Similarity-based denoising of point-sampled surfaces *

Size: px
Start display at page:

Download "Similarity-based denoising of point-sampled surfaces *"

Transcription

1 Wang et al. / J Zhejang Unv Sc A 008 9(6): Journal of Zhejang Unversty SCIENCE A ISSN X (Prnt); ISSN (Onlne) E-mal: jzus@zju.edu.cn Smlarty-based denosng of pont-sampled surfaces * Ren-fang WANG,, Wen-zh CHEN, San-yuan ZHANG, Yn ZHANG, Xu-z YE ( School of Computer Scence and Technology, Zhejang Unversty, Hangzhou 3007, Chna) ( Faculty of Computer Scence and Informaton Technology, Zhejang Wanl Unversty, Nngbo 3500, Chna) E-mal: wrfwpln@6.com; syzhang@cs.zju.edu.cn Receved Sept., 007; revson accepted Mar. 8, 008 Abstract: A non-local denosng (NLD) algorthm for pont-sampled surfaces (PSSs) s presented based on smlartes, ncludng geometry ntensty and features of sample ponts. By usng the trlateral flterng operator, the dfferental sgnal of each sample pont s determned and called geometry ntensty. Based on covarance analyss, a regular grd of geometry ntensty of a sample pont s constructed, and the geometry-ntensty smlarty of two ponts s measured accordng to ther grds. Based on mean shft clusterng, the PSSs are clustered n terms of the local geometry-features smlarty. The smoothed geometry ntensty,.e., offset dstance, of the sample pont s estmated accordng to the two smlartes. Usng the resultng ntensty, the nose component from PSSs s fnally removed by adjustng the poston of each sample pont along ts own normal drecton. Expermental results demonstrate that the algorthm s robust and can produce a more accurate denosng result whle havng better feature preservaton. Key words: Pont-sampled surfaces (PSSs), Smlarty, Geometry ntensty, Geometry feature, Non-local flterng do:0.63/jzus.a07465 Document code: A CLC number: TP39.7 INTRODUCTION Pont-sampled models wthout topologcal connectvty are normally generated by samplng the boundary surface of physcal 3D objects wth 3D-scannng devces. Despte the steady mprovement n scannng accuracy, undesrable nose s nevtably ntroduced from varous sources such as local measurements and algorthmc errors. Thus, nosy models need to be denosed or smoothed before performng any subsequent geometry processng such as smplfcaton, reconstructon and parameterzaton. It remans a challenge to remove the nevtable nose whle preservng the underlyng surface features n computer graphcs. In partcular, fne features are often lost f no specal treatment s provded. Correspondng author * Project supported by the H-Tech Research and Development Program (863) of Chna (Nos. 007AA0Z3 and 007AA04ZA5), and the Research Fund for the Doctoral Program of Hgher Educaton of Chna (No ) In recent years, a varety of pont-based denosng and smoothng approaches have been ntroduced, and they can be roughly categorzed nto the followng four groups: () Spectral technques. The technques of spectral flters n mage settng were generalzed to pont-based surfaces, for example, Pauly and Gross (00) created a spectral decomposton of a pont cloud and denosed t by manpulaton of the spectral coeffcents. () Interpolaton or approxmaton approaches. The raw pont clouds are nterpolated or approxmated wth smooth surfaces such as extremal surfaces (Amenta and Kl, 004), mplct surfaces (Carr et al., 00; Samozno et al., 006; Danels II et al., 007), movng least-squares (MLS) surfaces (Mederos et al., 003; Weyrch et al., 004; Dey and Sun, 005; Lpman et al., 006; Danels II et al., 007) and local parametrc surfaces (Pauly et al., 003), etc. (3) Statstcal technques. Based on robust statstcs, the smoothng technques for pont-sampled surfaces (PSSs) were ntroduced. Pauly et al.(004)

2 808 Wang et al. / J Zhejang Unv Sc A 008 9(6): proposed a framework for analyzng shape uncertanty and varablty n pont-sampled geometry based on statstcal data analyss, whch can be appled to reconstruct surfaces n the presence of nose. Schall et al.(005) developed a method for robust flterng of a gven nosy pont set usng a mean shft based clusterng procedure. Postons on a smooth surface were found by movng every sample to maxmum lkelhood postons. Jenke et al.(006) showed how to generate a smooth pont cloud from a gven nosy one usng Bayesan statstcs. (4) Extenson of D flters to 3D ones. The D flters were drectly extended nto 3D settngs to smooth PSSs by applyng local poston estmatng teratvely or non-teratvely, sotropcally or ansotropcally, based on statstcs, dfferental geometry theory, approxmaton theory, etc. The last group s more attractve snce t s smple and straghtforward. Technques for mage smoothng such as Laplacan, blateral and trlateral flterng, moreover, commonly act as foundatons for 3D surface denosng algorthms. In ths paper, we ntroduce a non-local denosng (NLD) method for PSSs nspred by a non-local algorthm for mage denosng (Buades et al., 005) whch presents remarkable results. Unlke the non-local mage flter, our flter computes the denosed poston of a vertex as a weghted average of the vertces n ts vcnty wth smlar geometry features whch are determned by mean shft clusterng. The non-local approach defnes the ntensty smlarty of two ponts by comparng regons of the surface around the vertces rather than usng only ther postons and sometmes normals locally. Ths yelds a more accurate denosng result of the surface and mproves the removal of hgher-level noses compared to prevous state-of-the- art flterng technques. Moreover, fne geometry features are better preserved. An example of the effectveness of our approach s presented n Fg.. In ths paper all the pont models are rendered by usng a pont-based renderng technque. The rest of ths paper s organzed as follows. In Secton, we brefly revew the related works. In Secton 3, we brefly overvew the non-local means approach of Buades, Coll, and Morel. We present our smlarty-based method for denosng PSSs n Secton 4. We compare our method wth two denosng technques n Secton 5. Secton 6 concludes the paper. (a) (c) Fg. Denosng Buste model wth our method. (a) Nosy model; (b) Model denosed by our method; (c) Nosy model colored by mean curvature; (d) Denosed model colored by mean curvature. Notce that hgh-frequency nose s properly removed, whle fne detals n har regon are accurately preserved RELATED WORKS Earler methods such as Laplacan (Pauly et al., 00b) for denosng PSSs are sotropc, whch result commonly n pont drftng and oversmoothng. So the ansotropc methods were ntroduced. Clarenz et al.(004) presented a PDE-based surface farng applcaton wthn the framework of processng pontbased surface va PDEs. Lange and Polther (005) proposed a new method for ansotropc farng of a pont-sampled surface based on the concept of ansotropc geometrc mean curvature flow. Based on dynamc balanced flow, Xao et al.(006) presented a novel approach for farng pont-sampled geometry. Other methods have also been proposed for denosng the PSSs. Algorthms that recently attracted the nterest of many researchers are MLS approaches. Based on MLS or ts varants, the algorthms for denosng PSSs have also been ntroduced (Mederos et al., 003; Weyrch et al., 004; Dey and Sun, 005; (b) (d) Max Mn

