[lecture notes in computer science] computer vision – eccv 2012 volume 7572 ||
TRANSCRIPT
Lecture Notes in Computer Science 7572Commenced Publication in 1973Founding and Former Series Editors:Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David HutchisonLancaster University, UK
Takeo KanadeCarnegie Mellon University, Pittsburgh, PA, USA
Josef KittlerUniversity of Surrey, Guildford, UK
Jon M. KleinbergCornell University, Ithaca, NY, USA
Alfred KobsaUniversity of California, Irvine, CA, USA
Friedemann MatternETH Zurich, Switzerland
John C. MitchellStanford University, CA, USA
Moni NaorWeizmann Institute of Science, Rehovot, Israel
Oscar NierstraszUniversity of Bern, Switzerland
C. Pandu RanganIndian Institute of Technology, Madras, India
Bernhard SteffenTU Dortmund University, Germany
Madhu SudanMicrosoft Research, Cambridge, MA, USA
Demetri TerzopoulosUniversity of California, Los Angeles, CA, USA
Doug TygarUniversity of California, Berkeley, CA, USA
Gerhard WeikumMax Planck Institute for Informatics, Saarbruecken, Germany
Andrew Fitzgibbon Svetlana LazebnikPietro Perona Yoichi SatoCordelia Schmid (Eds.)
Computer Vision –ECCV 2012
12th European Conference on Computer VisionFlorence, Italy, October 7-13, 2012Proceedings, Part I
13
Volume Editors
Andrew FitzgibbonMicrosoft Research Ltd., Cambridge, CB3 0FB, UKE-mail: [email protected]
Svetlana LazebnikUniversity of North Carolina, Dept. of Computer ScienceChapel Hill, NC 27599, USAE-mail: [email protected]
Pietro PeronaCalifornia Institute of TechnologyPasadena, CA 91125, USAE-mail: [email protected]
Yoichi SatoThe University of Tokyo, Institute of Industrial ScienceTokyo 153-8505, JapanE-mail: [email protected]
Cordelia SchmidINRIA, 38330 Montbonnot, FranceE-mail: [email protected]
ISSN 0302-9743 e-ISSN 1611-3349ISBN 978-3-642-33717-8 e-ISBN 978-3-642-33718-5DOI 10.1007/978-3-642-33718-5
Springer Heidelberg Dordrecht London New York
Library of Congress Control Number: 2012947663
CR Subject Classification (1998): I.4.6, I.4.8, I.4.1-5, I.4.9, I.5.2-4, I.2.10, I.3.5, F.2.2
LNCS Sublibrary: SL 6 – Image Processing, Computer Vision, Pattern Recognition,and Graphics
© Springer-Verlag Berlin Heidelberg 2012
This work is subject to copyright. All rights are reserved, whether the whole or part of the material isconcerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting,reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publicationor parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965,in its current version, and permission for use must always be obtained from Springer. Violations are liableto prosecution under the German Copyright Law.The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,even in the absence of a specific statement, that such names are exempt from the relevant protective lawsand regulations and therefore free for general use.
Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Foreword
The European Conference on Computer Vision is one of the top conferencesfor researchers in this field and is held biennially in alternation with the Inter-national Conference on Computer Vision. It was first held in 1990 in Antibes(France) with subsequent conferences in Santa Margherita Ligure (Italy) in 1992,Stockholm (Sweden) in 1994, Cambridge (UK) in 1996, Freiburg (Germany) in1998, Dublin (Ireland) in 2000, Copenhagen (Denmark) in 2002, Prague (CzechRepublic) in 2004, Graz (Austria) in 2006, Marseille (France) in 2008, and Her-aklion (Greece) in 2010. To our great delight, the 12th conference was held inFlorence, Italy.
ECCV has an established tradition of very high scientific quality and anoverall duration of one week. ECCV 2012 began with a keynote lecture from thehonorary chair, Tomaso Poggio. The main conference followed over four dayswith 40 orals, 368 posters, 22 demos, and 12 industrial exhibits. There werealso 9 tutorials and 21 workshops held before and after the main event. For thisevent we introduced some novelties. These included innovations in the reviewpolicy, the publication of a conference booklet with all paper abstracts and thefull video recording of oral presentations.
This conference is the result of a great deal of hard work by many people,who have been working enthusiastically since our first meetings in 2008. We areparticularly grateful to the Program Chairs, who handled the review of about1500 submissions and co-ordinated the efforts of over 50 area chairs and about1000 reviewers (see details of the process in their preface to the proceedings). Weare also indebted to all the other chairs who, with the support of our researchteams (names listed below), diligently helped us manage all aspects of the mainconference, tutorials, workshops, exhibits, demos, proceedings, and web presence.Finally we thank our generous sponsors and Consulta Umbria for handling theregistration of delegates and all financial aspects associated with the conference.
