Noisy Depth Maps Fusion for Multiview Stereo Via Matrix Completion

This paper introduces a general framework to fuse noisy point clouds from multiview images of the same object. We solve this classical vision problem using a newly emerging signal processing technique known as matrix completion. With this framework, we construct the initial incomplete matrix from th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE journal of selected topics in signal processing Jg. 6; H. 5; S. 566 - 582
Hauptverfasser: Deng, Yue, Liu, Yebin, Dai, Qionghai, Zhang, Zengke, Wang, Yao
Format: Journal Article
Sprache:Englisch
Veröffentlicht: IEEE 01.09.2012
Schlagworte:
ISSN:1932-4553, 1941-0484
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract This paper introduces a general framework to fuse noisy point clouds from multiview images of the same object. We solve this classical vision problem using a newly emerging signal processing technique known as matrix completion. With this framework, we construct the initial incomplete matrix from the observed point clouds by all the cameras, with the invisible points by any camera denoted as unknown entries. The observed points corresponding to the same object point are put into the same row. When properly completed, the recovered matrix should have rank one, since all the columns describe the same object. Therefore, an intuitive approach to complete the matrix is by minimizing its rank subject to consistency with observed entries. In order to improve the fusion accuracy, we propose a general noisy matrix completion method called log-sum penalty completion (LPC), which is particularly effective in removing outliers. Based on the majorization-minimization algorithm (MM), the non-convex LPC problem is effectively solved by a sequence of convex optimizations. Experimental results on both point cloud fusion and MVS reconstructions verify the effectiveness of the proposed framework and the LPC algorithm.
AbstractList This paper introduces a general framework to fuse noisy point clouds from multiview images of the same object. We solve this classical vision problem using a newly emerging signal processing technique known as matrix completion. With this framework, we construct the initial incomplete matrix from the observed point clouds by all the cameras, with the invisible points by any camera denoted as unknown entries. The observed points corresponding to the same object point are put into the same row. When properly completed, the recovered matrix should have rank one, since all the columns describe the same object. Therefore, an intuitive approach to complete the matrix is by minimizing its rank subject to consistency with observed entries. In order to improve the fusion accuracy, we propose a general noisy matrix completion method called log-sum penalty completion (LPC), which is particularly effective in removing outliers. Based on the majorization-minimization algorithm (MM), the non-convex LPC problem is effectively solved by a sequence of convex optimizations. Experimental results on both point cloud fusion and MVS reconstructions verify the effectiveness of the proposed framework and the LPC algorithm.
Author Wang, Yao
Zhang, Zengke
Deng, Yue
Liu, Yebin
Dai, Qionghai
Author_xml – sequence: 1
  givenname: Yue
  surname: Deng
  fullname: Deng, Yue
  organization: Automation department, Tsinghua National Laboratory for information science and technology (TNList), Tsinghua University, Beijing, China
– sequence: 2
  givenname: Yebin
  surname: Liu
  fullname: Liu, Yebin
  organization: Automation department, Tsinghua National Laboratory for information science and technology (TNList), Tsinghua University, Beijing, China
– sequence: 3
  givenname: Qionghai
  surname: Dai
  fullname: Dai, Qionghai
  organization: Automation department, Tsinghua National Laboratory for information science and technology (TNList), Tsinghua University, Beijing, China
– sequence: 4
  givenname: Zengke
  surname: Zhang
  fullname: Zhang, Zengke
  organization: Automation department, Tsinghua National Laboratory for information science and technology (TNList), Tsinghua University, Beijing, China
– sequence: 5
  givenname: Yao
  surname: Wang
  fullname: Wang, Yao
