Stereo Depth Continuity

Size: px
Start display at page:

Download "Stereo Depth Continuity"

Transcription

1 Stereo Depth Contnuty Steven Damond Jessca Taylor March 17, Abstract We tackle the problem of producng depth maps from stereo vdeo. Some algorthms for dong ths usng correspondences produce nosy output. Chan et al. (011) [] use convex optmzaton technques to reduce nose and also smooth out the depth maps over tme. We start wth the nosy output of OpenCV and apply ths technque to smooth t, developng some refnements along the way. Specfcally, we gnore pxels wth the hghest possble depth value (as ths ndcates that no correspondence was found for the pxel), and we make the smoothng less aggressve f the pxels dffer by color. These refnements mprove the smoothng to yeld more accurate depth estmates. We test the algorthm on multple vdeos and fnd that t generally performs well as long as OpenCV s output was somewhat reasonable. Revew of prevous work Gven stereo mages, t s possble to retreve depth maps by fndng correspondences between the left and rght mages. Ths technque s used by OpenCV ( trunk/doc/py_tutorals/py_calb3d/py_depthmap/py_depthmap.html). We ran OpenCV s StereoBM algorthm wth a SADWndowSze parameter of 5, whch ncreases accuracy by matchng smaller regons but ncreases nose. It s straghtforward to fnd depth maps for stereo vdeo by smply fndng maps for each frame ndependently. However, n addton to nhertng all the problems of the orgnal algorthm, the result s usually choppy, lackng smooth changes over tme. Chan et al. (011) [] apply a smoothng algorthm to sttch together frame-by-frame depth maps, whch succeeds n both reducng the nose of each depth map and ncreasng contnuty over tme. They solve the followng optmzaton problem: mnmze µ f,j,k g,j,k +,j,k,j,k β x (f +1,j,k f,j,k ) β y (f,j+1,k f,j,k ) β t (f,j,k+1 f,j,k ) where g s a 3D matrx contanng depth estmates from a frame-by-frame algorthm, f s a 3D matrx of smoothed depth estmates, and µ, β x, β y, and β t are non-negatve parameters. 1

2 Steven Damond, Jessca Taylor 4 SUMMARY OF TECHNICAL SOLUTION 3 Improvements on Prevous Work We found that OpenCV produced very nosy output. Addtonally, many pxels (n some cases the majorty) were labeled as the furthest possble depth value. Usually, these pxels dd not actually have such hgh depths; nstead, they were scattered around the mage at many dfferent depth values. It s lkely that OpenCV found no correspondences for these pxels, and gave them the hghest depth value as a default. Therefore, we adjusted the smoothng algorthm to gnore these pxels depth values. Ths way, the algorthm s free to vary the smoothed depth estmate correspondng to these pxels n order to produce more contnuty. Addtonally, we changed the smoothng to place less emphass on smoothness for pxels wth smlar colors. Ths captures the ntuton that depth changes should happen where color 1 changes happen. To do ths, we created a weght for each pxel proportonal to 1+, where s a vector combnng color dfferences n the x, y, and tme drectons. As explaned later, Chan et al. [] nclude smoothness constrants between pxels at opposte ends of the mage, or between the fnal and frst frame, just as there are smoothness constrants between neghborng pxels or frames. Includng these constrants s necessary to make optmzaton much more effcent. We do not drop these constrants, but we reduce the weght of pxels on edges that correspond to these wrapped constrants (to 1/3 of the orgnal weght) to make these constrants less cumbersome. 4 Summary of Techncal Soluton To obtan a smoothed depth map for the vdeo, we solve a convex optmzaton problem. We regard the vdeo as a tme space volume, where each pxel has an x,y, and t coordnate. The x axs corresponds to the pxel column, the y axs to the pxel row, and the t axs to the pxel frame. Let g be a 3D matrx of pxel depth estmates ndexed by x,y, and t. g s obtaned by estmatng a depth map for each frame of the vdeo ndependently. Some pxels may not be assgned a depth by the frame-by-frame algorthm. We call such pxels unlabeled. Let E be the set of labeled pxels n g. Let f be a matrx varable representng the smoothed pxel depth estmates. Then our optmzaton problem s gven by mnmze µ (,j,k) E f,j,k g,j,k + w,j,k,j,k β x (f +1,j,k f,j,k ) β y (f,j+1,k f,j,k ) β t (f,j,k+1 f,j,k ) w, µ, β x, β y, and β t are non-negatve parameters. The optmzaton problem fnds an f that s close to the orgnal depth estmates g but has fewer sharp changes n depth. Specfcally, the term (,j,k) E f,j,k g,j,k penalzes f s devaton from labeled pxels n g whle the term,j,k w,j,k β x (f +1,j,k f,j,k ) β y (f,j+1,k f,j,k ) β t (f,j,k+1 f,j,k )