3 Wang et al. / J Zhejang Unv Sc A 008 9(6): Lpman et al., 006; Danels II et al., 007). The man problem of MLS-based methods s that promnent shape features are blurred whle smoothng PSSs. All of the above methods dd not take nto consderaton the smlarty of geometry features of sample ponts. Hu et al.(006) ntroduced a mean shft based denosng algorthm for PSSs. Durng the course of denosng PSSs, they respected the knd of smlarty va a 3D mean shft procedure. However, all the above-mentoned methods determne the ntensty smlarty of two ponts locally usng only ther postons and sometmes normals. Concernng these ssues, ths paper defnes the ntensty smlarty n a non-local fashon and takes nto account the smlarty of geometry features whle denosng PSSs. We ntroduce the followng technques to acheve the non-local denosng algorthm: () By usng the trlateral flterng operator, the geometry ntensty of each sample pont s determned; () Based on covarance analyss, a regular grd of geometry ntensty s constructed for each pont and the geometry-ntensty smlarty between two ponts s measured; (3) Based on mean shft clusterng, the PSSs are clustered accordng to the surface-features smlarty. NON-LOCAL IMAGE DENOISING The non-local algorthm for mage denosng was proposed by Buades et al.(005). The basc dea behnd the non-local method s very smple: for a gven pxel, ts denosed ntensty value s estmated as a weghted average of the other mage pxels wth weghts reflectng the smlarty between local neghborhoods of the pxel beng processed and the other pxels. More precsely, f an mage Ω={I(u) u P} s gven, where u=(x, y) s a pxel and I(u) s the ntensty value at u, the smoothed pxel ntensty I'(u) can be computed as the average of all pxel ntenstes n the mage weghted by a smlarty factor Φ(u,v), Φ( uv, ) I( v) P I ( u)=, Φ( uv, ) v v P () where Φ(u,v)=exp( S u,v /h ). The parameter h acts as a degree of flterng, and the smlarty S u,v between u and v s measured as Suv, = Ga ( o ) I( u+ o) I( v+ o ), () o whch depends on the pxel-wse ntensty dfference of two square neghborhoods centered at the pxels u and v. The vector o denotes the offset between the center pxel and an arbtrary neghborhood pxel. The nfluence of a pxel par on the smlarty falls wth ncreasng Eucldean dstance to the center of the neghborhoods. For the dstance weghtng a Gaussan kernel G a ( ) wth a standard devaton a>0 s used. NON-LOCAL DENOISING ALGORITHM FOR PSSs For the reason that mage pxels are usually algned on a regular and equspaced grd, whch s n general not true for a pont-sampled model, the man dffculty of extendng the non-local method to PSSs conssts of how to determne the ntensty smlarty of two ponts. Our strategy for solvng ths ssue s as follows. By usng the trlateral flterng operator, the dfferental value of each sample pont s determned and called geometry ntensty as a counterpart to the ntensty value of an mage. Based on covarance analyss, the local reference plane s defned on whch a regular grd of geometry ntensty s then constructed for each pont, and accordng to geometry ntenstes on the correspondng grds, the geometryntensty smlarty of two ponts s fnally measured. Computaton of geometry ntensty Although the trlateral flter presented by Choudhury and Tumbln (003) for hgh contrast mages and meshes can be extended to PSSs, t does not consder the curvature-related nformaton. We desgn the followng trlateral flter wth the curvature-related term to compute the geometry ntensty δ of each pont p δ = w <, >, ( ) j n qj p w qj N p qj N( p ) j wj = wc( qj p ) ws( < n, qj p > ) wh( Hj H ), (3) where n s the surface normal at the pont p, N(p ) s the neghborhood of p, H s the mean curvature and w(x) s a Gaussan kernel: w ( x)=exp[ x /( σ )], c c

4 80 Wang et al. / J Zhejang Unv Sc A 008 9(6): w ( x)=exp[ x /( σ )] and w h (x)=exp[ exp( H j s s H /)]. The term w h (x) denotes that the nfluence on the weght w j ncreases wth an ncrease n the curvature dfference (H j H ) so that the hgh gradent regons can be effcently smoothed. In ths paper we take the parameter σ c as σ c =r/, where r s the radus of the enclosng sphere of N(p ), and σ s as the standard devaton of the projectons of the vector (q j p ) onto n. Because our trlateral flterng operator consders not only the pont postons and normals but curvatures, the geometry ntensty computed by t can reflect the local geometry feature and descrbe the dfferental property at each pont more effcently. Fg.b demonstrates the geometry-ntensty value vsualzaton of the ponts for the Face model. Fg.e llustrates the palette for vsualzaton of mean curvature, geometry ntensty and the dstance between two ponts. (a) (b) Measurng the geometry-ntensty smlarty of two ponts By covarance analyss, a local frame can be constructed based on the tangent plane and normal for each sample pont. The covarance matrx for the pont set P s defned as p p p p p p p p C =, p p k p p k where p k s the kth neghborng pont around p and p= p / k s the centrod of the neghborhood. k Snce C s symmetrc and postve sem-defnte, all egenvalues λ (=0,, ) are of real value and the egenvectors v (=0,, ) form an orthogonal bass. The egenvalue λ measures the varaton of the local pont set along the drecton of the correspondng egenvector. Assumng λ 0 λ λ, let the plane (x p ) v 0 =0 through p mnmze the sum of the squared dstance to the neghborng ponts of p, then the normal v 0 of ths plane can be regarded as the normal of the local surface at p. In ths partcular case, λ 0 expresses the varaton of the surface along the normal v 0. Pauly et al.(00a) defned T σ = σ ( p ) = λ ( λ + λ + λ ) (4) k k 0 0 (c) Mn (e) Fg. (a) Nosy Face model; (b) Geometry-ntensty value vsualzaton of the ponts; (c) Our mean-shft clusterng; (d) The pont set of local modes; (e) Palette for vsualzaton of geometry ntensty (d) Max as the surface varaton at p assumng a neghborhood of sze k. It s also observed that σ k (p ) s closely related to the local curvature. To ensure a consstent orentaton of the normal vectors, we use a method based on the mnmum spannng tree (Hoppe et al., 99). For each pont, we construct the local frame (v, v, v 0 ) wth the local orgn p and on the plane v p v, a regular geometry-ntensty grd G wth the sze of M M centered at p can be bult by an nterpolaton method. Let N R (p )={q j q j P n, q j p M E } be the set of the neghbors of p whose elements are wthn a fxed radus R bound centered at p, where E = n r mn n s the average edge length of the sample =0

