Monocular template-based 3D surface reconstruction: Convex inextensible and nonconvex isometric methods
Identifieur interne : 004051 ( Main/Curation ); précédent : 004050; suivant : 004052Monocular template-based 3D surface reconstruction: Convex inextensible and nonconvex isometric methods
Auteurs : F. Brunet [France] ; A. Bartoli [France] ; R. I. Hartley [Australie]Source :
- Computer vision and image understanding : (Print) [ 1077-3142 ] ; 2014.
Descripteurs français
- Pascal (Inist)
- Image tridimensionnelle, Reconstruction image, Traitement image, Vision ordinateur, Reconstruction surface, Forme libre, Processeur pipeline, Vision monoculaire, Surface convexe, Surface lisse, Initialisation, Analyse non convexe, Isométrie, Programmation conique, Modélisation, Méthode moindre carré, Algorithme de Levenberg Marquardt, Minimisation, Résultat expérimental, Méthode raffinement.
English descriptors
- KwdEn :
- Computer vision, Convex surface, Experimental result, Free form, Image processing, Image reconstruction, Initialization, Isometry, Least squares method, Levenberg Marquardt algorithm, Minimization, Modeling, Monocular vision, Non convex analysis, Pipeline processor, Refinement method, Smooth surface, Surface reconstruction, Tridimensional image, conic programming.
Abstract
We study the 3D reconstruction of an isometric surface from point correspondences between a template and a single input image. The template shows the surface flat and fronto-parallel. We propose three new methods. The first two use a convex relaxation of isometry to inextensibility. They are formulated as Second Order Cone Programs (SOCP). The first proposed method is point-wise (it reconstructs only the input point correspondences) while the second proposed method uses a smooth and continuous surface model, based on Free-Form Deformations (FFD). The third proposed method uses the 'true' nonconvex isometric constraint and the same continuous surface model. It is formulated with Nonlinear Least-Squares and can thus be solved with the efficient Levenberg-Marquardt minimization method. The proposed approaches may be combined in a single pipeline whereby one of the convex approximations is used to initialize the nonconvex method. Our contributions solve two important limitations of current state of the art: our convex methods are the first ones to handle noise in both the template and image points, and our nonconvex method is the first one to use 'true' isometric constraints. Our experimental results on simulated and real data show that our convex point-wise method and our nonconvex method outperform respectively current initialization and refinement methods in 3D reconstructed surface accuracy.
Links toward previous steps (curation, corpus...)
- to stream PascalFrancis, to step Corpus: Pour aller vers cette notice dans l'étape Curation :000147
- to stream PascalFrancis, to step Curation: Pour aller vers cette notice dans l'étape Curation :005D06
- to stream PascalFrancis, to step Checkpoint: Pour aller vers cette notice dans l'étape Curation :000188
- to stream Main, to step Merge: Pour aller vers cette notice dans l'étape Curation :004112
Links to Exploration step
Pascal:15-0004489Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">Monocular template-based 3D surface reconstruction: Convex inextensible and nonconvex isometric methods</title>
<author><name sortKey="Brunet, F" sort="Brunet, F" uniqKey="Brunet F" first="F." last="Brunet">F. Brunet</name>
<affiliation wicri:level="3"><inist:fA14 i1="01"><s1>Ubleam</s1>
<s2>Toulouse</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14>
<country>France</country>
<placeName><region type="region">Occitanie (région administrative)</region>
<region type="old region">Midi-Pyrénées</region>
<settlement type="city">Toulouse</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Bartoli, A" sort="Bartoli, A" uniqKey="Bartoli A" first="A." last="Bartoli">A. Bartoli</name>
<affiliation wicri:level="3"><inist:fA14 i1="02"><s1>ISIT, UMR 6284 CNRS-UdA</s1>
<s2>Clermont-Ferrand</s2>
<s3>FRA</s3>
<sZ>2 aut.</sZ>
</inist:fA14>
<country>France</country>
<placeName><region type="region">Auvergne-Rhône-Alpes</region>
<region type="old region">Auvergne (région administrative)</region>
<settlement type="city">Clermont-Ferrand</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Hartley, R I" sort="Hartley, R I" uniqKey="Hartley R" first="R. I." last="Hartley">R. I. Hartley</name>
<affiliation wicri:level="1"><inist:fA14 i1="03"><s1>ANU-NICTA</s1>
<s2>Canberra</s2>
<s3>AUS</s3>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>Australie</country>
<wicri:noRegion>ANU-NICTA</wicri:noRegion>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">15-0004489</idno>
<date when="2014">2014</date>
<idno type="stanalyst">PASCAL 15-0004489 INIST</idno>
<idno type="RBID">Pascal:15-0004489</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000147</idno>
<idno type="wicri:Area/PascalFrancis/Curation">005D06</idno>
<idno type="wicri:Area/PascalFrancis/Checkpoint">000188</idno>
<idno type="wicri:explorRef" wicri:stream="PascalFrancis" wicri:step="Checkpoint">000188</idno>
<idno type="wicri:doubleKey">1077-3142:2014:Brunet F:monocular:template:based</idno>
<idno type="wicri:Area/Main/Merge">004112</idno>