We hope you enjoyed ECCV 2012. Benvenuti a Firenze!
October 2012 Roberto CipollaCarlo Colombo
Alberto Del Bimbo
Preface
Welcome to the proceedings of the 2012 European Conference on ComputerVision in Florence, Italy! We received 1437 complete submissions, the largestnumber of submissions in the history of ECCV. Forty papers were selected fororal presentation and 368 papers for poster presentation, resulting in acceptancerates of 2.8% for oral, 25.6% for poster, and 28.4% in total.
The following is a brief description of the review process. After the submis-sion deadline, each paper was assigned to one of 54 area chairs (28 from Europe,21 from the USA and Canada, and 4 from Asia) with the help of the Toronto Pa-per Matching System (TMS). TMS, developed by Laurent Charlin and RichardZemel, is beginning to be used by an increasing number of conferences, includingNIPS, ICML, and CVPR. To ensure the best possible assignment of papers toarea chairs, the program chairs manually selected several area chair candidatesfor each paper based on the suggestions generated by TMS. After automatic loadbalancing and conflict resolution, each AC was finally assigned approximately30 papers closely matching their expertise.
Area chairs then made reviewer suggestions (an average of seven per paper),which were load-balanced and conflict-resolved, giving 3 reviewers for each pa-per and a maximum of 11 papers per reviewer. The ACs were assisted in thisprocess by TMS, which was also used for automatically selecting potential re-viewers, matching each submitted paper based on the reviewers’ representativepublications. These suggestions came from a pool of potential reviewers com-posed from names of people who have reviewed for recent vision conferences,self-nominations (any member of the community could fill out a form on theECCV website asking to be a reviewer), and nominations by ACs. From an ini-tial pool of 863 reviewers, 638 ended up reviewing at least one paper. This wasthe first time that TMS had been used this extensively in the review processfor a vision conference (CVPR 2012 used a restricted version of the system forassigning papers to area chairs), and in the end, we were very pleased with itsperformance. An important improvement over previous conferences was that ini-tial reviewer suggestions were generated entirely in parallel by the ACs, withoutthe “race” for good reviewers that the previous methods have implicitly encour-aged. Area chairs were then given the opportunity to correct infelicities in theload balancing before the final list was generated. We extend our heartfelt thanksto the area chairs, who participated vigorously in this process, to maximize thequality of the review assignments.
For the decision process, we introduced one major innovation. We replacedthe physical area chair meeting and the conventional AC buddy system with vir-tual meetings of AC triplets (this system was first tried out for BMVC 2011 andfound to work very well). After the conclusion of the review, rebuttal, and discus-sion periods, the AC triplets met on the phone or on Skype (and, in just one case,
VIII Preface
in person), jointly discussed all their papers, and made acceptance/rejection de-cisions. Thus, the reviews and consolidation reports for each paper were carefullyexamined by three ACs, ensuring a fair and thorough assessment. A programchair assisted in each AC triplet meeting to maintain the consistency in the de-cision process and to provide any necessary support. Furthermore, each tripletrecommended a small number of top-ranked papers (typically one to three) fororal presentation, and the program chairs took these candidates and made thefinal oral vs. poster decisions.
Double-blind reviewing policies were strictly maintained throughout the en-tire process – neither the area chairs nor the reviewers knew the identity of theauthors, and the authors did not know the identity of the reviewers and ACs.Based on feedback from authors, reviewers, and area chairs, we believe we suc-cessfully maintained the integrity of the paper selection process, and we are veryexcited about the quality of the resulting program.
We wish to thank everyone involved for their time and dedication to makingthe ECCV 2012 program possible. The success of ECCV 2012 entirely relied onthe time and effort invested by the authors into producing high-quality research,on the care taken by the reviewers in writing thorough and professional reviews,and on the commitment by the area chairs to reconciling the reviews and writingdetailed and precise consolidation reports. We also wish to thank the generalchairs, Roberto Cipolla, Carlo Colombo, and Alberto Del Bimbo, and the otherorganizing committee members for their top-notch handling of the event.
Finally, we would like to commemorate Mark Everingham, whose untimelydeath has shocked and saddened the entire vision community. Mark was an areachair for ECCV and also an organizer for one of the workshops; his hard work anddedication were absolutely essential in enabling us to put together a high-qualityconference program. We salute his record of exemplary service and intellectualcontributions to the discipline of computer vision. Mark, you will be missed!