5 Wang et al. / J Zhejang Unv Sc A 008 9(6): pont set P n, n= P n and r mn s the dstance between p and ts nearest pont. Assumng that the vector v g s the projecton of the vector (q j p ) onto the plane v p v and δ g s the geometry ntensty of the pont q j, the geometry ntensty δ m of each node g m (0 m M M ) among G s estmated as δ m = vg N ( gm ) vg N ( gm ) wg( vg) δg, w ( v ) g g s an adaptve nonparametrc estmator of the densty at locaton x n the feature space. The functon k(x) (0 x ) s called the profle of the kernel, and the normalzaton constant c k,d assures that K(x) ntegrates to one. The functon g(x)= k'(x) can always be defned when the dervatve of the kernel profle k(x) exsts. Usng g(x) as the profle, the kernel G(x) s defned as G(x)=c g,d g( x ). By takng the gradent of Eq.(5) the followng property can be proven where N(g m ) s the neghborhood of g m and w g (x) s a Gaussan kernel: w g (v g )=exp[ v g g m /(σ g )]. In ths paper, we take agan the parameter σ g as σ g =r/, where r s the radus of the enclosng crcle of N(g m ). When the regular geometry-ntensty grd G of p s regarded as a counterpart to the square neghborhood of the pxel u, we can, as a result, measure the geometry-ntensty smlarty of two ponts by means of Eq.(). m ( x)= C ˆ f ( x) fˆ ( x ), G K G where C s a postve constant and = d + ( ( x x )/ h ) ( ( x x)/ h ) n x g d + = h mg ( x) = x n g h (6) Mean shft clusterng for PSSs Unlke the non-local mage flterng algorthm, we do not sum over all pont postons to flter a pont but over a local neghborhood of ths pont whch s determned by a mean shft clusterng method. The mean shft algorthm s a nonparametrc clusterng technque for the analyss of a complex multmodal feature space and the delneaton of arbtrarly shaped clusters (Comancu and Meer, 00), and t has a wde varety of applcatons n the felds of computer vson and pattern recognton. Recently t has been extended to the feld of dgtal geometry processng (Yamauch et al., 005; Hu et al., 006; Shamr et al., 006). In the followng we frst present a short revew of the adaptve mean shft technque and then descrbe how to apply t to the pont model. Assume that each data pont x ú d (=,,,n) s assocated wth a bandwdth value h >0. The sample pont estmator (Georgescu et al., 003; Yamauch et al., 005) n f ˆ ( x K )= K( ( )/ h d ) nh x x (5) = based on a sphercally symmetrc kernel K wth bounded support satsfyng ( x)= kd, ( x ) > 0, x K c k s called the mean shft vector pontng toward the drecton of the maxmum ncrease n the densty. A gradent-ascent process wth an adaptve step sze y = m ( y ), j=0,,,... (7) [ j+ ] [ j] G consttutes the core of the mean shft clusterng procedure. For clusterng S={x, x,, x n } wth mean shft, the followng two steps are performed on each x S: () Intalze y [0] wth x ; () Compute y [] j accordng to Eq.(7) untl convergence. It s shown n (Comancu and Meer, 00) that under some general assumptons the sequences { [] j y } converge to the ponts where fˆ K ( x ) defned by Eq.(5) attans ts local maxma (mode). Accordngly, the ponts that converge to the same mode are assocated wth the same cluster. One smple extenson of the above clusterng procedure conssts of dealng wth a set S, each element of whch has two components of a dfferent nature, S={x =(c,q ) c C, q Q}. In such a stuaton, t s convenent to use the mean shft clusterng procedure wth separable kernels ( ) ( ) n f ˆ ( x )= k ( c c )/ h k ( q q )/ h. K d d nh h =