<idno type="wicri:Area/Main/Curation">004051</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">Monocular template-based 3D surface reconstruction: Convex inextensible and nonconvex isometric methods</title>
<author><name sortKey="Brunet, F" sort="Brunet, F" uniqKey="Brunet F" first="F." last="Brunet">F. Brunet</name>
<affiliation wicri:level="3"><inist:fA14 i1="01"><s1>Ubleam</s1>
<s2>Toulouse</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14>
<country>France</country>
<placeName><region type="region">Occitanie (région administrative)</region>
<region type="old region">Midi-Pyrénées</region>
<settlement type="city">Toulouse</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Bartoli, A" sort="Bartoli, A" uniqKey="Bartoli A" first="A." last="Bartoli">A. Bartoli</name>
<affiliation wicri:level="3"><inist:fA14 i1="02"><s1>ISIT, UMR 6284 CNRS-UdA</s1>
<s2>Clermont-Ferrand</s2>
<s3>FRA</s3>
<sZ>2 aut.</sZ>
</inist:fA14>
<country>France</country>
<placeName><region type="region">Auvergne-Rhône-Alpes</region>
<region type="old region">Auvergne (région administrative)</region>
<settlement type="city">Clermont-Ferrand</settlement>
</placeName>
</affiliation>
</author>
<author><name sortKey="Hartley, R I" sort="Hartley, R I" uniqKey="Hartley R" first="R. I." last="Hartley">R. I. Hartley</name>
<affiliation wicri:level="1"><inist:fA14 i1="03"><s1>ANU-NICTA</s1>
<s2>Canberra</s2>
<s3>AUS</s3>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>Australie</country>
<wicri:noRegion>ANU-NICTA</wicri:noRegion>
</affiliation>
</author>
</analytic>
<series><title level="j" type="main">Computer vision and image understanding : (Print)</title>
<title level="j" type="abbreviated">Comput. vis. image underst. : (Print)</title>
<idno type="ISSN">1077-3142</idno>
<imprint><date when="2014">2014</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
<seriesStmt><title level="j" type="main">Computer vision and image understanding : (Print)</title>
<title level="j" type="abbreviated">Comput. vis. image underst. : (Print)</title>
<idno type="ISSN">1077-3142</idno>
</seriesStmt>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Computer vision</term>
<term>Convex surface</term>
<term>Experimental result</term>
<term>Free form</term>
<term>Image processing</term>
<term>Image reconstruction</term>
<term>Initialization</term>
<term>Isometry</term>
<term>Least squares method</term>
<term>Levenberg Marquardt algorithm</term>
<term>Minimization</term>
<term>Modeling</term>
<term>Monocular vision</term>
<term>Non convex analysis</term>
<term>Pipeline processor</term>
<term>Refinement method</term>
<term>Smooth surface</term>
<term>Surface reconstruction</term>
<term>Tridimensional image</term>
<term>conic programming</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr"><term>Image tridimensionnelle</term>
<term>Reconstruction image</term>
<term>Traitement image</term>
<term>Vision ordinateur</term>
<term>Reconstruction surface</term>
<term>Forme libre</term>
<term>Processeur pipeline</term>
<term>Vision monoculaire</term>
<term>Surface convexe</term>
<term>Surface lisse</term>
<term>Initialisation</term>
<term>Analyse non convexe</term>
<term>Isométrie</term>
<term>Programmation conique</term>
<term>Modélisation</term>
<term>Méthode moindre carré</term>
<term>Algorithme de Levenberg Marquardt</term>
<term>Minimisation</term>
<term>Résultat expérimental</term>
<term>Méthode raffinement</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">We study the 3D reconstruction of an isometric surface from point correspondences between a template and a single input image. The template shows the surface flat and fronto-parallel. We propose three new methods. The first two use a convex relaxation of isometry to inextensibility. They are formulated as Second Order Cone Programs (SOCP). The first proposed method is point-wise (it reconstructs only the input point correspondences) while the second proposed method uses a smooth and continuous surface model, based on Free-Form Deformations (FFD). The third proposed method uses the 'true' nonconvex isometric constraint and the same continuous surface model. It is formulated with Nonlinear Least-Squares and can thus be solved with the efficient Levenberg-Marquardt minimization method. The proposed approaches may be combined in a single pipeline whereby one of the convex approximations is used to initialize the nonconvex method. Our contributions solve two important limitations of current state of the art: our convex methods are the first ones to handle noise in both the template and image points, and our nonconvex method is the first one to use 'true' isometric constraints. Our experimental results on simulated and real data show that our convex point-wise method and our nonconvex method outperform respectively current initialization and refinement methods in 3D reconstructed surface accuracy.</div>
</front>
</TEI>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Asie/explor/AustralieFrV1/Data/Main/Curation
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 004051 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Curation/biblio.hfd -nk 004051 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Asie |area= AustralieFrV1 |flux= Main |étape= Curation |type= RBID |clé= Pascal:15-0004489 |texte= Monocular template-based 3D surface reconstruction: Convex inextensible and nonconvex isometric methods }}
This area was generated with Dilib version V0.6.33. |