Pixel-Perfect Structure-From-Motion With Featuremetric Refinement
Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this artic...
Saved in:
| Published in: | IEEE transactions on pattern analysis and machine intelligence Vol. 47; no. 5; pp. 3298 - 3309 |
|---|---|
| Main Authors: | , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
United States
IEEE
01.05.2025
|
| Subjects: | |
| ISSN: | 0162-8828, 1939-3539, 2160-9292, 1939-3539 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this article, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale. |
|---|---|
| AbstractList | Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this article, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale. Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this article, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale.Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this article, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale. |
| Author | Pollefeys, Marc Sarlin, Paul-Edouard Larsson, Viktor Lindenberger, Philipp |
| Author_xml | – sequence: 1 givenname: Paul-Edouard orcidid: 0000-0001-6230-266X surname: Sarlin fullname: Sarlin, Paul-Edouard email: psarlin@inf.ethz.ch organization: Department of Computer Science, ETH Zurich, Zürich, Switzerland – sequence: 2 givenname: Philipp orcidid: 0000-0003-4112-6542 surname: Lindenberger fullname: Lindenberger, Philipp email: plindenbe@ethz.ch organization: Department of Mathematics, ETH Zurich, Zürich, Switzerland – sequence: 3 givenname: Viktor orcidid: 0000-0001-8427-7215 surname: Larsson fullname: Larsson, Viktor email: viktor.larsson@math.lth.se organization: Lund University, Lund, Sweden – sequence: 4 givenname: Marc surname: Pollefeys fullname: Pollefeys, Marc email: marc.pollefeys@inf.ethz.ch organization: Department of Computer Science, ETH Zurich, Zürich, Switzerland |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/37021895$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9kUFv1DAQhS1URLeFP4AQ2iOXLPY4duLjqmKhUitWUMRxNPFOVKNsstiOgH9P0t0ixIHTyJr3zTzPuxBn_dCzEC-VXCkl3du77fr2egUS9EqDrsC6J2IBysrCgYMzsZDKQlHXUJ-Li5S-SalKI_Uzca4rCap2ZiHW2_CTu2LLsWWfl59zHH0eIxebOOyL2yGHoV9-Dfl-uWGaG3vOMfjlJ25DPz36_Fw8balL_OJUL8WXzbu7qw_Fzcf311frm8KXtc0F6KZ1lpVWjfdlK6myWnuljTHNrnEVgdq1jZlctdYbMhZcaUsCXzPIxjX6UtBxbvrBh7HBQwx7ir9woICHIWbqMHJiiv4euxET46Tqgqf5Cwm11tYzMeodlVhaY5FK7xFAEZnWgtJy2vHmuOMQh-8jp4z7kDx3HfU8jAmhctV8w6qepK9P0rHZ8-6PncfTToL6KPBxSClyiz7kBzM5UuhQSZxTxIcUcU4RTylOKPyDPk7_L_TqCAVm_guQqi6l078BHLGpAw |
| CODEN | ITPIDJ |
| CitedBy_id | crossref_primary_10_3390_s24206595 |
| Cites_doi | 10.1109/TPAMI.1986.4767838 10.1007/978-3-642-15552-9_3 10.1109/TPAMI.2011.256 10.1109/CVPR46437.2021.00326 10.1109/CVPR.2019.00863 10.1109/CVPR42600.2020.00629 10.1109/TPAMI.2017.2658577 10.1007/978-3-319-54190-7_20 10.1007/978-3-030-58452-8_39 10.1007/978-3-030-01237-3_47 10.1109/ICRA.2017.7989525 10.5244/C.2.23 10.1023/B:VISI.0000011205.11775.fd 10.1109/LRA.2020.2965031 10.1109/ICCV.2013.70 10.1109/CVPR.2019.00471 10.1109/CVPR46437.2021.00566 10.1109/CVPR.2017.272 10.1109/cvpr.1994.323831 10.1109/TIP.2015.2431445 10.1109/CVPR46437.2021.01431 10.1109/TPAMI.2009.77 10.1007/s11263-007-0107-3 10.1109/CVPR.2019.01300 10.1109/CVPR.2018.00897 10.1007/978-3-030-58545-7_35 