6 8 Wang et al. / J Zhejang Unv Sc A 008 9(6): In ths paper, we consder the sample ponts {p } equpped wth the normals {n } and the mean curvature {H } as scattered data S={x =(c,q ) c p, q (n, H )} n ú 7. For both c and q, we use the normal kernel. For the bandwdth values h, there are numerous methods to defne them, most of whch use a plot densty estmate. The smplest way to obtan the plot densty estmate s by nearest neghbors. To accelerate the mean shft computaton, we construct a k-d tree for the pont set {c }. Accordng to the k-nearest neghbors N k (c ) of c, we can adaptvely take h = c c,k, where c,k s the k-nearest neghbor of c, and h =max{ q q,, q q,,, q q,k }. After clusterng for PSSs by usng ths mean shft technque, the geometry features of the ponts n the same cluster, whch contan the pont postons, normals and mean curvatures, are locally smlar, respectvely. Fg.c demonstrates the mean shft clusterng of the Face model, and ts pont set of local modes s llustrated n Fg.d. RESULTS AND DISCUSSION In our experments, we use Mcrosoft Vsual C++ programmng language on a personal computer wth a Pentum IV.8 GHz CPU and GB man memory. We have mplemented our non-local denosng (NLD) and two state-of-the-art denosng technques: the Blateral denosng (BIL) (Fleshman et al., 003) and the Mean Shft denosng (MST) (Hu et al., 006) to compare ther denosng results. We use three models n our comparson: a nosy Buste model wth 5 83 sample ponts (Fg.a), a nosy Face model wth sample ponts (Fg.a) and a nosy Dragon-head model wth sample ponts (Fg.3). For these models, Table presents the related statstcs and parameter settngs used for our method and our mplementatons of BIL and MST. For BIL and MST, we tred to choose the parameter settngs that produce the best results. Non-local denosng of PSSs From the geometry-ntensty smlartes of two ponts and Eq.(), we can compute the denosed geometry ntensty δ of p as C ( Φ δ ) δ = ( p, q ) Φ( p, q ), qj ( p ) j j qj C( p ) j where C(p ) s the cluster to whch p belongs,.e., a local neghborhood of p determned by our mean shft clusterng method. The smlarty factor Φ(p, q j ) s computed by Φ S h ( p, qj )=exp( p, / ), qj where the geometry-ntensty smlarty S, between p and q j s measured by the method descrbed p q j n the subsecton Measurng the geometry-ntensty smlarty of two ponts. Accordng to the offset dstance δ ', the smoothed poston p ' s gven by p '=p +δ 'n. Snce the pont p s moved along ts normal drecton, our denosed method wll not ntroduce undesrable ponts drftng over the surface. Moreover, ths method s more effectve and robust than the local denosng methods as t consders not only the geometry-features smlarty between two ponts but also the geometry-ntensty smlarty between them. (a) (b) Fg.3 (a) Nosy Dragon-head model colored by mean curvature; (b) Closer vew of ts upper jaw Table Parameter settng and the related statstcs Fg. Method Iters. Sm. k Max. error ( 0 4 ) Avg. error ( 0 4 ) T MS (s) T D (s) BIL MST NLD BIL MST NLD BIL MST NLD Iters. stands for the number of teratons. Sm. s the sze of the regular grd consdered to measure the geometry-ntensty smlarty of two ponts. For BIL, k s the number of the neghbors, and for NLD and MST, k s the number of the sample ponts n bandwdth wndow for mean shft clusterng. Max. error s the maxmum of dstances between the orgnal (nosy) ponts and ther correspondng denosed ponts, and Avg. error s the average of dstances. T MS ndcates the tme of mplementng the mean shft clusterng and T D denotes the pont estmatng tme for one teraton

7 Wang et al. / J Zhejang Unv Sc A 008 9(6): We use two vsualzaton schemes to compare the technques wth our method. The frst scheme conssts of colorng by the mean curvature. The second one measures the dfference between the orgnal and denosed pont model,.e., we vsualze the dfferences n the postons of the correspondng sample nosy denosed ponts of the models p p. In Fg.4, we demonstrate a comparson of the denosed Face models by BIL, MST and NLD. The denosed models are llustrated n the top row of Fg.4, and ther correspondng mean curvature vsualzatons n the bottom row. As seen n Fg.4, our NLD removes the hgh-frequency nose properly and acheves a more accurate result than BIL or MST does. Fg.5 shows a comparson of BIL, MST and NLD concernng feature preservaton. Note that our NLD preserves sharp features more accurately than BIL or MST does whle producng a smooth result, as shown n the closer vews of the upper jaw of the denosed model. In Fg., we show the denosng effcency of our approach on the nosy Buste model (Fg.a), whch s produced by addng zero-mean Gaussan nose wth σ nose =0.3 E to the orgnal model. It can be notced that the hgh-frequency nose s properly removed, whle fne detals n har, mouth and ear regons are accurately preserved. At the same tme, we demonstrate that our NLD presents the best performance accordng to the entropy of the dfferences between the nosy and denosed models, as shown n the bottom row of Fg.6. From the Max. and Avg. errors n Table we can also notce that our method outperforms ts two rvals. As a result, our method produces the lowest oversmoothng when compared wth the other two denosng technques. Due to our regon-based defnton of the geometry-ntensty smlarty measure whch adds more geometrc nformaton nto the denosng process, our algorthm removes the hgh-frequency nose properly and acheves a more accurate result than BIL or MST does. Furthermore, our method has a better feature preservaton than BIL or MST does whle producng a smooth result, manly because the method utlzes the trlateral flter n the denosng process. From the executon tme lsted n Table, we notce that our method s slower than BIL or MST snce our method needs to construct the regular geometry-ntensty grd for each sample pont and measure the geometry-ntensty smlarty of two ponts. Max Mn (a) (b) (c) Fg.4 Denosng the nosy Face model (Fg.a). Top: the denosed models. Bottom: the correspondng denosed model colored by mean curvature. Mean curvature colorng helps us to compare ther correspondng fne detals. (a) BIL; (b) MST; (c) NLD

8 84 Wang et al. / J Zhejang Unv Sc A 008 9(6): (a) (b) (c) Fg.5 Denosng the nosy Dragon-head model (Fg.3). Top: the denosed model colored by mean curvature. Bottom: a closer vew of the upper jaw of the correspondng denosed model. (a) BIL; (b) MST; (c) NLD Max (a) (b) (c) Mn Fg.6 Denosng the nosy Buste model (Fg.a). Top: the denosed model colored by mean curvature. Bottom: the correspondng denosed model colored accordng to the entropy of the dfferences between the nosy and denosed models. (a) BIL; (b) MST; (c) NLD CONCLUSION In ths paper, we presented an NLD algorthm for PSSs by extendng the non-local technque for mage denosng to a pont-sampled model. Snce the orgnal non-local method reles heavly on the mage structure regularty, the man dffculty n extendng t to PSSs s how to determne the ntensty smlarty of two sample ponts. We frst compute the geometry ntensty of the sample pont by usng the trlateral flterng operator. Based on covarance analyss, a regular geometry-ntensty grd as a counterpart to the