3 Steven Damond, Jessca Taylor 5 DETAILS OF TECHNICAL SOLUTION penalzes large depth gradents n f. The balance between these terms s controlled by the parameter µ. The parameter w s a matrx wth a non-negatve weght for each pxel. The weght w,j,k determnes how heavly the gradent at f,j,k s penalzed. To choose w we followed the prncple that neghborng pxels wth smlar colors or ntenstes are lkely to belong to the same surface and thus be at smlar depths. Conversely, neghborng pxels wth very dfferent colors or ntenstes lkely belong to dfferent objects and thus could have very dfferent depths. β x, β y, and β t are global weghts appled to the component of the gradent n x,y, and t dmensons, respectvely. The β parameters allow changes n depth n certan dmensons to be weghted more (or less) heavly. Emprcally, settng β t less than β x and β y mproves performance. Ths makes sense gven that movng a sngle frame typcally makes more of a dfference to depth than movng a sngle pxel. 5 Detals of Techncal Soluton Our optmzaton problem s extremely hgh dmensonal, snce a vdeo contans mllons of pxels and we estmate a depth for every pxel. There s no generc solver for convex optmzaton problems capable of solvng our problem n a reasonable amount of tme. Hence, we mplemented a custom solver for our problem based on the work by Chan et al. (011) []. Lke Chan et al., our solver uses an teratve algorthm where each teraton runs n O(n log n). Here n s the total number of pxels n the vdeo. Ths runtme s an mportant dstncton of our algorthm, snce a generc convex optmzaton algorthm would run n O(n ) or O(n 3 ) []. We obtan an O(n log n) runtme per teraton by explotng specal structure n our problem. 5.1 The ADMM Algorthm Our algorthm uses the Alternatng Drecton Method of Multplers (ADMM). ADMM s an teratve method desgned for convex optmzaton problems too large for standard nteror pont methods [1]. ADMM solves problems of the form mnmze f(x) + g(z) subject to Ax + Bz = c where f, g are convex functons. ADMM operates on the so-called augmented Lagrangan L ρ (x, z, y) = f(x) + g(x) + y T (Ax + Bz c) + (ρ/) Ax + Bz c 3

4 Steven Damond, Jessca Taylor 5 DETAILS OF TECHNICAL SOLUTION where y s the dual varable. ADMM conssts of the followng algorthm: Data: A,B,c Result: Prmal optmal x, z and dual optmal y Intalze x (0), z (0), y (0) ; whle The stoppng condtons are not met do x (k+1) = arg mn x L ρ (x, z (k), y (k) ); z (k+1) = arg mn z L ρ (x (k+1), z, y (k) ); y (k+1) = y (k) + ρ(ax (k+1) + Bz (k+1) c); end The ntalzaton depends on the problem. ADMM wll converge faster f t s ntalzed near the soluton. The stoppng condton s based on the L norms of the prmal resdual and dual resdual r (k+1) = Ax (k+1) + Bz (k+1) c s (k+1) = ρa T B(z (k+1) z (k) ) When r (k+1) and s (k+1) are both small t ndcates that ADMM s near the correct soluton. Boyd et al. [1] recommend usng both a relatve and absolute metrc for measurng convergence. Specfcally, they suggest termnatng when r (k+1) ɛ pr and s (k+1) ɛ dual where ɛ pr = pɛ abs + ɛ rel max { Ax (k), Bz (k), c } ɛ dual = nɛ abs + ɛ rel A T y (k) p s the number of rows n A and n s the number of columns. ɛ abs and ɛ rel are chosen based on the desred accuracy and the problem scale. 5. Applyng ADMM To use ADMM we must reformulate our problem to match the ADMM standard form. Let f R C R F. Denote N = CRF. We defne new varables r R C R F, u (x), u (y), u (t) R N. Let vec(f) denote the vectorzed verson of f. Specfcally, to vectorze f we flatten each frame n row-major order and concatenate the frame vectors. We now defne matrces D x, D y, D t R N N. D x, D y, D t are forward shft operators n the x, y, and t dmensons respectvely. [D x vec(f)] s the dfference between the depth at the pxel one space forward n x from pxel and the depth at pxel. [D y vec(f)] and [D t vec(f)] are defned smlarly. Formally, f pxel occurs at locaton x, y, t: [D x vec(f)] = f(x + 1 mod C, y, t) f(x, y, z) [D y vec(f)] = f(x, y + 1 mod R, t) f(x, y, z) [D t vec(f)] = f(x, y, t + 1 mod F ) f(x, y, z) 4