10.1023/A:1014554110407 10.1007/978-3-319-46466-4_28 10.1109/3DV50981.2020.00107 10.1109/IROS.2013.6696650 10.1109/TPAMI.2010.147 10.1007/978-3-030-58548-8_36 10.1007/11744023_34 10.1109/CVPR46437.2021.00048 10.1109/CVPR.2019.00828 10.1109/CVPR.2007.383115 10.1109/CVPR.2009.5206587 10.1109/CVPR.2017.179 10.1080/03610927708827533 10.1109/tpami.2021.3103980 10.1109/CVPR.2016.445 10.5244/C.26.76 10.1023/b:visi.0000029664.99615.94 10.1007/978-3-642-15561-1_27 10.1007/s11263-020-01385-0 10.1007/978-3-642-33718-5_2 10.1007/978-3-030-58536-5_42 10.1007/11744023_32 10.5244/C.22.14 10.1109/3DV.2016.48 10.1109/CVPR.2014.193 10.1007/s11263-020-01399-8 10.1023/A:1014573219977 10.1145/2001269.2001293 10.1109/cvpr.2017.402 10.1109/CVPR46437.2021.00881 10.1109/CVPR42600.2020.00499 10.1007/978-3-540-45243-0_31 10.1109/CVPR42600.2020.00202 10.1002/9781118186435 10.1109/CVPR.2005.354 10.1109/ICCVW.2017.105 10.1109/CVPRW.2018.00060 10.1109/CVPR.2018.00931 10.1007/978-3-319-10605-2_54 10.1109/CVPR46437.2021.00464 10.1109/CVPR.2016.592 10.1090/qam/10666 10.1109/CVPR.2019.00022 10.1007/978-3-030-01237-3_18 10.1109/LRA.2020.3039216 10.1109/CVPR.2009.5206848 10.1109/TPAMI.2019.2952114 10.1109/CVPR.2019.00567 10.1109/ICCV.2017.260 10.1109/ICCV48922.2021.00593 10.1007/3-540-44480-7_21 |
| ContentType | Journal Article |
| CorporateAuthor | Computer Vision and Machine Learning Lunds universitets profilområden Naturvetenskapliga fakulteten Faculty of Engineering, LTH Lunds Tekniska Högskola LU Profile Area: Natural and Artificial Cognition LTH Profile areas ELLIIT: the Linköping-Lund initiative on IT and mobile communication Strategiska forskningsområden (SFO) Matematikcentrum LTH profilområde: AI och digitalisering Matematik LTH Research groups at the Centre for Mathematical Sciences Lunds universitet LTH profilområden Profile areas and other strong research environments LU profilområde: Naturlig och artificiell kognition Faculty of Science Lund University Lund University Profile areas Datorseende och maskininlärning LTH Profile Area: AI and Digitalization Centre for Mathematical Sciences Forskargrupper vid Matematikcentrum Mathematical Imaging Group Strategic research areas (SRA) Mathematics (Faculty of Engineering) Profilområden och andra starka forskningsmiljöer |
| CorporateAuthor_xml | – name: Naturvetenskapliga fakulteten – name: Strategiska forskningsområden (SFO) – name: LTH Profile Area: AI and Digitalization – name: Research groups at the Centre for Mathematical Sciences – name: Strategic research areas (SRA) – name: Lunds Tekniska Högskola – name: Datorseende och maskininlärning – name: Faculty of Engineering, LTH – name: Lund University Profile areas – name: Lund University – name: LTH profilområde: AI och digitalisering – name: Matematik LTH – name: Computer Vision and Machine Learning – name: Profile areas and other strong research environments – name: Matematikcentrum – name: ELLIIT: the Linköping-Lund initiative on IT and mobile communication – name: Mathematics (Faculty of Engineering) – name: LU Profile Area: Natural and Artificial Cognition – name: Lunds universitet – name: Faculty of Science – name: Lunds universitets profilområden – name: Profilområden och andra starka forskningsmiljöer – name: LTH Profile areas – name: LTH profilområden – name: Centre for Mathematical Sciences – name: LU profilområde: Naturlig och artificiell kognition – name: Forskargrupper vid Matematikcentrum – name: Mathematical Imaging Group |