  organization: Dept. of Electrical and Computer Engineering, Polytechnic Institute of NYU, Brooklyn
BookMark eNp9kMtOwzAQRS1UJNrCD8DGP5DicWzHXkKhPNQCUgvbKHEcYZTWke0C_XsSiliwYDWzuGfm6ozQYOM2BqFTIBMAos7vl6vl04QSoBMKirOMHqAhKAYJYZIN-j2lCeM8PUKjEN4I4ZkANkSXD86GHb4ybXzFi6INeLYN1m1w7TxebJto3635wMtovHH4xRZdKHr7iadu3TYmdtFjdFgXTTAnP3OMnmfXq-ltMn-8uZtezBNNRRYTzgWtNa9AyVIYWkJZ6bTUkimZGlaVjJZdPTCaGMIEVJIQWivgWtSgVQHpGNH9Xe1dCN7UeevtuvC7HEjeW8i_LeS9hfzHQgfJP5C2sehrR1_Y5n_0bI9aY8zvLwEyE1KlXx-5bMc
CODEN IJSTGY
CitedBy_id crossref_primary_10_1109_TCYB_2016_2547941
crossref_primary_10_1007_s13042_017_0665_9
crossref_primary_10_1016_j_neucom_2016_11_068
crossref_primary_10_1109_TMM_2018_2839911
crossref_primary_10_1186_s12938_015_0116_3
crossref_primary_10_3389_feart_2023_1285622
crossref_primary_10_1109_ACCESS_2019_2894533
crossref_primary_10_1016_j_neucom_2015_08_032
crossref_primary_10_1109_TCSVT_2024_3385360
crossref_primary_10_1109_TNNLS_2016_2522401
crossref_primary_10_1109_TIP_2014_2298976
crossref_primary_10_1016_j_jvcir_2016_01_009
crossref_primary_10_1109_TII_2015_2404299
crossref_primary_10_1109_TIP_2017_2781425
crossref_primary_10_1049_iet_ipr_2013_0098
crossref_primary_10_1109_LSP_2014_2364612
crossref_primary_10_1007_s11042_016_3621_x
crossref_primary_10_1016_j_neucom_2015_09_109
crossref_primary_10_1016_j_jvcir_2016_11_005
crossref_primary_10_1109_TCYB_2013_2272321
crossref_primary_10_1109_TIP_2016_2588330
crossref_primary_10_1109_TNNLS_2012_2235082
crossref_primary_10_1371_journal_pone_0092069
crossref_primary_10_1109_TCYB_2014_2300192
Cites_doi 10.1109/CVPR.2008.4587792
10.1109/ICCV.2007.4408933
10.1007/s10107-009-0306-5
10.1561/2200000016
10.1109/TPAMI.2009.161
10.1109/CVPR.2005.227
10.1145/138859.138867
10.1109/CVPR.2010.5540138
10.1109/CVPR.2007.383195
10.1145/1970392.1970395
10.1109/CVPR.2006.199
10.1109/ICCV.2009.5459424
10.2307/1390605
10.1109/ACC.2010.5531594
10.1109/ISIT.2010.5513535
10.1109/CVPR.2005.238
10.1109/TVCG.2009.88
10.1109/TIP.2011.2109729
10.1007/s10208-009-9045-5
10.1090/S0025-5718-1962-0136527-4
10.1109/CVPR.2008.4587706
10.1109/CVPR.2010.5540004
10.1111/j.1467-9868.2005.00503.x
10.1109/ICCV.2007.4408984
10.1145/1531326.1531335
10.1109/CVPR.2005.84
10.1109/ICCV.2007.4409000
10.1090/S0025-5718-2012-02598-1
10.1145/1149283.1149286
10.1109/JPROC.2009.2035722
10.1109/TPAMI.2010.116
10.1007/s00041-008-9045-x
10.1137/080716542
10.1145/1102351.1102441
10.1109/CVPR.2006.19
ContentType Journal Article
DBID 97E
RIA
RIE
AAYXX
CITATION
DOI 10.1109/JSTSP.2012.2195472
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE/IET Electronic Library
CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE/IET Electronic Library
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISSN 1941-0484
EndPage 582
ExternalDocumentID 10_1109_JSTSP_2012_2195472
6187689
Genre orig-research
GroupedDBID -~X
0R~
29I
4.4
5GY
5VS
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACIWK
AENEX
AETIX
AGQYO
AGSQL
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
EJD
F5P
HZ~
IFIPE
IPLJI
JAVBF
LAI
M43
O9-
OCL
RIA
RIE
RNS
AAYXX
CITATION
ID FETCH-LOGICAL-c267t-5562fc5d198b6e2b1bdc3bc84983e4db42b5531ec0e0461d8002f915c6f1c9a13
IEDL.DBID RIE
ISICitedReferencesCount 38
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000307842100014&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1932-4553
IngestDate Sat Nov 29 03:55:51 EST 2025
Tue Nov 18 22:18:55 EST 2025
Tue Aug 26 17:19:20 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 5
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c267t-5562fc5d198b6e2b1bdc3bc84983e4db42b5531ec0e0461d8002f915c6f1c9a13
PageCount 17
ParticipantIDs crossref_primary_10_1109_JSTSP_2012_2195472
crossref_citationtrail_10_1109_JSTSP_2012_2195472
ieee_primary_6187689
PublicationCentury 2000
PublicationDate 2012-09-01
PublicationDateYYYYMMDD 2012-09-01
PublicationDate_xml – month: 09
  year: 2012
  text: 2012-09-01
  day: 01
PublicationDecade 2010
PublicationTitle IEEE journal of selected topics in signal processing
PublicationTitleAbbrev JSTSP
PublicationYear 2012
Publisher IEEE
Publisher_xml – name: IEEE
References ref13
ref12
ref15
ref14
ref10
brox (ref39) 2004
ref17
ref16
(ref51) 0
ref19
ref18
toh (ref29) 2010; 6
fazel (ref4) 2002
(ref45) 0
ref46
zach (ref26) 2008
ref48
ref42
ref41
ref44
ref43
kazhdan (ref50) 2006
ref7
lin (ref8) 2011
ref3