October 2012 Andrew FitzgibbonSvetlana Lazebnik
Pietro PeronaYoichi Sato
Cordelia Schmid
Organization
General Chairs
Roberto Cipolla University of Cambridge, UKCarlo Colombo University of Florence, ItalyAlberto Del Bimbo University of Florence, Italy
Program Coordinator
Pietro Perona California Institute of Technology, USA
Program Chairs
Andrew Fitzgibbon Microsoft Research, Cambridge, UKSvetlana Lazebnik University of Illinois at Urbana-Champaign, USAYoichi Sato The University of Tokyo, JapanCordelia Schmid INRIA, Grenoble, France
Honorary Chair
Tomaso Poggio Massachusetts Institute of Technology, USA
Tutorial Chairs
Emanuele Trucco University of Dundee, UKAlessandro Verri University of Genoa, Italy
Workshop Chairs
Andrea Fusiello University of Udine, ItalyVittorio Murino Istituto Italiano di Tecnologia, Genoa, Italy
Demonstration Chair
Rita Cucchiara University of Modena and Reggio Emilia, Italy
Industrial Liaison Chair
Bjorn Stenger Toshiba Research Europe, Cambridge, UK
Web Chair
Marco Bertini University of Florence, Italy
X Organization
Publicity Chairs
Terrance E. Boult University of Colorado at Colorado Springs, USATat Jen Cham Nanyang Technological University, SingaporeMarcello Pelillo University Ca’ Foscari of Venice, Italy
Publication Chair
Massimo Tistarelli University of Sassari, Italy
Video Processing Chairs
Sebastiano Battiato University of Catania, ItalyGiovanni M. Farinella University of Catania, Italy
Travel Grants Chair
Luigi Di Stefano University of Bologna, Italy
Travel Visa Chair
Stefano Berretti University of Florence, Italy
Local Committee Chair
Andrew Bagdanov MICC, Florence, Italy
Local Committee
Lamberto BallanLaura BenassiMarco FanfaniAndrea FerracaniClaudio GuidaLea Landucci
Giuseppe LisantiIacopo MasiFabio PazzagliaFederico PerniciLorenzo SeidenariGiuseppe Serra
Area ChairsSimon Baker Microsoft Research, USAHorst Bischof Graz University of Technology, AustriaMichael Black Max Planck Institute, GermanyRichard Bowden University of Surrey, UKMichael S. Brown National University of Singapore, SingaporeJoachim Buhmann ETH Zurich, SwitzerlandAlyosha Efros Carnegie Mellon University, USAMark Everingham University of Leeds, UKPedro Felzenszwalb Brown University, USA
Organization XI
Rob Fergus New York University, USAVittorio Ferrari ETH Zurich, SwitzerlandDavid Fleet University of Toronto, CanadaDavid Forsyth University of Illinois at Urbana-Champaign, USAKristen Grauman University of Texas at Austin, USAMartial Hebert Carnegie Mellon University, USAAaron Hertzmann University of Toronto, CanadaDerek Hoiem University of Illinois at Urbana-Champaign, USAKatsushi Ikeuchi The University of Tokyo, JapanMichal Irani The Weizmann Institute of Science, IsraelDavid Jacobs University of Maryland, USASing Bing Kang Microsoft Research, USADavid Kriegman University of California, San Diego, USAKyros Kutulakos University of Toronto, CanadaChristof Lampert Institute of Science and Technology, AustriaIvan Laptev INRIA, FranceVictor Lempitsky Yandex, RussiaSteve Lin Microsoft Research, ChinaJitendra Malik University of California, Berkeley, USAJirı Matas Czech Technical University, Czech RepublicYasuyuki Matsushita Microsoft Research, ChinaTomas Pajdla Czech Technical University, Czech RepublicPatrick Perez Thomson-Technicolor, FranceMarc Pollefeys ETH Zurich, SwitzerlandJean Ponce Ecole Normale Superieure, FranceLong Quan Hong Kong Univ. of Science and Technology, ChinaDeva Ramanan University of California, Irvine, USAStefan Roth TU Darmstadt, GermanyCarsten Rother Microsoft Research, UKYoav Schechner Technion, IsraelBernt Schiele Max Planck Institute, GermanyChristoph Schnorr University of Heidelberg, GermanyStan Sclaroff University of Boston, USAJosef Sivic Ecole Normale Superieure, FrancePeter Sturm INRIA, FranceCarlo Tomasi Duke University, USAAntonio Torralba Massachusetts Institute of Technology, USATinne Tuytelaars University of Leuven, BelgiumJakob Verbeek INRIA, FranceYair Weiss The Hebrew University of Jerusalem, IsraelChristopher Williams University of Edinburgh, UKRamin Zabih Cornell University, USALihi Zelnik Technion, IsraelAndrew Zisserman University of Oxford, UKLarry Zitnick Microsoft Research, USA
