208 {yingwu,ingyu,ganghua}@ece.nwu.edu bsrac

More information

CENG 477 Introduction to Computer Graphics. Modeling Transformations

CENG 477 Introduction to Computer Graphics. Modeling Transformations CENG 477 Inroducion o Compuer Graphics Modeling Transformaions Modeling Transformaions Model coordinaes o World coordinaes: Model coordinaes: All shapes wih heir local coordinaes and sies. world World

More information

CONTEXT MODELS FOR CRF-BASED CLASSIFICATION OF MULTITEMPORAL REMOTE SENSING DATA

CONTEXT MODELS FOR CRF-BASED CLASSIFICATION OF MULTITEMPORAL REMOTE SENSING DATA ISPRS Annals of he Phoogrammery, Remoe Sensing and Spaial Informaion Sciences, Volume I-7, 2012 XXII ISPRS Congress, 25 Augus 01 Sepember 2012, Melbourne, Ausralia CONTEXT MODELS FOR CRF-BASED CLASSIFICATION

More information

Shortest Path Algorithms. Lecture I: Shortest Path Algorithms. Example. Graphs and Matrices. Setting: Dr Kieran T. Herley.

Shortest Path Algorithms. Lecture I: Shortest Path Algorithms. Example. Graphs and Matrices. Setting: Dr Kieran T. Herley. Shores Pah Algorihms Background Seing: Lecure I: Shores Pah Algorihms Dr Kieran T. Herle Deparmen of Compuer Science Universi College Cork Ocober 201 direced graph, real edge weighs Le he lengh of a pah

More information

EECS 487: Interactive Computer Graphics

EECS 487: Interactive Computer Graphics EECS 487: Ineracive Compuer Graphics Lecure 7: B-splines curves Raional Bézier and NURBS Cubic Splines A represenaion of cubic spline consiss of: four conrol poins (why four?) hese are compleely user specified

More information