9 Wang et al. / J Zhejang Unv Sc A 008 9(6): square neghborhood of the pxel s then constructed. Fnally, the geometry-ntensty smlarty of two ponts s measured accordng to ther grds. Furthermore, the neghborhood of the sample pont s adaptvely selected n the denosng process by means of our mean shft method so as to produce a more accurate denosng result. Our expermental results demonstrate that ths proposed algorthm s robust, and can produce more accurate denosng results than the two state-of-theart smoothng technques, the Blateral denosng and the Mean Shft denosng, whle havng better feature preservaton. ACKNOWLEDGEMENTS Models are courtesy of Yutake Ohtake (Nosy Face), Thous Jones (Nosy Dragon Head) and EU AIM@Shape project (Buste). References Amenta, N., Kl, Y.J., 004. Defnng pont set surfaces. ACM Trans. on Graph., 3(3): [do:0.45/ ] Buades, A., Coll, B., Morel, J.M., 005. A Non-Local Algorthm for Image Denosng. Proc. IEEE Computer Socety Int. Conf. on Computer Vson and Pattern Recognton, p [do:0.09/cvpr ] Carr, J.C., Beatson, R.K., Cherre, J.B., Mtchell, T.J., Frght, W.R., McCallum, B.C., Evans, T.R., 00. Reconstructon and Representaton of 3D Objects wth Radal Bass Functons. Proc. ACM SIGGRAPH, p [do:0. 45/ ] Choudhury, P., Tumbln, J., 003. The Trlateral Flter for Hgh Contrast Images and Meshes. Int. Conf. on Computer Graphcs and Interactve Technques, p [do:0. 45/ ] Clarenz, U., Rumpf, M., Telea, A., 004. Farng of Pont Based Surfaces. Proc. Computer Graphcs Internatonal, p [do:0.09/cgi ] Comancu, D., Meer, P., 00. Mean shft: a robust approach toward feature space analyss. IEEE Trans. on Pattern Anal. Machne Intell., 4(5): [do:0.09/ ] Danels II, J., Ha, L.K., Ochotta, T., Slva, C.T., 007. Robust Smooth Feature Extracton from Pont Clouds. Proc. Shape Modelng Internatonal, p [do:0.09/ SMI.007.3] Dey, T.K., Sun, J., 005. An Adaptve MLS Surface for Reconstructon wth Guarantees. Proc. Symp. on Geometry Processng, p Fleshman, S., Dror, I., Cohen-Or, D., 003. Blateral mesh denosng. ACM Trans. on Graph., (3): [do:0.45/ ] Georgescu, B., Shmshon, I., Meer, P., 003. Mean Shft Based Clusterng n Hgh Dmensons: A Texture Classfcaton Example. ICCV, p Hoppe, H., DeRose, T., Duchamp, T., McDonald, J., Stuetzle, W., 99. Surface Reconstructon from Unorganzed Ponts. Proc. 9th Annual Conf. on Computer Graphcs and Interactve Technques, p [do:0.45/ ] Hu, G.F., Peng, Q.S., Forrest, A.R., 006. Mean shft denosng of pont-sampled surfaces. The Vsual Computer, (3): [do:0.007/s ] Jenke, P., Wand, M., Bokeloh, M., Schllng, A., Strasser, W., 006. Bayesan pont cloud reconstructon. Computer Graphcs Forum, 5(3): [do:0./j x] Lange, C., Polther, K., 005. Ansotropc smoothng of pont sets. Computer Aded Geometrc Desgn, (7): [do:0.06/j.cagd ] Lpman, Y., Cohen-Or, D., Levn, D., 006. Error Bounds and Optmal Neghborhoods for MLS Approxmaton. Proc. Eurographcs, p [do:0.3/sgp/sgp06/07-080] Mederos, B., Velho, L., de Fgueredo, L.H., 003. Robust Smoothng of Nosy Pont Clouds. Proc. SIAM Conf. on Geometrc Desgn and Computng, p.-3. Pauly, M., Gross, M., 00. Spectral Processng of Pont- Sampled Geometry. Proc. ACM SIGGRAPH, p [do:0.45/ ] Pauly, M., Gross, M., Kobbelt, L.P., 00a. Effcent Smplfcaton of Pont-Sampled Surfaces. Proc. IEEE Vsualzaton, p Pauly, M., Kobbelt, L.P., Gross, M., 00b. Multresoluton Modelng of Pont-Sampled Geometry. Techncal Report, CS #379, ETH, Zurch. Pauly, M., Keser, R., Kobblet, L.P., Gross, M., 003. Shape modelng wth pont-sampled geometry. ACM Trans. on Graph., (3): [do:0.45/ ] Pauly, M., Mtra, N.J., Gubas, L.J., 004. Uncertanty and Varablty n Pont Cloud Surface Data. Proc. Eurographcs Symp. on Pont-Based Graphcs, p Samozno, M., Alexa, M., Allez, P., Yvnec, M., 006. Reconstructon wth Vorono Centered Radal Bass Functons. Proc. Eurographcs, p [do:0.3/sgp/ SGP06/05-060] Schall, O., Belyaev, A., Sedel, H.P., 005. Robust Flterng of Nosy Scattered Pont Data. Eurographcs Symp. on Pont-Based Graphcs, p [do:0.3/spbg/ SPBG05/07-077] Shamr, A., Shapra, L., Cohen-Or, D., 006. Mesh analyss usng geodesc mean-shft. The Vsual Computer, (): [do:0.007/s ] Weyrch, T., Pauly, M., Keser, R., Henzle, S., Scandella, S., Gross, M., 004. Post-processng of Scanned 3D Surface Data. Proc. Eurographcs, p Xao, C.X., Mao, Y.W., Lu, S., Peng, Q.S., 006. A dynamc balanced flow for flterng pont-sampled geometry. The Vsual Computer, (3):0-9. [do:0.007/s ] Yamauch, H., Lee, S., Lee, Y., Ohtake, Y., Belyaev, A., Sedel, H.P., 005. Feature Senstve Mesh Segmentaton wth Mean Shft. Proc. Shape Modelng Internatonal, p [do:0.09/smi.005.]