XII Organization
Reviewers
Vitaly AblavskyLourdes AgapitoSameer AgarwalAmit AgrawalKarteek AlahariKarim AliSaad AliS. Ali EslamiDaniel AliagaNeil AlldrinMarina AltermanJose M. AlvarezBrian AmbergCosmin AncutiJuan AndradeMykhaylo AndrilukaAnton AndriyenkoElli AngelopoulouRoland AngstRelja ArandjelovicHelder AraujoPablo ArbelaezAntonis ArgyrosKalle AstromVassilis AthitsosJosep AulinasShai AvidanTamar AvrahamYannis AvrithisYusuf AytarLuca BallanLamberto BallanAtsuhiko BannoYinzge BaoAdrian BarbuNick BarnesJoao Pedro BarretoAdrien BartoliArslan BasharatDhruv BatraSebastiano BattiatoJean-Charles BazinFethallah Benmansour
Alexander BergTamara BergHakan BilenMatthew BlaschkoMichael BleyerLiefeng BoDaniele BorghesaniTerrance BoultLubomir BourdevY-Lan BoureauKevin BowyerEdmond BoyerSteven BransonMathieu BredifWilliam BrendelMichael BronsteinGabriel BrostowMatthew BrownThomas BroxMarcus BrubakerDarius BurschkaTiberio CaetanoBarbara CaputoStefan CarlssonGustavo CarneiroJoao CarreiraYaron CaspiCarlos CastilloJan CechTurgay CelikAyan ChakrabartiTat Jen ChamAntoni ChanManmohan ChandrakerMing-Ching ChangLin ChenXilin ChenDaozheng ChenWen-Huang ChengYuan ChengTat-Jun ChinHan-Pang ChiuMinsu Cho
Tae ChoeOndrej ChumAlbert C.S. ChungJohn CollomosseTim CootesFlorent Couzine-DevyDavid CrandallKeenan CraneAntonio CriminisiShengyang DaiDima DamenLarry DavisAndrew DavisonFernando De la TorreJoost de WeijerTeofilo deCamposVincent DelaitreAmael DelaunoyAndrew DelongDavid DemirdjianJia DengJoachim DenzlerKonstantinos DerpanisChaitanya DesaiThomas DeselaersFrederic DevernayThang DinhSantosh Kumar DivvalaPiotr DollarJustin DomkeGianfranco DorettoMatthijs DouzeTom DrummondLixin DuanOlivier DuchenneZoran DuricPinar DuyguluCharles DyerSandra EbertMichael EladJames ElderEhsan ElhamifarIan Endres
Organization XIII
Olof EnqvistSergio EscaleraJialue FanBin FanGabriele FanelliYi FangAli FarhadiRyan FarrellRaanan FattalPaolo FavaroRogerio FerisSanja FidlerRobert FisherPierre Fite-GeorgelBoris FlachFrancois FleuretWolfgang ForstnerAndrea FossatiCharless FowlkesJan-Michael FrahmJean-Sebastien FrancoFriedrich FraundorferWilliam FreemanOren FreifeldMario FritzYasutaka FurukawaAndrea FusielloAdrien GaidonJuergen GallAndrew GallagherSimone GaspariniPeter GehlerYakup GencLeifman GeorgeGuido GerigChristopher GeyerAbhijeet GhoshAndrew GilbertRoss GirshickMartin GodecRoland GoeckeMichael GoeseleSiome GoldensteinBastian GoldlueckeShaogang Gong
German GonzalezRaghuraman GopalanAlbert GordoLena GorelickPaulo GotardoStephen GouldHelmut GrabnerEtienne GrossmannMatthias GrundmannJinwei GuSteve GuLi GuanPeng GuanMatthieu GuillauminJean-Yves GuillemautRuiqi GuoGuodong GuoAbhinav GuptaMohit GuptaTony HanBohyung HanMei HanEdwin HancockJari HannukselaKenji HaraTatsuya HaradaDaniel HarariZaid HarchaouiStefan HarmelingSøren HaubergMichal HavlenaJames HaysXuming HeKaiming HeVarsha HedauNicolas HeessYong HeoAdrian HiltonStefan HinterstoisserMinh HoaiJesse HoeyAnthony HoogsJoachim HorneggerAlexander HornungEdward Hsiao
Wenze HuChangbo HuGang HuaXinyu HuangRui HuangWonjun HwangIchiro IdeJuan IglesiasIvo IhrkeNazli Ikizler-CinbisSlobodan IlicIgnazio InfantinoMichael IsardHerve JegouC.V. JawaharRodolphe JenattonHueihan JhuangQiang JiJiaya JiaHongjun JiaYong-Dian JianHao JiangZhuolin JiangShuqiang JiangSam JohnsonAnne JorstadNeel JoshiArmand JoulinFrederic JurieIoannis KakadiarisZdenek KalalJoni-K. KamarainenKenichi KanataniAtul KanaujiaAshish KapoorJorg KappesLeonid KarlinskyKevin Karschkoray kavukcuogluRei KawakamiHiroshi KawasakiVerena KaynigQifa KeIra Kemelmacher-