5 Steven Damond, Jessca Taylor 5 DETAILS OF TECHNICAL SOLUTION We use perodc boundary condtons, meanng that for pxels on the boundares of f we look at the depth dfference between that pxel and the next pxel n the flattened verson of f. For example, f x = C 1 we look at the dfference f(x + 1 mod C, y, t) f(x, y, z) = f(0, y + 1, t) f(x, y, z) These boundary condtons are not natural, but they are necessary for the matrces D x, D y, D t to be block-crculant. A matrx s defned as block-crculant f every kth column of the matrx s the frst column rotated k 1 steps forward. Vsually the matrx has the form below c 0 c n 1... c 1 c 1 c 0... c c c 1... c c n 1 c n... c 0 Let c be the frst column of a block crculant matrx C. Multplyng a vector x by C s equvalent to a crcular convoluton of c wth x [3]: Cx = c x In the Fourer doman, crcular convoluton s equvalent to elementwse multplcaton. Thus applyng the dscrete fourer transform operator F, F(c x) = F(c) F(x) Remark 5.1. We can thus use the DFT to solve lnear equatons Cx = b where C s a blockcrculant matrx n tme O(n log n). Observe, F(Cx) = F(b) F(c) F(x) = F(b) ( ) F(b) x = F 1 F(c) Returnng to our formulaton of the optmzaton problem, we make two more defntons to smplfy notaton u (x) D x u = u (y), D = D y u (t) D t We can now express our optmzaton problem n ADMM form: mnmze µ (,j,k) E subject to r = f g u = Dvec(f) r,j,k + w u (x) u (y) u (t) 5

6 Steven Damond, Jessca Taylor 5 DETAILS OF TECHNICAL SOLUTION The augmented Lagrangan s L ρ (f, u, r, y, z) = µ (,j,k) E r,j,k + w u (x) u (y) u (t) z T vec(r f + g) y (x) + (ρ/) vec(r f + g) yt (u Dvec(f)) + (ρ/) u Dvec(f) where y = y (y) R 3N, z R N are dual varables. y (t) 5.3 Computng ADMM Iteratons We now explan how we compute each teraton of ADMM. The structure of our problem allows us to compute each step of the teraton n at most tme O(n log n) where n s the number of pxels n the vdeo. Wthout takng advantage of structure, an teraton would take tme O(n ), whch s far less feasble gven that we have mllons of pxels. For our problem the ADMM algorthm takes the followng form: Data: g Result: Prmal optmal f, u, r and dual optmal y, z Intalze f (0) = g, u (0) = Dvec(g), r (0) = z (0) = y (0) = 0; whle The stoppng condtons are not met do u (k+1) = arg mn u L ρ (f (k), u, r (k), y (k), z (k) ); r (k+1) = arg mn r L ρ (f (k), u (k+1), r, y (k), z (k) ); f (k+1) = arg mn f L ρ (f, u (k+1), r (k+1), y (k), z (k) ); y (k+1) = y (k) ρ(u (k+1) Dvec(f (k+1) )); z (k+1) = z (k) ρvec(r (k+1) f (k+1) + g); end Intalzng f (0) = g, u (0) = Dvec(f (0) ), r (0) = f (0) g = 0 s useful because startng ADMM near ts eventual soluton decreases the number of teratons needed. There s no obvous startng estmate for the dual varables, so we ntalze them to 0. We next explan how the u,r, and f updates are computed u update Recall the u update s the soluton to the problem u (x) u = arg mn w u (y) u y T (u Dvec(f)) + (ρ/) u Dvec(f) u (t) The u update can be expressed n closed form usng the shrnkage formula defned by L [4]. We defne v (x) = β x D x vec(f) + 1 ρ y(x) 6