| DBID | 97E RIA RIE AAYXX CITATION NPM 7X8 ADTPV AGCHP AOWAS D8T D95 ZZAVC |
| DOI | 10.1109/TPAMI.2023.3237269 |
| DatabaseName | IEEE Xplore (IEEE) IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed MEDLINE - Academic SwePub SWEPUB Lunds universitet full text SwePub Articles SWEPUB Freely available online SWEPUB Lunds universitet SwePub Articles full text |
| DatabaseTitle | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering Computer Science |
| EISSN | 2160-9292 1939-3539 |
| EndPage | 3309 |
| ExternalDocumentID | oai_portal_research_lu_se_publications_3336ceae_3da4_4656_a4cc_221aa5f62130 37021895 10_1109_TPAMI_2023_3237269 10018409 |
| Genre | orig-research Journal Article |
| GrantInformation_xml | – fundername: ETH Zurich Postdoctoral Fellowship – fundername: Huawei |
| GroupedDBID | --- -DZ -~X .DC 0R~ 29I 4.4 53G 5GY 5VS 6IK 97E 9M8 AAJGR AARMG AASAJ AAWTH ABAZT ABFSI ABQJQ ABVLG ACGFO ACGFS ACIWK ACNCT ADRHT AENEX AETEA AETIX AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 E.L EBS EJD F5P FA8 HZ~ H~9 IBMZZ ICLAB IEDLZ IFIPE IFJZH IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNI RNS RXW RZB TAE TN5 UHB VH1 XJT ~02 AAYXX CITATION NPM RIG 7X8 ADTPV AGCHP AOWAS D8T D95 ZZAVC |
| ID | FETCH-LOGICAL-c486t-23bf96e131bcc4f0a7633c13555bdb97a21dfb5189f6c5a5629464a2c8e20b9b3 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 6 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001465416300050&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0162-8828 1939-3539 |
| IngestDate | Sun Nov 23 03:11:00 EST 2025 Sat Sep 27 22:52:52 EDT 2025 Mon Jul 21 05:19:24 EDT 2025 Tue Nov 18 22:17:20 EST 2025 Sat Nov 29 08:02:39 EST 2025 Wed Aug 27 02:04:40 EDT 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 5 |
| Language | English |
| License | https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037 |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c486t-23bf96e131bcc4f0a7633c13555bdb97a21dfb5189f6c5a5629464a2c8e20b9b3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0001-8427-7215 0000-0001-6230-266X 0000-0003-4112-6542 |
| PMID | 37021895 |
| PQID | 2797145078 |
| PQPubID | 23479 |
| PageCount | 12 |
| ParticipantIDs | proquest_miscellaneous_2797145078 pubmed_primary_37021895 crossref_citationtrail_10_1109_TPAMI_2023_3237269 crossref_primary_10_1109_TPAMI_2023_3237269 swepub_primary_oai_portal_research_lu_se_publications_3336ceae_3da4_4656_a4cc_221aa5f62130 ieee_primary_10018409 |
| PublicationCentury | 2000 |
| PublicationDate | 2025-05-01 |
| PublicationDateYYYYMMDD | 2025-05-01 |
| PublicationDate_xml | – month: 05 year: 2025 text: 2025-05-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | IEEE transactions on pattern analysis and machine intelligence |
| PublicationTitleAbbrev | TPAMI |
| PublicationTitleAlternate | IEEE Trans Pattern Anal Mach Intell |