Shlizerman
XIV Organization
Aditya KhoslaTae-Kyun KimJaechul KimSeon Joo KimKris KitaniJyri KivinenHedvig KjellstromJan KnoppKevin KoeserPushmeet KohliNikos KomodakisKurt KonoligeFilip KorcAndreas KoschanAdriana KovashkaJosip KrapacDilip KrishnanZuzana KukelovaNeeraj KumarM. Pawan KumarJunghyun KwonDongjin KwonJunseok KwonFlorent LafargeShang-Hong LaiJean-Francois LalondeMichael LangerDouglas LanmanDiane LarlusLongin Jan LateckiErik Learned-MillerSeungkyu LeeKyong Joon LeeHonglak LeeYong Jae LeeBastian LeibeIdo LeichterFrank LenzenMatt LeottaVincent LepetitAnat LevinMaxime LhuillierRui LiStan LiHongsheng Li
Ruonan LiHongdong LiFeng LiYunpeng LiFuxin LiLi-Jia LiZicheng LiaoShengcai LiaoJongwoo LimJoseph LimYen-Yu LinDahua LinDaniel LinHaibin LingJames LittleCe LiuXiaobai LiuMing-Yu LiuXiaoming LiuTyng-Luh LiuYunlong LiuWei LiuJingen LiuMarcus LiwickiLiliana Lo PrestiRoberto Lopez-SastreJiwen LuZheng LuLe LuSimon LuceyJulien MairalMichael MaireSubhransu MajiYasushi MakiharaDimitrios MakrisTomasz MalisiewiczJiri MatasIain MatthewsStefano MattocciaThomas MauthnerSteven MaybankWalterio Mayol-CuevasScott McCloskeyStephen McKennaGerard Medioni
Jason MeltzerTalya MeltzerHeydi Mendez-VazquezThomas MensinkFabrice MichelBranislav MicusikKrystian MikolajczykNiloy MitraAnurag MittalPhilippos MordohaiFrancesc Moreno-NoguerGreg MoriBryan MorseYadong MuYasuhiro MukaigawaLopamudra MukherjeeAndreas MullerJane MulliganDaniel MunozA. MurilloCarlo MuttoHajime NagaharaVinay NamboodiriSrinivasa NarasimhanFabian NaterShawn NewsamKai NiFeiping NieJuan Carlos NieblesClaudia NieuwenhuisKo NishinoSebastian NowozinJean-Marc OdobezPeter O’DonovanSangmin OhTakeshi OishiTakahiro OkabeTakayuki OkataniAude OlivaCarl OlssonBjorn OmmerEng-Jon OngAnton OsokinMatthew O’TooleMustafa Ozuysal
Organization XV
Maja PanticCaroline PantofaruGeorge PapandreouToufiq ParagVasu ParameswaranDevi ParikhSylvain ParisMinwoo ParkDennis ParkIoannis PatrasIoannis PavlidisNadia PayetKim PedersenOfir PeleShmuel PelegYigang PengAmitha PereraFlorent PerronninAdrian PeterMaria PetrouPatrick PeursumTomas PfisterJames PhilbinJustus PiaterHamed PirsiavashRobert PlessThomas PockGerard Pons-MollRonald PoppeFatih PorikliMukta PrasadAndrea PratiJerry PrinceNicolas PugeaultNovi QuadriantoVincent RabaudRahul RaguramSrikumar RamalingamNarayanan RamanathanMarc’Aurelio RanzatoKonstantinos
RapantzikosNikhil RasiwasiaMohammad RastegariJames Rehg
Erik ReinhardXiaofeng RenChristoph RhemannAntonio Robles-KellyEmanuele RodolaMikel RodriguezAntonio Rodriguez-
SanchezMarcus RohrbachJavier RomeroCharles RosenbergBodo RosenhahnSamuel Rota BuloPeter RothAmit Roy-ChowdhuryDmitry RudoyOlga RussakovskyBryan RussellChris RussellRadu RusuMichael RyooMohammad SadeghiKate SaenkoAmir SaffariAlbert SalahMathieu SalzmannDimitris SamarasAswin
SankaranarayananBenjamin SappRadim SaraScott SatkinImari SatoEric SaundDaniel ScharsteinWalter ScheirerKevin ScheltenRaimondo SchettiniKonrad SchindlerJoseph SchlechtFrank SchmidtUwe SchmidtFlorian SchroffRodolphe SepulchreUri Shalit
Shiguang ShanLing ShaoAbhishek SharmaEli ShechtmanYaser SheikhAlexander ShekhovtsovIlan ShimshoniTakaaki ShiratoriJamie ShottonNitesh ShroffZhangzhang SiLeonid SigalNathan SilbermanKaren SimonyanVivek SinghVikas SinghManeesh SinghSudipta SinhaGreg SlabaughArnold SmeuldersCristian SminchisescuWilliam A. P. SmithKevin SmithNoah SnavelyCees SnoekMichal SofkaQi SongXuan SongAnuj SrivastavaMichael StarkBjorn StengerYu SuYusuke SuganoJu SunMin SunDeqing SunJian SunDavid SuterYohay SwirskiRick SzeliskiYuichi TaguchiYu-Wing TaiJun TakamatsuHugues TalbotRobby Tan