ref6
fu (ref38) 2011; 21
ref5
ref40
kolmogorov (ref11) 2002
liu (ref24) 2009
ref35
liu (ref49) 2012
ref34
ref37
ref36
lin (ref47) 2011
ref31
ref30
ref33
ref32
campbell (ref22) 2008; 1
ref2
ref1
ref23
ref25
ref20
ref21
ref28
ref27
deng (ref9) 2012
References_xml – ident: ref25
  doi: 10.1109/CVPR.2008.4587792
– start-page: 82
  year: 2002
  ident: ref11
  article-title: Multi-camera scene reconstruction via graph cuts
  publication-title: Eur Conf Comput Vis
– ident: ref14
  doi: 10.1109/ICCV.2007.4408933
– year: 2008
  ident: ref26
  article-title: Fast and high quality fusion of depth maps
  publication-title: Proc Int Symp 3D Data Process Visualiz Transmiss (3DPVT)
– ident: ref1
  doi: 10.1007/s10107-009-0306-5
– ident: ref7
  doi: 10.1561/2200000016
– ident: ref13
  doi: 10.1109/TPAMI.2009.161
– ident: ref44
  doi: 10.1007/s10107-009-0306-5
– ident: ref16
  doi: 10.1109/CVPR.2005.227
– ident: ref31
  doi: 10.1145/138859.138867
– ident: ref37
  doi: 10.1109/CVPR.2010.5540138
– year: 2002
  ident: ref4
  publication-title: Matrix Rank Minimization With Applications
– volume: 1
  start-page: 766
  year: 2008
  ident: ref22
  article-title: Using multiple hypotheses to improve depth-maps for multi-view stereo
  publication-title: Proc Eur Conf Comput Vis
– year: 0
  ident: ref51
– start-page: 2121
  year: 2009
  ident: ref24
  article-title: Continuous depth estimation for multi-view stereo
  publication-title: Proc IEEE Conf Comput Vis Pattern Recogn CVPR '09
– ident: ref15
  doi: 10.1109/CVPR.2007.383195
– ident: ref3
  doi: 10.1145/1970392.1970395
– ident: ref20
  doi: 10.1109/CVPR.2006.199
– year: 2012
  ident: ref49
  article-title: Robust recovery of subspace structures by low-rank representation
  publication-title: IEEE Trans Pattern Anal Mach Intell
– ident: ref35
  doi: 10.1109/ICCV.2009.5459424
– ident: ref6
  doi: 10.2307/1390605
– ident: ref33
  doi: 10.1109/ACC.2010.5531594
– ident: ref30
  doi: 10.1109/ISIT.2010.5513535
– ident: ref12
  doi: 10.1109/CVPR.2005.238
– ident: ref18
  doi: 10.1109/TVCG.2009.88
– ident: ref36
  doi: 10.1109/TIP.2011.2109729
– ident: ref28
  doi: 10.1007/s10208-009-9045-5
– ident: ref43
  doi: 10.1090/S0025-5718-1962-0136527-4
– ident: ref46
  doi: 10.1109/CVPR.2008.4587706
– ident: ref21
  doi: 10.1109/CVPR.2010.5540004
– ident: ref41
  doi: 10.1111/j.1467-9868.2005.00503.x
– ident: ref19
  doi: 10.1109/ICCV.2007.4408984
– ident: ref34
  doi: 10.1145/1531326.1531335
– ident: ref23
  doi: 10.1109/CVPR.2005.84
– ident: ref40
  doi: 10.1109/ICCV.2007.4409000
– year: 0
  ident: ref45
– ident: ref48
  doi: 10.1090/S0025-5718-2012-02598-1
– ident: ref32
  doi: 10.1145/1149283.1149286
– volume: 21
  start-page: 2369
  year: 2011
  ident: ref38
  article-title: Adaptive compressed sensing recovery utilizing the property of signals autocorrelations
  publication-title: IEEE Trans Image Process
– year: 2011
  ident: ref8
  publication-title: The augmented Lagrange multiplier method for exact recovery of corrupted low-rank matrices
– ident: ref2
  doi: 10.1109/JPROC.2009.2035722
– volume: 6
  start-page: 15
  year: 2010
  ident: ref29
  article-title: An accelerated proximal gradient algorithm for nuclear norm regularized linear least squares problems
  publication-title: Pac J Optimiz
– ident: ref17
  doi: 10.1109/TPAMI.2010.116
– ident: ref5
  doi: 10.1007/s00041-008-9045-x
– ident: ref42
  doi: 10.1137/080716542
– year: 2011
  ident: ref47
  article-title: Linearized alternating direction method with adaptive penalty for low-rank representation
  publication-title: Proc Adv Neural Inf Process Syst
– start-page: 61
  year: 2006
  ident: ref50
  article-title: Poisson surface reconstruction
  publication-title: Proc 4th Eurographics Symp Geometry Process Ser SGP'06
– start-page: 25
  year: 2004
  ident: ref39
  publication-title: High Accuracy Optical Flow Estimation Based on A Theory For Warping
– ident: ref27
  doi: 10.1145/1102351.1102441
– ident: ref10
  doi: 10.1109/CVPR.2006.19
– year: 2012
  ident: ref9
  article-title: Low-rank structure learning via log-sum heuristic recovery
SSID ssj0057614
Score 2.1793826
Snippet This paper introduces a general framework to fuse noisy point clouds from multiview images of the same object. We solve this classical vision problem using a...