Feature-Preserving Denoising of Point-Sampled Surfaces

Feature-Preserving Denoising of Point-Sampled Surfaces Feature-Preservng Denosng of Pont-Sampled Surfaces Jfang L College of Computer Scence and Informaton Technology Zhejang Wanl Unversty Nngbo 315100 Chna Abstract: Based on samplng lkelhood and feature ntensty,

More information

Robust Denoising of Point-Sampled Surfaces

Robust Denoising of Point-Sampled Surfaces Jfang L, Renfang Wang Robust Denosng of Pont-Sampled Surfaces Jfang L, Renfang Wang College of Computer Scence and Informaton Technology Zhejang Wanl Unversty Nngbo 315100, Chna Abstract: - Based on samplng

More information

High-Boost Mesh Filtering for 3-D Shape Enhancement

High-Boost Mesh Filtering for 3-D Shape Enhancement Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,

More information

Feature-Preserving Mesh Denoising via Bilateral Normal Filtering

Feature-Preserving Mesh Denoising via Bilateral Normal Filtering Feature-Preservng Mesh Denosng va Blateral Normal Flterng Ka-Wah Lee, Wen-Png Wang Computer Graphcs Group Department of Computer Scence, The Unversty of Hong Kong kwlee@cs.hku.hk, wenpng@cs.hku.hk Abstract

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Image Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline

Image Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline mage Vsualzaton mage Vsualzaton mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and Analyss outlne mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Fitting: Deformable contours April 26 th, 2018

Fitting: Deformable contours April 26 th, 2018 4/6/08 Fttng: Deformable contours Aprl 6 th, 08 Yong Jae Lee UC Davs Recap so far: Groupng and Fttng Goal: move from array of pxel values (or flter outputs) to a collecton of regons, objects, and shapes.

More information

Active Contours/Snakes

Active Contours/Snakes Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng

More information

The Theory and Application of an Adaptive Moving Least. Squares for Non-uniform Samples

The Theory and Application of an Adaptive Moving Least. Squares for Non-uniform Samples Xanpng Huang, Qng Tan, Janfe Mao, L Jang, Ronghua Lang The Theory and Applcaton of an Adaptve Movng Least Squares for Non-unform Samples Xanpng Huang, Qng Tan, Janfe Mao*, L Jang, Ronghua Lang College

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

LECTURE : MANIFOLD LEARNING

LECTURE : MANIFOLD LEARNING LECTURE : MANIFOLD LEARNING Rta Osadchy Some sldes are due to L.Saul, V. C. Raykar, N. Verma Topcs PCA MDS IsoMap LLE EgenMaps Done! Dmensonalty Reducton Data representaton Inputs are real-valued vectors

More information

Hierarchical clustering for gene expression data analysis

Hierarchical clustering for gene expression data analysis Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally

More information

Machine Learning: Algorithms and Applications

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

More information

TN348: Openlab Module - Colocalization

TN348: Openlab Module - Colocalization TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification Introducton to Artfcal Intellgence V22.0472-001 Fall 2009 Lecture 24: Nearest-Neghbors & Support Vector Machnes Rob Fergus Dept of Computer Scence, Courant Insttute, NYU Sldes from Danel Yeung, John DeNero

More information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge Detection in Noisy Images Using the Support Vector Machines Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and

More information

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

Face Recognition University at Buffalo CSE666 Lecture Slides Resources: Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural

More information

SIGGRAPH Interactive Image Cutout. Interactive Graph Cut. Interactive Graph Cut. Interactive Graph Cut. Hard Constraints. Lazy Snapping.

SIGGRAPH Interactive Image Cutout. Interactive Graph Cut. Interactive Graph Cut. Interactive Graph Cut. Hard Constraints. Lazy Snapping. SIGGRAPH 004 Interactve Image Cutout Lazy Snappng Yn L Jan Sun Ch-Keung Tang Heung-Yeung Shum Mcrosoft Research Asa Hong Kong Unversty Separate an object from ts background Compose the object on another

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents

More information

Simplification of 3D Meshes

Simplification of 3D Meshes Smplfcaton of 3D Meshes Addy Ngan /4/00 Outlne Motvaton Taxonomy of smplfcaton methods Hoppe et al, Mesh optmzaton Hoppe, Progressve meshes Smplfcaton of 3D Meshes 1 Motvaton Hgh detaled meshes becomng

More information

THE PULL-PUSH ALGORITHM REVISITED

THE PULL-PUSH ALGORITHM REVISITED THE PULL-PUSH ALGORITHM REVISITED Improvements, Computaton of Pont Denstes, and GPU Implementaton Martn Kraus Computer Graphcs & Vsualzaton Group, Technsche Unverstät München, Boltzmannstraße 3, 85748

More information

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

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

More information

Lobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide

Lobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.

More information

Polyhedral Surface Smoothing with Simultaneous Mesh Regularization

Polyhedral Surface Smoothing with Simultaneous Mesh Regularization olyhedral Surface Smoothng wth Smultaneous Mesh Regularzaton Yutaka Ohtake The Unversty of Azu Azu-Wakamatsu Cty Fukushma 965-8580 Japan d800@u-azu.ac.jp Alexander G. Belyaev The Unversty of Azu Azu-Wakamatsu

More information

Lecture #15 Lecture Notes

Lecture #15 Lecture Notes Lecture #15 Lecture Notes The ocean water column s very much a 3-D spatal entt and we need to represent that structure n an economcal way to deal wth t n calculatons. We wll dscuss one way to do so, emprcal

More information

S1 Note. Basis functions.

S1 Note. Basis functions. S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type

More information

Hermite Splines in Lie Groups as Products of Geodesics

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

More information

Robust Mean Shift Tracking with Corrected Background-Weighted Histogram

Robust Mean Shift Tracking with Corrected Background-Weighted Histogram Robust Mean Shft Trackng wth Corrected Background-Weghted Hstogram Jfeng Nng, Le Zhang, Davd Zhang and Chengke Wu Abstract: The background-weghted hstogram (BWH) algorthm proposed n [] attempts to reduce

More information

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

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

More information

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity

Corner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent

More information

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

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

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

More information

An Image Fusion Approach Based on Segmentation Region

An Image Fusion Approach Based on Segmentation Region Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned

More information

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples

More information

Local Quaternary Patterns and Feature Local Quaternary Patterns

Local Quaternary Patterns and Feature Local Quaternary Patterns Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents

More information

An efficient method to build panoramic image mosaics

An efficient method to build panoramic image mosaics An effcent method to buld panoramc mage mosacs Pattern Recognton Letters vol. 4 003 Dae-Hyun Km Yong-In Yoon Jong-Soo Cho School of Electrcal Engneerng and Computer Scence Kyungpook Natonal Unv. Abstract

More information

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

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

More information

A Robust Method for Estimating the Fundamental Matrix

A Robust Method for Estimating the Fundamental Matrix Proc. VIIth Dgtal Image Computng: Technques and Applcatons, Sun C., Talbot H., Ourseln S. and Adraansen T. (Eds.), 0- Dec. 003, Sydney A Robust Method for Estmatng the Fundamental Matrx C.L. Feng and Y.S.