XVI Organization
Xiaoou TangMarshall TappenJonathan TaylorChristian TheobaltTai-Peng TianJoseph TigheRadu TimofteSinisa TodorovicFederico TombariAkihiko ToriiDuan TranTali TreibitzBill TriggsNhon TrinhIvor TsangYanghai TsinAggeliki TsoliZhuowen TuPavan TuragaAmbrish TyagiMartin UrschlerRaquel UrtasunJan van GemertDaniel VaqueroAndrea VedaldiAshok VeeraraghavanOlga VekslerAlexander VezhnevetsSara VicenteSudheendra
VijayanarasimhanPascal VincentCarl VondrickChaohui WangYang WangJue WangHanzi Wang
Song WangGang WangHongcheng WangJingdong WangLu WangYueming WangRuiping WangKai WangAlexander WeissAndreas WendelManuel WerlbergerTomas WernerGordon WetzsteinYonatan WexlerOliver WhyteRichard WildesOliver WilliamsThomas WindheuserDavid WipfKwan-Yee K. WongJohn WrightShandong WuYi WuChangchang WuJianxin WuYing WuJonas WulffJing XiaoJianxiong XiaoWei XuLi XuYong XuYi XuYasushi YagiTakayoshi YamashitaMing YangMing-Hsuan Yang
Qingxiong YangJinfeng YangWeilong YangRuigang YangJianchao YangYi YangBangpeng YaoAngela YaoMohammad YaqubLijun YinKuk-Jin YoonTianli YuQian YuLu YuanXiaotong YuanChristopher ZachStefanos ZafeiriouAndrei ZaharescuMatthew ZeilerYun ZengGuofeng ZhangLi ZhangLei ZhangXinhua ZhangShaoting ZhangJianguo ZhangYing ZhengS. Kevin ZhouChangyin ZhouShaojie ZhuoTodd ZicklerDarko ZikicHenning ZimmerDaniel ZoranSilvia Zuffi
Sponsoring Companies and Institutions
Gold Sponsors
Silver Sponsors
Bronze Sponsors
Institutional Sponsors
Table of Contents
Poster Session 1
Lie Bodies: A Manifold Representation of 3D Human Shape . . . . . . . . . . . 1Oren Freifeld and Michael J. Black
Worldwide Pose Estimation Using 3D Point Clouds . . . . . . . . . . . . . . . . . . 15Yunpeng Li, Noah Snavely, Dan Huttenlocher, and Pascal Fua
Improved Reconstruction of Deforming Surfaces by Cancelling AmbientOcclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Thabo Beeler, Derek Bradley, Henning Zimmer, and Markus Gross
On the Statistical Determination of Optimal Camera Configurations inLarge Scale Surveillance Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Junbin Liu, Clinton Fookes, Tim Wark, and Sridha Sridharan
The Scale of Geometric Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Geoffrey Oxholm, Prabin Bariya, and Ko Nishino
Efficient Articulated Trajectory Reconstruction Using DynamicProgramming and Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Jack Valmadre, Yingying Zhu, Sridha Sridharan, and Simon Lucey
Object Co-detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86Sid Yingze Bao, Yu Xiang, and Silvio Savarese
Morphable Displacement Field Based Image Matching for FaceRecognition across Pose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Shaoxin Li, Xin Liu, Xiujuan Chai, Haihong Zhang,Shihong Lao, and Shiguang Shan
Combining Per-frame and Per-track Cues for Multi-person ActionRecognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Sameh Khamis, Vlad I. Morariu, and Larry S. Davis
Joint Image and Word Sense Discrimination for Image Retrieval . . . . . . . 130Aurelien Lucchi and Jason Weston
Script Data for Attribute-Based Recognition of Composite Activities . . . 144Marcus Rohrbach, Michaela Regneri, Mykhaylo Andriluka,Sikandar Amin, Manfred Pinkal, and Bernt Schiele
Undoing the Damage of Dataset Bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158Aditya Khosla, Tinghui Zhou, Tomasz Malisiewicz,Alexei A. Efros, and Antonio Torralba
XX Table of Contents
Dog Breed Classification Using Part Localization . . . . . . . . . . . . . . . . . . . . . 172Jiongxin Liu, Angjoo Kanazawa, David Jacobs, and Peter Belhumeur