| PublicationYear | 2025 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| References | ref13 ref12 ref15 ref14 ref58 ref53 Tang (ref81) ref52 ref96 ref11 ref10 ref54 ref17 ref16 Li (ref48) Revaud (ref6) ref92 ref51 ref95 ref94 ref91 Ono (ref57); 31 ref46 ref45 ref89 ref47 Simonyan (ref93) ref42 ref41 ref85 ref44 ref88 ref43 ref87 Förstner (ref50) Tang (ref56) ref49 ref8 ref7 ref4 ref3 Woodford (ref69) ref5 Lucas (ref24) Christiansen (ref55) 2019 ref82 ref40 ref84 ref83 ref80 ref35 ref79 ref34 ref78 ref37 Rocco (ref39) ref36 ref31 ref75 ref30 ref74 ref33 ref77 ref76 ref2 ref1 Choy (ref38) ref71 ref70 ref73 ref72 ref68 ref23 ref67 ref26 ref25 ref20 ref64 ref63 ref22 ref66 ref21 ref65 Germain (ref19) Tyszkiewicz (ref18) ref28 ref27 ref29 Eichhardt (ref59) Heinly (ref9) ref60 ref62 ref61 |
| References_xml | – ident: ref51 doi: 10.1109/TPAMI.1986.4767838 – ident: ref21 doi: 10.1007/978-3-642-15552-9_3 – ident: ref22 doi: 10.1109/TPAMI.2011.256 – ident: ref82 doi: 10.1109/CVPR46437.2021.00326 – ident: ref54 doi: 10.1109/CVPR.2019.00863 – ident: ref41 doi: 10.1109/CVPR42600.2020.00629 – volume-title: Proc. Int. Conf. Learn. Representations ident: ref56 article-title: Neural outlier rejection for self-supervised keypoint learning – start-page: 3287 volume-title: Proc. Conf. Comput. Vis. Pattern Recognit. ident: ref9 article-title: Reconstructing the world* in six days *(as captured by the Yahoo 100 million image dataset) – ident: ref25 doi: 10.1109/TPAMI.2017.2658577 – ident: ref63 doi: 10.1007/978-3-319-54190-7_20 – ident: ref23 doi: 10.1007/978-3-030-58452-8_39 – ident: ref45 doi: 10.1007/978-3-030-01237-3_47 – start-page: 2414 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref38 article-title: Universal correspondence network – ident: ref72 doi: 10.1109/ICRA.2017.7989525 – volume-title: Proc. Int. Conf. Learn. Representations ident: ref81 article-title: BA-Net: Dense bundle adjustment network – ident: ref1 doi: 10.5244/C.2.23 – start-page: 626 volume-title: Proc. Eur. Conf. Comput. Vis. ident: ref19 article-title: S2DNet: Learning accurate correspondences for sparse-to-dense feature matching – ident: ref60 doi: 10.1023/B:VISI.0000011205.11775.fd – ident: ref76 doi: 10.1109/LRA.2020.2965031 – ident: ref30 doi: 10.1109/ICCV.2013.70 – volume: 31 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref57 article-title: LF-Net: Learning local features from images – ident: ref73 doi: 10.1109/CVPR.2019.00471 – ident: ref92 doi: 10.1109/CVPR46437.2021.00566 – ident: ref68 doi: 10.1109/CVPR.2017.272 – start-page: 281 volume-title: Proc. ISPRS Intercommission Conf. Fast Process. Photogrammetric Data ident: ref50 article-title: A fast operator for detection and precise location of distinct points, corners and centres of circular features – ident: ref65 doi: 10.1109/cvpr.1994.323831 – ident: ref78 doi: 10.1109/TIP.2015.2431445 – ident: ref31 doi: 10.1109/CVPR46437.2021.01431 – ident: ref37 doi: 10.1109/TPAMI.2009.77 – ident: ref27 doi: 10.1007/s11263-007-0107-3 – start-page: 674 volume-title: Proc. Int. Joint Conf. Artif. Intell. ident: ref24 article-title: An iterative image registration technique with an application to stereo vision – ident: ref33 doi: 10.1109/CVPR.2019.01300 – ident: ref34 doi: 10.1109/CVPR.2018.00897 – ident: ref47 doi: 10.1007/978-3-030-58545-7_35 – ident: ref53 doi: 10.1023/A:1014554110407 – start-page: 367 volume-title: Proc. Brit. Mach. Vis. Conf. ident: ref69 article-title: Large scale photometric bundle adjustment – ident: ref17 doi: 10.1007/978-3-319-46466-4_28 – ident: ref77 doi: 10.1109/3DV50981.2020.00107 – ident: ref61 doi: 10.1109/IROS.2013.6696650 – ident: ref36 doi: 10.1109/TPAMI.2010.147 – ident: ref40 doi: 10.1007/978-3-030-58548-8_36 – ident: ref16 doi: 10.1007/11744023_34 – ident: ref84 doi: 10.1109/CVPR46437.2021.00048 – ident: ref5 doi: 10.1109/CVPR.2019.00828 – ident: ref29 doi: 10.1109/CVPR.2007.383115 – ident: ref11 doi: 10.1109/CVPR.2009.5206587 – start-page: 17346 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref48 article-title: Dual-resolution correspondence networks – volume-title: Proc. Int. Conf. Learn. Representations ident: ref93 article-title: Very deep convolutional networks for large-scale image recognition – start-page: 12414 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref6 article-title: R2D2: Repeatable and reliable detector and descriptor – ident: ref43 doi: 10.1109/CVPR.2017.179 – ident: ref87 doi: 10.1080/03610927708827533 – start-page: 1658 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref39 article-title: Neighbourhood consensus networks – ident: ref58 doi: 10.1109/tpami.2021.3103980 – ident: ref28 doi: 10.1109/CVPR.2016.445 – ident: ref95 doi: 10.5244/C.26.76 – ident: ref2 doi: 10.1023/b:visi.0000029664.99615.94 – ident: ref15 doi: 10.1007/978-3-642-15561-1_27 – ident: ref89 doi: 10.1007/s11263-020-01385-0 – ident: ref10 doi: 10.1007/978-3-642-33718-5_2 – ident: ref7 doi: 10.1007/978-3-030-58536-5_42 – ident: ref3 doi: 10.1007/11744023_32 – ident: ref70 doi: 10.5244/C.22.14 – ident: ref71 doi: 10.1109/3DV.2016.48 – ident: ref66 doi: 10.1109/CVPR.2014.193 – ident: ref96 doi: 10.1007/s11263-020-01399-8 – ident: ref52 doi: 10.1023/A:1014573219977 – ident: ref14 doi: 10.1145/2001269.2001293 – ident: ref79 doi: 10.1109/cvpr.2017.402 – ident: ref42 doi: 10.1109/CVPR46437.2021.00881 – ident: ref8 doi: 10.1109/CVPR42600.2020.00499 – ident: ref88 doi: 10.1007/978-3-540-45243-0_31 – start-page: 14254 volume-title: Proc. Int. Conf. Neural Inf. Process. Syst. ident: ref18 article-title: DISK: Learning local features with policy gradient – ident: ref80 doi: 10.1109/CVPR42600.2020.00202 – year: 2019 ident: ref55 article-title: UnsuperPoint: End-to-end unsupervised interest point detector and descriptor – volume-title: Proc. Brit. Mach. Vis. Conf. ident: ref59 article-title: Optimal multi-view correction of local affine frames – ident: ref83 doi: 10.1002/9781118186435 – ident: ref91 doi: 10.1109/CVPR.2005.354 – ident: ref26 doi: 10.1109/ICCVW.2017.105 – ident: ref4 doi: 10.1109/CVPRW.2018.00060 – ident: ref44 doi: 10.1109/CVPR.2018.00931 – ident: ref62 doi: 10.1007/978-3-319-10605-2_54 – ident: ref49 doi: 10.1109/CVPR46437.2021.00464 – ident: ref13 doi: 10.1109/CVPR.2016.592 – ident: ref85 doi: 10.1090/qam/10666 – ident: ref64 doi: 10.1109/CVPR.2019.00022 – ident: ref74 doi: 10.1007/978-3-030-01237-3_18 – ident: ref75 doi: 10.1109/LRA.2020.3039216 – ident: ref94 doi: 10.1109/CVPR.2009.5206848 – ident: ref46 doi: 10.1109/TPAMI.2019.2952114 – ident: ref67 doi: 10.1109/CVPR.2019.00567 – ident: ref12 doi: 10.1109/ICCV.2017.260 – ident: ref35 doi: 10.1109/ICCV48922.2021.00593 – ident: ref20 doi: 10.1007/3-540-44480-7_21 |
| SSID | ssj0014503 |
| Score | 2.5274084 |
| Snippet | Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects... |