More information

Fair Triangle Mesh Generation with Discrete Elastica

Fair Triangle Mesh Generation with Discrete Elastica Far Trangle Mesh Generaton wth Dscrete Elastca Shn Yoshzawa, and Alexander G. Belyaev, Computer Graphcs Group, Max-Planck-Insttut für Informatk, 66123 Saarbrücken, Germany Phone: [+49](681)9325-414 Fax:

More information

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches Proceedngs of the Internatonal Conference on Cognton and Recognton Fuzzy Flterng Algorthms for Image Processng: Performance Evaluaton of Varous Approaches Rajoo Pandey and Umesh Ghanekar Department of

More information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

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

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

More information

IMPROVING AND EXTENDING THE INFORMATION ON PRINCIPAL COMPONENT ANALYSIS FOR LOCAL NEIGHBORHOODS IN 3D POINT CLOUDS

IMPROVING AND EXTENDING THE INFORMATION ON PRINCIPAL COMPONENT ANALYSIS FOR LOCAL NEIGHBORHOODS IN 3D POINT CLOUDS IMPROVING AND EXTENDING THE INFORMATION ON PRINCIPAL COMPONENT ANALYSIS FOR LOCAL NEIGHBORHOODS IN 3D POINT CLOUDS Davd Belton Cooperatve Research Centre for Spatal Informaton (CRC-SI) The Insttute for

More information

Scale and Orientation Adaptive Mean Shift Tracking

Scale and Orientation Adaptive Mean Shift Tracking Scale and Orentaton Adaptve Mean Shft Trackng Jfeng Nng, Le Zhang, Davd Zhang and Chengke Wu Abstract A scale and orentaton adaptve mean shft trackng (SOAMST) algorthm s proposed n ths paper to address

More information

COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL

COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL COMPLEX WAVELET TRANSFORM-BASED COLOR INDEXING FOR CONTENT-BASED IMAGE RETRIEVAL Nader Safavan and Shohreh Kasae Department of Computer Engneerng Sharf Unversty of Technology Tehran, Iran skasae@sharf.edu

More information

Detection of an Object by using Principal Component Analysis

Detection of an Object by using Principal Component Analysis Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,

More information

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION 1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute

More information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

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

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like: Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A

More information

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity ISSN(Onlne): 2320-9801 ISSN (Prnt): 2320-9798 Internatonal Journal of Innovatve Research n Computer and Communcaton Engneerng (An ISO 3297: 2007 Certfed Organzaton) Vol.2, Specal Issue 1, March 2014 Proceedngs

More information

Reading. 14. Subdivision curves. Recommended:

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

More information

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

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

More information

Dynamic wetting property investigation of AFM tips in micro/nanoscale

Dynamic wetting property investigation of AFM tips in micro/nanoscale Dynamc wettng property nvestgaton of AFM tps n mcro/nanoscale The wettng propertes of AFM probe tps are of concern n AFM tp related force measurement, fabrcaton, and manpulaton technques, such as dp-pen

More information

Feature Reduction and Selection

Feature Reduction and Selection Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components

More information

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010

Simulation: Solving Dynamic Models ABE 5646 Week 11 Chapter 2, Spring 2010 Smulaton: Solvng Dynamc Models ABE 5646 Week Chapter 2, Sprng 200 Week Descrpton Readng Materal Mar 5- Mar 9 Evaluatng [Crop] Models Comparng a model wth data - Graphcal, errors - Measures of agreement

More information

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User Authentication Based On Behavioral Mouse Dynamics Biometrics User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA

More information

Efficient Mean-shift Clustering Using Gaussian KD-Tree

Efficient Mean-shift Clustering Using Gaussian KD-Tree Pacfc Graphcs 2010 P. Allez, K. Bala, and K. Zhou (Guest Edtors) Volume 29 (2010), Number 7 Effcent Mean-shft Clusterng Usng Gaussan KD-Tree Chunxa Xao Meng Lu The School of Computer, Wuhan Unversty, Wuhan,

More information

Object-Based Techniques for Image Retrieval

Object-Based Techniques for Image Retrieval 54 Zhang, Gao, & Luo Chapter VII Object-Based Technques for Image Retreval Y. J. Zhang, Tsnghua Unversty, Chna Y. Y. Gao, Tsnghua Unversty, Chna Y. Luo, Tsnghua Unversty, Chna ABSTRACT To overcome the

More information

Unsupervised Learning

Unsupervised Learning Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and

More information

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

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

More information

Hybridization of Expectation-Maximization and K-Means Algorithms for Better Clustering Performance

Hybridization of Expectation-Maximization and K-Means Algorithms for Better Clustering Performance BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 2 Sofa 2016 Prnt ISSN: 1311-9702; Onlne ISSN: 1314-4081 DOI: 10.1515/cat-2016-0017 Hybrdzaton of Expectaton-Maxmzaton

More information

A 3D Paper-Cutting Oriented Mesh Trimming Algorithm

A 3D Paper-Cutting Oriented Mesh Trimming Algorithm ISSN 1000-9825, CODEN RUXUEW E-mal: jos@scasaccn Journal of Software, Vol17, Supplement, November 2006, pp169 175 http://wwwjosorgcn 2006 by of Journal of Software All rghts reserved Tel/Fax: +86-10-62562563

More information

Computer Animation and Visualisation. Lecture 4. Rigging / Skinning

Computer Animation and Visualisation. Lecture 4. Rigging / Skinning Computer Anmaton and Vsualsaton Lecture 4. Rggng / Sknnng Taku Komura Overvew Sknnng / Rggng Background knowledge Lnear Blendng How to decde weghts? Example-based Method Anatomcal models Sknnng Assume