A Dictionary Learning Approach for Classification: Separating theParticularity and the Commonality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Shu Kong and Donghui Wang
Learning to Efficiently Detect Repeatable Interest Points in DepthData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Stefan Holzer, Jamie Shotton, and Pushmeet Kohli
Effective Use of Frequent Itemset Mining for Image Classification . . . . . . 214Basura Fernando, Elisa Fromont, and Tinne Tuytelaars
Efficient Discriminative Projections for Compact Binary Descriptors . . . . 228Tomasz Trzcinski and Vincent Lepetit
Descriptor Learning Using Convex Optimisation . . . . . . . . . . . . . . . . . . . . . 243Karen Simonyan, Andrea Vedaldi, and Andrew Zisserman
Bottom-Up Perceptual Organization of Images into Object PartHypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Maruthi Narayanan and Benjamin Kimia
Match Graph Construction for Large Image Databases . . . . . . . . . . . . . . . . 272Kwang In Kim, James Tompkin, Martin Theobald, Jan Kautz, andChristian Theobalt
Modeling Complex Temporal Composition of Actionlets for ActivityPrediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Kang Li, Jie Hu, and Yun Fu
Learning Human Interaction by Interactive Phrases . . . . . . . . . . . . . . . . . . 300Yu Kong, Yunde Jia, and Yun Fu
Learning to Recognize Daily Actions Using Gaze . . . . . . . . . . . . . . . . . . . . . 314Alireza Fathi, Yin Li, and James M. Rehg
Gait Recognition by Ranking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328Raul Martın-Felez and Tao Xiang
Semi-intrinsic Mean Shift on Riemannian Manifolds . . . . . . . . . . . . . . . . . . 342Rui Caseiro, Joao F. Henriques, Pedro Martins, and Jorge Batista
Efficient Nonlocal Regularization for Optical Flow . . . . . . . . . . . . . . . . . . . . 356Philipp Krahenbuhl and Vladlen Koltun
Fast Fusion Moves for Multi-model Estimation . . . . . . . . . . . . . . . . . . . . . . . 370Andrew Delong, Olga Veksler, and Yuri Boykov
Table of Contents XXI
Approximate MRF Inference Using Bounded Treewidth Subgraphs . . . . . 385Alexander Fix, Joyce Chen, Endre Boros, and Ramin Zabih
Recursive Bilateral Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399Qingxiong Yang
Accelerated Large Scale Optimization by Concomitant Hashing . . . . . . . . 414Yadong Mu, John Wright, and Shih-Fu Chang
Graph Degree Linkage: Agglomerative Clustering on a DirectedGraph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
Wei Zhang, Xiaogang Wang, Deli Zhao, and Xiaoou Tang
Supervised Earth Mover’s Distance Learning and Its Computer VisionApplications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Fan Wang and Leonidas J. Guibas
Global Optimization of Object Pose and Motion from a Single RollingShutter Image with Automatic 2D-3D Matching . . . . . . . . . . . . . . . . . . . . . 456
Ludovic Magerand, Adrien Bartoli, Omar Ait-Aider, andDaniel Pizarro
Online Learning of Linear Predictors for Real-Time Tracking . . . . . . . . . . 470Stefan Holzer, Marc Pollefeys, Slobodan Ilic, David Joseph Tan, andNassir Navab
Online Learned Discriminative Part-Based Appearance Models forMulti-human Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484
Bo Yang and Ram Nevatia
Exposure Stacks of Live Scenes with Hand-Held Cameras . . . . . . . . . . . . . 499Jun Hu, Orazio Gallo, and Kari Pulli
Dual-Force Metric Learning for Robust Distracter-Resistant Tracker . . . . 513Zhibin Hong, Xue Mei, and Dacheng Tao
Shape and Reflectance from Natural Illumination . . . . . . . . . . . . . . . . . . . . 528Geoffrey Oxholm and Ko Nishino
Frequency Analysis of Transient Light Transport with Applications inBare Sensor Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
Di Wu, Gordon Wetzstein, Christopher Barsi, Thomas Willwacher,Matthew O’Toole, Nikhil Naik, Qionghai Dai, Kyros Kutulakos, andRamesh Raskar