SourceID crossref
ieee
SourceType Enrichment Source
Index Database
Publisher
StartPage 566
SubjectTerms Cameras
Compressive sensing
Estimation
fusion
matrix completion
multiview stereo (MVS)
Noise
Noise measurement
point cloud
Robustness
Three dimensional displays
Vectors
Title Noisy Depth Maps Fusion for Multiview Stereo Via Matrix Completion
URI https://ieeexplore.ieee.org/document/6187689
Volume 6
WOSCitedRecordID wos000307842100014&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE/IET Electronic Library
  customDbUrl:
  eissn: 1941-0484
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0057614
  issn: 1932-4553
  databaseCode: RIE
  dateStart: 20070101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEA61eNCDryrWFzl4022TTfaRo6_iQUuhWnpbNsksLshuabei_94k3ZYKInhblgks3yaZL5mZbxC6NHxNQkC4FyvO7dUNeEIT7ZnThdaaRZl0Skyjp6jfj8djMWig61UtDAC45DPo2EcXy9elmtursm5IzdqNxQbaiKJwUau13HUNbaZ1BNn3eBCwZYEMEV0zxYcDm8Xld3wrcBb5P5zQWlcV51R6u__7nD20U5NHfLP42_uoAcUB2l6TFGyh236Zz77wPUyqN_ycTma4N7cXYtiQU-yqbW0sAA8NnFDiUZ4ao2qaf2K7MVgh7rI4RK-9h5e7R6_uk-ApP4wqLzAcJlOBpiKWIfiSSq2YVDEXMQOuJfelAYSCImDl1bXliJmggQozqkRK2RFqFmUBxwhz46w4y4hiLOUpSaXx_xSItG2n0hhEG9ElcImqRcRtL4v3xB0miEgc2IkFO6nBbqOr1ZjJQkLjT-uWRXplWYN88vvrU7RlBy-Svs5Qs5rO4Rxtqo8qn00v3Az5BksLt-Y
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LS8NAEB5qFdSDb7E-9-BNY7PJps0efZWKtQitxVvI7k6wIE1pU9F_706algoieAthEsK3m51vZ3a-ATi3fE1h4Aon1EJQ6AYdaVzj2N2FMcavJypXYuq16u12-Poqn0twOa-FQcT88Ble0WWeyzepnlCorFrj9t8N5RIsU-esolprtu5a4syLHLLniCDwZyUyrqzaSd55pnNc3pVHEmd174cbWuirkruVxub_PmgLNgr6yK6n470NJRzswPqCqOAu3LTT_viL3eEwe2NP8XDMGhMKiTFLT1leb0vZANaxgGLKev3YGmWj_iejpYGkuNPBHrw07ru3TafolOBor1bPnMCymEQHhstQ1dBTXBntKx0KGfoojBKesoBw1C6SwLohlphIHuhawrWMub8P5UE6wANgwror4Seu9v1YxG6sLAPg6CpqPBWHKCvAZ8BFupARp24W71G-nXBllIMdEdhRAXYFLubPDKciGn9a7xLSc8sC5MPfb5_BarP71IpaD-3HI1ijF02PgB1DORtN8ARW9EfWH49O89nyDQ8euy8
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Noisy+Depth+Maps+Fusion+for+Multiview+Stereo+Via+Matrix+Completion&rft.jtitle=IEEE+journal+of+selected+topics+in+signal+processing&rft.au=Deng%2C+Yue&rft.au=Liu%2C+Yebin&rft.au=Dai%2C+Qionghai&rft.au=Zhang%2C+Zengke&rft.date=2012-09-01&rft.issn=1932-4553&rft.eissn=1941-0484&rft.volume=6&rft.issue=5&rft.spage=566&rft.epage=582&rft_id=info:doi/10.1109%2FJSTSP.2012.2195472&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_JSTSP_2012_2195472
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1932-4553&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1932-4553&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1932-4553&client=summon