More information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1 4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:

More information

SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS

SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS J.H.Guan, F.B.Zhu, F.L.Ban a School of Computer, Spatal Informaton & Dgtal Engneerng Center, Wuhan Unversty, Wuhan, 430079,

More information

Lecture 4: Principal components

Lecture 4: Principal components /3/6 Lecture 4: Prncpal components 3..6 Multvarate lnear regresson MLR s optmal for the estmaton data...but poor for handlng collnear data Covarance matrx s not nvertble (large condton number) Robustness

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

MOTION BLUR ESTIMATION AT CORNERS

MOTION BLUR ESTIMATION AT CORNERS Gacomo Boracch and Vncenzo Caglot Dpartmento d Elettronca e Informazone, Poltecnco d Mlano, Va Ponzo, 34/5-20133 MILANO boracch@elet.polm.t, caglot@elet.polm.t Keywords: Abstract: Pont Spread Functon Parameter

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution

Real-time Motion Capture System Using One Video Camera Based on Color and Edge Distribution Real-tme Moton Capture System Usng One Vdeo Camera Based on Color and Edge Dstrbuton YOSHIAKI AKAZAWA, YOSHIHIRO OKADA, AND KOICHI NIIJIMA Graduate School of Informaton Scence and Electrcal Engneerng,

More information

Structure from Motion

Structure from Motion Structure from Moton Structure from Moton For now, statc scene and movng camera Equvalentl, rgdl movng scene and statc camera Lmtng case of stereo wth man cameras Lmtng case of multvew camera calbraton

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

Mesh Editing in ROI with Dual Laplacian

Mesh Editing in ROI with Dual Laplacian Mesh Edtng n ROI wth Dual Laplacan Luo Qong, Lu Bo, Ma Zhan-guo, Zhang Hong-bn College of Computer Scence, Beng Unversty of Technology, Chna lqngng@sohu.com, lubo@but.edu.cn,mzgsy@63.com,zhb@publc.bta.net.cn

More information

Adaptive Fairing of Surface Meshes by Geometric Diffusion

Adaptive Fairing of Surface Meshes by Geometric Diffusion Adaptve Farng of Surface Meshes by Geometrc Dffuson Chandrajt L. Bajaj Department of Computer Scences, Unversty of Texas, Austn, TX 78712 Emal: bajaj@cs.utexas.edu Guolang Xu State Key Lab. of Scentfc

More information

Development of an Active Shape Model. Using the Discrete Cosine Transform

Development of an Active Shape Model. Using the Discrete Cosine Transform Development of an Actve Shape Model Usng the Dscrete Cosne Transform Kotaro Yasuda A Thess n The Department of Electrcal and Computer Engneerng Presented n Partal Fulfllment of the Requrements for the

More information

GSLM Operations Research II Fall 13/14

GSLM Operations Research II Fall 13/14 GSLM 58 Operatons Research II Fall /4 6. Separable Programmng Consder a general NLP mn f(x) s.t. g j (x) b j j =. m. Defnton 6.. The NLP s a separable program f ts objectve functon and all constrants are

More information

PCA Based Gait Segmentation

PCA Based Gait Segmentation Honggu L, Cupng Sh & Xngguo L PCA Based Gat Segmentaton PCA Based Gat Segmentaton Honggu L, Cupng Sh, and Xngguo L 2 Electronc Department, Physcs College, Yangzhou Unversty, 225002 Yangzhou, Chna 2 Department

More information

Load Balancing for Hex-Cell Interconnection Network

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

More information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng, Natonal Unversty of Sngapore {shva, phanquyt, tancl }@comp.nus.edu.sg

More information

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information

An Improved Stereo Matching Algorithm Based on Guided Image Filter

An Improved Stereo Matching Algorithm Based on Guided Image Filter nd Internatonal Conference on Modellng, Identfcaton and Control (MIC 015 An Improved Stereo Matchng Algorthm Based on Guded Image Flter Rudong Gao, Yun Chen, Lna Yan School of nstrumentaton Scence and

More information

A Gradient Difference based Technique for Video Text Detection

A Gradient Difference based Technique for Video Text Detection 2009 10th Internatonal Conference on Document Analyss and Recognton A Gradent Dfference based Technque for Vdeo Text Detecton Palaahnakote Shvakumara, Trung Quy Phan and Chew Lm Tan School of Computng,

More information

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

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

More information

Registration of Expressions Data using a 3D Morphable Model

Registration of Expressions Data using a 3D Morphable Model Regstraton of Expressons Data usng a 3D Morphable Model Curzo Basso DISI, Unverstà d Genova, Italy Emal: curzo.basso@ds.unge.t Thomas Vetter Departement Informatk, Unverstät Basel, Swtzerland Emal: thomas.vetter@unbas.ch

More information

A fast algorithm for color image segmentation

A fast algorithm for color image segmentation Unersty of Wollongong Research Onlne Faculty of Informatcs - Papers (Arche) Faculty of Engneerng and Informaton Scences 006 A fast algorthm for color mage segmentaton L. Dong Unersty of Wollongong, lju@uow.edu.au

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 15

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 15 CS434a/541a: Pattern Recognton Prof. Olga Veksler Lecture 15 Today New Topc: Unsupervsed Learnng Supervsed vs. unsupervsed learnng Unsupervsed learnng Net Tme: parametrc unsupervsed learnng Today: nonparametrc

More information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive Semi-definite Programming Localization in Wireless Sensor Networks Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer

More information

Accounting for the Use of Different Length Scale Factors in x, y and z Directions

Accounting for the Use of Different Length Scale Factors in x, y and z Directions 1 Accountng for the Use of Dfferent Length Scale Factors n x, y and z Drectons Taha Soch (taha.soch@kcl.ac.uk) Imagng Scences & Bomedcal Engneerng, Kng s College London, The Rayne Insttute, St Thomas Hosptal,

More information

Image Alignment CSC 767

Image Alignment CSC 767 Image Algnment CSC 767 Image algnment Image from http://graphcs.cs.cmu.edu/courses/15-463/2010_fall/ Image algnment: Applcatons Panorama sttchng Image algnment: Applcatons Recognton of object nstances

More information