7 Steven Damond, Jessca Taylor 5 DETAILS OF TECHNICAL SOLUTION wth v (y), v (t) defned analogously. We further defne v R N such that v = (v (x) ) + (v (y) ) + (v (t) ) for all. Then for each pxel wth gradent weght w the updated value of u (x) s gven by u (x) The updates for u (y), u (t) are analogous r update { = max v w } ρ, 0 v(x) v Recall the r update s the soluton to the problem r = arg mn µ r,j,k z T vec(r f + g) + (ρ/) vec(r f + g) r (,j,k) E For a pxel (, j, k) / E, the value r,j,k does not change. For a pxel (, j, k) E, a smlar shrnkage formula gves a closed form expresson for the r update. We use l to ndcate the ndex of (, j, k) n the vectorzed mage. { r,j,k = max f,j,k g,j,k + 1 ρ z l µ } ( ρ, 0 sgn f,j,k g,j,k + 1 ) ρ z l f update Recall the f update s the soluton to the problem f = arg mn(ρ/) vec(r f + g) + (ρ/) u Dvec(f) + yt Dvec(f) + z T vec(f) f The optmal f satsfes the normal equatons ρ(i + D T D)vec(f) = ρg + (ρr z) + D T (ρu y) ρ(i + D T x D x + D T y D y + D T t D t )vec(f) = ρg + (ρr z) + D T (ρu y) Recall that D x, D y, D t are block-crculant matrces. Ths means Dx T, Dy T, Dt T are also block crculant. The dentty matrx I s block crculant as well. Thus by Remark 5.1 we can solve the normal equatons n tme O(n log n) usng the DFT operator F and the FFT algorthm. Denote M = ρ(i + Dx T D x + Dy T D y + Dt T D t ). Then applyng F to Mvec(f) yelds F(Mvec(f)) = ρ(f(i) + F(D T x ) F(D x ) + F(D T y ) F(D y ) + F(D T t ) F(D t )) F(vec(f)) = F(M) F(vec(f)) 7

8 Steven Damond, Jessca Taylor 6 EXPERIMENTS Here F(I), F(D T x ), F(D x ), etc., mean the DFT operator appled to the frst column of each matrx. F(M) need only be computed once, before the ADMM loop begns. We use F(M) to solve the normal equatons Convergence Mvec(f) = ρg + (ρr z) + D T (ρu y) F(Mvec(f)) = F(ρg + (ρr z) + D T (ρu y)) F(M) F(vec(f)) = F(ρg + (ρr z) + D T (ρu y)) vec(f) = F 1 ( F(ρg + (ρr z) + D T (ρu y)) F(M) We use the stoppng crtera descrbed n Secton 5.1 wth ɛ abs = 1e 4 and ɛ rel = 1e 4. Ths s dfferent from the crtera used by Chan et al. []. Our stoppng crtera results n more ADMM teratons, but t corresponds much better to the accuracy of the ADMM soluton. We were stll able to get convergence n relatvely few teratons by scalng g so all entres were n [0, 1]. We used a consstent value of ρ = 1 rather than updatng ρ each teraton as done by Chan et al. []. We found that updatng ρ prevented ADMM from convergng. ) 6 Experments We tested ths algorthm on a vdeo also used n [], specfcally the Horse vdeo found at We scaled the vdeo down to be 180 by 10 pxels and used all 140 frames. We optmzed over the 3.09 mllon varables n 980 seconds, or 7 seconds per frame. As there s no ground truth provded for ths dataset, qualtatve comparson s necessary. We compare depth maps of a sngle frame of the vdeo (orgnal, OpenCV, and our algorthm): The smoothed depth map s much less nosy than the orgnal depth map, contanng clearly defned regons of smlar depth. Black pxels n the orgnal depth map (representng unlabeled pxels wth no depth estmate) have been smoothed over wth neghborng depth values, mprovng accuracy. Although t s dffcult to show ths n paper form, the depth maps are choppy when they are not smoothed over tme (only smoothed over pxel locatons). Smoothng over tme reduces choppness wthout sacrfcng accuracy. 8