| SourceID | swepub proquest pubmed crossref ieee |
| SourceType | Open Access Repository Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 3298 |
| SubjectTerms | Bundle adjustment Computer and Information Sciences Computer graphics and computer vision Costs Data- och informationsvetenskap (Datateknik) Datorgrafik och datorseende Estimation Feature extraction feature matching featuremetric optimization Geometry Image reconstruction Location awareness Natural Sciences Naturvetenskap Optimization structure-from-Motion visual localization |
| Title | Pixel-Perfect Structure-From-Motion With Featuremetric Refinement |
| URI | https://ieeexplore.ieee.org/document/10018409 https://www.ncbi.nlm.nih.gov/pubmed/37021895 https://www.proquest.com/docview/2797145078 |
| Volume | 47 |
| WOSCitedRecordID | wos001465416300050&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 Electronic Library (IEL) customDbUrl: eissn: 2160-9292 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014503 issn: 0162-8828 databaseCode: RIE dateStart: 19790101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3di9QwEB-8wwd98PQ8tX4cFXyT7rVJmzSPi7jowx2Lnrj4EibplFtYdmU_jvvzzaTtsogKvhXaNGFmkplJMr8fwDvnvCpdhVlDiFlZ5ZjVqskznaM3lS5KLV0km9BXV_VsZqZ9sXqshSGiePmMRvwYz_Kbld_xVtkF4wVxQnIER1rrrlhrf2QQ-pEdkHeY4iGPGCpkcnNxPR1ffh4xUfhICqmFYqxQqdm9Ma_EgUOKDCt_CjZ_QxKN3mdy8p_jfgyP-jAzHXd28QTu0fIUTgYKh7Sf0afw8ACP8CmMp_M7WmRTWvMlj_RrxJbdrSmbrMNSehkJf9Lv8-1NypFj3FpkgP_0C7XhJzyIM_g2-Xj94VPWcyxkvqzVNhPStUZRIQvnfdnmGNYb6YsQhVSucUajKJrWVUFQrfIVhmjJlKpE4WsSuTNOPoPj5WpJLyClwjlUKPO2Cn7fSMRWkmlEXeS1b0gnUAyCtr4HIGcejIWNiUhubNSTZT3ZXk8JvN-3-dnBb_zz6zMW_sGXndwTeDso1IbZw0ciuKTVbmOFNpotR9cJPO80vW89GEgCPzrV798wJHeXHdkekunGLnZ2E3o-2Gu1UkrlCcnKBkvLwHQWS--tEAVi1SoR4oeXfxnyK3ggmGk4Xq18DcdB4fQG7vvb7XyzPg_2P6vPo_3_ApVQ_9w |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3di9QwEB_0FNQHT89T62cF36R7bZImzeMiLnd4uyy64uFLSNIpt7Dsyn6If76ZtF0WUcG3QpsmzGQyM0nm9wN465yXwpU2q9HaTJS5zSpZ55nKrdelKoTiLpJNqMmkurrS065YPdbCIGK8fIYDeoxn-fXK72ir7IzwgighuQm3SiFY0ZZr7Q8NQk-8hfIORh4yib5GJtdns-lwfDEgqvABZ1wxSWihXJGDI2aJA5cUOVb-FG7-hiUa_c_o-D9H_gDud4FmOmxnxkO4gcsTOO5JHNLOpk_g3gEi4SMYTuc_cZFNcU3XPNLPEV12t8ZstA6L6ThS_qRf59vrlGLHuLlIEP_pJ2zCT2gQp_Bl9GH2_jzrWBYyLyq5zRh3jZZY8MJ5L5rchhWH-yLEIaWrnVaWFXXjyiCoRvrShnhJCyks8xWy3GnHH8PRcrXEp5Bi4ZyVludNGTy_5tY2HHXNqiKvfI0qgaIXtPEdBDkxYSxMTEVybaKeDOnJdHpK4N2-zfcWgOOfX5-S8A--bOWewJteoSbYDx2K2CWudhvDlFY0c1SVwJNW0_vW_QRJ4Fur-v0bAuVu8yPTgTJdm8XObELPB7uthnMuPVo0vLbCEDSdscJ7w1hhbdlIFiKIZ38Z8mu4cz4bX5rLi8nH53CXEe9wvGj5Ao6C8vEl3PY_tvPN-lW0gl_9pAJK |
| 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=Pixel-Perfect+Structure-From-Motion+With+Featuremetric+Refinement&rft.jtitle=IEEE+transactions+on+pattern+analysis+and+machine+intelligence&rft.au=Sarlin%2C+Paul-Edouard&rft.au=Lindenberger%2C+Philipp&rft.au=Larsson%2C+Viktor&rft.au=Pollefeys%2C+Marc&rft.date=2025-05-01&rft.issn=0162-8828&rft.eissn=2160-9292&rft.volume=47&rft.issue=5&rft.spage=3298&rft.epage=3309&rft_id=info:doi/10.1109%2FTPAMI.2023.3237269&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_TPAMI_2023_3237269 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0162-8828&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0162-8828&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0162-8828&client=summon |