Nonuniform Lattice Regression for Modeling the Camera ImagingPipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 556
Hai Ting Lin, Zheng Lu, Seon Joo Kim, and Michael S. Brown
XXII Table of Contents
Context-Based Automatic Local Image Enhancement . . . . . . . . . . . . . . . . . 569Sung Ju Hwang, Ashish Kapoor, and Sing Bing Kang
Segmentation with Non-linear Regional Constraints via Line-SearchCuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583
Lena Gorelick, Frank R. Schmidt, Yuri Boykov,Andrew Delong, and Aaron Ward
Hausdorff Distance Constraint for Multi-surface Segmentation . . . . . . . . . 598Frank R. Schmidt and Yuri Boykov
Background Subtraction Using Low Rank and Group SparsityConstraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612
Xinyi Cui, Junzhou Huang, Shaoting Zhang, andDimitris N. Metaxas
Free Hand-Drawn Sketch Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 626Zhenbang Sun, Changhu Wang, Liqing Zhang, and Lei Zhang
Auto-Grouped Sparse Representation for Visual Analysis . . . . . . . . . . . . . 640Jiashi Feng, Xiaotong Yuan, Zilei Wang, Huan Xu, andShuicheng Yan
Oral Session 1: Geometry: Theory and application
A QCQP Approach to Triangulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 654Chris Aholt, Sameer Agarwal, and Rekha Thomas
Reconstructing the World’s Museums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668Jianxiong Xiao and Yasutaka Furukawa
Poster Session 2
Background Inpainting for Videos with Dynamic Objects and aFree-Moving Camera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682
Miguel Granados, Kwang In Kim, James Tompkin, Jan Kautz, andChristian Theobalt
Optimal Templates for Nonrigid Surface Reconstruction . . . . . . . . . . . . . . 696Markus Moll and Luc Van Gool
Learning Domain Knowledge for Facade Labelling . . . . . . . . . . . . . . . . . . . . 710Dengxin Dai, Mukta Prasad, Gerhard Schmitt, and Luc Van Gool
Simultaneous Shape and Pose Adaption of Articulated Models UsingLinear Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 724
Matthias Straka, Stefan Hauswiesner, Matthias Ruther, andHorst Bischof
Table of Contents XXIII
Robust Fitting for Multiple View Geometry . . . . . . . . . . . . . . . . . . . . . . . . . 738Olof Enqvist, Erik Ask, Fredrik Kahl, and Kalle Astrom
Improving Image-Based Localization by Active CorrespondenceSearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 752
Torsten Sattler, Bastian Leibe, and Leif Kobbelt
From Meaningful Contours to Discriminative Object Shape . . . . . . . . . . . . 766Pradeep Yarlagadda and Bjorn Ommer
A Particle Filter Framework for Contour Detection . . . . . . . . . . . . . . . . . . . 780Nicolas Widynski and Max Mignotte
TriCoS: A Tri-level Class-Discriminative Co-segmentation Method forImage Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 794
Yuning Chai, Esa Rahtu, Victor Lempitsky, Luc Van Gool, andAndrew Zisserman
Multi-view Discriminant Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 808Meina Kan, Shiguang Shan, Haihong Zhang, Shihong Lao, andXilin Chen
Multi-scale Patch Based Collaborative Representation for FaceRecognition with Margin Distribution Optimization . . . . . . . . . . . . . . . . . . 822
Pengfei Zhu, Lei Zhang, Qinghua Hu, and Simon C.K. Shiu
Object Detection Using Strongly-Supervised Deformable Part Models . . . 836Hossein Azizpour and Ivan Laptev
Efficient Misalignment-Robust Representation for Real-Time FaceRecognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 850
Meng Yang, Lei Zhang, and David Zhang
Monocular Object Detection Using 3D Geometric Primitives . . . . . . . . . . 864Peter Carr, Yaser Sheikh, and Iain Matthews
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 879