9 Steven Damond, Jessca Taylor 7 CONCLUSION We also tred some alternatve datasets. Specfcally, we used stereo vdeos found at hc.wr.un-hedelberg.de/benchmarks/document/challengng_data_for_stereo_and_optcal_ Flow/, whch do not contan ground truth. The algorthm performed well on some of these vdeos but not others, mostly n lne wth how OpenCV s algorthm dd on these vdeos. It performed relatvely well on the CrossngCars vdeo: Our algorthm succeeded n fllng n black regons and handlng occlusons over tme. When an occluded car becomes vsble agan, t mantans roughly the same depth that t had orgnally. Addtonally, we tred usng the Indoor 1 dataset found at com/en-us/downloads/9d bce1/ because ths dataset contaned ground truth. Unfortunately, none of the frame-by-frame algorthms we tested produced reasonable results on ths dataset. We tred algorthms other than OpenCV s StereoBM, but none was very accurate for the datasets we used. These algorthms were OpenCV s StereoSGBM, the MRF mnmzaton-based algorthm found at and the Mcrosoft Research StereoMatcher package at (wth a patch from edu/stereo/code/). None of these algorthms produced reasonable depth maps for our datasets. For a more quanttatve evaluaton, we downloaded depth maps from top-ranked algorthms at and ran our algorthm on them. We counted bad pxels n the same way the webste does (proporton of non-occluded pxels that are more than 4 ntensty unts away from the ground truth value) on the teddy mage. Algorthm Orgnal Bad Pxels New Bad Pxels Percent Change TSGO % % % AdaptngBP % % % DoubleBP % % % ADCensus % % % CoopRegon % % % In general, smoothng made very lttle dfference to the accuracy of the depth maps. Ths s most lkely because the output of the top algorthms s already relatvely smooth. 7 Concluson Our mprovements to the smoothng algorthm n [] make t approprate for use wth OpenCV s depth maps. The algorthm succeeds n smoothng depth maps over space and tme as long 9

10 Steven Damond, Jessca Taylor REFERENCES as the orgnal depth maps are reasonable. Our extensons to the algorthm handle unlabeled pxels and mprove performance around edges. We mplement the ADMM algorthm, whch greatly mproves performance to O(n log n) per teraton, allowng entre vdeos to be smoothed. Due to ts qualty and reasonable performance, our algorthm s approprate for practcal use n smoothng nosy but relatvely accurate depth maps. In partcular, our algorthm works well when only a small proporton of pxels (such as only pxels near edges) ntally have depth estmates and others are unlabeled. References [1] S. Boyd, N. Parkh, E. Chu, B. Peleato, and J. Ecksten. Dstrbuted optmzaton and statstcal learnng va the alternatng drecton method of multplers. Foundatons and Trends R n Machne Learnng, 3(1):1 1, 011. [] S. Chan, R. Khoshabeh, K. Gbson, P. Gll, and T. Nguyen. An augmented lagrangan method for total varaton vdeo restoraton. Image Processng, IEEE Transactons on, 0(11): , Nov 011. [3] G. H. Golub and C. F. Van Loan. Matrx computatons, volume 3. JHU Press, 01. [4] C. L. An effcent algorthm for total varaton regularzaton wth applcatons to the sngle pxel camera and compressve sensng. PhD thess, Cteseer,

Wavefront Reconstructor

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

More information

Discriminative Dictionary Learning with Pairwise Constraints

Discriminative Dictionary Learning with Pairwise Constraints Dscrmnatve Dctonary Learnng wth Parwse Constrants Humn Guo Zhuoln Jang LARRY S. DAVIS UNIVERSITY OF MARYLAND Nov. 6 th, Outlne Introducton/motvaton Dctonary Learnng Dscrmnatve Dctonary Learnng wth Parwse

More information

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

Biostatistics 615/815

Biostatistics 615/815 The E-M Algorthm Bostatstcs 615/815 Lecture 17 Last Lecture: The Smplex Method General method for optmzaton Makes few assumptons about functon Crawls towards mnmum Some recommendatons Multple startng ponts

More information

Alternating Direction Method of Multipliers Implementation Using Apache Spark

Alternating Direction Method of Multipliers Implementation Using Apache Spark Alternatng Drecton Method of Multplers Implementaton Usng Apache Spark Deterch Lawson June 4, 2014 1 Introducton Many applcaton areas n optmzaton have benefted from recent trends towards massve datasets.

More information

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

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

More information

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

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

CS 534: Computer Vision Model Fitting

CS 534: Computer Vision Model Fitting CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust

More information

Programming in Fortran 90 : 2017/2018

Programming in Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Exercse 1 : Evaluaton of functon dependng on nput Wrte a program who evaluate the functon f (x,y) for any two user specfed values

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

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

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

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

More information

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

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

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

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

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

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

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

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

More information

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

5 The Primal-Dual Method

5 The Primal-Dual Method 5 The Prmal-Dual Method Orgnally desgned as a method for solvng lnear programs, where t reduces weghted optmzaton problems to smpler combnatoral ones, the prmal-dual method (PDM) has receved much attenton

More information

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

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

More information

Mathematics 256 a course in differential equations for engineering students

Mathematics 256 a course in differential equations for engineering students Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

Real-time Joint Tracking of a Hand Manipulating an Object from RGB-D Input

Real-time Joint Tracking of a Hand Manipulating an Object from RGB-D Input Real-tme Jont Tracng of a Hand Manpulatng an Object from RGB-D Input Srnath Srdhar 1 Franzsa Mueller 1 Mchael Zollhöfer 1 Dan Casas 1 Antt Oulasvrta 2 Chrstan Theobalt 1 1 Max Planc Insttute for Informatcs

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

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

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

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

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

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming CS 4/560 Desgn and Analyss of Algorthms Kent State Unversty Dept. of Math & Computer Scence LECT-6 Dynamc Programmng 2 Dynamc Programmng Dynamc Programmng, lke the dvde-and-conquer method, solves problems

More information

Support Vector Machines. CS534 - Machine Learning

Support Vector Machines. CS534 - Machine Learning Support Vector Machnes CS534 - Machne Learnng Perceptron Revsted: Lnear Separators Bnar classfcaton can be veed as the task of separatng classes n feature space: b > 0 b 0 b < 0 f() sgn( b) Lnear Separators

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

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

Optimizing Document Scoring for Query Retrieval

Optimizing Document Scoring for Query Retrieval Optmzng Document Scorng for Query Retreval Brent Ellwen baellwe@cs.stanford.edu Abstract The goal of ths project was to automate the process of tunng a document query engne. Specfcally, I used machne learnng

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

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

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

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

More information

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

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

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

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

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 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

CS246: Mining Massive Datasets Jure Leskovec, Stanford University

CS246: Mining Massive Datasets Jure Leskovec, Stanford University CS46: Mnng Massve Datasets Jure Leskovec, Stanford Unversty http://cs46.stanford.edu /19/013 Jure Leskovec, Stanford CS46: Mnng Massve Datasets, http://cs46.stanford.edu Perceptron: y = sgn( x Ho to fnd

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

Multi-view 3D Position Estimation of Sports Players

Multi-view 3D Position Estimation of Sports Players Mult-vew 3D Poston Estmaton of Sports Players Robbe Vos and Wlle Brnk Appled Mathematcs Department of Mathematcal Scences Unversty of Stellenbosch, South Afrca Emal: vosrobbe@gmal.com Abstract The problem

More information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

y and the total sum of

y and the total sum of Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton

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

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

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.

More information

A Scalable Projective Bundle Adjustment Algorithm using the L Norm

A Scalable Projective Bundle Adjustment Algorithm using the L Norm Sxth Indan Conference on Computer Vson, Graphcs & Image Processng A Scalable Projectve Bundle Adjustment Algorthm usng the Norm Kaushk Mtra and Rama Chellappa Dept. of Electrcal and Computer Engneerng

More information

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming CEE 60 Davd Rosenberg p. LECTURE NOTES Dualty Theory, Senstvty Analyss, and Parametrc Programmng Learnng Objectves. Revew the prmal LP model formulaton 2. Formulate the Dual Problem of an LP problem (TUES)

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

Robust Low-Rank Regularized Regression for Face Recognition with Occlusion

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

More information

Efficient Distributed Linear Classification Algorithms via the Alternating Direction Method of Multipliers

Efficient Distributed Linear Classification Algorithms via the Alternating Direction Method of Multipliers Effcent Dstrbuted Lnear Classfcaton Algorthms va the Alternatng Drecton Method of Multplers Caoxe Zhang Honglak Lee Kang G. Shn Department of EECS Unversty of Mchgan Ann Arbor, MI 48109, USA caoxezh@umch.edu

More information

Problem Set 3 Solutions

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

More information

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

Classification / Regression Support Vector Machines

Classification / Regression Support Vector Machines Classfcaton / Regresson Support Vector Machnes Jeff Howbert Introducton to Machne Learnng Wnter 04 Topcs SVM classfers for lnearly separable classes SVM classfers for non-lnearly separable classes SVM

More information

AMath 483/583 Lecture 21 May 13, Notes: Notes: Jacobi iteration. Notes: Jacobi with OpenMP coarse grain

AMath 483/583 Lecture 21 May 13, Notes: Notes: Jacobi iteration. Notes: Jacobi with OpenMP coarse grain AMath 483/583 Lecture 21 May 13, 2011 Today: OpenMP and MPI versons of Jacob teraton Gauss-Sedel and SOR teratve methods Next week: More MPI Debuggng and totalvew GPU computng Read: Class notes and references

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

SVM-based Learning for Multiple Model Estimation

SVM-based Learning for Multiple Model Estimation SVM-based Learnng for Multple Model Estmaton Vladmr Cherkassky and Yunqan Ma Department of Electrcal and Computer Engneerng Unversty of Mnnesota Mnneapols, MN 55455 {cherkass,myq}@ece.umn.edu Abstract:

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

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

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

Chapter 6 Programmng the fnte element method Inow turn to the man subject of ths book: The mplementaton of the fnte element algorthm n computer programs. In order to make my dscusson as straghtforward

More information

Stitching of off-axis sub-aperture null measurements of an aspheric surface

Stitching of off-axis sub-aperture null measurements of an aspheric surface Sttchng of off-axs sub-aperture null measurements of an aspherc surface Chunyu Zhao* and James H. Burge College of optcal Scences The Unversty of Arzona 1630 E. Unversty Blvd. Tucson, AZ 85721 ABSTRACT

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

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

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

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

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

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

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

CMPS 10 Introduction to Computer Science Lecture Notes

CMPS 10 Introduction to Computer Science Lecture Notes CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not

More information

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following.

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following. Complex Numbers The last topc n ths secton s not really related to most of what we ve done n ths chapter, although t s somewhat related to the radcals secton as we wll see. We also won t need the materal

More information

High resolution 3D Tau-p transform by matching pursuit Weiping Cao* and Warren S. Ross, Shearwater GeoServices

High resolution 3D Tau-p transform by matching pursuit Weiping Cao* and Warren S. Ross, Shearwater GeoServices Hgh resoluton 3D Tau-p transform by matchng pursut Wepng Cao* and Warren S. Ross, Shearwater GeoServces Summary The 3D Tau-p transform s of vtal sgnfcance for processng sesmc data acqured wth modern wde

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

Analysis of Continuous Beams in General

Analysis of Continuous Beams in General Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Intro. Iterators. 1. Access

Intro. Iterators. 1. Access Intro Ths mornng I d lke to talk a lttle bt about s and s. We wll start out wth smlartes and dfferences, then we wll see how to draw them n envronment dagrams, and we wll fnsh wth some examples. Happy

More information

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Supervsed vs. Unsupervsed Learnng Up to now we consdered supervsed learnng scenaro, where we are gven 1. samples 1,, n 2. class labels for all samples 1,, n Ths s also

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

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between

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

X- Chart Using ANOM Approach

X- Chart Using ANOM Approach ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are

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

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law)

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law) Machne Learnng Support Vector Machnes (contans materal adapted from talks by Constantn F. Alfers & Ioanns Tsamardnos, and Martn Law) Bryan Pardo, Machne Learnng: EECS 349 Fall 2014 Support Vector Machnes

More information

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints TPL-ware Dsplacement-drven Detaled Placement Refnement wth Colorng Constrants Tao Ln Iowa State Unversty tln@astate.edu Chrs Chu Iowa State Unversty cnchu@astate.edu BSTRCT To mnmze the effect of process

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

Private Information Retrieval (PIR)

Private Information Retrieval (PIR) 2 Levente Buttyán Problem formulaton Alce wants to obtan nformaton from a database, but she does not want the database to learn whch nformaton she wanted e.g., Alce s an nvestor queryng a stock-market

More information

Meta-heuristics for Multidimensional Knapsack Problems

Meta-heuristics for Multidimensional Knapsack Problems 2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,

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

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

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

More information

Categories and Subject Descriptors B.7.2 [Integrated Circuits]: Design Aids Verification. General Terms Algorithms

Categories and Subject Descriptors B.7.2 [Integrated Circuits]: Design Aids Verification. General Terms Algorithms 3. Fndng Determnstc Soluton from Underdetermned Equaton: Large-Scale Performance Modelng by Least Angle Regresson Xn L ECE Department, Carnege Mellon Unversty Forbs Avenue, Pttsburgh, PA 3 xnl@ece.cmu.edu

More information